Handbook of
NON-INVASIVE METHODS and the
SKIN S E C O N D
E D I T I O N Edited by
Jorgen Serup Gregor B.E. Jemec Ga...
67 downloads
1607 Views
33MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Handbook of
NON-INVASIVE METHODS and the
SKIN S E C O N D
E D I T I O N Edited by
Jorgen Serup Gregor B.E. Jemec Gary L. Grove
Boca Raton London New York
A CRC title, part of the Taylor & Francis imprint, a member of the Taylor & Francis Group, the academic division of T&F Informa plc.
Handbook of
NON-INVASIVE METHODS and the
SKIN S E C O N D
E D I T I O N
Published in 2006 by CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2006 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-8493-1437-2 (Hardcover) International Standard Book Number-13: 978-0-8493-1437-7 (Hardcover) Library of Congress Card Number 2005045688 This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Handbook of non-invasive methods and the skin / edited by Jørgen Serup, Gregor B.E. Jemec, Gary L. Grove.--2nd ed. p. ; cm. Includes bibliographical references and index. ISBN 0-8493-1437-2 (alk. paper) 1. Skin--Diseases--Diagnosis. 2. Skin--Imaging. 3. Diagnosis, Noninvasive. 4. Bioengineering. I. Serup, Jørgen. II. Jemec, B.E. III. Grove, Gary L. (Gary Lee) [DNLM: 1. Skin Physiology. 2. Biomedical Engineering. 3. Skin Diseases--diagnosis. WR 102 H236 2005] RL105.H34 2005 616.5’075--dc22
2005045688
Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com Taylor & Francis Group is the Academic Division of Informa plc.
and the CRC Press Web site at http://www.crcpress.com
This handbook is dedicated to Professor Albert M. Kligman, who searched truth and gave so much to so many
Preface This second edition of the Handbook of Non-Invasive Methods and the Skin contains over 100 chapters on bioinstrumental examination of skin including classical reviews and 75 entirely new and updated chapters. The first edition published in 1995 included 89 chapters. In the meantime, much has happened in the field of imaging methods and computer-based techniques, and a number of advanced instruments were introduced. This development as well as the ongoing development of the classical instruments is covered in the second edition. Dr. Gary Grove has joined as co-editor of the second edition. The main purpose of the book is to review important techniques and present key information of practical importance as a guide both to the young and the senior researcher. It is aimed to be a useful guide in academia as well as product development. Taylor & Francis, along with Professor Howard I. Maibach as the main editor, also publishes a number of monographs on specific methods. Another important source of information is the Skin Research and Technology journal (http://blackwellpublishing.com/cservices), official publication of the International Society for Bioengineering and the Skin (ISBS), the International Society for Skin Imaging (ISSI) and the International Society for Digital Imaging of Skin (ISDIS). In October 2005, the ISBS and ISSI decided to merge and form the International Society of Biophysics and Imaging of the Skin (ISBS). This indexed journal publishes original research in the field. In the year 2000, Professsor Pierre Agache, Besancon published the handbook Physiologie de la Peau et Explorations Fonctionelles Cutanées with extensive reviews on bio-instrumental methods written by himself and many French authors. It is to our deep sorrow that Pierre Agache is no longer with us.
The second edition would not have been possible without the generous support and patience of the many contributors, who were invited as internationally recognized experts in the field. The editors wish to extend their gratitude to the contributors. Taylor & Francis has done tremendous work with the book. The support of Barbara Norwitz, Susan Fox-Greenberg and Erika Dery is acknowledged. The second edition is dedicated to Professor Albert M. Kligman of University of Pennsylvania, the Duhring Laboratory, as was the first edition. A whole generation of bio-instrumentalists have been inspired by his visions, sharpness, courage and dedication to truth, and many all over the world enjoy his generosity and personal friendship. Soon 90 and beyond aging, Albert remains at full speed. Jørgen Serup MD, DMSc, Chief-editor of the handbook Professor of Dermatology Linköping University, Sweden Bispebjerg Hospital, Copenhagen, Denmark Gregor BE Jemec MD, DMSc, Co-editor of the handbook Associate Professor of Dermatology University of Copenhagen Roskilde Hospital, Denmark Gary Grove PhD, Co-editor of the handbook Chief Scientist cyberDERM, Inc, Broomall, Pennsylvania
Contributors P. Åberg Karolinska Institutet Medical Engineering Novum Research Park Huddinge, Sweden Pierre G. Agache Department of Functional Dermatology University Hospital Besançon, France Peter J. Altmeyer (Deceased) Dermatological Clinic Ruhr University Bochum Bochum, Germany Peter Andersen Optics and Plasma Research Department Risø National Laboratory Roskilde, Denmark Lars Arndt-Nielsen Center for Sensory-Motor Interaction Laboratory for Experimental Pain Research Aalborg University Aalborg, Denmark Jorge E. Arrese Department of Dermatopathology University Hospital Sart Tilman Liège, Belgium Sitke Aygen Institut für Zentrale Analytik und Strukturanalyse Universität Witten/Herdecke Witten, Germany Lora Bankova Department of Dermatology Friedrich Schiller University Jena, Germany J.C. Barbenel Bioengineering Unit University of Strathclyde Glasgow, Scotland
André O. Barel Faculty of Physical Education and Physiotherapy Vrije Universiteit Brussel Brussels, Belgium Julian H. Barth Department of Chemical Pathlology General Infirmary at Leeds Leeds, United Kingdom T. Bauermann Institut für Zentrale Analytik und Strukturanalyse Universität Witten/Herdecke Witten, Germany Claus Bay Mathematical Statistical Department LEO Pharma Ballerup, Denmark Gianni Belcaro Irvine Laboratory for Cardiovascular Investigation and Research St. Mary’s Hospital Medical School London, United Kingdom Eva Benfeldt Department of Dermatology University of Copenhagen Bispebjerg Hospital Copenhagen, Denmark Enzo Berardesca Department of Dermatology Istituto Dermatologico di S. Maria e S. Gallicano Rome, Italy Andreas J. Bircher Department of Dermatology University Hospital Basel, Switzerland Ulrike Blume-Peytavi Department of Dermatology Hospital Charité Humboldt University Berlin, Germany
C.H. Chang Department of Dermatology College of Medicine Tzu-Chi Medical University Hualien, Taiwan
Shabtay Dikstein Unit of Cell Pharmacology School of Pharmacy The Hebrew University of Jerusalem Jerusalem, Israel
Steen Christiansen Technical University of Denmark Institute of Manufacturing Engineering Lyngby, Denmark
Peter Dykes Cutest Systems Ltd. Cardiff, United Kingdom
E. Claridge School of Computer Science University of Birmingham Birmingham, United Kingdom Peter Clarys Faculty of Physical Education and Physiotherapy Vrije Universiteit Brussel Brussels, Belgium Pierre Corcuff Laboratoires de Recherche de L’Oreal Aulnay Sous Bois, France
E. Anne Eady Department of Microbiology University of Leeds Leeds, United Kingdom C. Edwards Cardiff Biometrics Ltd. Cardiff, United Kingdom Howell G.M. Edwards Chemical and Forensic Sciences School of Pharmacy University of Bradford Bradford, United Kingdom
S. Cotton Astron Clinica The Mount Cambridge, United Kingdom
Jan Efsen Novo Nordisk A/S Bagsvaerd, Denmark
W. Courage Courage + Khazaka Electronic GmbH Köln, Germany
Mariko Egawa Bioengineering Research Labs Shiseido Co., Ltd. Yokohama, Japan
Pierre Creidi Laboratoire de Biologie et d’Ingénierie Cutanée Besançon, France W.J. Cunliffe Leeds Foundation for Dermatological Research Leeds General Infirmary Leeds, United Kingdom John Damia cyberDERM, Inc. Broomall, Pennsylvania David de Berker Bristol Dermatology Centre Bristol Royal Infirmary Bristol, United Kingdom Mitsuhiro Denda Shiseido Research Center Yokohama, Japan
Claudia El Gammal Dermatological Clinic Hospital Bethesda Freudenberg, Germany Stephan El Gammal Dermatological Clinic Hospital Bethesda Freudenberg, Germany Helmut Ermert Institute for High Frequency Techniques Ruhr University Bochum Bochum, Germany Jan Faergemann Department of Dermatology University of Gotenburg Sahlgren’s Hospital Gothenburg, Sweden
Nadia Farinelli Department of Dermatology University of Pavia Pavia, Italy Joachim Fluhr Department of Dermatology Friedrich Schiller University Jena, Germany Bo Forslind (Deceased) Department of Medicine Biochemistry and Biophysics Karolinska Institute Stockholm, Sweden Ann Fullerton LEO Pharma Ballerup, Denmark Bernard Gabard Egerkingen, Switzerland Johannes Gassmueller BioSkin Institut für Dermatologische Forschung und Entwicklung GmbH Hamburg, Germany Tijani Gharbi Laboratoire d’Optique P.M. Duffieux University of Franche-Comté Besançon, France Yolanda Gilaberte-Calzada San Jorge Hospital Huesca, Spain
Chee Leok Goh National Skin Centre Singapore Salvador González Dermatology Service Memorial Sloan-Kettering Cancer Center New York, New York and Department of Dermatology Massachusetts General Hospital Harvard Medical School Boston, Massachusetts and Clinica La Luz Madrid, Spain Costantino Grana Department of Computer Engineering University of Modena and Reggio Emilia Modena, Italy Lotte Groth LEO Pharma Ballerup, Denmark Gary Lee Grove cyberDERM, Inc. Broomall, Pennsylvania Mary Jo Grove cyberDERM, Inc. Broomall, Pennsylvania
Norm V. Gitis Center for Tribology, Inc. Campbell, California
P.N. Hall Department of Plastic Surgery Addenbrooke’s Hospital Cambridge, United Kingdom
Francesca Giusti Department of Dermatology University of Modena and Reggio Emilia Modena, Italy
Allan Halpern Dermatology Service Memorial Sloan-Kettering Cancer Center New York, New York
Monika Gniadecka Department of Dermatology Bispebjerg Hospital Copenhagen, Denmark
Hans Nørgaard Hansen Copenhagen, Denmark
Robert Gniadecki Department of Dermatoloy Bispebjerg Hospital Copenhagen, Denmark
C.W. Hargens Philadelphia, Pennsylvania Roland Hartwig Dermatological Practice Wuppertal, Germany
Stacy S. Hawkins Unilever Research, U.S. Edgewater, New Jersey
H. Irving Katz Minnesota Clinical Study Center Fridley, Minnesota
Andreas Herpens Research Bioengineering — Biophysics Beiersdorf AG Hamburg, Germany
Andrei Kecskés Schering AG Berlin, Germany
Jutta Hofmann proDERM Institute for Applied Dermatological Research Schenefeld/Hamburg, Germany K. Hoffmann Dermatologische Klinik der Ruhr Universität Bochum Bochum, Germany Elisabeth A. Holm Department of Dermatology Roskilde Hospital University of Copenhagen Roskilde, Denmark Takeshi Horio Department of Dermatology Kansai Medical University Osaka, Japan Sidney B. Hornby Neutrogena Corporation Los Angeles, California Philippe Humbert Laboratoire de Biologie et d’Ingénierie Cutanée Besançon, France Pedro Jaén-Olasold Dermatology Service Ramon y Cajal Hospital and Clinica La Luz Madrid, Spain Peter Jahn Schering AG, Diagnostika Koordination Berlin, Germany Gregor B.E. Jemec Department of Dermatology Roskilde Hospital University of Copenhagen Roskilde, Denmark
Jens Keiding LEO Pharma Ballerup, Denmark Nis Kentorp Department of Dermatology Bispebjerg Hospital Copenhagen, Denmark Albert M. Kligman Department of Dermatology University of Pennsylvania Philadelphia, Pennsylvania Jürgen Lademann Center of Experimental and Applied Cutaneous Physiology (CCP) Department of Dermatology University Hospital Charité Humboldt University Berlin, Germany Nicholas Lange Biometric and Field Studies Branch National Institute of Neurological Disorders and Stroke National Institutes of Health Bethesda, Maryland C.H. Lee Department of Dermatology College of Medicine Kaohsiung Medical University Kaohsiung, Taiwan Jean-Luc Lévêque Laboratoires de Recherche de L’Oreal Aulnay Sous Bois, France Jane S. Lindholm Minnesota Clinical Study Center Fridley, Minnesota Sophie Mac-Mary Laboratoire de Biologie et d’Ingénierie Cutanée Besançon, France
Howard I. Maibach Department of Dermatology School of Medicine University of California San Francisco, California
Andrew N. Nicolaides Irvine Laboratory for Cardiovascular Investigation and Research St. Mary’s Hospital Medical School London, United Kingdom
Mette Midttun Department of Medical Physiology The Panum Institute University of Copenhagen Copenhagen, Denmark
Mikkel Noerreslet The Danish University of Pharmaceutical Sciences Copenhagen, Denmark
Jean Mignot Laboratoire de Métrologie des Interfaces Techniques Institut Universitaire de Technologie Besançon, France David L. Miller Bionet Incorporated Dallas, Texas Otto H. Mills, Jr. Hill Top Research, Inc. University of Medicine and Dentistry of New Jersey Robert Wood Johnson Medical School East Brunswick, New Jersey Susanne Møller Mathematical Statistical Department LEO Pharma Ballerup, Denmark M. Moncrieff Department of Plastic Surgery Addenbrooke’s Hospital Cambridge, United Kingdom P.S. Mortimer St. George’s and Royal Marsden Hospitals London, United Kingdom I. Nicander Department of Dermatology Huddinge University Hospital Huddinge, Sweden Carsten N. Nickelsen Hvidovre Hospital University of Copenhagen Hvidovre, Denmark
J. Nuutinen Delfin Technologies Ltd. Kuopio, Finland Ken-ichiro O’goshi Department of Dermatology Bispebjerg Hospital Copenhagen, Denmark Motoki Oguri Shiseido Research Center Yokohama-shi, Japan Chil Hwan Oh Department of Dermatology School of Medicine Korea University Seoul, Korea Hans Öhman Department of Dermato-Venereology University of Linköping Linköping, Sweden S. Ollmar Karolinska Institutet Medical Engineering Novum Research Park Huddinge, Sweden Constantin E. Orfanos Department of Dermatology University Medical Center Steglitz The Free University of Berlin Berlin, Germany Patricia García Ortiz Department of Dermatology University of Copenhagen Gentofte Hospital Hellerup, Denmark Alessandra Pagnoni Philadelphia, Pennsylvania
Giovanni Pellacani Department of Dermatology University of Modena and Reggio Emilia Modena, Italy Adeline Petitjean Laboratoire de Biologie et d’Ingénierie Cutanée Besançon, France Claudine Piérard-Franchimont Department of Dermatopathology University Hospital Sart Tilman Liège, Belgium Gérald E. Piérard Department of Dermatopathology University Hospital Sart Tilman Liège, Belgium J. Pinnagoda Ministry of Public Health Singapore Pascale Quatresooz Department of Dermatopathology University Hospital Sart Tilman Liège, Belgium Bernard Querleux Laboratoires de Recherche de L’Oréal Aulnay-sous-bois, France Milind Rajadhyaksha Dermatology Service Memorial Sloan-Kettering Cancer Center New York, New York E.F.J. Ring Medical Imaging Research Group School of Computing University of Glamorgan Pontypridd, United Kingdom Jeffrey S. Roth Department of Dermatology College of Physicians and Surgeons Columbia University New York, New York D. Hugh Rushton School of Pharmacy and Biomedical Sciences University of Portsmouth Portsmouth, United Kingdom
Iqbal Sadiq S.K.I.N., Inc. Conshohocken, Pennsylvania Jean-Marie Sainthillier Laboratoire de Biologie et d’Ingénierie Cutanée Besançon, France S. Schagen Research Bioengineering — Biophysics Beiersdorf AG Hamburg, Germany Harald Schatz Dermatologische Klinik der Ruhr Universität Bochum Bochum, Germany S. Scheede Research Bioengineering — Biophysics Beiersdorf AG Hamburg, Germany Richard K. Scher Department of Dermatology College of Physicians and Surgeons Columbia University New York, New York Stefania Seidenari Department of Dermatology University of Modena and Reggio Emilia Modena, Italy Per Sejrsen Department of Medical Physiology The Panum Institute University of Copenhagen Copenhagen, Denmark Jørgen Serup Department of Dermatology Linköping University Linköping, Sweden and Department of Dermatology Bispebjerg Hospital Copenhagen, Denmark Raja K. Sivamani Department of Dermatology School of Medicine University of California San Francisco, California
Tracy Stoudemayer S.K.I.N., Inc. Conshohocken, Pennsylvania Hachiro Tagami Department of Dermatology Tohoku University School of Medicine Sendai, Japan
Abel Torres Division of Dermatology Loma Linda University Hospital Loma Linda, California R.A. Tupker Department of Dermatology St. Antonius Hospital Nieuwegein, The Netherlands
Motoji Takahashi Shiseido Research Center Yokohama-shi, Japan and Bioengineering Research Labs Shiseido Co., Ltd. Yokohama, Japan
Andreas Tycho OCT Innovation ApS. Roskilde, Denmark
Hirotsugu Takiwaki Department of Dermatology The University of Tokushima School of Medicine Tokushima, Japan
Michael Vogt Institute for High Frequency Techniques Ruhr University Bochum Bochum, Germany
M. Tanaka Department of Bioengineering and Robotics Graduate School of Engineering Tohoku University Sendai, Japan
Karin Wårdell Department of Biomedical Engineering Linköping University Linköping, Sweden
J.P. Taylor Leeds Foundation for Dermatological Research Leeds General Infirmary Leeds, United Kingdom Roderick A. Thomas Snell International Tyn-Y-Coed Pontardulais, Swansea, United Kingdom Steven G. Thomas Optiscan Pty. Ltd. Notting Hill, Victoria, Australia Lars Thrane Optics and Plasma Research Department Risø National Laboratory Roskilde, Denmark Merete Thyme Quality Assurance Department Scantox (part of LAB Research International) LI. Skensved, Denmark
D. Van Neste Skinterface Tournai, Belgium
Wiete Westerhof Department of Dermatology Academic Medical Center University of Amsterdam Amsterdam, The Netherlands Martin A. Weinstock Dermatoepidemiology Unit VA Medical Center Roger Williams Medical Center and Brown University Providence, Rhode Island Elizabeth Grove Wickersheim cyberDERM, Inc. Broomall, Pennsylvania R. Randall Wickett College of Pharmacy University of Cincinnati Cincinnati, Ohio Klaus-Peter Wilhelm proDERM Institute for Applied Dermatological Research Schenefeld/Hamburg, Germany
Ximena Wortsman Servicio de Imagenologia Hospital del Profesor Santiago, Chile Gabriel Wu Department of Dermatology School of Medicine University of California San Francisco, California Hans Christian Wulf Department of Dermatology Bispebjerg Hospital University of Copenhagen Copenhagen, Denmark Toyonobu Yamashita Bioengineering Research Labs Shiseido Co., Ltd. Yokohama, Japan
H.S. Yu Department of Dermatology College of Medicine National Taiwan University Hospital Taipei, Taiwan H. Zahouani Laboratoire de Tribologie et Dynamique des Systems Ecully, France A. Zemtsov University Dermatology Center, P.C. Muncie, Indiana Charles Zerweck cyberDERM, Inc. Broomall, Pennsylvania Burton Zweiman University of Pennsylvania School of Medicine Philadelphia, Pennsylvania
Table of Contents SECTION I
General Introduction
Chapter 1 Personal Perspectives on Bioengineering and the Skin: The Successful Past and the Brilliant Future ...........................3 Albert M. Kligman Chapter 2 How to Choose and Use Non-Invasive Methods ...............................................................................................................9 Jørgen Serup Chapter 3 A Practical Guide to Resources on the Internet for the Skin Researcher .......................................................................15 Elizabeth Grove Wickersheim and Gary Lee Grove Chapter 4 The Skin Integument: Variation Relative to Sex, Age, Race, and Body Region ............................................................27 Nadia Farinelli and Enzo Berardesca Chapter 5 Seasonal Variations and Environmental Influences on the Skin ......................................................................................33 Chee Leok Goh Chapter 6 Non-Invasive Methods and Assessment of Skin Diseases ...............................................................................................37 Stefania Seidenari, Francesca Giusti, and Giovanni Pellacani Chapter 7 Standards for Good Clinical Practice (GCP)....................................................................................................................47 Merete Thyme Chapter 8 Statistical Analysis of Sensitivity, Specificity, and Predictive Value of a Diagnostic Test .............................................53 Nicholas Lange and Martin A. Weinstock Chapter 9 Sample Size Calculation ...................................................................................................................................................63 Claus Bay and Susanne Møller Chapter 10 Implementation of a Quality Management System in a Contract Laboratory Working with Non-Invasive Methods.......................................................................................................................................................67 Klaus-Peter Wilhelm and Jutta Hofmann
Chapter 11 Ethical Considerations.......................................................................................................................................................73 Mikkel Noerreslet and Gregor B.E. Jemec
SECTION II
Technique, Application, and Validation Skin Surface, Epidermal Structure, and Function
Clinical Photography, Surface Imaging Techniques, and Computerized Image Analysis Chapter 12 General Aspects in Medical/Clinical Photography...........................................................................................................81 Nis Kentorp Chapter 13 Use of Compact Digital Camera for Snap Photography..................................................................................................89 Ken-ichiro O’goshi Chapter 14 Computerized Image Analysis of Clinical Photos............................................................................................................95 Stacy S. Hawkins Chapter 15 Magnifying Lens — Non-Invasive Oil Immersion Examination of the Skin ...............................................................101 H. Irving Katz and Jane S. Lindholm Chapter 16 Dermatoscopy..................................................................................................................................................................109 Wiete Westerhof Chapter 17 Fiber-Optic Microscopy System for Skin Surface Imaging...........................................................................................125 Iqbal Sadiq and Tracy Stoudemayer Chapter 18 Automated Assessment of Pigment Distribution and Color Areas for Melanoma Diagnosis ......................................135 Stefania Seidenari, Giovanni Pellacani, and Costantino Grana Skin Surface Contour and Roughness Assessment Chapter 19 Skin Replication for Light and Scanning Electron Microscopy ....................................................................................147 Bo Forslind Chapter 20 Skin Surface Replica Image Analysis of Furrows and Wrinkles...................................................................................155 Pierre Corcuff and Jean-Luc Lévêque Chapter 21 Stylus Method for Skin Surface Contour Measurement ................................................................................................163 Johannes Gassmueller, Andrei Kecskés, and Peter Jahn
Chapter 22 Laser Profilometry...........................................................................................................................................................169 Jan Efsen, Steen Christiansen, Hans Nørgaard Hansen, and Jens Keiding Chapter 23 Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief ....................................................................179 Jean Mignot Chapter 24 The Morphological Tree of the Cutaneous Network of Lines.......................................................................................195 H. Zahouani and Philippe Humbert Chapter 25 Comparison of Methodologies for Evaluation of Skin Surface Contour and Wrinkles: Advantages and Limitations............................................................................................................................................205 Motoji Takahashi and Motoki Oguri Skin Surface Friction Chapter 26 Tribological Studies on Skin: Measurement of the Coefficient of Friction ..................................................................215 Raja K. Sivamani, Gabriel Wu, Howard I. Maibach, and Norm V. Gitis Chapter 27 Skin Friction Evaluation by Unidirectional Stress Using a Friction Tester ..................................................................225 Mariko Egawa and Motoji Takahashi Chapter 28 Haptic Finger...................................................................................................................................................................233 M. Tanaka Epidermis Structure Chapter 29 Cyanoacrylate Biopsy for Cytologic Evaluation of the Epidermis................................................................................239 Jorge E. Arrese, Pascale Quatresooz, Claudine Piérard-Franchimont, and Gérald E. Piérard Chapter 30 High-Resolution Sonography of the Epidermis In Vivo.................................................................................................245 Stephan El Gammal, Claudia El Gammal, Peter J. Altmeyer, Michael Vogt, and Helmut Ermert Chapter 31 Optical Coherence Tomography in Dermatology...........................................................................................................257 Peter Andersen, Lars Thrane, Andreas Tycho, and Gregor B.E. Jemec Chapter 32 In Vivo Reflectance Mode Confocal Microscopy in Clinical and Surgical Dermatology.............................................267 Salvador González, Yolanda Gilaberte-Calzada, Pedro Jaén-Olasold, Milind Rajadhyaksha, Abel Torres, and Allan Halpern
Chapter 33 In Vivo Reflectance Mode Confocal Laser Microscopy of Melanocytic Skin Lesions.................................................277 Giovanni Pellacani and Stefania Seidenari Chapter 34 In Vivo Confocal Microscopy of the Skin Surface Using Fluorescent Markers ...........................................................285 Steven G. Thomas Chapter 35 In Vivo Confocal Microscopy Application in Product Research and Development......................................................297 Toyonobu Yamashita and Motoji Takahashi Chapter 36 Nuclear Magnetic Resonance (NMR) Examination of the Epidermis In Vivo..............................................................307 Bernard Querleux Chapter 37 Spectrophotometric Intracutaneous Imaging (SIAscopy): Method and Clinical Applications .....................................315 E. Claridge, S. Cotton, M. Moncrieff, and P.N. Hall Epidermis Hydration Chapter 38 Epidermal Hydration: Measurement of High-Frequency Electrical Conductance ........................................................329 Hachiro Tagami Chapter 39 Measurement of Epidermal Capacitance ........................................................................................................................337 André O. Barel and Peter Clarys Chapter 40 Bioimpedance as a Non-Invasive Method for Measuring Changes in Skin..................................................................345 I. Nicander, P. Åberg, and S. Ollmar Chapter 41 Comparison of Commercial Electrical Measurement Instruments for Assessing the Hydration State of the Stratum Corneum.......................................................................................................................351 Bernard Gabard, Peter Clarys, and André O. Barel Desquamation Chapter 42 Methods to Determine Desquamation Rate....................................................................................................................361 C. Edwards Chapter 43 Application of Adhesive Techniques to Harvest Stratum Corneum Material ...............................................................371 David L. Miller
Chapter 44 Dry Skin and Scaling Evaluated by D-Squames and Image Analysis ..........................................................................375 Harald Schatz, Peter J. Altmeyer, and Albert M. Kligman Barrier Functions and Gradients Chapter 45 Measurement of Transepidermal Water Loss by Semiopen Systems ............................................................................383 R.A. Tupker and J. Pinnagoda Chapter 46 Measurement of Transepidermal Water Loss by Closed-Chamber Systems .................................................................393 J. Nuutinen Chapter 47 Measurement of Transcutaneous Oxygen Tension .........................................................................................................397 Hirotsugu Takiwaki Chapter 48 Measurement of Transcutaneous PCO2 ............................................................................................................................407 Carsten N. Nickelsen Chapter 49 Skin Surface pH: Mechanism, Measurement, Importance.............................................................................................411 Joachim Fluhr, Lora Bankova, and Shabtay Dikstein Chapter 50 The pH Gradient in the Epidermis Evaluated by Serial Tape Stripping .......................................................................421 Hans Öhman Chapter 51 Techniques for Visualization of Ionic Gradation in Human Epidermis.........................................................................429 Mitsuhiro Denda Chapter 52 Skin Chamber Techniques...............................................................................................................................................433 Burton Zweiman Chapter 53 Microdialysis Methodology for Sampling in the Skin...................................................................................................443 Lotte Groth, Patricia García Ortiz, and Eva Benfeldt Skin Surface Microflora Chapter 54 Sampling the Bacteria of the Skin..................................................................................................................................457 E. Anne Eady Chapter 55 Mapping the Fungi of the Skin.......................................................................................................................................467 Jan Faergemann
Dermis Structure and Function Dermis Structure Chapter 56 High-Frequency Ultrasound Examination of Skin: Introduction and Guide.................................................................473 Jørgen Serup, Jens Keiding, Ann Fullerton, Monika Gniadecka, and Robert Gniadecki Chapter 57 Ultrasound B-Mode Imaging and In Vivo Structure Analysis .......................................................................................493 Stefania Seidenari Chapter 58 Ultrasound Assessment of Dermal Water and Edema In Vivo .......................................................................................507 Monika Gniadecka Chapter 59 Ultrasound Assessment of Skin Aging ...........................................................................................................................511 Giovanni Pellacani, Francesca Giusti, and Stefania Seidenari Chapter 60 Ultrasound Imaging of Subcutaneous Tissue and Adjacent Structures .........................................................................515 Ximena Wortsman, Elisabeth A. Holm, and Gregor B.E. Jemec Chapter 61 Magnetic Resonance Spectroscopy of the Skin .............................................................................................................531 A. Zemtsov Chapter 62 Nuclear Magnetic Resonance Examination of Skin Disorders......................................................................................537 Stephan El Gammal, Roland Hartwig, Sitke Aygen, T. Bauermann, K. Hoffmann, and Peter J. Altmeyer Chapter 63 Raman Spectroscopy of Skin..........................................................................................................................................551 Howell G.M. Edwards Mechanical Properties Chapter 64 Identification of Langer’s Lines......................................................................................................................................565 J.C. Barbenel Chapter 65 Suction Chamber Method for Measuring Skin Mechanical Properties: The Dermaflex®.............................................571 Monika Gniadecka and Jørgen Serup Chapter 66 Suction Chamber Method for Measurement of Skin Mechanics: The Cutometer® ......................................................579 Ken-ichiro O’goshi
Chapter 67 Suction Chamber Method for Measurement of Skin Mechanics: The New Digital Version of the Cutometer ................................................................................................................................................583 André O. Barel, W. Courage, and Peter Clarys Chapter 68 Suction Chamber Method for Measurement of Skin Mechanics: The DermaLab........................................................593 Gary Lee Grove, John Damia, Mary Jo Grove, and Charles Zerweck Chapter 69 Twistometry Measurement of Skin Elasticity ................................................................................................................601 Pierre G. Agache Chapter 70 Levarometry.....................................................................................................................................................................613 Shabtay Dikstein and Joachim Fluhr Chapter 71 Indentometry....................................................................................................................................................................617 Shabtay Dikstein and Joachim Fluhr Chapter 72 The Gas-Bearing Electrodynamometer...........................................................................................................................621 C.W. Hargens Chapter 73 Ballistometry ...................................................................................................................................................................627 C.W. Hargens
The Cutaneous Vasculature Skin Color and Blood Vessels Chapter 74 Colorimetry......................................................................................................................................................................635 Wiete Westerhof Chapter 75 Quasi-L*a*b* Color Measurement from Digital Images...............................................................................................649 Hirotsugu Takiwaki Chapter 76 Practical Color Calibration for Dermatoscopic Images .................................................................................................653 Constantino Grana, Giovanni Pellacani, and Stefania Seidenari Chapter 77 Measurement of Erythema and Melanin Indices............................................................................................................665 Hirotsugu Takiwaki
Chapter 78 Dynamic Capillaroscopy .................................................................................................................................................673 H.S. Yu, C.H. Lee, and C.H. Chang Chapter 79 Capillaroscopy and Videocapillaroscopy Assessment of Skin Microcirculation: Dermatological and Cosmetic Approaches......................................................................................................................................................679 Philippe Humbert, Jean-Marie Sainthillier, Sophie Mac-Mary, Adeline Petitjean, Pierre Creidi, and Tijani Gharbi Blood Flow, Vasomotion, and Vascular Functions Chapter 80 Laser Doppler Measurement of Skin Blood Flux: Variation and Validation.................................................................691 Andreas J. Bircher Chapter 81 Examination of Periodic Fluctuations in Cutaneous Blood Flow..................................................................................697 Robert Gniadecki, Monika Gniadecka, and Jørgen Serup Chapter 82 Laser Doppler Flowmetry: Principles of Technology and Clinical Applications..........................................................709 Gianni Belcaro and Andrew N. Nicolaides Chapter 83 Laser Doppler Imaging of Skin ......................................................................................................................................717 Karin Wårdell Chapter 84 The Heat Wash-In and Heat Wash-Out Technique for Quantitative, Non-Invasive Measurement of Cutaneous Blood Flow Rate................................................................................................................723 Per Sejrsen and Mette Midttun Chapter 85 The 133Xenon Wash-Out Technique for Quantitative Measurement of Cutaneous and Subcutaneous Blood Flow Rates ....................................................................................................................................733 Per Sejrsen Chapter 86 Evaluation of Lymph Flow .............................................................................................................................................741 P.S. Mortimer Temperature and Thermoregulation Chapter 87 Sensors and Handheld Devices for Surface Measurement of Skin Temperature ..........................................................753 Roderick A. Thomas Chapter 88 Thermal Imaging of Skin Temperature ..........................................................................................................................769 E.F.J. Ring
Neural Supply Chapter 89 Assessment of Cutaneous Pain .......................................................................................................................................787 Lars Arendt-Nielsen
Sweat Gland Distribution and Function Chapter 90 Classical Techniques for the Localization of Sweat Glands..........................................................................................805 Peter Dykes Chapter 91 Micro-Sensor Mapping of Sudoral Activity and Skin Surface Hydration.....................................................................811 Jean-Luc Lévêque Chapter 92 Methods for the Collection of Eccrine Sweat ................................................................................................................817 Julian H. Barth Chapter 93 Methods for the Collection of Apocrine Sweat..............................................................................................................821 Julian H. Barth
Sebaceous Glands and Sebum Excretion Chapter 94 The Follicular Biopsy......................................................................................................................................................825 Otto H. Mills, Jr. Chapter 95 Measurement of Excreted Sebum Using Sebum-Absorbent Film and an Optical Reader: The TapeAnalyzer ...........................................................................................................................................................831 David L. Miller Chapter 96 Quantification of Sebum Output Using Sebum-Absorbent Tapes (Sebutapes®)............................................................835 Claudia El Gammal, Stephan El Gammal, Alessandra Pagnoni, and Albert M. Kligman Chapter 97 Optical Measurement of Sebum Excretion Using Opalescent Film Imprint: The Sebumeter® ....................................841 Ken-ichiro O’goshi Chapter 98 Gravimetric Technique for Measuring Sebum Excretion Rate (SER) ...........................................................................847 W.J. Cunliffe and J.P. Taylor Chapter 99 Fluorescence Photography of Sebaceous Follicles.........................................................................................................853 Andreas Herpens, S. Schagen, and S. Scheede
Chapter 100 Methods for Assessment of Follicular Transport in Ex Vivo and In Vivo......................................................................861 Jürgen Lademann
Hair, Physical Properties, and Growth Rate Chapter 101 Measurement of Hair Growth .........................................................................................................................................869 Julian H. Barth and D. Hugh Rushton Chapter 102 Microscopy of the Hair: The Trichogram.......................................................................................................................875 Ulrike Blume-Peytavi and Constantin E. Orfanos Chapter 103 Photographic and Computerized Techniques for Quantification of Hair Growth .........................................................883 D. Van Neste Chapter 104 Measurement of the Mechanical Strength of Hair .........................................................................................................895 R. Randall Wickett Chapter 105 Evaluating the Strength of Human Hair .........................................................................................................................903 Sidney B. Hornby
Nail Structure and Growth Chapter 106 Methods for Nail Assessment: An Overview .................................................................................................................911 David de Berker Chapter 107 Measurement of Longitudinal Nail Growth ...................................................................................................................919 Jeffrey S. Roth and Richard K. Scher Chapter 108 Measurement of Nail Thickness .....................................................................................................................................923 Gregor B.E. Jemec Chapter 109 Image Analysis of the Nail Surface................................................................................................................................925 Claudine Piérard-Franchimont and Gérald E. Piérard
SECTION III
Clinical Experimentation, Evaluation, and Quantification
Chapter 110 General Guidelines for Assessment of Skin Diseases....................................................................................................931 Elisabeth A. Holm and Gregor B.E. Jemec
Chapter 111 Sodium Lauryl Sulfate (SLS) Testing: ESCD Application and Reading Standards .....................................................943 R.A. Tupker Chapter 112 Instrumental and Computer-Based Methods for Measurement of Surface Area Afflicted with Disease .....................957 Chil Hwan Oh Chapter 113 Instrumental Evaluation of Wheal-and-Flare Reactions.................................................................................................967 D. Van Neste Chapter 114 Instrumental Evaluation of Occluded Patch Test Reactions ..........................................................................................973 Stefania Seidenari, Francesca Giusti, and Giovanni Pellacani Chapter 115 Light Sources, Sunlight, and Radiation Dosimetry........................................................................................................981 Hans Christian Wulf Chapter 116 Phototesting: Phototoxicity and Photoallergy.................................................................................................................991 Takeshi Horio Index ...............................................................................................................................................................................997
Section I General Introduction
Perspectives on 1 Personal Bioengineering and the Skin: The Successful Past and the Brilliant Future Albert M. Kligman Department of Dermatology, University of Pennsylvania, Philadelphia, Pennsylvania
CONTENTS 1.1 Some Current Issues..................................................................................................................................................4 1.2 Perspectives on the Future.........................................................................................................................................5 1.3 Contact Dermatitis .....................................................................................................................................................6 1.4 Diagnosis of Predisease.............................................................................................................................................6 1.5 Photoaging .................................................................................................................................................................6 1.6 Epilogue .....................................................................................................................................................................6 References ...........................................................................................................................................................................7
The International Society for Bioengineering and the Skin, now 25 years of age, the brainchild of Ronald Marks and Harvey Blank, has passed through its adolescent phase in good health and is now poised to reap the fruits of adulthood. We now have a cornucopia of powerful instruments that can reveal the mysteries of skin in health and disease, all without touching the surface, a triumph of bioengineering creativity. Moreover, unlike traditional histological studies, which give a static picture at one point in time, ruining the site for further study, bioengineering techniques present a moving picture of sequential events in real time, a delight to watch and a gratifying experience for investigators whose daily work is generally pretty dull and dreary. These technical wonders were unthinkable at the end of my residency in 1950. The reigning doctrine at that time was that dermatology was a choice field of study because the skin was so accessible to the eye and to the finger. Everything was laid out before your eyes, unlike for internists and surgeons, who needed x-rays to see what was going on inside. I think that this entrenched belief in the diagnostic powers of touching and looking at the skin kept dermatology as a backwater, purely descriptive specialty for more than a hundred years. Dermatologists were the butt of many scornful jokes and dermatology was derided as a skin game. Medical graduates who chose dermatology as a career came from the bottom of the class.
In my essay on the invisible dermatoses, written 25 years ago, I opined that the most important early events in the pathogenesis of skin disease were largely invisible, hidden beneath the surface.1 Visible signs were a late stage in the disease process, obscuring what had gone on before, useful for classification but providing no insights regarding pathogenesis. Moreover, dermatologists were mistaken when they thought, as many still do, that clinical clearing of lesions is a reliable means to declare therapeutic success. The fact is that in chronic diseases such as psoriasis, the visibly cleared site remains abnormal for many months, evidenced by histologic study. Besides, the subjective estimate of say 50 to 75% clearing brings little satisfaction to patients who want to be free of lesions. Subjective estimates of improvement still plague us to this very day. The beauty of bioengineering and imaging techniques is that clinical changes are measurable and quantitative, based on objective techniques. Stopping treatment upon clearing is a sure invitation to recurrence, which in psoriasis occurs precisely at the “cleared” site. It is biological nonsense to think that the skin is a transparent window through which the pawing and peering dermatologist could ascertain the underlying changes. The professorate of my day loved the window imagery. I prophesied that the golden age of dermatology would begin when a totally blind student would not be barred from applying for a residency in dermatology.2 I now allow myself the conceit of foreseeing that new instruments would be 3
4
Handbook of Non-Invasive Methods and the Skin, Second Edition
developed that would far surpass the human eye for diagnosing and treating disease, as well as providing an increased understanding of how diseases evolved and resolved. The blind student can now “see” what is really important. I argued in my invisible dermatology thesis that occult, subclinical events were far more important in diagnosing and understanding disease than what was revealed to the naked eye. I gave examples of different disorders that illustrated how much dermatologists were missing by depending so much on the eye. To be sure, some masters in the early 20th century had an inkling of the invisible dermatology concept. For example, Gougerot observed that the normal-appearing skin of patients with leprosy showed granulomatous lesions histologically. I emphasized that the skin had a long memory of previous insults, for example, allergic contact dermatitis. I sensitized a volunteer to gold chloride who reacted strongly to a patch test. The inflamed site healed nicely. Then, when this person was given a gold salt orally for the treatment of rheumatoid arthritis, the patch test site flared up brilliantly. Another memorable experience happened to a colleague to whom we mistakenly gave 8 MEDs to a 1-cm circle. Now, every time she takes a hot bath, the site turns briefly red. The last flare was 11 years after the irradiation. The skin does not forget. Imaging techniques in both instances reveal that the cleared sites are not normal. The enterprise of developing novel, ingenious, noninvasive measuring and imaging devices has been a remarkable success story. The number and variety of these have been so great that Ronnie Marks has remarked that “one could well argue that it is time for a pause. We are confronted with a plethora of instrumentation which is simply overwhelming.” I take this to be the weary utterance of an aging dermatologist, who he himself has described as GOD (grumpy old dermatologist), struggling to stay afloat in the rushing stream of new inventions that stagger the mind. These are as thrilling as they are overwhelming. The creations of the bioengineering community are coming so fast that the instrument one buys today will be obsolete in 5 years. Why should we pause when each month seems to provide us with a novel technique to measure functional and structural changes that we could not have dreamed about 10 years ago. In his 1989 text, Lévêque was impressed by the availability of non-invasive devices, which he estimated to be about 20.3 In my introduction to the first edition of this volume I estimated that the inventory of new instruments was about 50. The number is now approaching 300 (Jørgen Serup, personal communication). Many of these are expensive, still experimental, and not commercially available. Most are highly specialized and are useful only to investigators with a specific focus, for example, in the diagnosis and treatment of malignancies. Still, there is no question that their power to provide important new images
and accurate measurements will compel investigators in all areas of the broad domain of skin to get on board this train, which is moving so swiftly to the future. It is now 10 years since the first publication of this classic text, Handbook of Non-Invasive Methods and the Skin. This new edition may be viewed as a celebration of the enormous advances that have taken place in one decade. A mere glance at the impressive list of authors of this text, comprising all the major players in the field, reflects the progress that has been made, for which the term fabulous is not an exaggeration.
1.1 SOME CURRENT ISSUES Lest one become inebriated by our dizzying success story, I bring attention briefly to some issues that are relevant to our growth and future status. Bioengineering has still not made a great impact on the practice of dermatology. At least in the U.S., very few departments possess more than one bioengineering instrument, and most not even that. This is regrettable since these non-invasive techniques have great potential for enhancing teaching, research, and patient care. We have still not penetrated the groves of academia where the current emphasis is on molecular biology, genomics, immunology, proteonomics, etc. These are lofty, important, fashionable subjects that are the favorites of national funding and granting agencies. No one denies that these are cutting-edge areas that deserve support. On the other hand, imaging and bioengineering fall into the same category. To my knowledge, the National Institutes of Health has not funded any study centered on this fast-growing field, which has enormous diagnostic potential. This is extremely shortsighted. It is regrettable that there has been virtually no effort to demonstrate to chairmen of researchoriented departments of dermatology the usefulness of bioengineering techniques for the practice of what is now called evidence-based medicine. We need to publish in more mainstream clinical journals and to find more places on national and international meetings of dermatologic societies. At recent meetings of the American Academy of Dermatology and the European Academy of Dermatology and Venereology, where at each venue about a thousand posters were exhibited, I found no more than two at either meeting that had bioengineering as a central theme, although some used such techniques in various clinical and experimental studies. We are not a service industry but a scientific enterprise that needs standing in its own right. We need to be less parochial as a society and to spend more time in educating the professors who mentor the young. At the same time, we must praise industry and its subsidiaries for investing in and developing the sophisticated technologies now in our hands. The skin care
Personal Perspectives on Bioengineering and the Skin: The Successful Past and the Brilliant Future
industry has shown both vision and financial courage in putting so much money and effort into developing new non-invasive models. It is also appropriate to pay tribute to the confederacy of engineers, material scientists, physicists, chemists, even mathematicians, that have come together to produce instruments that are marvels of ingenuity and precision. Who else but industry, often viewed as basely commercial, would have created such powerful tools as fluorescent confocal microscopy, optothermal coherent tomography, fringe projection imaging, capacitive pixelsensing technology, magnetic resonance imaging, ultrasonography, and many others. The latest inventions increasingly deal with dynamic physiological functions in addition to structure, for example, tracking in real time the complex events that take place during the inflammatory process.5,6 Non-invasive technology has the prospect of fulfilling the age-old dream of preventing rather than treating diseases. What we shall be able to do is constrained only by our imagination, not by technical limitations. The time is not far off when we will be tracking and visualizing structures at the molecular level, also establishing stratum corneum gradients of water and elements such as calcium, magnesium, and sodium. An aging population is now seeking help from cosmetic surgeons who are using a variety of antiaging surgical procedures, which include myriads of lasers, chemical peels, light therapies, microdermabrasion, etc., which can improve the appearance and quality of photoaged skin. Cosmetic dermatologists are beseiged by manufacturers of a huge array of these devices, whose efficacy is hawked to practitioners by intense advertising and marketing, with scanty science to back up the therapeutic benefits. The only way to differentiate among these procedures is by the use of noninvasive techniques. One will have to look hard to find how these devices compare in efficacy to each other for a given indication. Some cautions are in order since every good thing can be subverted toward unworthy and ignoble ends. A case in point is the burgeoning of hundreds of antiwrinkle topical creams and formulations now flooding the marketplace in a multi-billion dollar industry. The marketers of such products often exploit non-invasive technology to substantiate “scientifically” claims of efficacy, taking advantage of the fact that cosmetic products that promise to restore a youthful appearance are not regulated by the Food and Drug Agency, providing ample room for hype, gross exaggeration, and preposterous claims, served up to credulous consumers who are unable to distinguish fact from fiction. Physicians are often in the same predicament. For the unscrupulous manufacturer, non-invasive methodologies can be mobilized that can parade before the profession a host of “objective” data to support claims of efficacy. It is an easy matter to provide the “right” answers required by the sponsor by selecting the “right”
5
equipment. It is well to keep in mind that the bedrock reputation of bioengineering is based upon its credo of objectivity of providing unbiased information that is credible and truthful. Unfortunately, there are rampant exceptions to this state of detachment in the highly competitive antiaging marketplace. Anyone can select a number of instruments that seemingly, but falsely, support claims. It is an indispensable requirement to know what is being measured and whether it is truly relevant to the intended purpose. Most measurements are one-dimensional and focus on only one factor of a complex phenomenon in which many coexisting factors are interacting. One needs to know what the measurement actually means in order to interpret whether the result is good or bad. Not everything that is measurable is worth measuring. Many measurements, backed up by formidable statistics, are simply meaningless. One of my widely quoted maxims applies here: “A fool with a tool is still a fool.” One might add that a rogue with a probe is still a rogue. So much for the misuse of bioengineering methods to promote the current rash of products, some of which are marginally effective or even harmful. I am joined by JeanLuc Lévêque, who in the introduction to his 1989 text remarked that “an abundance of apparatuses with uncertain functions and measurements whose significance is often doubtful, debauched and contradictory will not serve the science of dermatology.”3 On the positive side are the publications created by members of our society aimed at establishing international guidelines for conducting non-invasive tests (which also happen to be nonaversive and patient friendly). These serve the extremely important function of standardizing procedures so that investigators in widely separated laboratories can obtain reproducible results and come to similar conclusions. These carefully constructed documents have been invaluable in precisely defining guidelines that ensure that contradictory or controversial results are not due to technical details. We know that subtle differences can have large effects. Panels of experienced investigators have now given us explicit guidelines for measuring transepidermal water loss, laser Doppler imaging, dermoscopy, stratum corneum hydration, tristimulus colorimetry, optical profilometry, and others in an ongoing international collaboration.
1.2 PERSPECTIVES ON THE FUTURE Ample tools are now in place with regard to clinical applications concerning the classification, diagnosis, pathogenesis, and treatment of acute and chronic dermatoses. Endless possibilities jump out before our eyes if we consider what can be done with confocal microscopy, which makes it possible to optically cut through the skin in horizontal sections 1 to 3 μ thick. Cells and organelles with different reflective properties can then be identified
6
Handbook of Non-Invasive Methods and the Skin, Second Edition
from the surface down to the papillary dermis. Corneocytes, keratinocytes, melanocytes, Langerhans cells, and dermal papillae come into view.4 It is a feast for the eye to see red blood cells coursing through capillaries accompanied by emigration of leukocytes in various inflammatory conditions. With fluorescent confocal microscopy it is possible to see terminal axons coursing through the epidermis, which will make it possible to show how nerves play a central role in what is now called neurogenic inflammatory processes.5 Non-invasive imaging of skin tumors is already extremely valuable in differentiating malignant melanomas. I cite the following possibilities from my own experience.
1.3 CONTACT DERMATITIS Patch testers find it exceedingly difficult to distinguish irritant from allergic reactions, especially when the responses are mild. Although allergic reactions come under the category of delayed hypersensitivity states, we have seen early microvascular events such as swelling of vascular endothelial cells and trafficking of leukocytes into the tissue as early as 1 hour after applying contact allergens such as nickel sulfate in sensitized patients. By contrast, in irritant reactions, such as those produced by sodium lauryl sulfate, we see very little dermal change after 2 to 3 hours except for swelling of corneocytes. Here, timing is of the essence since at the end of 24 hours, irritant and allergic reactions become indistinguishable. False positive irritant patch tests are a huge problem in the diagnoses of allergic contact dermatitis.
1.4 DIAGNOSIS OF PREDISEASE Chronic dermatoses do not spring up overnight, but may be preceded by occult changes, even years before the disorder becomes clinically visible. Being able to recognize the invisible early stages makes it possible to begin effective treatment long before the signs and symptoms of disease cause overt suffering. The ancient ideal of clinicians can then be realized, namely, to prevent disease rather than to start therapy after it has become clinically visible, by which time it may have caused irreversible damage. I cite two cases in which the strategy of prevention has succeeded. In high-risk young persons whose family history shows relatives with classic rosacea, but who show no signs of a red face, videomicroscopy with polarized light can reveal a network of invisible telangiectatic vessels, as early as age 10. Mild topical treatments starting then will likely greatly retard the later emergence of clinical rosacea, with its distressing psychosocial consequences.
In the case of high-risk prepubertal youngsters whose parents have experienced scarring acne in adolescence, microcomedones, the precursor of all later manifestations, can be demonstrated as early as age 8 in females using optical coherent tomography. Optical coherent tomography furnishes transverse images perpendicular to the surface, analogous to traditional histologic sectioning. One sees multiple follicles distended with horny material (microcomedones), completely invisible to the naked eye. Topical retinoids are very effective in eliminating these horn-filled follicles, preventing progression to the visible manifestations of acne, namely, comedones and papulo-pustules. Clinicians in all specialties are recognizing the great benefits of diagnosing predisease so that prevention can finally replace treatment; the latter is the paradigm physicians have followed for centuries.
1.5 PHOTOAGING By taking biopsies of the face throughout the entire life span of white persons, I showed four decades ago that there were striking alterations of the dermal matrix in young subjects, ages 10 to 15, with lovely, unblemished complexions. They all showed a moderate amount of branched, thickened, curled masses of abnormal elastic fibers (elastosis). Again, topical retinoids can reverse these recondite changes and, along with other protective measures against exposure to solar radiation, prevent the inevitable, and detested, dreadful signs of photoaging. It now turns out that we can estimate the degree of elastosis by using high-intensity blue light, in the 400- to 500-nm spectral range. Abnormally thickened and increased lastic fibers exhibit fluorescence, the amount of which can be captured and graded by appropriate imaging. We can then educate young people to take proper precautions against the destructive effects of heedless exposure to mid-day sunlight. Showing patients the difference between their chronological age and their photoage has greater impact on protective behaviors than warning against actinically induced cancers, such as squamous cell cancer and malignant melanoma.
1.6 EPILOGUE We shall know that non-invasive methodologies have finally come of age when volumes like this one can be found on the shelves of dermatologic libraries everywhere, along with the classic texts that are required reading for informed practitioners. My only criticism of this voluminous text, edited by the indefatiguable Jørgen Serup, is that it may contain much more information than you care to know about.
Personal Perspectives on Bioengineering and the Skin: The Successful Past and the Brilliant Future
The future of non-invasive technology is not only bright but positively radiant.
REFERENCES 1. AM Kligman. The invisible dermatoses. Arch Dermatol 127:1375, 1991. 2. AM Kligman. Blind man dermatology. J Soc Cosm Chem 17:505, 1966. 3. JL Lévêque. Cutaneous Investigations in Health and Disease. Marcel Dekker, New York, 1989.
7
4. M Rajadhyaksha, S Gonzalez, JM Zavislan, RR Anderson, and RH Webb. In vivo confocal scanning laser microscopy of human skin. II. Advances in instrumentation and comparison with histology. J Invest Dermatol 113:293, 1999. 5. LD Swindle, SG Thomas, M Freeman, and PN Delancy. View of normal human skin in vivo as observed using fluorescent fiber-optic confocal microscopic imaging. J Invest Dermatol 121:706, 2003. 6. E Ruocco, F Arganziano, G Pellacani, and I Seidenari. Non-invasive imaging of skin tumors. Am Soc Dermatol Surg 30:301, 2004.
to Choose and Use Non-Invasive 2 How Methods Jørgen Serup Department of Dermatology, Linköping University, Linköping, Sweden Department of Dermatology, Bispebjerg Hospital, Copenhagen, Denmark
CONTENTS 2.1 Introduction................................................................................................................................................................9 2.2 Choice of Method and Instrument ............................................................................................................................9 2.3 The Legal or Authoritative Reference behind Study by Non-Invasive Methods ...................................................11 References .........................................................................................................................................................................12
2.1 INTRODUCTION Professor Albert Kligman is cited for the provocative statement “A fool with a tool is still a fool.” The complete tool in a study is much more than the instrument used for the measurement or scanning. It is also the study design, including the purpose and the declared endpoint, the sample, the statistical method, the legal reference, the guideline or operation procedure, the competence of the researcher and his allied, the ethical aspect and the resource, capacity and economy. It is not easy to be up to ideal standard in every aspect. It may actually be very difficult to avoid foolishness in research, and foolishness might even turn out to be innovative. Innovation and progress can be strangled by rigid research principles and burocracy. Use of methods can be so critical that even potential Nobel Prize winner projects are not started or brought to a conclusion. Projects range from the very first observation, which may be a result of serendipity, to the final confirmation and proof conducted under the strict requirements of an authoritative body. Nevertheless, academic use of noninvasive methods is to master precision and handle the signal-to-noise ratio in a sound, honest, and balanced way, and the need for precise and accurate measurement is paramount and not dependent on the project type and state, as Professor Kligman’s statement implicates. A fool is a fool, and we must be well prepared and careful with the methods and the projects.
2.2 CHOICE OF METHOD AND INSTRUMENT There is always an idea, a vision, or something broader behind the penciled endpoint in a protocol. The very first
question is simply to make a move in the right direction and enter the right door. We have very little tradition for qualitative research methods to outline the contour of a problem and find special facets or domains of major importance worthy for a detailed study by quantitative research methods, such as the non-invasive methods. Typically, we start with a somewhat occasional discussion among colleagues and sponsors, a discussion kept quite open to subjective preferences. However, we learned from evidence-based medicine that such personal opinion has variable value. Discussion about project perspective, idea, study endpoint, and choice of method can be improved and made more professional if some of the methods or principles developed in qualitative research are adopted. Especially, consumer-oriented studies can benefit from a prestudy assessment finding the right type of point to measure. Methods such as focus group interviews are established for such a purpose.1 The non-invasive methods are by nature quantitative and narrow. We follow the ground principle of natural sciences expressed by the French physiologist Claude Bernard, to measure and to make objects measurable. Today, we are digitalizing vigorously. The methods are narrow because on the sensor side, they are based on a single physical modality such as sound, light or ray, electricity, etc. Signal processing is with modern electronics highly complex, and every component of the process has its own linearity, cutoff, and upper limit. The resultant measured value is in comparison with the in vivo biological object, an artifact that only becomes valid through validation study, comparisons with reference methods, calibrations, etc. Computerized imaging is quantitative and involves filtering, manipulation, and measurement of small picture elements or pixels, but the biological sample imaged consists of surfaces, tissue, and cells not built up of 9
10
Handbook of Non-Invasive Methods and the Skin, Second Edition
digital squares, each of intensity 0 to 255. The image perceived with the naked eyes is a very complex and nonlinear comprehension, often purely qualitative and often individual or personal. Choice of method in a study to verify a hypothesis is a critical decision where the character of the endpoint and the usefulness and validation of the method are of special importance. In applied medicine and in product substantiation, the real endpoint may be identical to the protocol endpoint or broader in nature. It may be feeling of improvement of disease, impression of younger skin, or some other feature or change noticed with the eyes, felt with the finger, or otherwise perceived by a person. The ideal validation of a non-invasive method is only achieved if perception by lay individuals or special observers is established as the gold reference. In pure research, validation and reference to standard are, of course, much simpler, as exemplified by measurement of size and position of an object in the dermis by high-frequency ultrasound. The following 30-point prestudy checklist is proposed for systematic use when a study is initiated (for definitions regarding instrument performance and validation, see Table 2.1): TABLE 2.1 Instrument Validation, Terms and Definitions Accuracy: Degree of similarity between the value that is accepted either as a conventional true value (in-house or local standard) or as an accepted reference value (international standard) and the mean value found by performing the test procedure a certain number of times. Provides an indication of systemic error. Precision: Degree of similarity (degree of scatter) among a series of measurements obtained from multiple sampling of the same homogenous sample under prescribed conditions, expressed as repeatability and reproducibility. Repeatability: Expresses the situation under the same conditions, that is, same operator, same apparatus, short time interval, identical samples. Reproducibility: Expresses the situation under different conditions, that is, different laboratories, different samples, different operators, different days, different instruments from different manufacturers. Range: The interval between the upper and lower levels for which the procedure has been demonstrated as applicable with precision, accuracy, and linearity. Linearity: Ability of the procedure (within a given range) to obtain test results directly proportional to true values. Sensitivity: Capacity of the procedure to record small variations or differences within the defined range. Limit of detection: Lowest change above zero that is detectable. Limit of quantification: Lowest change above zero that can be quantitatively determined (not only detected) with defined precision and accuracy under the stated experimental conditions. Ruggedness: Evaluates the effects of small changes in the test procedure on measuring performance.
1. What is the study idea and what is the precise study endpoint supporting the idea? 2. Is the study endpoint truly quantitative in nature, narrow enough for specific study, and truly suited to support the idea? 3. Stratification of endpoints into primary, secondary, tertiary, etc.? 4. Shall one or more instruments be applied (monoinstrumental or multi-instrumental design)? 5. Function of the measurements and the instrument in the study: support, description, exclusion, comparison, validation during study, etc.? 6. Which structure or function is actually being measured? 7. Range, linearity, and expected change of variables during study? 8. When should measurements be performed? 9. Interperson, intraperson, and intralesion variation, and if possible, variability data from normal and healthy skin? 10. Influence of gender, age, and race? 11. Statistical evaluation of the design and the size of the sample studied? 12. Studies and literature validating the instruments applied? 13. If the target or measured area is small, do measurements need be repeated to overcome local site variation? 14. Existing in-house standards or recommendations, standard operating procedure (SOP)? 15. Guidelines and legal requirements, including ethical aspects? 16. Output from the instrument and source data — can these be handled and stored safely? 17. Is the laboratory room and are the laboratory conditions up to good standard, or are improvements needed? 18. Is a backup situation prepared if unexpected breakdown occurs of the device, laboratory, etc.? 19. Are ambient conditions such as temperature and humidity under control and expected to remain constant during the study period? 20. Needs for preconditioning of study subjects? 21. Are technicians well educated, trained, and well prepared for the specific study, and who is responsible for what? 22. Are various types of bias identified and, if possible, eliminated? 23. How is it ensured or monitored that the study develops as planned, and what are the requirements for constancy and the consequences of inconstancy?
How to Choose and Use Non-Invasive Methods
24. Calibration, maintenance, and control of instruments before, during, and at end of study? 25. Events and circumstances that exclude measurements from being performed or invalidate results? 26. How to conclude and report the study? 27. Timetable for the study — is it realistic and satisfactory? 28. Resources involved — are they available from start to end? 29. Is the study at an academic level where conclusion and interpretation are independent of economic interest, even if the study is conducted in an industrial environment? 30. Is the study documented and prepared for a special situation if some accusation about fraud would come up? The typical pitfalls are:
11
2.3 THE LEGAL OR AUTHORITATIVE REFERENCE BEHIND STUDY BY NONINVASIVE METHODS There are numerous references, some strictly technical, others more general (Figure 2.1). References may be institutional, national, regional, or global. References undergo constant change and update. It is up to the researchers to ensure that their study is in accordance with the relevant references at a given time. Good laboratory practices (GLPs) are standards mainly used in the pharmaceutical industry for preclinical safety testing in vitro and for experimental study of animals. Good clinical practices (GCPs) are standards used in the pharmaceutical industry for the good conduct of clinical trials from phase I to IV. The International Conference on Harmonization (ICH) GCP guideline (ICH-GCP) in 1996 introduced a tripartite GCP system recognized in the U.S., Europe, and Japan. GCP is particularly relevant for non-invasive measurement in humans and is covered in Chapter 7. However, GCP in a milder version was in 2001
1. Strategic error: poor design and study plan 2. Technological error: poor choice of variable and method 3. Technical error: measuring device not functioning, inconstant, or unstable 4. Performance error: imprecise or not competent use of the device 5. Inadequate measuring conditions: laboratory facilities not acceptable 6. Object-related error: poor selection and preconditioning of study subjects 7. Data error: wrong data acquisition, handling, and storing 8. Malconclusion: conclusion not really supported by data, or inconvenient conclusion suppressed 9. Explanatory mistake: inferior reporting and publication policy The soundness of research is a dimension in itself, difficult to describe and variable from project to project. Clarity and simplicity are at least two important technical elements. Honesty and goodwill are two other elements at another level. Wisdom and knowledge are not identical. A sound project has wisdom built in and, of course, a solid base of knowledge. Maybe soundness is illustrated best with the words of the English author Rudyard Kipling: “I keep six honest serving men. They taught me all I know. Their names are WHAT and WHY and WHEN and HOW and WHERE and WHO.”
FIGURE 2.1 Searching the literature. Cartoon by Storm P, 1944, Copenhagen, Denmark. (Published with permission of the Storm P Museum, Jens Bing, Copenhagen.)
12
Handbook of Non-Invasive Methods and the Skin, Second Edition
according to a European directive implemented in public institutions and hospitals for any clinical trial on patients. The strong emphasis on regulation of clinical studies for the last decades and the increased need for resource and economy to conduct studies has reduced the medical profession’s involvement and activity in clinical research initiated and conducted by the clinicians themselves, independent of sponsorship. Serendipity, as opener of major achievement and based on clinical observation, may soon belong to history. The development of standards and requirements has moved development away from innovation and in the direction of strictly scheduled confirmative research. Fortunately, many portable, accurate, easy-tooperate non-invasive devices were developed and commercialized, and the bioengineer is relatively privileged in comparison with other fields of clinical research. The Declaration of Helsinki of 1964 is the classical global standard for clinical study, with emphasis on ethical conduct of study in humans. This declaration introduced rules for information of study subjects, the informed consent. It became adopted in many countries and legally implemented. The original version and recent updates are found on the website of the World Medical Association (WMA). The device itself shall of course do no harm to the person studied. In the U.S., the Food and Drug Administration (FDA) has formulated rules for safety classification and marking of medicotechnical instruments. In Europe, the CE marking is a similar procedure. It is as a rule necessary that authorized technicians make a safety check of research instruments resulting in approval before the instrument is applied. Regarding instrument-related guidelines and standards, the manufacturer’s manual is a core document practically and legally. With or without reference to GLPs or GCPs, it is common that laboratories develop their own in-house standard operating procedure (SOP) or instruction.2 The Skin Research and Technology journal (Blackwell Publishing; blackwellpublishing.com), official journal of the societies active in the field, is fully indexed and the leading source of information about original academic publication on new non-invasive methods, their validation, and application. The active societies are the International Society for Bioengineering and the Skin (ISBS), the International Society for Skin Imaging (ISSI), and the International Society for Digital Imaging of Skin (ISDIS). These societies organize annual or biannual congresses in different parts of the world and work well together. The Standardization Group of the European Society for Contact Dermatitis in the journal Contact Dermatitis published a number of instrumental guidelines, i.e., on measurement of transepidermal water loss (TEWL),3 on laser Doppler measurement of blood flow,4 on measurement of skin color,5 on laser Doppler scanning of
FIGURE 2.2 Selection of monographs on non-invasive methods and the skin, including the first edition of this handbook, 1995, published by CRC Press, Boca Raton, FL. The method-specific monographs are referenced in the literature list (9–13).
cutaneous blood flow,5 and on assessment of irritant skin reactions elicited by the experimental standard irritant sodium lauryl sulfate (SLS).6 Authors of some of these standardization papers have contributed with chapters to this handbook of non-invasive methods. The European Group for Efficacy Measurements on Cosmetics and Other Topical Products (EEMCO) has produced a number of guidance papers and introductory reviews on use of non-invasive methods for efficacy documentation of cosmetics. These are found on the website under PubMed/EEMCO (see Chapter 3). An EEMCO guidance on standard scoring schemes for assessment of dry skin is not included in this index.8 Regarding evaluation of cosmetics, the European Cosmetic Toiletry and Perfumery Association, named COLIPA, published various recommendations of this organization on behalf of the industry. The COLIPA published a widely used standard on determination of sun protection factor (SPF) of sun blockers parallel to FDA and German industrial standards (DIN). The information landscape in the field of non-invasive methods and the skin thus includes very different sources of information, some very detailed, some broad and educational, some independent, and some partial. The method-specific books9–13 published by CRC Press (Figure 2.2), the present and updated second edition of The Handbook of Non-Invasive Methods, and the French handbook Physiologie de la peau et explorations fonctionelles cutanées,14 edited by Professor Pierre Agache in 2000, are all instruments for overview and introduction.
REFERENCES 1. Hill CE, Thompson BJ, Williams EN. A guide to conducting consensual qualitative research. The Counselling Psychologist 25:517–572, 1997.
How to Choose and Use Non-Invasive Methods
2. Serup J. Bioengineering and the skin: standardization. Clinics in Dermatology 13:293–297, 1995. 3. Pinnagoda J, Tupker RA, Agner T, Serup J. Guidelines for transepidermal water loss (TEWL) measurement. Contact Dermatitis 22:164–178, 1990. 4. Bircher A, de Boer EM, Agner T, Serup J. Guidelines for measurement of cutaneous blood flow by laser Doppler flowmetry. Contact Dermatitis 30:65–72, 1994. 5. Fullerton A, Fisher T, Lahti A, Wilhelm K-P, Takiwaki H, Serup J. Guidelines for measurement of skin colour and erythema. Contact Dermatitis 35:1–10, 1996. 6. Fullerton A, Stücker M, Wilhelm K-P, Wårdell K, Andersson C, Fischer T, Nilsson GE, Serup J. Guidelines for visualization of cutaneous blood flow by laser Doppler perfusion imaging. Contact Dermatitis 46:129–140, 2002. 7. Tupker RA, Willis C, Berardesca E, Lee CH, Fartasch M, Agner T, Serup J. Guidelines on sodium lauryl sulphate (SLS) exposure tests. Contact Dermatitis 37:78–81, 1997.
13
8. Serup J. EEMCO guidance for assessment of dry skin (xerosis) and ichthyosis. Skin Research Technology 1:109–114, 1995. 9. Berardesca E, Elsner P, Wilhelm K-P, Maibach HI, Eds. Bioengineering of the Skin: Methods and Instrumentation. CRC Press, Boca Raton, FL, 1995. 10. Berardesca E, Elsner P, Wilhelm K-P, Maibach HI, Eds. Bioengineering of the Skin: Cutaneous Blood Flow and Erythema. CRC Press, Boca Raton, FL, 1995. 11. Berardesca E, Elsner P, Wilhelm K-P, Maibach HI, Eds. Bioengineering of the Skin: Water and the Stratum Corneum. CRC Press, Boca Raton, FL, 1994. 12. Berardesca E, Elsner P, Wilhelm K-P, Maibach HI, Eds. Bioengineering of the Skin: Skin Surface Imaging and Analysis. CRC Press, Boca Raton, FL, 1997. 13. Berardesca E, Elsner P, Wilhelm K-P, Maibach HI, Eds. Bioengineering of the Skin: Skin Biomechanics. CRC Press, Boca Raton, FL, 2002. 14. Agache P, Ed. Physiologie de la peau et explorations fonctionelles cutanées. Editions Médicales Internationales, Cachan Cedex, France, 2000.
Practical Guide to Resources on the 3 AInternet for the Skin Researcher Elizabeth Grove Wickersheim and Gary Lee Grove cyberDERM, Inc., Broomall, Pennsylvania
CONTENTS 3.1 3.2 3.3 3.4 3.5 3.6 3.7
The Internet and Its History ....................................................................................................................................15 Using URLs to Search for Specific Web Pages......................................................................................................15 Using Selected Subject Directories to Search for Related Web Pages ..................................................................16 Online Literature Searches ......................................................................................................................................19 Online Tutorials and Educational Courses..............................................................................................................20 Search Engines ........................................................................................................................................................21 Closing Remarks......................................................................................................................................................24
3.1 THE INTERNET AND ITS HISTORY The Internet is an international network of computers that are linked up through various types of connections to exchange information. Through an Internet service provider, one can access the “Net” and a variety of related services, such as e-mail, web pages, online commerce, etc. Once you are connected, your computer can “talk” to any other computer anywhere in the world that is also connected. There are certain hardware and software requirements that need to be carefully considered, but these are beyond the scope of this chapter. For sure, a high-speed connection is desirable and a high-power processor will greatly facilitate dealing with images. The Internet is often thought to be of recent vintage, but it actually dates back to the Cold War in the 1960s, when the U.S. Department of Defense in response to Sputnik formed the Advanced Research Projects Agency (ARPA) as the world’s first decentralized computer network. In the next few years, other government agencies and various universities, such as UCLA, MIT, Stanford, and Harvard, flocked to join the Net. By the early 1970s the network had crossed the Atlantic to include English and Norwegian universities. The 1970s also saw the introduction of electronic mail, File Transfer Protocols, Telnet, and the Usenet newsgroup. The 1980s brought further improvements with one of the more important being Transmission Control Protocol/Internet Protocol (TCP/IP), which enables computers to talk the Net’s language. Thanks to intense media coverage from the 1990s onward, the Internet has become so well known that even
preschoolers can have an Internet savvy that will amaze their grandparents. The most successful part of the Internet is the World Wide Web (also known as WWW, W3, or simply the Web), a concept originally developed at CERN Laboratories in Geneva by particle physicists so they could share information throughout Europe. Over the years it has grown into a user-friendly point-and-click way of navigating through what now should be considered the world’s biggest library. The Web can be best defined as a distributed multimedia hypermedia system. Distributed refers to the various locations that one might find information among the computer systems around the world, multimedia means that the information can include text, graphics, sound clips, and video, while hypermedia means that it is possible to move to new information through highlighted phrases or images by clicking the mouse. There are many excellent resources that can be found on the Web. These include online journals, databases, textbooks, atlases, instrumental specifications, department and societal pages, etc. Since the content of the Web is constantly changing and expanding, this chapter will primarily deal with how to conduct a search of the Web rather than just presenting lists of useful web pages.
3.2 USING URLS TO SEARCH FOR SPECIFIC WEB PAGES What is a URL? A URL is a uniform resource locator, or in other words, the address for a specific web page. Think of it as a networked extension of the standard filename 15
16
Handbook of Non-Invasive Methods and the Skin, Second Edition
concept: not only can you point to a file in a directory, but now that file and that directory can exist on any machine on the network, can be served via any of several different methods, and might not even be something as simple as a file. To reach that page, you only need to carefully enter its URL into your browser’s location or address bar, which is usually located directly under the menu. When I begin searching for a specific organization’s website, I often start by trying to guess the central URL for that organization. With Netscape and Internet Explorer, leave off http://. Begin with the common WWW, add the name or acronym of the organization, and end with the appropriate domain ending. Some of the more common domain endings include: com for commercial edu for educational institutions org for other organizations gov for U.S. federal government mil for U.S. military net for Internet service providers and networks Sometimes this brings up the organization you want, other times it does not. For example, any skin researcher should want to visit the web page of the International Society for Bioengineering and the Skin. At the time this chapter was written, if you input www.ISBS.com into your browser’s location or address bar, you will find yourself on the web page for International Specialized Book Services. If you use the org domain ending or www.ISBS.org, you will find yourself on the web page for the International Society of Biomechanics in Sports. It is only by using www.I-S-B-S.org that you will be brought to the home page of the International Society for Bioengineering and the Skin (Figure 3.1). You can bookmark it and other frequently visited websites so that you will not have any problems in the future recalling their unique URLs. Another web page that any serious skin researcher should have bookmarked is that of the International Society for Skin Imaging (ISSI). This is another case in which just guessing the URL as www.ISSI.com does not work, as this is the home page of Integrated Silicon Solution, Inc. Instead you should use www.ISSI.de, where the domain ending refers to a national domain, which in this case is Deutschland, or Germany. This home page is shown in Figure 3.2. On both the ISBS and ISSI home pages you will find links not only to other pages within the site being hosted by that organization, but also to related pages at other sites strewn throughout the Web. The language of the Internet is Hypertext Markup Language (HTML), which allows the possibility of correlating a phrase or an image with a specific URL. By clicking on such links, one arrives at another site, which might link to even more related sites.
For example, if one goes to the home page for ISBS and clicks on JOURNAL within the left-side menu, you will be taken instantly to another page (Figure 3.3), which is still within that site. If one then clicks on the phrase Skin Research and Technology found in the text on the righthand side of the page, you will be transported to a page (Figure 3.4) that is outside the ISBS site and has the URL http://www.blackwellpublishing.com/journal.asp?ref= 0909-752X. Fortunately, you do not have to remember this URL, but only take advantage of the HTML textual or graphical links. If you are a member of either the ISBS or the ISSI, then you can access this journal online by logging on with your membership number and password. You may also check to see if this journal is available to y o u a s p a r t o f y o u r l o c a l l i b r a r y ’s o n l i n e subscriptions.
3.3 USING SELECTED SUBJECT DIRECTORIES TO SEARCH FOR RELATED WEB PAGES If the reader returns to the ISSI home page, we can introduce another simple search strategy, namely, selected subject directories. These are human-constructed lists that provide various links that are arranged and classified in a meaningful way. If one clicks on WWW LINKS on the lefthand side of the ISSI home page, you will find such a list of links that point to sites sorted by societies, institutions, events, journals dealing with skin imaging, medical news, and tools. Unless the webmaster is extremely zealous, some of these links may no longer work. When this occurs, try chopping off parts of the URL starting on the right-hand side and stopping at every slash (/). This essentially takes you back through the hierarchy of the site and may help you find related information at a new URL. Most webmasters appreciate learning about their “dead” links, and a good citizen of the World Wide Web should report them. As you travel over the Web you will invariably find many very useful sites that serve as gateways or portals because of their extensive selected subject directories or databases. In some cases a fee is charged, but fortunately, others are sponsored by pharmaceutical companies who offer free access to these databases. One of the better ones is www.DermNet.com (Figure 3.5), which is sponsored through an unrestricted educational grant from Dermik Laboratories. It is a great site to visit for several reasons. For one, it provides an extensive library of digital images of various skin diseases that is updated weekly and truly deserving of its moniker “The Dermatologists’ Image Resource.” It also contains an extensive LINKS page that points to most of the journals, organizations, medical departments, and educational materials that would be of interest to a dermatologist. Unfortunately, there is very
A Practical Guide to Resources on the Internet for the Skin Researcher
17
FIGURE 3.1 Home page of the International Society for Bioengineering and the Skin, with its URL being www.I-S-B-S.org.
FIGURE 3.2 Home page of the International Society for Skin Imaging, URL www.ISSI.de, as an example of a national domain, namely, Germany/Deutschland.
little about bioengineering methods for studying skin structure and function non-invasively, but this is a great place to start learning more about the pathological processes and underlying diseases that may be worth studying with our instrumental approach.
Another great site is www.Medscape.com. It requires users to register in order to access the site, but it does not involve any fees (Figure 3.6). You will encounter some advertisement blurbs as you navigate the site, but they are not at all intrusive.
18
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 3.3 Example of HTML link based on a printed phrase. Clicking on JOURNAL will transfer you to the web page for the Skin Research and Technology journal.
FIGURE 3.4 Home page of the Skin Research and Technology journal, which can be assessed online by subscribers with the proper password.
A Practical Guide to Resources on the Internet for the Skin Researcher
19
FIGURE 3.5 Home page of DermNet.com, which is an example of a selected subject directory that provides links to websites of interest to the clinical dermatologist.
Although it covers the full breadth of all the medical specialties (physicians, pharmacists, dentists, etc.), you can fine-tune it to deal more with skin topics by selecting Dermatology as your specialty home page (Figure 3.7). By doing so, you can easily browse recently posted content of interest to dermatologists, such as Food and Drug Administration (FDA) advisory meetings’ notes and conference coverage of all the major dermatological meetings, e.g., the AAD and the Society of Pediatric Dermatology, which are archived back to 2000. It also permits you to read selected articles from a limited number of journals, including the British Journal of Dermatology, Dermatology Online, SKINmed, and Wounds. Again, these articles are archived back for several years and easily accessed through a simple click of your mouse.
3.4 ONLINE LITERATURE SEARCHES Medscape.com also allows one to do a free literature search using Medline, which is the most widely used medical database. However, I find doing literature searches using PubMed, which is offered as a free service of the National Library of Medicine, often works better for me because it is a relational database that allows one to easily find other articles that are directly related to a specific abstract. It can be accessed at http://www.ncbi. nlm.nih.gov/entrez/query.fcgi (Figure 3.8). Time does not
permit a full discussion of how best to do these types of searches, but fortunately the reference librarians at the University of Florida Health Science Center Libraries have provided an excellent online tutorial on PubMed, which can be viewed at http://www.library.health.ufl.edu/ PubMed/PubMed2/ (Figure 3.9). It is extremely well done, and you will be able to easily learn how to use and navigate the many search features of PubMed. You can either go through the entire tutorial from top to bottom or choose from an index for specific topics. I have found their explanation of how to use Boolean operators, truncations, and phrase terms to be very useful for search engines in general. Also, PubMed has some excellent help pages of its own that can be found by clicking the Help and Tutorial links on the PubMed home page. Since the “hits” do differ, you should use both Medscape and PubMed to do a complete literature search. The returns can be both awesome and intimidating. For example, if you input “TEWL,” you will have more than 450 returns with PubMed, while the Medscape list is limited to 200 by design. With Medline, the list is in order of how relevant that paper is to the search criteria. With PubMed, they are displayed in reverse entrez date order: last in, first out. The entrez date is the date that a record was initially added to PubMed and should not be confused with the publication date, which is the date an article was published. Thus, it is not surprising that there is very little
20
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 3.6 Home page of Medscape.com, which allows one to do a free literature search using Medline.
overlap between the two lists, and this is why it is imperative that for a serious search you utilize both Medline and PubMeb.
3.5 ONLINE TUTORIALS AND EDUCATIONAL COURSES The University of Florida’s tutorial on PubMed is just one example of how one can easily learn over the Internet. Indeed, there is considerable interest in establishing “the worldwide classroom,” and many universities offer distance learning courses and remote tutoring via the Internet. So much so that in some cases, one can earn a degree without ever setting foot on campus. In addition to these curricular connections being sponsored by various colleges and universities, there are a number of commercially sponsored sites with high educational value. One of the best introductions to skin science is “The World of Skin Care” by Dr. John Gray, which c a n b e v i ew e d a t w w w. p g . c o m / s c i e n c e / s k i n care/Skin_tws_toc.htm. Although sponsored by P&G Skin Care Science Center, there are no commercial overtones.
Instead, what you will find is an excellent primer that provides a very clear-cut explanation of the more important aspects of skin structure and function that is 110 pages long. It is a must-read for anyone entering the field. Another very good treatise is the Medical Student Core Curriculum that appears as a graphical link on the home page for the American Academy of dermatology at www.AAD.org. This is a Web-based textbook for students interested in dermatology. It is more advanced than The World of Science and covers both basic science and clinical topics. Each chapter is up-to-date and written by experts who are thoroughly familiar with the topic. To the best of my knowledge, there are no educational courses or tutorials that deal with bioengineering and the skin directly. It is possible to locate many academic institutions, course descriptions, and faculty members with research interests in these areas, but no real educational materials per se. Results improve if we broaden our search to more basic subjects such as physics. One of the best examples is CUPOL, an acronym for the Clemson University Physics Online Laboratories, available at http:// phoenix.phys.clemson.edu/labs/cupol/. These labs include
A Practical Guide to Resources on the Internet for the Skin Researcher
21
FIGURE 3.7 Same Medscape site as in Figure 3.6, but now with Dermatology selected as the specialties home page.
six different university-level courses, which range from a survey of general physics to calculus-based waves, optics, and modern physics. All include tutorials, learning guides, and actual laboratory experiments that may be performed wherever the Internet is available. It is not designed to eliminate the traditional, hands-on laboratory experience, but rather complement it. A very interesting feature of these courses is the extensive use of video clips and input boxes, which makes you feel like you are actually having a real-time laboratory experience without really being in the lab. This site also has very well presented “Physics Lab Tutorials” under the heading of Suppliments (sic), which deal with oscilloscope basics, EXCEL spreadsheets, precision and accuracy, etc. It also has a neat stopwatch for timing the experiments and an interactive color chart for determining the value of resistors. Although there are many other examples that could be given, it is best not to do so. The Web is constantly changing and what was present at the time this was written may not exist now. Hopefully it has been upgraded or replaced by something even better. That means that it is extremely important you learn how to search the Web for the resources that you need.
3.6 SEARCH ENGINES We briefly introduced the notion of using search engines when we described doing literature searches using either Medscape or PubMed. To keep abreast of the ever-expanding Web, it is important that you learn more about search engines in general. Search engines attempt to find and index as many sites as possible that match certain search criteria. The capabilities of the various search engines vary greatly, as does the actual scope, size, and accuracy of the databases that they are exploring. Although search engines will include millions of pages in their databases, none of them come close to indexing the entire Web, much less the entire Internet. There are a number of search engines that have become very popular over the years, such as AltaVista, Ask Jeeves, MSN Search, and Yahoo. When asked to name a search engine, many mention Google. Not only was Google.com the first search engine to index more than a billion pages, but it also displays highly relevant search results faster than any other. Additionally, it has one of the largest databases, which includes many different file types, including Portable Document Format (PDF) and
22
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 3.8 Home page for PubMed website of the National Library of Medicine, which allows free literature searches of this database.
images. It is very easy to use, and I have found the relevancy of its hits to be excellent. To fully describe how a search engine uses special software robots called spiders to crawl through the Web and create indices to find all of the information on the hundreds of million web pages that exist is beyond the scope of this chapter. Again, the reader can find a very good description of search engines at http://computer.howstuffworks.com/search-engine1.htm. Indeed, the “howstuffworks” website is an excellent place to start to learn about how anything works. I especially like its coverage of electronic gadgets like digital cameras, cell phones, GPS, etc. It is an extremely good example of how well a selected subject directory can work and clearly should be bookmarked in anyone’s favorite list. Let us get back to search engines, especially Google.com (Figure 3.10). For most users, executing a search consists of simply typing in a number of descriptive terms and clicking a button. Google searches are not case
sensitive. All letters, regardless of how you enter them, are understood as lowercase. For example, searches for “Albert Kligman,” “albert kligman,” and “ALBERT KLIGMAN” all return the same results. Note that this example used a quoted string as a query, which forces an exact match of that phrase. If you just enter a series of search terms without any punctuation, you are asking Google to return all results that contained all of your search terms. There is no need to include the Boolean operator “AND,” as this is the default. Google also supports the Boolean operator “OR” to return pages that contain at least one of the search terms and the Boolean operator “NOT” (indicated by a minus sign before the term) to exclude pages that include that term. Google also has the capability to recognize certain numerical inputs as being unique search terms, with one of the more practical being the tracking number for a UPS or FedEx package. Just input the tracking code without any spaces and you
A Practical Guide to Resources on the Internet for the Skin Researcher
23
FIGURE 3.9 Online tutorial describing how best to use PubMed, prepared by the librarians at the University of Florida.
FIGURE 3.10 Basic search page for Google.com. Note that the language preference has been set to Elmer Fudd.
should be able to learn its current status in the respective system. It also recognizes and maps addresses and has a reverse phone number lookup function. In searching the Web, Google uses very sophisticated text-matching techniques to find pages that are both impor-
tant and relevant to your search in order of their Google PageRank. Google analyzes not only the candidate page, but also the pages linking into it to determine the value of the candidate page for your search. Google also prefers pages in which your query terms are coded as headings.
24
Handbook of Non-Invasive Methods and the Skin, Second Edition
The Google’s results page is full of information and includes links to all of the pages that match your query in order of their relevancy according to their PageRank algorithms. Depending upon the nature of your search terms, you may also have sponsored links, which are paid advertisements that will appear at the top and along the right-hand side of the results page. These commercial pages are separate from and do not influence the PageRank of the search results. Each listing provides one or more relevant excerpts (snippets) that show how your search terms are used in context on that page. In the excerpt, your search terms are displayed in bold text so that you can quickly determine if that result is from a page you want to visit. To view a page listed in your search results, click on the page title, the first line in each result. Any of your query words that appear in the title will be in boldface, and the title will be underlined, i.e., it is a link to the web page. Note that when you position your mouse pointer on the title, the URL for the web page will appear in your browser’s status bar, at the bottom of many browsers. To get the full benefit of Google, you must learn how to really use it and fully understand what it is really capable of doing. Such an in-depth treatment is well beyond the scope of this introductory chapter. Again, we can turn to the Web for help in learning these advanced techniques. Although there is considerable information and support within the Google site itself, I found that the “Google Guide” prepared by Nancy Blachman is well worth browsing through. The home page URL is www.googleguide.com and includes tutorials for both the novice and experienced user. It is neither affiliated nor endorsed by Goggle, but the underlying philosophy is clearly very positive about Google. It should also be mentioned because of the potential for Google to dominate the Web and privacy concerns over how it issues cookies and tracks individual searches, there are several watch guard groups, such as http://www.google-watch.org/, that maintain pages that place Google in a less favorable light. From the home page for www.Google.com (Figure 3.10), one can click on About Google to access the official Help and How to Search guide for Google. From the home page you can also bring up the Advanced Search menu (Figure 3.11). This form allows one to specify required words and exact phrases to include or exclude in the query. You can also specify the language of the web pages to return, file formats, and web page domains. In addition, you can specify a date range for when the web pages were updated and ranges of numbers that appeared in the web pages. You can even specify where in the web pages the search terms must appear, such as title, URL, body text, etc. This allows you to really fine-tune your search with very little effort. One of the more fascinating aspects of Google can be found under Language Tools. As mentioned before, it is
possible to limit your search to web pages written only in a specific language or geographic area. Once these pages are found, it is possible to translate those written in either French, German, Italian, Portuguese, or Spanish into English through the Google language tool. It is also possible to set the Google home page, messages, and buttons to read in your preferred native language. The language choices are both amazing and amusing. Currently, there are more than 100, which range from Afrikaans to Zulu and include all the major languages and several nontraditional ones as well. Indeed, if you are a Trekkie and want to use Klingon as your language of preference, it can easily be done. Same goes for Pig Latin or Elmer Fudd. Indeed, if you do not see your native language, you can help the Google translators create it. The Language Tools page also contains a very nice display of the national flags of all the local domains that contain a Google site. Although I occasionally amuse myself by seeing how the I’m Feeling Lucky button on Google’s home page appears in different languages, I do not use it very often. This button, instead of showing you a list of web pages, sends you immediately to the one result that may be most relevant to your query that is not a paid advertisement. Although the I’m Feeling Lucky button can save you some time, it is really a matter of only a few seconds with a high-speed connection, and I much prefer a broader search list with its snippets to select from.
3.7 CLOSING REMARKS A few years ago, just a passing knowledge of the Internet and the World Wide Web was enough for you to be considered a guru and land you a very nice IT job. Now in many fields it has become mandatory that you know how to surf the Web, and Google has become a verb familiar to many laypersons. Indeed, one of the problems confronting today’s physicians is that their patients can easily obtain information on their disease that the physician may not be aware of. Unfortunately, there are no authorities that guarantee the truthfulness of the information that can be found on the Net. Anyone with a computer and a connection can “publish” material without editorial oversight or peer review. Although as shown by the examples cited in this chapter there are many superbly written, wellillustrated, and highly informative resources, one can just as easily encounter pages with incorrectly labeled images and patently false statements. Often these sites are touting the benefits of some dubious skin care treatment. We will not be able to free the Internet of such commercial influences, but should strive to ensure that we are providing accurate and interesting information in the web pages that we do have control over. It is the responsibility of the user to evaluate information gathered from the Internet for its accuracy and worth.
A Practical Guide to Resources on the Internet for the Skin Researcher
25
FIGURE 3.11 Advanced search page for Google.com.
Some basic questions that need to be critically considered include but certainly are not limited to:
Currency: When was the information last updated? Are the links up-to-date?
Authorship: Who is the author? Is he or she a recognized authority? What experience does he or she have? What are his or her credentials? Affiliations: Is this a personal web page or is the material posted sanctioned by an organization or institution? Does that organization have a basis or a commercial conflict of interest? Content: Are there errors in spelling and grammar? Is the presentation logical? Are the facts and quotations properly cited and referenced in a bibliography?
In conclusion, we would like to emphasize that the Internet offers a wealth of information, much of which is neither monitored nor edited. Searching the Internet offers both significant challenges and amazing benefits. If you are just starting to surf the Internet, we feel that the best advice that we can give you is to just jump in and get started. The sooner you jump in, the sooner you will benefit from the many resources and learning opportunities offered through the Internet. We also hope that by sharing our basic strategies and philosophies, we have been able to enhance your Internet experiences.
Skin Integument: Variation 4 The Relative to Sex, Age, Race, and Body Region Nadia Farinelli Department of Dermatology, University of Pavia, Pavia, Italy
Enzo Berardesca Department of Dermatology, Instituto Dermatologico di S Maria e Gallicano, Rome, Italy
CONTENTS 4.1 4.2 4.3 4.4
Introduction..............................................................................................................................................................27 Sex............................................................................................................................................................................27 Age and Body Region .............................................................................................................................................27 Race..........................................................................................................................................................................29 4.4.1 Physiological Differences ............................................................................................................................29 4.4.2 Percutaneous Absorption .............................................................................................................................29 4.4.3 Biomechanical Properties, TEWL, and Susceptibility to Irritants .............................................................29 References .........................................................................................................................................................................30
4.1 INTRODUCTION The skin is not a uniform sheath covering the body, but a specialized organ with several functions changing from site to site. These regional variations are of great importance because they can influence skin behavior and thus susceptibility to disease. The major anatomical differences related to site involve stratum corneum thickness, distribution of appendages and melanocytes, variation in the structure of the dermoepidermal junction and of the dermis, and changes in blood supply.1 Anatomical changes often induce functional changes that can be quantified with combined non-invasive techniques that allow the assessment of skin function relative to sex, age, and race.2
4.2 SEX According to the HANES survey,3 skin pathology is consistently more prevalent among males than females. Most of this higher prevalence among males is accounted for by the higher prevalence of dermatophytoses and skin tumors. This difference, however, is felt to be related to differences in behavior (hygiene and occupation) and not in skin characteristics.
There is evidence of greater skin irritability in females than males, although this difference is not reported to be large.4,5 The skin irritability to sodium lauryl sulfate in males and females has been studied by measurement of skin water vapor loss:6 the study showed that mean values of unirritated skin in females were significantly lower than those in male volunteers. The differences seem due to a lower basal metabolic rate in females. There was no significant difference between the mean values of irritated skin of male and female volunteers, and the irritation index was significantly lower in males than females. The study concluded that female skin is more prone to irritation. In spite of this, some other reports confirm that there is no difference in reactivity between the sexes.5–7 However, variations in skin reactivity during the menstrual cycle seem to occur: changes in skin extensibility and increased proneness to develop strong irritant reactions are reported during the menstrual phase.8,9
4.3 AGE AND BODY REGION Skin aging, more or less a physiological event, is characterized by several biological and histopathological 27
28
Handbook of Non-Invasive Methods and the Skin, Second Edition
changes. Transepidermal water loss (TEWL) and skin hydration both decrease during the aging process, maintaining a directly proportional relationship. The decrease of TEWL during life is conspicuous after the age of 60.10 Several factors may be responsible. The increased size of corneocytes and the increased thickness of the stratum corneum due to the greater accumulation of corneocytes related to an impaired desquamation11 are factors that should be considered. Similarly, corneum hydration is decreased in elderly subjects.12,13 Reduction in moisture content is more noticeable in exposed areas, where damage is the predominant factor accentuating aging.13 The simultaneous decrease of TEWL and water content of the corneum is a distinct feature of elderly skin. It confirms the decreased corneum hydration without impairment of the barrier function. Accordingly, dry skin in the elderly may be differentiated from pathologically dry skin since in the latter the barrier function is defective. One of the sites where xerotic changes occur more frequently in aged skin is the extensor aspects of the lower legs. Tagami and coworkers,14 using an evaporimeter to measure TEWL and a skin surface hygrometer15 to evaluate the hydration state of the skin, compared hydration of the lower legs in young and aged subjects. They found that skin of aged subjects is not particularly dry, compared with that of young adults. A lowered TEWL in elderly subjects was confirmed by other groups.10 A comparative study between aged individuals and children further substantiated that skin surface in aged people is not necessarily dehydrated, indicating that aging of the skin itself does not induce any marked derangement in stratum corneum function. On the other hand, stratum corneum moisturization detected by an impedance technique13 revealed differences between chronically sun-exposed skin and unexposed skin in the same individuals. Electrical skin impedance varies greatly in the different sites investigated in relation to age. Levels in elderly subjects were higher than those in young controls in both exposed and unexposed sites, except for the palms. Statistically significant differences were recorded only in chronically exposed areas, such as forehead and neck. The aging process varies greatly from site to site and from individual to individual, explaining the controversial data of the literature. Furthermore, senile xerosis represents a special pathologic condition affecting only certain subjects. Stratum corneum obtained from patients with senile dry skin showed a reduced capacity for secondary bound water, which plays an important role in maintaining corneum flexibility and suppleness.14 The skin of children is more easily irritated than that of adults.4 The skin of elderly persons is also more reactive than that of younger adults. This is attributed to the dry skin factor, which ensues with age.5
Many differences between a senile and young epidermis have been described, but a consistent interpretation of the aging process has proved difficult. In part, this is because the epidermis varies from site to site.16 Young skin from the back,17 like that from the scalp and axilla,18 has deep and complex rete ridges, whereas that of the face18 has a fairly flat dermoepidermal junction. It is widely agreed that in areas where the junction is corrugated in youth, it becomes flattened by age.18–23 Similarly, there are differences in epidermal thickness even in young skin. On the face or on the dorsum of the hand, for example, it is considerably greater than on the arms, legs, or trunk.18 In many areas the whole epidermis becomes thinner with age and the cells become less evenly aligned on the basement membrane and less regular in size, shape, and staining properties.18,22–25 The thickness of the stratum corneum is obviously not uniform over the whole body surface. In particular, the differentiation of the stratum corneum of the palms and soles is unlike that of the rest of the skin.26 Hammarlund et al.27 reported regional variations in newborns in the rate of TEWL. Comparing the abdomen, forearm, and buttocks, they found twofold TEWL on the buttocks. The findings were partially confirmed by Osmark et al.28 using a Meeco analyzer to detect TEWL. The technique allowed more precise measurements, not biased by air turbulence and ambient relative humidity over the probe. They recorded similar TEWLs on these three sites, but found a lower hydration level (obtained by inducing skin occlusion with plastic film for 1 hour) on the abdomen, reporting a decreased water-holding capacity on this site. In adults too, regional variations of moisturization reflecting differences of thickness and function of the corneum occur. Tagami et al.29 reported drier skin on the extremities than on the trunk. This correlates with the fact that clinically dry skin tends to develop more frequently on the limbs during winter. The reduction in TEWL with an increased thickness of the stratum corneum is not as much as would be expected.5 In fact, in some regions, especially on the palms and soles, TEWL increases with the thickness of the stratum corneum. The increase in thickness of the stratum corneum seems to be compensated by a corresponding increase in its diffusivity,5 resulting in skin with a relatively uniform steady-state TEWL over many parts of the body.26 However, not all differences in thickness are compensated in this manner, and some variations in the TEWL with the regional skin sites do exist. This regional variation in TEWL and permeability cannot be explained on the basis of differences in the chemical nature of the keratin molecule.30 The regional variations in the total lipid concentrations of the stratum corneum, however, may be the most critical factor determining the regional variations in TEWL and permeability.31
The Skin Integument: Variation Relative to Sex, Age, Race, and Body Region
Thus, barrier efficiency is not uniform over the whole body surface.32 The scrotum has long been known to be particularly permeable.33 The face, forehead, and dorsa of the hands may also be more permeable to water than the trunk, arms, and legs. The palms are practically impermeable, except to water and most water-soluble molecules. This may be the major reason why contact dermatitis is seen less often on the palms than on the dorsa of the hands.32
4.4 RACE 4.4.1 PHYSIOLOGICAL DIFFERENCES Although stratum corneum thickness is equal in blacks and whites,34,35 more tape strippings are required to remove the stratum corneum in blacks. Weigand and associates36 reported this is due to an increased number of corneocyte layers in black skin. Moreover, the same study revealed a great interindividual variability in the black race, whereas data from white subjects were more homogeneous. A correlation between the number of stratum corneum layers and the degree of pigmentation has not yet been demonstrated. The increased number of cell layers in the stratum corneum and the increased resistance to stripping could be related to lipid molecules in the intercellular matrix that increase cell cohesion. Indeed, Rienertson and Wheatley,37 investigating the stratum corneum lipid content, found higher values in blacks. Weigand and coworkers36 confirmed this result. Other parameters investigated in different studies, such as skin electrical resistance,38 were consistent with these findings.39
4.4.2 PERCUTANEOUS ABSORPTION In vitro penetration of fluocinolone acetonide through skin samples obtained from amputated black and white legs revealed an increased permeability in whites.40 In vitro water permeation through human skin did not reveal the racial differences that had been reported by Bronaugh and coworkers.41 In vivo studies show different patterns of penetration depending on the molecules tested.42 Tritiated diflorasone diacetate does not have different pharmacokinetics in blacks and whites.43 Wedig and Maibach44 applied Clabeled dipyrithione in different vehicles to stripped and unstripped skin of black and white volunteers and found 34% less absorption in blacks. A significantly lower penetration in blacks (47%) was also noted when a cosmetic vehicle (1:12:22:25:39 sodium lauryl sulfate:propylene glycol:stearyl alcohol:white petrolatum:distilled water) was compared with methyl alcohol on the forehead and when the methyl alcohol vehicle was compared with the shampoo vehicle on the scalp. The penetration of intact
29
vs. stripped skin by either the cosmetic cream or the shampoo vehicle was not different. Racial differences in methylnicotinate-induced vasodilatation in human skin were studied by Guy and associates; they induced vasodilatation by applying the substance to the skin and monitored the response with laser Doppler velocimetry,45 reporting statistically indistinguishable differences among the groups in the time-topeak response, the area under the response–time curve, and the time from the response to 75% decay. Only the magnitude of the peak response revealed some significant differences, with increased levels in young white subjects. However, no important differences seem to exist between black and white skin when tested with this chemical model.46
4.4.3 BIOMECHANICAL PROPERTIES, TEWL, SUSCEPTIBILITY TO IRRITANTS
AND
These parameters have been measured in whites, Hispanics, and blacks to assess whether the melanin content could induce changes in skin biophysical properties.47 Differences appear in skin conductance but are more marked in biomechanical parameters: skin extensibility, skin elastic modulus, and skin recovery. These relative variations of the parameters on dorsal and ventral sites are different according to the races and highlight the influence of solar irradiation on skin and the role of melanin in maintaining it unaltered. Skin lipids may play a role in modulating the relationship between stratum corneum water content and TEWL, resulting in higher conductance values in blacks and Hispanics. Previously, equal baseline TEWL on the back was reported48,49 among whites, blacks, and Hispanics. Moreover, TEWL revealed a different pattern of reaction in whites after chemical exposure to sodium lauryl sulfate, with blacks and Hispanics developing stronger irritant reactions after exposure to 2% sodium lauryl sulfate. Skin color and race are an influential factor determining skin reactivity; black skin (Negroid) is the least susceptible to irritants.50 Darkly pigmented individuals from the Mediterranean region are also less susceptible than light-complexioned individuals. It is thus reasonable to assume that fair-skinned persons of Celtic origin (Scottish–Irish– Welsh) have the most susceptible skin.51 All races show significant differences between the volar and dorsal forearms.47 These results are in apparent contrast with TEWL recordings. Indeed, the higher the stratum corneum water content, the higher the TEWL that may be expected.52 The data may be explained on the basis of different intercellular cohesion, lipid composition, or hair distribution. A greater cell cohesion with a normal TEWL could result in increased skin water content. The racial variability should be taken into account in terms of different skin responses to topical and environmental
30
Handbook of Non-Invasive Methods and the Skin, Second Edition
agents. Race provides a useful tool to investigate and compare the effects of lifetime sun exposure. It is clearly evident that melanin protection prevents sun damage; differences between sun-exposed and sun-protected areas are not detectable in races with dark skin.
REFERENCES 1. Ebling, F.J.G., Eady, R.A.J., and Leigh, I.M., Anatomy and organization of human skin, in Textbook of Dermatology, Rook, A.J., Wilkinson, D.S., and Ebling, F.J.G., Eds., Blackwell Scientific Publications, Oxford, 1992, p. 49. 2. Leveque, J.L., Ed., Cutaneous Investigation in Health and Disease, Noninvasive Methods and Instrumentation, Marcel Dekker, New York, 1989. 3. Stern, R.S., The epidemiology of cutaneous disease, in Dermatology in General Medicine, Vol. 1, 3rd ed., Fitzpatrick, T.B., Eisen, A.Z., Wolff, K., Freedberg, I.M., and Austen, K.F., Eds., McGraw-Hill, New York, 1987, p. 7. 4. Kligman, A.M. and Wooding, W.M., A method for the measurement and evaluation of irritants on human skin, J. Invest. Dermatol., 49, 78, 1967. 5. Pinnagoda, J., Transepidermal Water Loss: Its Role in the Assessment of Susceptibility to the Development of Irritant Contact Dermatitis, Ph.D. thesis, London University, July 1990. 6. Goh, C.L. and Chia, S.E., Skin irritability to sodium lauryl sulphate — as measured by skin water vapour loss — by sex and race, Clin. Exp. Dermatol., 13, 16, 1988. 7. Coenraads, P.J., Lee, J., and Pinnagoda, J., Changes in water vapour loss from the skin of metal industry workers monitored during exposure to oils, Scand. J. Work Environ. Health, 12, 494, 1986. 8. Berardesca, E., Gabba, P., Farinelli, N., Borroni, G., and Rabbiosi, G., Skin extensibility time in women. Changes in relation to sex hormones, Acta Derm. Venereol., 69, 431, 1989. 9. Agner, T., Damm, P., and Skouby, S., Menstrual cycle and skin reactivity, J. Am. Acad. Dermatol., 24, 566, 1991. 10. Leveque, J.L., Corcuff, P., de Rigal, J., and Agache, P., In vivo studies of the evolution of physical properties of the human skin with age, Int. J. Dermatol., 23, 322, 1984. 11. Nicholls, S., King, C.S., and Marks, R., The Influence of Corneocytes Area on Stratum Corneum Function (abstract), paper presented at the ESDR Annual Meeting, Amsterdam, 1980. 12. Berardesca, E. and Maibach, H.I., Bioengineering and the patch test, Contact Derm., 18, 3, 1988. 13. Borroni, G., Berardesca, E., Bellosta, M., Bernardi, L., and Rabbiosi, G., Evidence for regional variations in water content of the stratum corneum in senile skin: an electrophysiologic assessment, Ital. Gen. Rev. Derm., 19, 91, 1982.
14. Tagami, H., in Cutaneous Aging, Kligman, A.M. and Takase, Eds., University of Tokyo Press, Tokyo, 1988, p. 99. 15. Tagami, H., Kanamura, Y., Inoue, K., Suehisa, S., Inoue, F., Iwatsuki, K., Yoshikuni, K., and Yamada, M., Water sorption-desorption test of the skin in vivo for functional assessment of the stratum corneum, J. Invest. Dermatol., 78, 425, 1982. 16. Graham-Brown, R.A.C. and Ebling, F.J.G., The ages of man and their dermatoses, in Textbook of Dermatology, Vol. 4, 5th ed., Rook, A.J., Wilkinson, D.S., and Ebling, F.J.G., Eds., Blackwell Scientific, Oxford, 1992, p. 2877. 17. Eller, J.J. and Eller, W.D., Oestrogenic ointments: cutaneous effects of topical application of natural oestrogens with report of three hundred and twenty-one biopsies, Arch. Dermatol. Syphilol., 59, 449, 1949. 18. Montagna, W., Morphology of the aging skin: the cutaneous appendages, in Advances in Biology of Skin, Vol. 6, Aging, Montagna, W., Ed., Pergamon Press, Oxford, 1965, p. 1. 19. Christophers, E. and Kligman, A.M., Percutaneous absorption in aged skin, in Advances in Biology of Skin, Vol. 6, Aging, Montagna, W., Ed., Pergamon Press, Oxford, 1965, p. 163. 20. Hill, W.R. and Montgomery, H., Regional changes and changes caused by age in the normal skin, J. Invest. Dermatol., 3, 321, 1940. 21. Lavker, R.M., Zheng, P., and Dong, G., Morphology of aged skin, Dermatol. Clin., 4, 379, 1986. 22. Montagna, W. and Carlisle, K., Structural changes in ageing skin, J. Invest. Dermatol., 73, 47, 1979. 23. Montagna, W. and Carlisle, K., Structural changes in ageing skin, Br. J. Dermatol., 122 (Suppl. 35), 61, 1990. 24. Gilchrest, B.A., Skin and Ageing Processes, CRC Press, Boca Raton, FL, 1984. 25. Lavker, R.M., Structural alterations in exposed and unexposed aged skin, J. Invest. Dermatol., 73, 59, 1979. 26. Scheuplein, R.J. and Blank, I.H., Permeability of the skin, Physiol. Rev., 51, 702, 1971. 27. Hammarlund, K., Nilsson, G., Oberg, A., and Sedin, G., Transepidermal water loss in newborn infants: relation to ambient humidity and site of measurement and estimation of total transepidermal water loss, Acta Paediatr. Scand., 68, 371, 1979. 28. Osmark, K., Wilson, D., and Maibach, H.I., In vivo transepidermal water loss and epidermal occlusive hydration in newborn infants: anatomical regional variations, Acta Derm. Venereol., 60, 403, 1980. 29. Tagami, H., Masatoshi, O., Iwatsuki, K., Kanamaru, Y., Yamada, M., and Ichijo, B., Evaluation of skin surface hydration in vivo by electrical measurement, J. Invest. Dermatol., 75, 500, 1980. 30. Tregear, R.T., The structures which limit the penetrability of the skin, J. Soc. Cosmet. Chem., 13, 145, 1962. 31. Elias, P.M., Cooper, E.R., Core, A., and Brown, B.E., Percutaneous transport in relation to stratum corneum structure and lipid composition, J. Invest. Dermatol., 76, 297, 1981.
The Skin Integument: Variation Relative to Sex, Age, Race, and Body Region
32. Baker, H., The skin as a barrier, in Textbook of Dermatology, Rook, A., Ed., Blackwell Scientific, Oxford, 1986, p. 355. 33. Smith, J.G., Jr., Fisher, R.W., and Blank, I.H., The epidermal barrier: a comparison between scrotal and abdominal skin, J. Invest. Dermatol., 36, 337, 1961. 34. Freeman, R.G., Cockerell, E.G., Armstrong, J., and Knox, J.M., Sunlight as a factor influencing the thickness of the epidermis, J. Invest. Dermatol., 39, 295, 1962. 35. Thomson, M.L., Relative efficiency of pigment and horny layer thickness in protecting the skin of Europeans and Africans against solar ultraviolet radiation, J. Physiol. (London), 127, 236, 1955. 36. Weigand, D.A., Haygood, C., and Gaylor, J.R., Cell layers and density of Negro and Caucasian stratum corneum, J. Invest. Dermatol., 62, 563, 1974. 37. Rienertson, R.P. and Wheatley, V.R., Studies on the chemical composition of human epidermal lipids, J. Invest. Dermatol., 32, 49, 1959. 38. Johnson, L.C. and Corah, N.L., Racial differences in skin resistance, Science, 139, 766, 1963. 39. Berardesca, E. and Maibach, H.I., Skin color and proclivity to irritation, in Exogenous Dermatoses, Menne, T. and Maibach, H., Eds., CRC Press, Boca Raton, FL 1990, p. 65. 40. Stoughton, R.B., Bioassay methods for measuring percutaneous absorption, in Pharmacology of the Skin, Montagna, W., Stoughton, R.B., and Van Scott, E.J., Eds., Appleton-Century-Crofts, New York, 1969, p. 542. 41. Bronaugh, R.L., Stewart, F.R., and Simon, M., Methods for in vitro percutaneous absorption studies. VII. Use of excised human skin, J. Pharm. Sci., 75, 1094, 1986. 42. Berardesca, E. and Maibach, H.I., Physical anthropology and skin: a model for exploring skin function, in Models in Dermatology 4, Maibach, H.I. and Lowe N., Eds., Karger, Basel, 1989, p. 202.
31
43. Wickema-Sinha, W.J., Shaw, S.R., and Weber, O.J., Percutaneous absorption and excretion of tritium-labelled diflorasone diacetate, a new topical corticosteroid in the rat, monkey and man, J. Invest. Dermatol., 7, 373, 1978. 44. Wedig, J.H. and Maibach, H.I., Percutaneous penetration of dipyrithione in man: effect of skin color (race), Am. Acad. Dermatol., 5, 433, 1981. 45. Guy, R.H., Tur, E., and Bierke, S., Are there age and racial differences to methylnicotinate-induced vasodilatation in human skin?, J. Am. Acad. Dermatol., 12, 1001, 1985. 46. Berardesca, E. and Maibach, H.I., Contact dermatitis in blacks, Dermatol. Clin., 6, 363, 1988. 47. Berardesca, E., de Rigal, J., Leveque, J.L., and Maibach, H.I., In vivo biophysical characterization of skin physiological differences in races, Dermatologica, 182, 89, 1991. 48. Berardesca, E. and Maibach, H.I., Racial differences in sodium lauryl sulphate induced cutaneous irritation: black and white, Contact Derm., 18, 65, 1988. 49. Berardesca, E. and Maibach, H.I., Sodium lauryl sulphate induced cutaneous irritation: comparison of white and Hispanic subjects, Contact Derm., 19, 136, 1988. 50. Kligman, A.M., Assessment of mild irritants, in Principles of Cosmetics for the Dermatologist, Frost, P. and Horwitz, S.N., Eds., C.V. Mosby, St. Louis, MO, 1982, p. 265. 51. Frosch, P.J. and Kligman, A.M., Recognition of a chemically vulnerable and delicate skin, in Principles of Cosmetics for the Dermatologist, Frost, P. and Horwitz, S.N., Eds., C.V. Mosby, St. Louis, MO, 1982, p. 287. 52. Rietschel, R.L., A method to evaluate skin moisturizers in vivo, J. Invest. Dermatol., 70, 152, 1978.
Variations and 5 Seasonal Environmental Influences on the Skin Chee Leok Goh National Skin Centre, Singapore
CONTENTS 5.1 5.2
Introduction..............................................................................................................................................................33 Effect of Seasonal Variation on Skin Functions .....................................................................................................33 5.2.1 Solar Radiation ............................................................................................................................................33 5.2.1.1 Immediate Effects of Solar Radiation .........................................................................................34 5.2.1.2 Long-Term Effects of Solar Radiation ........................................................................................34 5.2.2 Effect of Temperature on the Skin ..............................................................................................................34 5.2.3 Effect of Humidity on the Skin...................................................................................................................35 5.3 Other Environmental Factors Influencing the Skin ................................................................................................35 References .........................................................................................................................................................................35
5.1 INTRODUCTION The skin is subject to the influence of solar radiation, temperature, humidity, domestic contactants, occupational contactants, therapeutic agents, and a host of environmental agents. All of these environmental agents may have adverse effects on the skin. The characteristic structure of the human skin places it as an important interface between human beings and their physical, chemical, and biological environment.1 It is a primary organ of defense and adaptation. The skin is the largest organ of the human body. It is also the largest organ that is exposed to all elements of the external environment. It is a vulnerable target of environmental agents.2 As a target organ, the skin is capable of responding in a variety of pathologic patterns.1 It is also an important portal of entry for potentially hazardous agents. The cell structure and appendages of the skin provide it with defenses against environmental elements. It has protective elements against physical trauma, thermal stress, solar radiation, chemical agents, fluid loss, and antimicrobial agents. The stratum corneum provides a barrier against the various physical agents and biological agents, the sweat gland activities against thermal stress, and the pigmentary system, including the melanocytes and melanin, against ultraviolet radiation. This chapter discusses the effect of seasonal variation and environmental influences on the skin.
5.2 EFFECT OF SEASONAL VARIATION ON SKIN FUNCTIONS The climatic and physical environmental conditions in different latitudes of the world differ. In temperate latitudes, seasonal variations occur during different periods of the year. Such climatic differences have an influence on the integrity and functional capacity of the skin. Certain skin disorders are more prevalent in different countries because of climatic differences, and similarly, certain skin disorders tend to occur during different seasons of the year. The following environmental factors, which change with different seasons and which have an effect on the skin, will be discussed: 1. Solar radiation 2. Temperature 3. Humidity
5.2.1 SOLAR RADIATION Solar radiation, in particular ultraviolet light (UVL), has immediate and long-term effects on the skin. The flux of solar radiation varies during different seasons in temperate countries. Skin pigment provides protection against actinic or ultraviolet radiation, which causes sunburn (ultraviolet erythema — immediate effect) and fragmentation and destruction of elastic tissue fibers, UV-induced 33
34
Handbook of Non-Invasive Methods and the Skin, Second Edition
aging of the skin, actinic keratoses, skin cancer, and alteration of the immune function of the skin (long-term effects). 5.2.1.1 Immediate Effects of Solar Radiation Sunburn from UVL can be elicited in all human beings, but photosensitivity is inversely related to the degree of melanin pigmentation. UVL exposure results in the immediate and delayed dilation of blood vessels in the dermis, which is usually confined to the irradiated sites.3 UVB is the major cause of sunburn from sunlight. Cutaneous reactions from UVB are influenced by environmental conditions, season of the year, latitude and time of day, altitude, atmospheric pollution, and time of exposure and skin thickness and pigmentation, as well as other factors.4 UVB erythema becomes visible within 2 to 6 hours following irradiation, reaches a maximum at 24 to 36 hours, fades in 72 to 120 hours, and is followed in most individuals by increased skin pigmentation (tanning).4 Increased pigmentation provides protection against further damage from UVB, since the pigment is a remarkably effective absorber of UVB. Although UVA is 1000-fold less potent than UVB in causing erythema, its predominance in the solar spectrum reaching the Earth’s surface (10- to 100-fold more than UVB) may account for its toxic effect on the skin. Highintensity UVA light sources, which may emit as much as five times more UVA than sunlight, widely used for cosmetic tanning, have effects on the skin and contribute to the long-term effects of UVL on the skin. The effect of UVB and UVA on the skin has been studied in detail by Gilchrest et al.5,6 5.2.1.2 Long-Term Effects of Solar Radiation UVA, in addition to UVB, is believed to contribute to the long-term effects of chronic sun exposure, including premature skin aging, actinic keratosis, and skin carcinogenesis.7 Solar radiation is the principal cause of skin cancer in humans. The most important wavelength responsible is UVB (290 to 320 nm). Recent studies have documented that the environmental flux of UVL radiation from the sun is increasing, especially over the North and South Poles.16–18 This is contributed by the liberation into the atmosphere of tons of chlorofluorocarbons by human activities, which eventually removes the protective ozone shield in the stratosphere.19,20 UVB irradiation has been shown to induce immunosuppression.21 Irradiation with high-dose UVB results in systemic immunosuppression, while exposure to low-dose UVB produces local immunosuppression.22,23 There is substantial evidence to implicate the effect of UVL on the epidermal Langerhans cells as the cause of the change in immunosuppression.21 The effect of UVL on Langerhans
cells studies has also been demonstrated by studies that found lower Langerhans cells density in the non-sunexposed skin than in chronically sun-exposed skin.24 The effect of seasonal variation in UVL flux, in particular UVB, may have an influence on the immunological response of the skin to contact allergens. The afferent and efferent limbs of allergic contact dermatitis in experimental animals may be suppressed by irradiation with UVB.25 Bruze26 found fewer positive patch tests per tested patient in Sweden during the summer months of June, July, and August than the other months. Similarly, Veien et al.27 in Denmark also found significantly lower patch test reactivity during the same period when compared to other months. Epidemiological evidence has also shown the association of UVL exposure, which differs in different latitudes and different seasons and skin cancers. Studies have identified that in the white populations, there was an inverse relationship between latitude and the incidence of skin cancers.28 Skin cancers showed a rising incidence with increasing dose of UVL exposure at different latitudes in North America.29 A linear relationship was also observed on the incidence of no-melanoma skin cancers in countries of different proximity to the equator. A linear relationship was observed between the incidence of nonmelanoma skin cancers and the annual ultraviolet solar radiation.30
5.2.2 EFFECT
OF
TEMPERATURE
ON THE
SKIN
The skin is an important thermoregulation organ. The rate of blood flow and sweating controls body temperature. The body temperature is maintained at a very constant temperature with minimal variations. This is vital to the function of the various body organs. One of the body responses to an increase in environmental temperature is increased rate of blood flow through dilatation of dermal capillaries and stimulation of the sweat glands to increase secretion of sweat. Sweating allows the evaporation process to occur, leading to loss of skin surface heat. Increased sweating is associated with increased hydration of the stratum corneum. An increased hydration of stratum corneum will enhance the penetration of chemical agents on the skin. Excessive sweating has a clinical impact on contact dermatitis. Olumide et al.8 reported a high incidence in Nigerian workers of contact allergy to clothing dyes from work uniforms, caused by enhanced dissolution of dyes from clothing in a hot environment. Increased sweating secondary to high environmental temperature also tends to provoke workers to discard protective clothing, and therefore expose workers’ skin to irritants and allergens.9 Excessive sweating in intertriginous areas leads to skin maceration and dermatitis. It predisposes the moist skin to colonization of fungus and superficial fungal infections.9
Seasonal Variations and Environmental Influences on the Skin
Excessive sweating due to high ambient temperature may lead to sweat duct swelling, resulting in obstruction, leading to miliaria. If severe, heat exhaustion and heat stroke may occur.
5.2.3 EFFECT
OF
HUMIDITY
ON THE
SKIN
The effect of high ambient humidity on the skin is similar to that of high temperature. High ambient humidity prevents the skin surface sweat from evaporating and leads to an increased hydration of the stratum corneum. Low humidity has been reported to cause skin symptoms. Rycroft et al.10–12 described a phenomenon of “lowhumidity occupational dermatoses” that affected office workers. Affected workers presented with itch and urticaria on covered parts of the body and scaly eczema on the face, scalp, and ears. The cause was believed to be due to low relative humidity in the work environment. The symptoms in these workers improved when the relative humidity of the work environment was raised above 45%. Gaul and Underwood13 reported skin chapping on the hands, lips, and face and ichthyosis on lower legs and arms in subjects who were exposed to low environmental humidity. Similar findings were also reported by Chernosky14 in patients living in air-conditioned houses when the environmental humidity was lowered drastically. However, the effect of low humidity on skin symptoms has been disputed. Andersen et al.15 could not evoke similar symptoms in subjects exposed to 78 hours in dry air. Subjects could not accurately judge whether they were exposed to low (20%) or high (70%) air humidity when the temperatures were held constant.15
5.3 OTHER ENVIRONMENTAL FACTORS INFLUENCING THE SKIN Other factors influencing the skin include environmental contactants. The effects of skin contactants in different countries depend on several factors. They are influenced by the prevailing type of industry,31 availability of topical medicaments and prescribing habits of physicians,32 cultural and traditional habits of individuals in the country, and also the fauna of the country.9 The type of prevailing industrial activities in a country influences the prevalence of the type of contact dermatitis. For example, contact allergy to chromates from cement (representing about 6% of patients attending its patch test clinic) was prevalent in Singapore between 1980 and 1985 and declined (to less than 1% of patients attending its patch test clinic) after 1985, when the construction industry experienced a slump. Similarly, the prevalence of contact irritant dermatitis from solvents and cutting fluids, which are widely used in the electronics and metals industries, recorded an increase after 1985 as the two industrial activities took more prominence.34
35
Topical medicament allergy is common in developing countries where relatively cheap and more sensitizing medicaments are more widely used than in developed countries. In India contact allergy to nitrofurazone cream and in Singapore contact allergy to proflavine lotion were common causes of contact allergy to topical medicament, respectively.33,35 These allergens are not known to cause problems in developed countries. Preservatives allergy also varies in different regions of the world. Formaldehyde and formaldehyde releasers are common preservatives used in cosmetics and are the common cause of contact allergies from preservatives in Europe. However, contact allergies to formaldehyde and formaldehyde releasers are relatively uncommon in Singapore and probably in Southeast Asia. One reason could be the widespread use of Japanese cosmetics in Singapore, in which formaldehyde and formaldehyde releasers are not used as preservatives.36 Allergies from plants differ in different regions of the world; e.g., primin allergies from primulas, which are common in Europe, are uncommon in tropical countries. Rhus allergy is uncommon in Southeast Asian countries, whereas rengas allergy is very common in Southeast Asian countries.37 The characteristic warm and humid tropical climate is unique compared to the temperate climates. The tropical climate varies minimally throughout the year. The average ambient temperature of 30˚C and humidity of more than 70% are about the same throughout the year. Other types of skin disorders associated with such climate and peculiar to the tropics include: 1. Acne estivalis, a condition described in patients who developed acneiform eruption after spending time in the tropical climate. The exact mechanism is unknown, but heat is believed to play a role. Heat and high humidity have been known to affect the pilosebaceous unit.37,38 2. Miliaria, a common disorder in the tropics resulting from swelling of the keratin lining of the sweat ducts due to heat and high humidity. 3. Skin infections, high heat, and humidity in the tropics promote sweating and sweat retention, especially on skin folds, resulting in skin maceration. Secondary bacterial infection and fungal infections on macerated skin are common.
REFERENCES 1. Suskind, R.R., The environment and the skin, Environ. Health Perspect., 20, 27, 1977. 2. Suskind, R.R., Environment and the skin, Med. Clin. North Am., 74, 307, 1990.
36
Handbook of Non-Invasive Methods and the Skin, Second Edition
3. Farr, P.M. and Diffey, B.C., The vascular response of human skin to ultraviolet radiation, Photochem. Photobiol., 44, 501, 1986. 4. Harber, L.C. and Bickers, D.R., Eds., Photosensitivity Diseases: Principles of Diagnosis and Treatment, 2nd ed., B.C. Decker, Toronto, 1989, p. 112. 5. Gilchrest, B.A., Soter, N.A., Stoff, J.S., et al., The human sunburn reaction: histologic and biochemical studies, J. Am. Acad. Dermatol., 4, 411, 1981. 6. Gilchrest, B.A., Soter, N.A., Hawk, J.L.M., et al., Histologic changes associated with ultraviolet A-induced erythema in normal human skin, J. Am. Acad. Dermatol., 9, 213, 1983. 7. Staberg, B., Wulf, H.C., Klemp, P., et al., The carcinogenic effect of UVA irradiation, J. Invest. Dermatol., 81, 517, 1983. 8. Olumide, Y., Oleru, G.U., and Enu, C.C., Cutaneous implications of excessive heat in the work-place, Contact Derm., 9, 360, 1983. 9. Goh, C.L., Exogenous dermatoses in the tropics, in Exogenous Eczema, Menne, T. and Maibach, H.I., Eds., CRC Press, Boca Raton, FL, 1990, p. 351. 10. Rycroft, R.J.G., Occupational dermatoses among office personnel, Occup. Med., 1, 323, 1986. 11. Rycroft, R.J.G., and Smith, W.D.L., Low humidity occupational dermatoses, Contact Derm., 6, 488, 1980. 12. White, I.R. and Rycroft, R.J.G., Low humidity occupational dermatoses: an epidemic, Contact Derm., 8, 287, 1982. 13. Gaul, L.E. and Underwood, G.B., Relation of dew point and barometric pressure to chapping of normal skin, J. Invest. Dermatol., 19, 9, 1952. 14. Chernosky, M.E., Pruritic skin disease and summer air conditioning, JAMA, 179, 1005, 1962. 15. Andersen, I.B., Lundqvist, G.R., Jensen, P.L., and Proctor, D., Human response to 78 hour exposure to dry air, Arch. Environ. Health, 29, 319, 1974. 16. Callis, L.B. and Natarajan, M., Ozone and nitrogen dioxide changes in the stratosphere during 1979–1984, Nature, 323, 772, 1986. 17. Solomon, S., Garcia, R.R., Rowland, F.S., and Wuebbles, D.J., On the depletion of the Antartic ozone, Nature, 321, 755, 1986. 18. Stolarski, R.S., Kreuger, A.J., Shcoeberl, M.R., McPeters, R.D., Newman, P.A., and Alper, J.C., Nimbus and satellite measurements of the springtime Antarctic ozone decrease, Nature, 322, 808, 1986. 19. Farman, J.C., Gardiner, B.G., and Shanklin, J.D., Large losses of total ozone in Antarctica reveal seasonal CLOx/Nox interactions, Nature, 315, 207, 1985. 20. McElroy, M.B., Salawitch, R.J., Wofsy, S.C., and Longan, J.A., Reductions of Antarctic ozone due to synergistic interaction of chlorine and bromine, Nature, 321, 759, 1986. 21. Crus, P.D. and Bergstresser, P.R., The low-dose model of UVB-induced immunosuppression, Photodermatology, 5, 151, 1988.
22. Bergstresser, P.R., Ultraviolet B radiation induces “local immunosuppression,” Curr. Prob. Dermatol., 15, 205, 1986. 23. Kripke, M.L. and Morison, W.L., Modulation of immune function by UV radiation, J. Invest. Dermatol., 85, 62s, 1985. 24. Czernielewski, J.M., Masouye, I., Pisani, A., Ferracin, J., Auvolat, D., and Ortonne, J.P., Effect of chronic sun exposure on human Langerhans cell densities, Photodermatology, 5, 116, 1988. 25. Sjovall, P. and Moller, H., The influence of locally administered ultraviolet light (UVB) on the allergic contact dermatitis in the mouse, Acta Derm. Venereol., 65, 465, 1985. 26. Bruze, M., Seasonal influence on routine patch test results, Contact Derm., 14, 184, 1986. 27. Veien, N.K., Hattel, T., and Laurberg, G., Is patch testing a less accurate tool during the summer months, Am. J. Contact Derm., 3, 35, 1992. 28. Scotto, J. and Fraumeni, J., Skin cancer (other than melanoma), in Cancer Epidemiology and Prevention, Schotterfeld, D. and Fraumeni, J., Eds., W.B. Saunders, Philadelphia, 1982, p. 996. 29. Russell Jones, R., Consequences for human health of stratospheric ozone depletion, in Ozone Depletion, Health and Environmental Consequences, Russel Jones, R. and Wigley, T., Eds., John Wiley & Sons, New York, 1989, p. 207. 30. Gordon, D. and Silverstone, H., Worldwide epidemiology of premalignant and malignant cutaneous lesions, in Cancer of the Skin, Andrade, R., Ed., W.B. Saunders, Philadelphia, 1976, p. 405. 31. Goh, C.L., Epidemiology of contact allergy in Singapore, Int. J. Dermatol., 27, 308, 1988. 32. Goh, C.L., Contact sensitivity to topical antimicrobials. 1. Epidemiology in Singapore, Contact Derm., 21, 46, 1989. 33. Goh, C.L., Contact sensitivity to topical medicaments, Int. J. Dermatol., 28, 25, 1989. 34. Goh, C.L., Occupational dermatitis in Singapore. Changing pattern: 1985–1989, Hifu (Skin Research), 33, 95, 1991. 35. Goh, C.L., Contact sensitivity to proflavine, Int. J. Dermatol., 25, 449, 1986. 36. Goh, C.L., Allergic contact dermatitis from cosmetics, J. Derm., 14, 248, 1987. 37. Goh, C.L., Occupational allergic contact dermatitis from Rengas wood, Contact Derm., 18, 300, 1988. 38. Lobitz, W.C. and Dobson, R.L., Miliaria, Arch. Environ. Health, 11, 460, 1965. 39. Taylor, J.S., The pilosebaceous unit, in Occupational and Industrial Dermatology, Maibach, H.I. and Gellin, G.A., Eds., Year Book Medical, Chicago, 1982, p. 125.
Methods and 6 Non-Invasive Assessment of Skin Diseases Stefania Seidenari, Francesca Giusti, and Giovanni Pellacani Department of Dermatology, University of Modena and Reggio Emilia, Modena, Italy
CONTENTS 6.1 6.2
Introduction..............................................................................................................................................................37 Instrumental Assessment of Psoriasis .....................................................................................................................38 6.2.1 Assessment of the Extent of the Disease (A in the PASI) .........................................................................38 6.2.2 Assessment of Erythema and Skin Blood Flow (E in the PASI)...............................................................38 6.2.3 Assessment of Induration and Desquamation (I and D in the PASI) ........................................................39 6.2.4 Recent Techniques .......................................................................................................................................39 6.2.5 Assessment of the Barrier in Psoriasis .......................................................................................................39 6.2.6 Assessment of Side Effects of Topical Drugs.............................................................................................40 6.3 Instrumental Assessment of Sclerotic Skin.............................................................................................................40 6.3.1 Localized Scleroderma ................................................................................................................................40 6.3.1.1 Skin Elasticity ..............................................................................................................................40 6.3.1.2 Ultrasound ....................................................................................................................................40 6.3.2 Systemic Sclerosis .......................................................................................................................................40 6.3.2.1 Skin Mechanical Properties .........................................................................................................41 6.3.2.2 Skin Thickness .............................................................................................................................41 6.3.2.3 Echogenicity Measurements ........................................................................................................41 6.3.2.4 Confocal Laser Scanning Microscopy.........................................................................................41 6.4 Instrumental Assessment of Atopic Dermatitis.......................................................................................................42 6.4.1 Transepidermal Water Loss in AD ..............................................................................................................42 6.4.2 Conductance and Capacitance as Parameters for Skin Hydration in AD ..................................................42 6.4.3 Reactivity to Irritants in AD Subjects .........................................................................................................43 6.4.4 Bioengineering Techniques and Topical Agents for AD ............................................................................43 References .........................................................................................................................................................................43
6.1 INTRODUCTION The severity of an inflammatory skin disease is a basic element both for guiding the diagnostic process and therapeutic choice and for making the prognosis. To get rid of avoidable mistakes, its evaluation should be as objective and reproducible as possible. Moreover, an estimate of the cost–benefit ratio of new treatments, in both economic and clinical terms, is increasingly required. Controlled clinical trials, where new therapies are evaluated in comparison to consolidated ones, represent a prerequisite for an objective assessment. Furthermore, a decisive loan in the definition and grading of skin alterations occurred with the introduction of standardized schemes to assess skin diseases. However, in spite of attempts to regulate clinical
judgment, there are still wide variations both in assessment rules and in the interpretation of their use, thus making intra- and interobserver variations unavoidable. In fact, even trained clinical observers’ opinions may greatly vary when attempting to repeatedly document the severity of the same clinical situation. Finally, since nearly all drugs do not lead to prompt and complete clearance of objective and subjective symptoms, but to a gradual reduction in clinical signs, there is a need for accurate measurements and for statistical analysis of the results to reach a valid conclusion concerning their real effectiveness. During the last decade several instrumental devices intended for in vivo measurement of skin anatomy and physiology have become commercially available. Because of their non-invasive approach, these techniques give out 37
38
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 6.1 Values Referring to a Psoriatic Plaque (Mean Values ± s.d. of 30 Lesions) Compared with Normal Skin (30 Control Subjects)
Normal skin Psoriatic plaque
TEWL
Erythema (a*)
Dermal Echogenicity (0–30 areas)
Epidermal Reflectivity (201–255 areas)
4.9 ± 1.5 21.1 ± 7.2
3.4 ± 1.1 14.4 ± 3.8
1629.5 ± 517.4 22780.1 ± 2460.3
409.13 ± 159.8 7822.81 ± 3042.88
information referring to different body sites at the same time of the investigation and follow-up examinations without influencing the natural course of the disease or the therapeutic effects. Moreover, the same skin sites can be simultaneously studied by different techniques, each assessing various aspects of the disease and of the healing process. For different diseases, biophysical parameters correlating with clinical data and suitable for the followup were identified. For disease monitoring, data referring to normal skin, to basal conditions of diseased skin, and to successive records should be compared.
6.2 INSTRUMENTAL ASSESSMENT OF PSORIASIS Psoriasis is a chronic disease characterized by hyperproliferation of the epidermis, inflammation and dilation, and elongation of capillaries. The severity of psoriasis has to be assessed both to establish treatment protocols and to evaluate the results of therapy. Psoriasis is also a reference disease for assessment of the potency and efficacy of topical corticosteroids. Clinical scoring systems, based on the gradation of erythema, scaling, and infiltration, are easy and quick to use, but are never objective and seldom reproducible. In spite of this, considering 122 studies evaluating the efficacy of psoriasis treatments performed from 2001 to 2004, bioengineering devices were employed by three authors alone.1–3 The most popular method for the assessment of psoriasis is the psoriasis area and severity index (PASI).4 It is calculated as follows: PASI = 0.1 (Eh + Ih + Dh)Ah + 0.3 (Et + It + Dt)At + 0.2 (Eu + Iu + Du)Au + 0.4 (El + Il + Dl)Al, where E = erythema, I = induration, D = desquamation, and A = area. In the different sites, head (h), trunk (t), upper extremities (u), and lower extremities (l), a numerical value is given to the extent of the lesions (1 < 10%, 2 = 10 to 29%, 3 = 30 to 49%, 4 = 50 to 69%, 5 = 70 to 89%, and 6 > 90%). E, I, and D are scored according to a 5-point scale: 0 = no lesions, 1 = slight, 2 = moderate, 3 = marked, and 4 = very marked. Even considering the most relevant clinical aspects of psoriasis severity, the subjective assessment of erythema, infiltration, desquamation, and extent of the psoriasis area unavoidably leads to intra- and interobserver variations in
the scores. Non-invasive techniques allow the objective and reproducible quantification of parameters of the PASI score system. Therefore, the evaluation of psoriasis should be based on both clinical and instrumental data. Table 6.1 shows typical values referring to a psoriatic plaque compared to those of normal skin.
6.2.1 ASSESSMENT OF THE EXTENT (A IN THE PASI)
OF THE
DISEASE
It has been shown that human eye assessment of psoriasis extent shows great variations and usually overestimates the skin surface area. Since error estimates have a significant effect on the PASI score, computer image analysis performed on pictures representing both compromised and healthy skin areas were introduced for the estimation of the involved skin area (A in the PASI). These techniques enable the evaluation of the affected area with great accuracy,5–7 even though they have the disadvantage of being time-consuming (photographing, processing of images) and technically demanding. Moreover, the cylindrical shape of the limbs produces shadows, sometimes leading to problems in calculating the surface of the involved skin areas. In view of the above, a simple horizontal averaging program enabling shade correction was recently developed to provide a quick and accurate assessment of disease extent.8
6.2.2 ASSESSMENT OF ERYTHEMA FLOW (E IN THE PASI)
AND
SKIN BLOOD
The erythematous component of psoriatic lesions was evaluated by colorimetry, spectrophotometry, and laser Doppler velocimetry.1,5,9–12 When desquamation is absent, instrumental values referring to erythema correlate well with the inflammatory component of the disease. On the contrary, in a psoriatic plaque surmounted by thick scales, erythema values approach those of healthy skin. When desquamation was scarce, both the E (Dermaspectrometer) and a* (Minolta Chromameter) parameters showed values that were twice those of normal skin, whereas they decreased as scales became thicker.9 By evaluating the activity of tacalcitol vs. placebo and betamethasone valerate on psoriatic plaques in two different patient groups, we observed a decrease in erythema parameters according
Non-Invasive Methods and Assessment of Skin Diseases
39
to the response to therapy in one group, whereas erythema increased, owing to the removal of the scaly layer, in the other. Final colorimetric values were similar to baseline ones and did not prove useful in the assessment of the response to treatment.12 In psoriasis changes in microvasculature of the upper dermis, with elongation and dilation of skin capillaries, are associated with increased blood flow. In psoriatic plaques prior to therapy, skin blood flow was significantly higher than that in uninvolved skin.13,14 During treatment skin blood flow progressively decreased, approaching that of uninvolved skin in 3 to 4 weeks. However, paradoxically, after descaling treatment, removal of scales can lead to an increase in blood flow owing to the proximity of deeper and larger vessels contributing to the signal, giving rise to increased values.
6.2.3 ASSESSMENT OF INDURATION AND DESQUAMATION (I AND D IN THE PASI) Induration corresponding to skin thickening can be measured by A- and B-scanning methods. By means of Ascanning procedures, Serup15 demonstrated an increase in skin thickness in psoriatic lesions with respect to healthy skin. This difference proved higher on the limbs than on the trunk. Thus, to study the efficacy of antipsoriatic treatments, it is advisable to select plaques localized to the limbs also because of fewer variations. High-frequency ultrasound providing high-resolution images of cutaneous structures has been employed as a method for assessing the severity of psoriasis plaques by numerous authors.10–12,16–18 Employing 20-MHz B-scanning, we confirmed that thickness measurement methods enable the evaluation of the disease’s progress, since this parameter gradually decreases according to clinical response to therapy.16 Moreover, we observed that psoriatic skin appears less echogenic than healthy skin, with a thick epidermal band, a hypoechogenic subepidermal band, corresponding to a combination of elongated papillae and inflammation, and numerous parallel acoustic shadows (Figure 6.1). Utilizing image analysis based on segmentation procedures, we also noticed a progressive reduction both in dermal echogenicity and in amplitude values corresponding to epidermal hyperreflectivity at skin sites treated with anthralin.16 In contrast with colorimetry, 20-MHz sonography enabled the distinction between the effects of tacalcitol, placebo, and betamethasone valerate in patients with psoriasis. 12 Colorimetry and ultrasound were also employed for the simultaneous assessment of the efficacy of different drugs in the same psoriatic patient in the psoriasis plaque test.11 Twenty-megahertz sonography is suitable for studying pathological changes and pharmacological effects in the dermis and the subcutaneous fat, but for investigating the epidermis, frequencies between 50 and 100 MHz are more
FIGURE 6.1 Echographic image of a psoriatic plaque. The entry echo is thickened and a hypoechogenic band is observable immediately below. Echogenicity of the dermis is decreased compared with unaffected skin (upper part of the image).
effective.18,19 By using 100-MHz ultrasound in psoriatic skin, El Gammal et al.19 found that the thickened horny layer appears echorich and its echodensity decreases after application of petrolatum. The acanthotic epidermis plus the dermis with the inflammatory infiltrate are represented as an echo-poor band, whose sonometric thickness proved to be correlated with histometric thickness.
6.2.4 RECENT TECHNIQUES Optical coherence tomography (OCT) was recently employed for the assessment of the response to therapy in psoriasis.20 OCT images of psoriatic plaques are characterized by a thickening of the entrance signal, which is composed of several parallel layers, with dilated signalfree cavities and significantly lower attenuation with respect to healthy skin. These changes revert to normal after treatment.
6.2.5 ASSESSMENT
OF THE
BARRIER
IN
PSORIASIS
Psoriatic lesions, which are generally thick and scaly, are known to produce high transepidermal water loss (TEWL) values compared to nonaffected skin. Thus, the altered stratum corneum in psoriasis has a defective water barrier function. Combined measurements of TEWL and stratum corneum hydration in psoriasis have shown an inverse relationship between water barrier and reservoir function; i.e., the higher the TEWL, the lower the skin hydration.21,22 However, these alterations are not directly correlated to the clinical situation and to response to therapy. Serup and Blichman21 reported that neither TEWL nor stratum
40
Handbook of Non-Invasive Methods and the Skin, Second Edition
corneum hydration was correlated to scaling, and they concluded that low conductance of psoriatic skin is probably a manifestation of abnormal keratinization.
6.2.6 ASSESSMENT DRUGS
OF
SIDE EFFECTS
OF
TOPICAL
Non-invasive bioengineering methods have also been employed to investigate the adverse effects of antipsoriatic drugs, such as calcipotriol23–25 and dithranol.26 Reactions to 48-hour patch tests with calcipotriol ointment resulted in redness, as evaluated by a Minolta Chromameter and laser Doppler flowmetry, whereas edema formation and skin thickening were observed in advanced reactions only by means of ultrasound.23 Clinically and instrumentally (TEWL, capacitance), Effendy et al.24 demonstrated that calcipotriol (0.005%) is less irritating than 1% sodium lauryl sulfate (SLS). We evaluated the irritant reaction induced by dithranol by means of visual assessment, colorimetry, and ultrasound and compared it to the 2% SLS response.26 In contrast with the SLS reaction, where a 24hour decrease in epidermal reflectivity was observable, responses to dithranol showed an accentuation of the hyperreflecting epidermal band at 24 to 96 hours.
6.3 INSTRUMENTAL ASSESSMENT OF SCLEROTIC SKIN 6.3.1 LOCALIZED SCLERODERMA Localized scleroderma (morphea) is a disorder of unknown cause characterized by localized dermal hardening. The early changes are superficial and deep perivascular and interstitial infiltrates of mixed inflammatory cells. As the infiltrates become progressively lymphoplasmacytic, collagen bundles in the dermis become thickened and packed together. In late lesions the infiltrates of lymphocytes and plasma cells, sometimes accompanied by eosinophils, wane and leave thickened bundles of collagen, located very close to one another and oriented parallel to the skin surface. Localized scleroderma has been studied by skin elasticity measurements; however, ultrasound measurements are particularly suitable for its assessment. 6.3.1.1 Skin Elasticity Employing a suction chamber method, Serup and Northeved27 observed that skin distensibility and stiffness decreased in every phase of the morphea plaque in contrast to thickening. In active plaques of morphea, skin thickening was found to be associated with decreased extensibility in comparison with symmetrical healthy control areas, whereas in regressing lesions, skin thickness did not differ from control areas, in spite of a persistent lower extensibility.28
6.3.1.2 Ultrasound A thickening of the skin at morphea plaques was observed using both A- and B-mode ultrasound devices.29–31 Using A-mode ultrasound, Serup29 showed increased thickening from plaques with slight alterations to plaques of clinically advanced scleroderma. The relative increase in thickness was higher in thinner skin, for example, on the limbs with respect to the trunk, and in patients with one or a few morphea plaques than in those with generalized morphea. Alterations of the dermis and hypodermis in sclerotic skin characterized by inflammation surrounding the vessels, edematous swelling of collagen bundles, and fibrosis not only correspond to modifications in skin thickness, but also bring about modifications in ultrasound reflections and echogenicity transformations referring to both intensity and echo distribution (Table 6.2). To describe a morphea plaque after image analysis, five different parameters can be used, namely, the absolute and relative extensions of isoechogenic areas and the number, size, and density of the objects (areas with definite shape and extension) composing the image.31 When assessment is performed with intensity bands covering the lower part of the amplitude scale, the image referring to sclerotic skin tissue appears relatively homogeneous with few, large objects within a thickened skin block, which occupy a more extensive image surface in comparison to normal skin images characterized by small and closely packed spots. On the contrary, images of localized scleroderma transformed by intermediate- to high-amplitude intervals appear with fewer objects of approximately the same size or smaller, which are less compressed with respect to healthy skin images. By employing a 13-MHz ultrasound device with a 60mm penetration depth for investigating morphea plaques, Cosnes et al.32 recently described undulations in the dermis, disorganization, loss of thickness and thickened hyperechoic bands in the hypodermis, and a characteristic dense image resembling a flattened yo-yo. A 92% sensitivity and a 100% specificity for localized scleroderma were found when at least four of these five signs were present.
6.3.2 SYSTEMIC SCLEROSIS Systemic sclerosis is a heterogeneous disease characterized by the overproduction of extracellular matrix by fibroblasts, damage of the endothelium of small vessels, and activation of the immune system, resulting in infiltration of the lower dermis and the upper subcutis. Widespread skin involvement is generally associated with internal organ involvement and bad prognosis.33 An objective evaluation of the extension of skin involvement is not only needed for prognostic purposes, but also to identify effective therapeutic interventions. The modified Rodnan total
Non-Invasive Methods and Assessment of Skin Diseases
41
TABLE 6.2 Numerical Values from Elaboration of Echographic Images of Sclerotic Skin and Normal Skin
Skin thickness (mm) Echogenicity values (0–30 areas)
PSS
HS
Morphea Plaque
Healthy Skin
1.5 ± 0.41a 12.500 ± 10.000 a
1.17 ± 0.13 6000 ± 4000
2.50 ± 0.92 b 14.000 ± 12.000 b
1.97 ± 0.67 3.300 ± 4.000
Note: PSS = mean values ± s.d. of sclerotic skin on the back of the hand referring to 18 subjects affected by systemic sclerosis. HS = mean values ± s.d. of healthy skin on the back of the hand in 20 healthy controls. Morphea plaque = mean values ± s.d. of 60 morphea plaques. Healthy skin = mean values ± s.d. of healthy controlateral skin in patients affected by morphea. a b
Significant with respect to healthy controls. Significant with respect to healthy skin.
skin thickness score, carried out by clinical palpation, is a commonly used outcome measure in trials of systemic sclerosis,34,35 but reproducibility represents a difficult task even for experienced clinicians. Bioengineering techniques provide a useful support to clinical judgment in this case too. 6.3.2.1 Skin Mechanical Properties Non-invasive suction devices were employed for the study of the elastic properties of the skin in systemic sclerosis.36,37 Investigating patients at a different stage of the disease (edematous and indurative phase), Dobrev37 observed that the edematous phase was characterized by significantly lower immediate distension, final distension, and higher viscoelastic-to-elastic ratio compared with the indurative phase, whereas low values in skin distensibility correlated with severe skin thickness or hidebound skin. 6.3.2.2 Skin Thickness The skin–phalanx distance on the digits and forearm skin thickness are increased in patients with acrosclerosis with respect to age-matched healthy controls.38–40 Ihn et al.41 demonstrated that skin thickening is also present at skin sites that are clinically uninvolved by the sclerotic process, for example, on the chest. Conversely, a skin thickening is not always observable at induration sites. As evaluated by 20-MHz B-scanning, forehead and cheek skin is thinner in systemic sclerosis than in healthy subjects.42 Therefore, results of different skin area measurements should be considered together when assessing the disease activity and as a measure of outcome, and each measure should be carefully estimated according to the affected skin site. Recently, Moore et al.43 described a 17-point dermal ultrasound scoring system, to be employed for evaluating skin thickness in systemic sclerosis, combining clinical and sonographic assessment. The method proved extremely reliable and was proposed as a useful measure
of outcome, showing acceptable intra- and interobserver variations. 6.3.2.3 Echogenicity Measurements Morphologic modifications of 20-MHz ultrasound images of the skin are observable in systemic sclerosis patients. When echographic images of the skin on the back of the hand, the forehead, and the cheek of patients with systemic sclerosis undergoing image analysis were compared to those referring to a population of healthy sex- and agematched subjects, marked differences in the echostructure of the tissue were observable.42 On the forehead and the cheek, the thinned skin appeared more echogenic with respect to the skin of healthy subjects, with smaller hyporeflecting objects and greater hyperreflecting areas, whereas the thickened skin on the back of the hand was less echogenic, with large hyporeflecting areas and small hyperreflecting objects. Numerical description of echographic images of patients with systemic sclerosis provides significantly different data from those of normal skin (Table 6.2). 6.3.2.4 Confocal Laser Scanning Microscopy In vivo confocal laser scanning microscopy is a new noninvasive microscopy technique with high resolution at a specific depth, allowing skin imaging at a cellular level in real time with a penetration depth of about 200 to 250 μm. Evaluating micromorphologic characteristics of the skin in patients with systemic sclerosis referring to melanization, epidermal hypotrophy, dislocation of capillaries, and collagen deposits in the papillary dermis, by confocal laser scanning microscopy, Sauermann et al.44 were able to identify histometric parameters that significantly differed with respect to healthy controls. Although at present expensive and time-consuming, this method may open new possibilities for the assessment and monitoring of systemic sclerosis.
42
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 6.3 Capacitance, Transepidermal Water Loss, pH, and Echogenicity Values in Healthy and Atopic Skin (Mean Values ± s.d. Referring to 200 AD Patients and 200 Healthy Subjects) Capacitance Healthy Subjects Cheek Volar forearm Antecubital fossa Back of the leg Affected skin areas
61 ± 53.7 ± 66.2 ± 53.6 ± —
9.7 9.1 7.6 8.1
TEWL
AD Patients 49 ± 13 56.3 ± 12.1 60.6 ± 12.9 51.6 ± 11.4 23–56
Healthy Subjects 6.59 4.06 5.82 5.00
± 2.21 ± 1.98 ± 2.43 ± 4.29 —
Echogenicity (0–30 areas)
pH AD Patients
8.61 ± 3.57 7.71 ± 4.13 10.09 ± 5.27 6.96 ± 3.81 8–55
Healthy Subjects 5.43 4.86 4.70 5.32
± 0.42 ± 0.45 ± 0.49 ± 0.48 —
AD Patients
Healthy Subjects
AD Patients
5.49 ± 0.57 5.23 ± 0.74 5.12 ± 0.73 5.55 ± 0.74 4.9–6.6
— 993 ± 345 — — —
— 1029 ± 581 — — 1500–4000
6.4 INSTRUMENTAL ASSESSMENT OF ATOPIC DERMATITIS
without clinical signs other than hand eczema in adult life, TEWL proved normal on the upper arm.65
For the evaluation of atopic dermatitis (AD), the introduction of standardized clinical scoring systems considering severity of skin lesions, extent of skin involvement, and subjective symptoms has represented a decisive breakthrough.45–47 These clinical methods should be integrated by the evaluation of the status of the barrier and the degree of inflammation at selected skin sites. Since epidemiological studies show that AD subjects have a high incidence of hand eczema induced by irritant substances, in both a domestic and an occupational environment, most experimental studies have concentrated on barrier assessment by conductance, capacitance,48–51 and TEWL measurements, referring to both baseline conditions52–54 and when the skin is challenged with model irritants.55–59 Table 6.3 shows instrumental data referring to AD patients.
6.4.2 CONDUCTANCE AND CAPACITANCE AS PARAMETERS FOR SKIN HYDRATION IN AD
6.4.1 TRANSEPIDERMAL WATER LOSS
IN
AD
Most authors report increased TEWL values in AD subjects, both adults and children, at eczematous and also at apparently unaffected skin areas.60–63 In a study performed on AD children and controls, we found significant alterations in TEWL, measured at different body sites on both involved and uninvolved skin.62,63 In patients with current eczema, TEWL values at healthy skin sites were higher than in patients without lesions,63 indicating that the presence of active eczematous lesions induces an impairment of the barrier at clinically healthy skin sites, and that TEWL values may vary according to the course of the disease. In AD patients, the skin barrier impairment appears reversible and the long-lasting absence of eczema makes water barrier restoration possible. Atopic individuals without active dermatitis for the past 2 years showed TEWL values that were similar to those of healthy volunteers.64 Also in patients with a past history of AD, but
Skin hydration in AD patients was first assessed by capacitance measurements by Werner.66 Lodén et al.67 observed lower capacitance values in atopics, especially with increasing degree of dryness. Reduced capacitance values were observed in both eczematous and uninvolved skin,48–50,62,63,66,67 but the reduction was particularly pronounced in severe AD.51 At healthy skin sites, these alterations were more marked in patients with active disease.63 Stratum corneum hydration depends on both the ability to bind and the ability to retain water.52 Investigating the hydration and water retention capacity of unaffected skin in patients with AD, Berardesca et al.50 reported that in atopic patients the stratum corneum water retention capacity, described by the skin surface water loss profile, was significantly reduced. Dynamic methods, like the sorption–desorption test (SDT) and the moisture accumulation test (MAT), were developed in order to study the horny layer hydration kinetics.54 When performed in children with AD, we observed that the stratum corneum of uninvolved atopic skin was less hydrated, but more easily hydratable, by water coming both from the deeper layers and from the environment, with respect to the skin of healthy subjects.54 On the contrary, the eczematous areas showed an increased avidity to retain water, but a reduced absorption capacity. In AD patients the barrier impairment coincides with marked alterations in the amount and composition of epidermal lipids.58,59,68,69 When investigating the relationship between different lipid classes and barrier impairment in 47 patients with AD,70 we found an inverse correlation between TEWL and ceramides and a direct correlation between the increase in free cholesterol and the reduction in ceramide 3 levels. Electrical impedence was reported
Non-Invasive Methods and Assessment of Skin Diseases
43
to be dependent on the lipid content of the stratum corneum. 71 Nicander and Ollmar 71 showed significant changes between baseline values of clinically normal atopic skin and healthy skin. Furthermore, impedence showed larger reactivity in AD patients after skin stripping and lipid extraction.
AD.79–84 In fact, certain moisturisers were shown to improve water barrier function, as reflected by TEWL, and skin susceptibility to irritants in atopic patients.80–82 Moreover, a significant relationship was noted between the reduction in TEWL and the clinical improvement of dryness.83 Lodén et al.83 performed an instrumental and clinical comparison of the effects of a cream containing 20% glycerine and a cream with 4% urea on the skin of AD patients, showing the superiority of one of the two preparations. For evaluating the effects of a ceramidedominant, physiologic lipid-based emollient for the treatment of AD children, SCORAD values proved to be less effective than TEWL, which was more sensitive for detecting subtle fluctuations in AD activity and for predicting potential relapse.84 Skin thickness ultrasound measurement showed that long-term tacrolimus ointment therapy in patients with AD is nonatrophogenic and reverses corticosteroid-induced skin atrophy.85 As evaluated by TEWL measurements, the same topical product was shown to enhance experimentally induced irritant contact dermatitis and not to accelerate healing of irritant contact dermatitis and UVB erythema.86
6.4.3 REACTIVITY
TO IRRITANTS IN
AD SUBJECTS
Both clinical and instrumental data demonstrate a cutaneous hyperreactivity in subjects with active AD, experimentally exposed to irritants, which is related to the degree of severity and the extension of the dermatitis. In an inactive phase of the disease, AD patients may or may not show enhanced skin reactions upon exposure to irritants with respect to nonatopics.64,72–75 Tupker et al.61 investigated skin irritability by repeated applications of different irritants and found increased TEWL values, both before and after exposure, in subjects with a history of AD with respect to subjects with a history of allergic contact dermatitis or controls. Agner60 challenged the skin of the flexor side of the upper arm with SLS for 24 hours and observed greater reactions in atopic patients than in controls, as assessed both clinically and instrumentally. Moreover, postexposure TEWL, correlating with baseline values, was significantly higher in atopics than in controls. After SLS challenge, we observed both an increase in TEWL and a decrease in capacitance, which were more marked in subjects with AD than in controls.76,77 When investigating skin reactions to 30 minutes of 0.5% SLS on the forearms in 20 healthy volunteers and on 34 subjects with localized eczema in a chronic phase, comprising 14 atopic patients and 20 individuals with contact dermatitis, the echographic assessment of SLS-exposed areas showed a significant decrease in epidermal reflectivity, indicating barrier function damage in atopic subjects, but not in contact dermatitis patients.78 Moreover, SLS pretreatment of nickel patch test sites induced an earlier and a more marked cutaneous damage in atopic nickel-sensitive patients with respect to nickel-sensitive nonatopics, followed by a more intense allergic response, indicating the role of barrier damage by irritants in the induction and elicitation of contact eczema in atopics.49 On the contrary, postexposure TEWL, capacitance, and echogenicity values did not differ between subjects with mucosal atopy and healthy volunteers.76
6.4.4 BIOENGINEERING TECHNIQUES AGENTS FOR AD
AND
TOPICAL
Objective monitoring of barrier impairment in AD is of considerable interest in studies evaluating the efficacy of anti-inflammatory drugs and moisturizing creams on
REFERENCES 1. Suh, D.H., Kwon, T.E., Kim, S.D., Park, S.B., Kwon, O.S., Eun, H.C., and Youn, J.I., Changes of skin blood flow and color on lesional and control sites during PUVA therapy for psoriasis, J. Am. Acad. Dermatol., 44, 987, 2001. 2. Stucker, M., Memmel, U., Hoffmann, M., Hartung, J., and Altmeyer, P., Vitamin B12 cream containing avocado oil in the therapy of plaque psoriasis, Dermatology, 203, 141, 2001. 3. Goh, C.L. and Khoo, L., Laser Doppler perfusion imaging (LDPI) and transepidermal water loss (TEWL) values in psoriatic lesions treated with narrow band UVB phototherapy. Dermal vascularity may be useful indicator of psoriatic activity, Ann. Acad. Med. Singapore, 33, 75, 2004. 4. Fredriksson, T. and Pettersson, U., Severe psoriasis-oral therapy with a new retinoid, Dermatologica, 157, 238, 1978. 5. Ormerod, A.D., Dwyer, C.M., and Weller, R., A comparison of subjective and objective measures of reduction of psoriasis with the use of ultrasound, reflectance colorimetry, computerized video image analysis, and nitric oxide production, J. Am. Acad. Dermatol., 37, 51, 1997. 6. Savolainen, L., Kontinen, J., Roning, J., and Oikarinen, A., Application of machine vision to assess involved surface in patients with psoriasis, Br. J. Dermatol., 137, 395, 1997.
44
Handbook of Non-Invasive Methods and the Skin, Second Edition
7. Savolainen, L., Kontinen, J., Alatalo, E., Roning, J., and Oikarinen, A., Comparison of actual psoriasis surface area and the psoriasis area and severity index by the human eye and machine vision methods in following the treatment of psoriasis, Acta Derm. Venereol., 78, 466, 1998. 8. Tanaka, M., Gaskell, S., Edwards, C., and Marks, R., Simple horizontal averaging programme enables shade correction for image analysis in psoriasis, Clin. Exp. Dermatol., 25, 323, 2000. 9. Takiwaki, H. and Serup, J., Measurement of color parameters of psoriatic plaques by narrow- and reflectance spectrophotometry and tristimulus colorimetry, Skin Pharmacol., 7, 145, 1994. 10. Hoffmann, M., Dirschka, T., Schwarze, H., el-Gammal, S., Matthes, U., Hoffmann, A., and Altmeyer, P., 20 MHz sonography, colorimetry and image analysis in the evaluation of psoriasis vulgaris, J. Dermatol. Sci., 9, 103, 1995. 11. Bangha, E. and Elsner, P., Evaluation of topical antipsoriatic treatment by chromametry, visiometry and 20 MHz ultrasound in the psoriasis plaque test, Skin Pharmacol., 9, 298, 1996. 12. Seidenari, S., Magni, R., and Giannetti, A., Assessment of the activity of tacalcitol on psoriatic plaques by means of colorimetry and high frequency ultrasound, Skin Pharmacol., 10, 40, 1997. 13. Staberg, B. and Klemp, P., Skin blood flow in psoriasis during Goeckerman or beach tar therapy, Acta Derm. Venereol., 64, 331, 1984. 14. Khan, A., Schall, L.M., Tur, E., Maibach, H.I., and Guy, R.J., Blood flow in psoriatic skin lesions: the effect of treatment, Br. J. Dermatol., 117, 193, 1987. 15. Serup, J., Non-invasive quantification of psoriasis plaques: measurement of skin thickness with 15 MHz pulsed ultrasound, Clin. Exp. Dermatol., 9, 502, 1984. 16. Di Nardo, A., Seidenari, S., and Giannetti, A., B-scanning evaluation with image analysis of psoriatic skin, Clin. Exp. Dermatol., 1, 121, 1992. 17. Gupta, A.K., Turnbull, D.H., and Harasiewics, K.A., The use of high-frequency ultrasound as a method of assessing the severity of a plaque of psoriasis, Arch. Dermatol., 132, 658, 1996. 18. Unholzer, A. and Korting, H.C., High-frequency ultrasound in the evaluation of pharmacological effects on the skin, Skin Pharmacol. Appl. Skin Physiol., 15, 71, 2002. 19. El Gammal, S., El Gammal, C., and Kaspar, K., Sonography of the skin at 100 MHz enables in vivo visualization of stratum corneum and viable epidermis in palmar skin and psoriatic plaques, J. Invest. Dermatol., 113, 821, 1999. 20. Welzel, J., Bruhns, M., and Wolff, H.H., Optical coherence otmography in contact dermatitis and psoriasis, Arch. Dermatol. Res., 295, 50, 2003. 21. Serup, J. and Blichman, C., Epidermal hydration of psoriasis plaques and the relation to scaling, Acta Derm. Venereol., 67, 357, 1987.
22. Berardesca, E., Fideli, D., Borroni, G., Rabbiosi, G., and Maibach, H.I., In vivo hydration and water retention capacity of the stratum corneum in clinically uninvolved skin in atopic and psoriatic patients, Acta Derm. Venereol., 70, 400, 1990. 23. Fullerton, A., Avnstorp, C., Agner, T., Dahl, J.C., Olsen, L.O., and Serup, J., Patch test study with calcipotriol ointment in different patient groups, including psoriatic patients with and without adverse dermatitis, Acta Derm. Venereol. (Stockh.), 76, 194, 1996. 24. Effendy, I., Kwangsukstith, C., Chiappe, M., and Maibach, H.I., Effects of calcipotriol on stratum corneum barrier function, hydration and cell renewal in humans, Br. J. Dermatol., 135, 545, 1996. 25. Fullerton, A., Benfeldt, E., Petersen, J.R., Jensen, S.B., and Serup, J., The calcipotriol dose-irritation relationship: 48 hour occlusive testing in healthy volunteers using Finn chambers, Br. J. Dermatol., 138, 259, 1998. 26. Schiavi, M.E., Belletti, B., and Seidenari, S., Ultrasound description and quantification of irritant reactions induced by dithranol at different concentrations. A comparison with visual assessment and colorimetric measurements, Contact Derm., 34, 272, 1996. 27. Serup, J. and Northeved, A., Skin elasticity in localized scleroderma (morphoea): introduction of a biaxial in vivo method, and the measurement of tensile distensibility, hysteresis and resilient distension of diseased and normal skin, J. Dermatol., 12, 318, 1985. 28. Kalis, B., De Rigai, l., Leonard, F., Léveque, l.L., Riche, O., Le Care, Y., and De Lacharriere, O., In vivo study of scleroderma by non-invasive techniques, Br. J. Dermatol., 122, 785, 1990. 29. Serup, J., Localized scleroderma (morphoea): thickness of sclerotic plaques as measured by 15 Mhz pulsed ultrasound, Acta Derm. Venereol. (Stockh.), 64, 214, 1984. 30. Hoffmann, K., Gerbaulet, U., El-Gammal, S., and Altmeyer, P., 20 MHz B-mode ultrasound in the monitoring of the course of localized scleroderma (morphea), Acta Derm. Venereol. Suppl. (Stockh.), 164, 3, 1991. 31. Seidenari, S., Conti, A., Pepe, P., and Giannetti, A., Quantitative description of echographic irnages of morphea plaques as assessed by computerized image analysis on 20 MHz B-scan recording, Acta Derm. Venereol. (Stockh.), 75, 442, 1995. 32. Cosnes, A., Anglade, M.C., Revuz, J., and Radier, C., Thirteen-megahertz ultrasound probe: its role in diagnosing localized scleroderma, Br. J. Dermatol., 148, 724, 2003. 33. Clements, P.J., Hurwitz, E.L., Wong, W.K, Seibold, J.R., Mayes, M., White, B., Wigley, F., Weisman, M., Barr, W., Moreland, L., Medsger, T.A., Jr., Steen, V.D., Martin, R.W., Collier, D., Weinstein, A., Lally, E., Varga, J., Weiner, S.R., Andrews, B., Abeles, M., and Furst, D.E., Skin thickness score as a predictor and correlate of outcome in systemic sclerosis: high-dose versus low-dose penicillamine trial, Arthritis Rheum., 43, 2445, 2000.
Non-Invasive Methods and Assessment of Skin Diseases
45
34. Rodnan, G.P., Lipinski, E., and Luksick, J., Skin thickness and collagen content in progressive systemic sclerosis and localized scleroderma, Arthritis Rheum., 22, 130, 1979. 35. Clements, P.J., Lachenbruch, P., Siebold, J., White, B., Weiner, S., Martin, R., Weinstein, A., Weisman, M., Mayes, M., and Collier, D., Inter- and intra-observer variability of total skin thickness score (mod Rodnan TSS) in systemic sclerosis, J. Rheumatol., 22, 1281, 1995. 36. Enomoto, D.N., Mekkes, J.R., Bossuyt, P.M., Hoekzena, R., and Bos, J.D., Quantification of cutaneous sclerosis with a skin elasticity meter in patients with generalized scleroderma, Am. Acad. Dermatol., 35, 381, 1996. 37. Dobrev, H.P., In vivo study of skin mechanical properties in patients with systemic sclerosis, J. Am. Acad. Dermatol., 40, 436, 1999. 38. Serup, J., Quantification of acrosclerosis: measurement of skin thickness and skin-phalanx distance in females with 15 MHz pulsed ultrasound, Acta Derm. Venereol. (Stockh.), 64, 35, 1984. 39. Akesson, A., Forsberg, L., Hederstrom, E., and Wollheim, F., Ultrasound examination of skin thickness in patients with progressive systemic sclerosis (scleroderma), Acta Radiol. Diagn., 27, 91, 1986. 40. Myers, S.L., Cohen, J.S., Sheets, P.W., and Bies, J.R., B-mode ultrasound evaluation of skin thickness in progressive systemic sclerosis, J. Rheumatol., 13, 577, 1986. 41. Ihn, H., Shimozuma, M., Fujimoto, M., Sato, S., Kikuchi, K., Igarashi, A., Soma, Y., Tamaki, K., and Takehara, K., Ultrasound measurement of skin thickness in systemic sclerosis, Br. J. Rheumatol., 34, 535, 1995. 42. Seidenari, S., Belletti, B., and Conti, A., A quantitative description of echographic images of sclerotic skin in patients with systemic sclerosis, assessed by computerized image analysis on 20 MHz B-scan recordings, Acta Derm. Venereol. (Stockh.), 76, 361, 1996. 43. Moore, T.L., Lunt, M., McManus, B., Anderson, M.E., and Herrick, A.L., Seventeen-point dermal ultrasound scoring system: a reliable measure of skin thickness in patients with systemic sclerosis, Rheumatology (Oxford), 42, 1559, 2003. 44. Sauermann, K., Gambichler, T., Jaspers, S., Radenhausen, M., Rapp, S., Reich, S., Altmeyer, P., Clemann, S., Teichmann, S., Ennen, J., and Hoffmann, K., Histometric data obtained by in vivo confocal laser scanning microscopy in patients with systemic sclerosis, BMC Dermatol., 2, 1, 2002. 45. Kunz, B., Oranje, A.P., Labreze, L., Stalder, J.F., Ring, J., and Taieb, A., Clinical validation and guidelines for the SCORAD index: consensus report of the European Task Force on Atopic Dermatitis, Dermatology, 195, 10, 1997. 46. Sprikkelman, A.B., Tupker, R.A., Burgerhof, H., Schouten, J.P., Brand, P.L., Heymans, H.S., and van Aalderen, W.M., Severity scoring of atopic dermatitis: a comparison of three scoring systems, Allergy, 52, 944, 1997.
47. Barbier, N., Paul, C., Luger, T., Allen, R., De Prost, Y., Papp, K., Eichenfield, L.F., Cherill, R., and Hanifin, J., Validation of the eczema area and severity index for atopic dermatitis in a cohort of 1550 patients from the pimecrolimus cream 1% randomized controlled clinical trials programme, Br. J. Dermatol., 150, 96, 2004. 48. Gollhausen, R., The phenomenon of irritable skin in atopic eczema, in Handbook of Atopic Eczema, Ruzicka, T., Ring, J., and Przybilla, B., Eds., SpringerVerlag, Berlin, 1991, p. 306. 49. Seidenari, S., Reactivity to nickel sulfate at sodium lauryl sulfate pre-treated sites is higher in atopics: an echographic evaluation by means of image analysis performed on 20 MHz B-scan recordings, Acta Derm. Venereol. (Stockh.), 74, 245, 1994. 50. Berardesca, E., Fideli, D., Borroni, G., Rabbiosi, G., and Maibach, H.I., In vivo hydration and water-retention capacity of stratum corneum in clinically uninvolved skin in atopic and psoriatic patients, Acta Derm. Venereol. (Stockh.), 70, 400, 1990. 51. Tanaka, M., Okada, M., Zhen, Y.X., Inamura, N., Kitano, T., Shirai, S., Sakamoto, K., Inamura, T., and Tagami, H., Decreased hydration state of the stratum corneum and reduced amino acid content of the skin surface in patients with seasonal allergic rhinitis, Br. J. Dermatol., 139, 618, 1998. 52. Tagami, H., Kanamaru, Y., and Inoue, K., Water sorption-desorption test of the skin in vivo for functional assessment of the stratum corneum, J. Invest. Dermatol., 78, 425, 1982. 53. Werner, Y., Lindberg, M., and Forslind, B., The waterbinding capacity of stratum corneum in dry non-eczematous skin of atopic eczema, Acta Derm. Venereol. (Stockh.), 62, 334, 1981. 54. Pellacani, G. and Seidenari, S., Water sorption-desorption test and moisture accumulation test for functional assessment of atopic skin in children, Acta Derm. Venereol. (Stockh.), 81, 100, 2001. 55. Imokawa, G., Akasaki, S., Minematsu, Y., and Kawai, M., Importance of intercellular lipids in water-retention properties of the stratum corneum: induction and recovery study of surfactant dry skin, Arch. Dermatol. Res., 281, 45, 1989. 56. Elias, P.M. and Menon, G.K, Structural and lipid biochemical correlates of the epidermal permeability barrier, Adv. Lipid. Res., 24, 1, 1991. 57. Grubauer, G., Elias, P.M., and Feingold, K.R., Transepidermal water loss: the signal for recovery of barrier structure and function, J. Lipid. Res., 30, 323, 1989. 58. Melnik, B., Hollmann, J., Hofmann, U., Yuh, M.S., and Plewig, G., Lipid composition of outer stratum corneum and nails in atopic and control subjects, Arch. Dermatol. Res., 282, 549, 1990. 59. Yamamoto, A., Serizawa, S., Ito, M., and Sato, Y., Stratum corneum lipid abnormalities in atopic dermatitis, Arch. Dermatol. Res., 283, 219, 1991. 60. Agner, T., Susceptibility of atopic dermatitis patients to irritant dermatitis caused by sodium lauryl sulphate, Acta Derm. Venereol. (Stockh.), 71, 296, 1991.
46
Handbook of Non-Invasive Methods and the Skin, Second Edition
61. Tupker, R.A., Pinnagoda, J., Coenraads, P.J., and Nater, J.P., Susceptibility to irritants: role of barrier function, skin dryness and history of atopic dermatitis, Br. J. Dermatol., 123, 199, 1990. 62. Seidenari, S. and Giusti, G., Objective assessment of the skin of children affected by atopic dermatitis: a study on pH, capacitance and TEWL in eczematous and clinically uninvolved skin, Acta Derm. Venereol. (Stockh.), 75, 429, 1995. 63. Giusti, G. and Seidenari, S., La barriera cutanea nei bambini con dermatite atopica: valutazione strumentale in 200 pazienti e 45 controlli, Riv. Ital. Pediatr., 24, 954, 1998. 64. Löffler, H. and Effendy, I., Skin susceptibility of atopic individuals, Contact Derm., 40, 239, 1999. 65. Agner, T., Skin susceptibility in uninvolved skin of hand eczema patients and healthy controls, Br. J. Dermatol., 125, 140, 1991. 66. Werner, Y., The water content of the stratum corneum in patients with atopic dermatitis. Measurement with the Corneometer CM 420, Acta Derm. Venereol. (Stockh.), 66, 281, 1986. 67. Lodén, M., Olsson, H., Axéll, T., and Werner Linde, Y., Friction, capacitance and transepidermal water loss (TEWL) in dry atopic and normal skin, Br. J. Dermatol., 126, 137, 1992. 68. Imokawa, G., Abe, A., Jin, K., Higaki, Y., Kawashima, M., and Hidano, A., Decreased level of ceramides in stratum corneum of atopic dermatitis: an etiologic factor in atopic dry skin? J. Invest. Dermatol., 96, 523, 1991. 69. Schäfer, L. and Kragballe, K., Abnormalities in epidermal lipid metabolism in patients with atopic dermatitis, J. Invest. Dermatol., 96, 10, 1991. 70. Di Nardo, A., Wertz, P., Giannetti, A., and Seidenari, S., Ceramide and cholesterol composition of the skin of patients with atopic dermatitis, Acta Derm. Venereol. (Stockh.), 78, 27, 1998. 71. Nicander, I. and Ollmar S., Clinically normal atopic skin vs. non-atopic skin as seen through electrical impedance, Skin Res. Technol., 10, 178, 2004. 72. Stolz, R., Hinnen, U., and Elsner, P., An evaluation of the relationship between ‘atopic skin’ and skin irritability in metalworkers trainees, Contact Derm., 36, 281, 1997. 73. Basketter, D.A., Miettinen, J., and Lahti, A., Acute irritant reactivity to sodium lauryl sulfate in atopics and nonatopics, Contact Derm., 38, 253, 1998. 74. Hannuksela, A. and Hannuksela, M., Irritant effects of a detergent in wash, chamber and repeated open application tests, Contact Derm., 34, 134, 1996.
75. Van der Valk, P.G.M., Nater, J.P., and Bleumink, E., Vulnerability of the skin to surfactants in different groups of eczema patients and controls as measured by water vapour loss, Clin. Exp. Dermatol., 10, 98, 1985. 76. Seidenari, S., Belletti, B., and Schiavi, M.E., Skin reactivity to sodium lauryl sulfate in patients with respiratory atopy, J. Am. Acad. Dermatol., 35, 47, 1996. 77. Seidenari, S., Skin sensitivity, interindividual factors: atopy, in The irritant Contact Dermatitis Syndrome, Van der Valk, P.G. and Maibach, H.I., Eds., CRC Press, Boca Raton, FL, 1996, p. 267. 78. Seidenari, S. and Di Nardo, A., B-scanning evaluation of irritant reactions with binary transformation and image analysis, Acta Derm. Venereol. (Stockh.), 175, 9, 1992. 79. Aalto-Korte, K., Improvement of skin barrier function during treatment of atopic dermatitis, J. Am. Acad. Dermatol., 33, 969, 1995. 80. Lodén, M., Barrier recovery and influence of irritant stimuli in skin treated with a moisturising cream, Contact Derm., 36, 256, 1997. 81. Held, E., Sveinsdottir, A., and Agner, T., Effect of longterm use of a moisturiser on skin hydration, barrier function and susceptibility to irritants, Acta Derm. Venereol. (Stockh.), 79, 49, 1999. 82. Lodén, M., Andersson, A.C., and Lindberg, M., Improvement in skin barrier function in patients with atopic dermatitis after treatment with a moisturizing cream (Canoderm®), Br. J. Dermatol., 140, 264, 1999. 83. Lodén, M., Andersson, A.C., Andersson, C., Frödin, T., Öman, H., and Lindberg, M., Instrumental and dermatological evaluation of the effect of glycerine and urea on dry skin in atopic dermatitis, Skin Res. Technol., 7, 209, 2001. 84. Chamlin, S.L., Kao, J., Frieden, I.J., Sheu, M.Y., Fowler, A.J., Fluhr, J.W., Williams, M.L., and Elias, P.M, Ceramide-dominant barrier repair lipids alleviate childhood atopic dermatitis: changes in barrier function provide a sensitive indicator of disease activity, J. Am. Acad. Dermatol., 47, 198, 2002. 85. Kyllönen, H., Remitz, A., Mandelin, J.M., Elg, P., and Reitamo, S., Effects of a 1-year intermittent treatment with topical tacrolimus monotherapy on skin collagen synthesis in patients with atopic dermatitis, Br. J. Dermatol., 150, 1174, 2004. 86. Fuchs, M., Schliemann-Willers, S., Heinemann, C., and Elsner, P., Tacrolimus enhances irritation in a 5-day human irritancy in vivo method, Contact Derm., 46, 290, 2002.
for Good Clinical Practice 7 Standards (GCP) Merete Thyme Quality Assurance Department, Scantox (part of LAB Research International), LI. Skensved, Denmark
CONTENTS 7.1 7.2 7.3 7.4 7.5 7.6 7.7
Introduction..............................................................................................................................................................47 Declaration of Helsinki ...........................................................................................................................................47 Background of Good Clinical Practice ...................................................................................................................47 Regulatory Requirements ........................................................................................................................................48 ICH-GCP Principles ................................................................................................................................................48 Essential Documents................................................................................................................................................49 Roles and Responsibilities.......................................................................................................................................49 7.7.1 Sponsor Responsibilities..............................................................................................................................49 7.7.2 Monitor Responsibilities..............................................................................................................................49 7.7.3 Investigator Responsibilities........................................................................................................................50 References .........................................................................................................................................................................52
7.1 INTRODUCTION Various national legislation places the responsibility for establishing the safety and efficacy of drugs on the manufacturer of the regulated product. The authorities are responsible for reviewing the test results and determining whether the product’s safety and efficacy can be demonstrated. The marketing of the product is permitted only when the agencies are satisfied that safety and efficacy have been established adequately.
7.2 DECLARATION OF HELSINKI The purpose of biomedical research involving human subjects is to improve diagnostic, therapeutic, and prophylactic procedures and the understanding of the etiology and pathogenesis of diseases. The World Medical Association has prepared recommendations as a guide to every physician in biomedical research involving human subjects. In 1964, the 18th World Medical Assembly adopted the Declaration of Helsinki. The declaration sets forth 12 basic principles for the protection of human subjects in both clinical and nonclinical research. The Declaration of Helsinki has been amended regularly during the years, with the latest revision in Edinburgh, Scotland, in October 2000. It must be stressed that the standards are only a
guidance to physicians all over the world. Physicians are not relieved from criminal, civil, and ethical responsibilities under the law of their own countries.
7.3 BACKGROUND OF GOOD CLINICAL PRACTICE Since the beginning of the 1960s the requirements concerning the conduct and documentation of clinical research have been increased heavily. The first inspection of clinical studies to ascertain the standards of conduct and record keeping took place by the Food and Drug Administration (FDA) of the U.S. in the early 1960s. During the early years of regulatory inspections a significant number of malpractices were discovered, which led to a number of legal prosecutions. Since the implementation of the FDA Clinical Inspection Programme in 1977, other authorities around the world issued their own guidelines concerning good clinical practice. With the requirement that clinical trials conducted in one country should be widely accepted internationally for registration purposes, the good clinical practice rules were developed internationally in order to agree on a common quality standard to be followed when conducting clinical trials. 47
48
Handbook of Non-Invasive Methods and the Skin, Second Edition
Good clinical practice is based on the philosophy that quality standards and principles should be harmonized between countries. This would ensure optimal use and protection of the subjects to be included in the clinical trials, thereby restricting the number of necessary subjects. It would not be necessary to repeat the trials in different countries, and the data would be reliable by following the same documentation practice. Last but not least, a goal was that the time for registration of the drug would be minimized. The definition of good clinical practice (GCP) as given by the GCP principles is “a standard for the design, conduct, performance, monitoring, auditing, recording, analyses, and reporting of clinical trials that provides assurance that the data and reported results are credible and accurate, and that the rights, integrity, and confidentiality of trial subjects are protected.” Compliance with GCP ensures that the rights, safety, and well-being of trial subjects are protected and that the clinical trial data are credible. GCP principles are not a term of science improvement; the purpose of GCP is to ensure a correct basis for drug registration.
7.4 REGULATORY REQUIREMENTS Based on the need for international harmonization of GCP, an International Conference on Harmonization (ICH) was held in 1991. The objective was to provide a unified standard of the European Union (EU), Japan, and the U.S. to facilitate the mutual acceptance of clinical data by the regulatory authorities in these jurisdictions. Regulatory authorities and the pharmaceutical industry from the EU, the U.S., and Japan participated in the conference. This resulted in the preparation of the ICHGCP guideline (ICH-GCP), which was finalized in 1996. Since 1997 it has been required that clinical trials that should be used as documentation to the authorities in connection with registration applications in the EU, the U.S., and Japan should be conducted in compliance with the ICHGCP. An exemption was investigator-initiated clinical trials that were not within the scope of this requirement. In May 2004 a new EU Clinical Trials Directive became effective. The directive applies to all phases of clinical trials and to academic as well as commercially driven studies. This means that investigator-initiated clinical trials are within the future scope of the requirements.
7.5 ICH-GCP PRINCIPLES The ICH-GCP describes in detail all perspectives related to:
• • • • • •
Protection of subjects (ethics committees, patient information, and informed consent) Roles and responsibilities of the involved parties (investigator, sponsor, monitor) Study design (randomization, blinding, statistical analysis) Quality assurance Data management Archiving of data
The ICH-GCP consists of the following stated 13 general principles, which give a good impression of the purpose of the guideline: 1. Clinical trials should be conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki, and that are consistent with GCP and the applicable regulatory requirement(s). 2. Before a trial is initiated, foreseeable risks and inconveniences should be weighed against the anticipated benefit for the individual trial subject and society. A trial should be initiated and continued only if the anticipated benefits justify the risks. 3. The rights, safety, and well-being of the trial subjects are the most important considerations and should prevail over interests of science and society. 4. The available non-clinical and clinical information on an investigational product should be adequate to support the proposed clinical trial. 5. Clinical trials should be scientifically sound and described in a clear, detailed protocol. 6. A trial should be conducted in compliance with the protocol that has received prior institutional review board (IRB)/independent ethics committee (IEC) approval/favorable opinion. 7. The medical care given to, and medical decisions made on behalf of, subjects should always be the responsibility of a qualified physician or, when appropriate, a qualified dentist. 8. Each individual involved in conducting a trial should be qualified by education, training, and experience to perform his or her respective task(s). 9. Freely given informed consent should be obtained from every subject prior to clinical trial participation. 10. All clinical trial information should be recorded, handled, and stored in a way that allows its accurate reporting, interpretation, and verification. 11. The confidentiality of records that could identify subjects should be protected, respecting the
Standards for Good Clinical Practice (GCP)
privacy and confidentiality rules in accordance with the applicable regulatory requirement(s). 12. Investigational products should be manufactured, handled, and stored in accordance with applicable good manufacturing practice (GMP). They should be used in accordance with the approved protocol. 13. Systems with procedures that ensure the quality of every aspect of the trial should be implemented.
7.6 ESSENTIAL DOCUMENTS The documents that serve to demonstrate the compliance of the investigator, sponsor, and monitor with GCP and with all applicable regulatory requirements have been defined as essential documents by ICH-GCP. These documents permit individually and collectively evaluation of the conduct of a clinical trial and the quality of the data produced. The minimum list of essential documents to be generated and maintained in a clinical trial has been specified in the guideline. A description is given of the purpose of each document, and whether it should be filed in either the investigator or sponsor files, or both. It is required that trial master files are established at the beginning of a trial, both at the investigator site and at the sponsor’s office. A final closeout of a trial can only be done when it has been confirmed that all necessary documents are available in the appropriate files. The essential documents have been grouped in three sections according to the stage of the trial during which they will normally be generated. The three stages are: 1. Before the clinical phase of the trial commences 2. During the clinical conduct of the trial 3. After completion or termination of the trial Upon request from the monitor, auditor, ethics committee, or competent authorities, the investigator/institution should make available for direct access all requested trial-related records.
7.7 ROLES AND RESPONSIBILITIES GCP defines roles and responsibilities of the different parties involved in the conduct of clinical trials. Besides the trial subjects there are regulatory authorities, ethics committees, data protection agencies, sponsors, monitors, investigators, and study staff at the investigator site, as well as at the sponsor site. The following sections give a summary of the roles and responsibilities to be undertaken by the sponsor, the monitor, and the investigator, who, apart from the subjects, are the main parties in a clinical trial.
49
7.7.1 SPONSOR RESPONSIBILITIES The sponsor is most often represented by the industry. The sponsor selects the investigators and should ensure that these have appropriate documented qualifications within the therapeutic area to be investigated. The sponsor is responsible for informing properly about the investigational medicinal products to be investigated. It is required that an investigator’s brochure is made available to the participating investigators prior to study start. An investigator’s brochure is a compilation of nonclinical and clinical data regarding the investigational medicinal product and includes all available safety data, e.g., nonclinical study results, reported adverse events. It is mandatory that written approval of the protocol has been obtained from regulatory authorities and ethics committees prior to initiation of a trial. It is the responsibility of the sponsor to ensure that all required approvals have been obtained. The sponsor must ensure that the investigational medicinal product has been manufactured according to good manufacturing practice (GMP), which is the quality assurance system applicable to drug manufacturing. Detailed accountability should be maintained for all investigational medicinal products from the production at the pharmaceutical company to the use by the individual subjects within the trial. At the end of the trial there should be a documented chain of custody from manufacturing to destruction of unused drug.
7.7.2 MONITOR RESPONSIBILITIES The monitor is a person appointed by the sponsor with the purpose of performing monitoring visits at the investigator site. Usually the monitor is an employee within the clinical research department at the sponsor site. The monitor should oversee the process of a clinical trial in order to verify that the rights and well-being of human subjects are protected, the reported trial data are accurate, complete, and verifiable from source documents, and the conduct of the trial is in compliance with the applicable regulatory requirements. The monitor should also verify that any adverse events or serious adverse events have been properly recorded, reported, and followed up. When performing data verification, the monitor needs direct access to the source data. This presumes that written informed consent has been obtained from each subject included in the trial. Data verification on all patient data is a requirement, and if a subject refuses to give access to the medical records, the subject cannot participate in the trial.
50
Handbook of Non-Invasive Methods and the Skin, Second Edition
7.7.3 INVESTIGATOR RESPONSIBILITIES The ICH-GCP defines several responsibilities to be undertaken by an investigator. The following gives an overview of the responsibilities and tasks to be performed by an investigator involved in a clinical trial: Qualification: The investigator should maintain curriculum vitae, documenting appropriate education, training, and experience, including knowledge in GCP, prior to involvement in a clinical trial. Investigational medicinal product (IMP): The investigator should be thoroughly familiar with the IMP as described in the investigators brochure provided by the sponsor. Furthermore, the investigator is responsible for the accountability of the IMP at the respective site. The investigator should maintain documentation concerning the IMP delivered at the site, the inventory at the site, the use by the individual subjects, and the return to the sponsor of unused products. It must be possible to reconcile all investigational medicinal products received from the sponsor. The investigator should ensure that the IMP is used in accordance with the approved protocol, and it should be documented that doses provided to the subjects were in accordance with the protocol. The IMP should be stored as specified by the sponsor and in accordance with applicable regulatory requirements with documented temperature monitoring during the storage period. Protocol compliance: The investigator should conduct the trial in compliance with the protocol, GCP, and applicable national legislation. To confirm the agreement, the investigator should sign the protocol, or an alternative contract. An investigator should never implement any deviations or changes to the protocol without prior written agreement with the sponsor and documented approval from the ethics committee. It should be stressed that if it becomes necessary to eliminate hazards to a trial subject, implementation of the change should be performed immediately at the discretion of the investigator. However, as soon as possible, the implemented deviation or change, the reason for it, and, if appropriate, a proposed protocol amendment should be submitted to the ethics committee, the sponsor, and the competent authorities, if required by national legislation. Monitoring: The investigator should permit monitoring and auditing by the sponsor and competent authorities and should be aware that the investigator is expected to be available for answering questions during the visits.
Delegation of duties: The investigator should maintain a list of persons to whom study-related duties have been delegated. It is the responsibility of the investigator to adequately inform any involved study staff at the investigator site about the study protocol, the investigational medicinal product, and the delegated duties. Resources: The investigator should demonstrate potential for recruiting the required number of subjects and have sufficient time to properly conduct and complete the trial within the agreed time period. Furthermore, there should be an adequate number of qualified personnel and facilities to conduct the trial safely and properly. Medical care: A qualified physician, who is an investigator or subinvestigator for the trial, must be responsible for all trial-related medical decisions. During and following a subject’s participation in a trial the investigator should ensure that adequate medical care is provided to the subject for any adverse events. The investigator must inform the subject when medical care is needed for intercurrent illness of which the investigator becomes aware. If a subject wishes to withdraw from the trial, the investigator should make reasonable efforts to ascertain the reasons — while fully respecting the subject’s rights. Ethics committees: Before initiation of a clinical trial the investigator must ensure that a written dated approval letter has been obtained from the ethics committee. The ethics committee approval must cover the trial period, the informed consent forms to be used in the trial, the subject recruitment procedures, and any other information to be provided to the subjects regarding the study. The investigator should provide a copy of the investigators brochure to the ethics committee and ensure submission of any updates regarding documentation relevant for the ethics committee during conduct of the trial. Randomization and blinding: When including subjects into a clinical trial, the investigator should follow the randomization procedure of the trial. The code is only to be broken in accordance with the protocol. If the trial is blinded, unblinding of the study may occur due to serious adverse events or by accident. In these cases, the investigator should promptly document and explain the unblinding to the sponsor. Informed consent of trial subjects: A critical task is to ensure that the subjects have been properly informed verbally as well as in writing and that informed consent has been obtained before inclusion in a trial. The informed consent form and any other trial-related written information to be
Standards for Good Clinical Practice (GCP)
provided to the subjects should be approved by the ethics committee prior to subject inclusion. The document should follow the guidelines of the hospital (if any) and should be revised whenever important new information becomes available that may be relevant to the consent of the subjects. All revised versions must be approved by the ethics committee prior to implementation. Information should be given in a timely manner should new information become available. The informed consent procedure as well as specification of the information to be provided to the subjects is detailed in the ICH-GCP. Records and reports: Records should be accurate, complete, legible, and timely pertinent to the data reported to the sponsor in the case report forms (CRFs) and other required documents. The data reported on the CRFs should be derived from source documents (e.g., medical records, x-rays, ECG) and should be consistent with the source documents, and all discrepancies should be explained. Any corrections to the entered data should be dated and initialed, explained, and should not obscure the original entry, whether the entry is written or electronic. Record retention requirements: The ICH-GCP requirement is that “essential documents should be retained until at least 2 years after the last marketing application in an ICH region and until there is no pending or contemplated marketing application in an ICH region or at least 2 years have elapsed since the formal documentation of clinical development of the investigational medicinal product.” This often means in practice “forever.” It is advisable to clarify the archiving procedures with the sponsor prior to involvement in a clinical trial in order for the investigator to become aware of the expectations concerning archiving at the investigator site. Surprises often arise when the investigator realizes that the hospital does not have sufficient archiving facilities or that hospital procedures require discarding of source documents after a defined period. Precautions should be taken in collaboration with the sponsor to ensure proper retaining of the source documents for the period required by ICH-GCP. Safety reporting: It is required that all serious adverse events (SAEs) are reported immediately to the sponsor except for those SAEs that the protocol or other document identifies as not needing immediate reporting. The immediate reports should be followed promptly by detailed written reports. The investigator should be aware of applicable regulatory requirements related to
51
reporting of unexpected SAEs to the competent authorities and the ethics committee. In addition to this, adverse events or laboratory abnormalities as identified in the protocol as critical to the safety evaluations should be reported to the sponsor within the periods specified in the protocol. In case of reported deaths, the investigator should supply the sponsor and the ethics committee with any additional requested information, e.g., autopsy reports, terminal medical reports. Premature termination or suspension of a clinical trial: If a trial is suspended or prematurely terminated for any reason, the investigator should promptly inform the trial subjects and should ensure appropriate therapy and follow-up where required. Depending on who decided to terminate the trial and for which reason, there are requirements to the reporting procedures. This is detailed in the ICH-GCP. Final reports: Upon completion of the trial, the investigator, where applicable, should inform the institution; the investigator/institution should provide the ethics committee with a summary of the trial outcome and the regulatory authorities with any reports required. The role of an investigator in a clinical trial is timeconsuming, often more than expected. Although some of the investigator duties as defined by GCP may be delegated to other staff members at the investigator site, the overall responsibility for the data at the site resides with the investigator. An investigator should be able to demonstrate active involvement in a trial; otherwise, the data may be refused by the authorities. One goal of GCP was to prevent or at least reduce the occurrence of fraud. It is unknown whether this has been obtained. The number of reported fraud cases is known, but the number of unreported and unknown cases of fraud is hard to discover. Although the time from development of a new drug until marketing has been prolonged during the last 30 years due to the increased requirements concerning the amount and quality of the documentation to be used for registration purposes, it is nice to realize that the number of subjects to be included in clinical trials sufficient to reach conclusive results has been reduced. The attitude of the authorities is generally that “if it is not documented, it never happened.” If a clinical trial is conducted in compliance with GCP, it should be possible many years hence to look at the records of the work and determine easily why, how, and by whom the work was done, who was in control, what equipment and methods were used, the results obtained, any problems that were encountered, and how they were overcome. It should
52
Handbook of Non-Invasive Methods and the Skin, Second Edition
be remembered that GCP is not simply a list of restrictive rules and regulations; much is common sense. It is important that the quality concept is integrated into every stage of the laborious clinical drug development process. It must be linked to every task undertaken and be included from the very first step of a project. Quality can never be introduced retrospectively.
REFERENCES 1. World Medical Association, Declaration of Helsinki. Recommendations Guiding Physicians in Biomedical Research Involving Human Subjects. Amended by the 52nd WMA General Assembly, Edinburgh, Scotland, October 2000. Available from: http:/www.wma.net/e/ policy/b3.htm. 2. CPMP/ICH/135/95 Step 5, Note for Guidance on Good Clinical Practice (CPMP adopted July 1996). Available from http://www.eudra.org/emea.html (juni1995).
Analysis of Sensitivity, 8 Statistical Specificity, and Predictive Value of a Diagnostic Test Nicholas Lange Biometric and Field Studies Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland
Martin A. Weinstock Dermatoepidemiology Unit, VA Medical Center, Roger Williams Medical Center, and Brown University, Providence, Rhode Island
CONTENTS 8.1 8.2 8.3
Introduction..............................................................................................................................................................53 Basic Concepts, Definitions, and Methods .............................................................................................................54 Limitations of Sensitivity, Specificity, and Predictive Value..................................................................................55 8.3.1 The Gold Standard.......................................................................................................................................55 8.3.2 The Clinical Context....................................................................................................................................56 8.3.3 The Artificial Dichotomy.............................................................................................................................56 8.4 Receiver Operating Characteristic (ROC) Curves ..................................................................................................56 8.4.1 Definition of the ROC Curve ......................................................................................................................56 8.4.2 Area under the ROC Curve .........................................................................................................................57 8.4.3 Correction of the ROC Curve for Verification Bias ...................................................................................58 8.4.4 Regression Methods.....................................................................................................................................60 8.5 Recommendations....................................................................................................................................................61 8.5.1 Considerations Other than Validity in Test Evaluation...............................................................................61 8.5.2 Specific Recommendations for Future Reports ..........................................................................................61 References ........................................................................................................................................................................61
8.1 INTRODUCTION In a perfect world, all of our diagnostic tests would be perfectly accurate — perfectly sensitive and specific — and 100% predictive of the disorder at issue. However, whereas perfect tests are all alike in that regard, every imperfect test is imperfect in its own way. Some will miss many cases, yet make few false diagnoses; others may miss few cases, yet falsely diagnose many. This chapter reviews appropriate methods for measuring that accuracy of tests. The example we use to illustrate these principles is the use of the aspartate aminotransferase (AST) test to diagnose hepatic fibrosis in patients receiving methotrexate therapy. Methotrexate is a very effective treatment
for severe recalcitrant psoriasis. However, its long-term use is limited by the occurrence of hepatic fibrosis, which can lead to cirrhosis of the liver and death. To avoid the adverse risk, a test is needed for the early stages of hepatic fibrosis so that the methotrexate can be stopped and clinical sequelae avoided. The present recommendation for monitoring includes periodic biopsies of the liver to determine the presence or absence of hepatic fibrosis. However, the biopsies themselves can have complications, so there is a need for a less invasive, safer procedure for determining whether hepatic fibrosis has developed. One such test is the AST test, a determination of the levels of aspartate aminotransferase in the blood. High levels suggest injury to the liver. 53
54
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 8.1 The Hepatic Fibrosis Example Hepatic Fibrosis? Elevated AST Level? Yes No Total
Yes
No
Total
15 8 23
3 24 27
18 32 50
TABLE 8.2 The Generic 2 × 2 Table True State Diagnostic Test Result Positive (d) – Negative (d) Total
Diseased (D)
Disease-Free – (D)
Total
a c a+c
b d b+d
a+b c+d n
The liver biopsy is viewed as the most accurate test for hepatic fibrosis. The AST test, either alone or in combination with other factors, is the alternative test for hepatic fibrosis used for the examples in this chapter. For simplicity, the numbers used in our examples are fictional. The reader is referred to an article by O’Conner and colleagues1 for actual observational data pertaining to the issues presented here.
8.2 BASIC CONCEPTS, DEFINITIONS, AND METHODS We begin by defining some basic terms and concepts. Many of these ideas are motivated by the generic 2 × 2 table. Table 8.1 gives an example of such a table, showing results for 50 fictitious patients cross-classified by their AST level and the presence of hepatic fibrosis. In order to understand more fully the properties of such a test, it is useful to abstract the clinical situation. Table 8.2 gives the generic form of the empirical cross-classification. In general, we use lowercase characters to denote observed quantities and events, and uppercase or Greek characters to denote theoretical quantities. For instance, p(·) is an observed probability or relative frequency, an empirical estimate of the theoretical probability P(·). In addition, the – symbol d denotes a positive test result, d a negative test – result, D a truly diseased state, and D a truly disease-free state. The observed prevalence of the disease is defined as p(D) = (a + c)/n, 0.46 or 46% in our example. Thus, the observed prevalence is a marginal probability, for this measure sums over the rows of the table, ignoring the test
result. If it can be assumed that the study population is representative of all such patient populations, then the observed quantities can serve as valid, accurate estimates of the corresponding true, theoretical quantities. In representative samples, the observed prevalence can be interpreted as an estimate of the probability that a randomly selected individual will have the disease, i.e., p(D) = P(D) on the average. If, on the other hand, the patient population under study cannot be assumed to be representative of all such patients, p(D) is a biased prevalence estimate, i.e., p(D) ≠ P(D) on the average. Selection bias would be present, for instance, if patients were either included or excluded according to criteria not accounted for in the cross-classification. In the following, however, until Section 8.3.1, we assume that the patients tested comprise a representative sample of all such patients. The observed sensitivity of the diagnostic test is defined as p(d|D) = a/(a + c), 0.65 in our example, the observed proportion of true positives among the diseased.* Similarly, the observed –– specificity of the diagnostic test is defined as p(d|D) = d/(b + d), 0.89 in our example, the observed proportion of true negatives.** Sensitivity and specificity are conditional probabilities as they include only those patients who are truly diseased or truly disease-free in their denominators, respectively. The overall accuracy of the diagnostic test is the sum of these two components weighted by the observed probabilities of the conditioning events, i.e., –– – p(d|D)p(D) + p(d|D)p(D). Inspection of the preceding conditional probabilities yields some useful relationships. A diagnostic test that is sensitive but not specific will correctly identify a large proportion of truly positive cases at the cost of labeling a large proportion of disease-free individuals as diseased. – In such a case, the proportion p(d|D) = b/(b + d) of false positives will be large. Conversely, a diagnostic test that is specific but not sensitive will correctly identify a large proportion of truly negative cases at the cost of not labeling a large proportion of diseased individuals as diseased; – the proportion p(d|D) = b/(b + d) of false negatives will be large. An extreme and unrealistic diagnostic test that declares every individual as diseased will not miss a single case and thus be completely sensitive, i.e., p(d|D) = 1, yet have a specificity of zero; similarly, a test that declares every individual as disease-free would be completely spe–– cific, i.e., p(d|D) = 1, yet have a sensitivity of zero. In the other extreme and more important case, a diagnostic – – test for which P(d|D) and P(d|D) are both zero, or both assumed to be zero, is called a gold standard for the disease in question. When using a diagnostic test that does * Read p(d|D) as “the observed proportion of patients testing positive given (‘|’) that they are truly diseased”, and similarly for other conditional probability statements. ** Note that the symbol d here denotes the number of disease-free patients that have a negative test result, not to be confused with the event d, a positive test result.
Statistical Analysis of Sensitivity, Specificity, and Predictive Value of a Diagnostic Test
not qualify as a gold standard, trade-offs are required between increasing sensitivity at the cost of decreasing specificity, and vice versa, in order to develop a test with optimal properties; see Section 8.4 on receiver operating characteristic curves for more detail on this point. The predictive value of a positive test result, or simply the positive predictive value, is defined as p(D|d) = a/(a + b). Similarly, the predictive value of a negative test – – result is defined as p(D|d) = d/(c + d). In our example, the positive predictive value is 0.83 and the negative predictive value is 0.75. Note the reversal of conditioning: given that the diagnostic test is positive, the positive predictive value is the proportion of truly positive cases, and similarly for negative test results and true negatives. A simple form of Bayes’ theorem relates the two types of conditional probabilities:
p (D d) =
p (d D) ⋅ p (D)
( ) ( )
p ( d D) ⋅ p (D) + p d D ⋅ p D
–– and similarly for p(D|d). This identity is verified trivially by inspection of Table 8.1 and Table 8.2.* In words, Bayes’ theorem applied here shows that positive predictive value =
edge of the test result. Equation 8.1 shows that this updating of prior information is given explicitly as posterior ∝ (likelihood) · (prior) with likelihood p(D|d) and constant of proportionality the denominator of Equation 8.1. This relationship can also be expressed in terms of odds ratios. Indeed, an odds ratio version of Bayes’ theorem is p (D d)
=
p ( d D) p ( D) ⋅ p dD p D
( ) ( ) ( )
p Dd
– with prior odds p(D)/p(D), 0.85 in our example, and like– lihood ratio p(d|D)/p(d|D), 5.9 in our example, and hence a posterior odds of 5.0.
8.3 LIMITATIONS OF SENSITIVITY, SPECIFICITY, AND PREDICTIVE VALUE In this section, we discuss the role of a definitive reference test, the clinical context, and the need to consider more than a simple normal/abnormal dichotomy.
8.3.1 THE GOLD STANDARD
( sensitivity ) ⋅ ( prevalence ) ( sensitivity ) ⋅ ( prevalence ) + (1 = specificity ) ⋅ (1–prevalence )
Equations 8.1 and 8.2 show exactly how the predictive value of a test depends on prevalence. Sensitivity and specificity, on the other hand, are conditional measures and thus do not depend on prevalence. The positive predictive value, p(D|d), is also called a posterior probability, as it is the probability estimate after the test result is known. It is thus an updated version of the prior probability p(D), the prevalence estimate, which is the a priori probability of disease prior to any knowl* Equation 8.1 is only the simplest, discrete form of Bayes’ theorem, however. When there are more than two possible true states (i.e., not – only D and D), the sum in Equation 8.1 gets more lengthy, until, in the limit, the state D is replaced by a parameter θ and the sum replaced by an integral over all possible parameter values, so that
p (θ d) =
55
p ( d θ) ⋅ p (θ)
∫
p ( d θ) ⋅ p (θ)
⋅ ∝ p ( d θ)
Although we do not develop this idea further here, the preceding relationship suggests Bayesian inference for the types of problems discussed in this chapter; interested readers should see the accessible article by Breslow2 on the subject.
Table 8.1 presupposes that we know who really has hepatic fibrosis, as Table 8.2 presupposes that for each person, we know whether the disease is actually present. The standard for these determinations is commonly named the gold standard or definitive reference test; i.e., the test that we assume is perfectly sensitive and specific. For hepatic fibrosis, the gold standard is generally the liver biopsy. The requirement for a gold standard raises several problems. Typically, there is some difficulty with the gold standard that motivates the search for a sensitive and specific alternative. This difficulty may also impede study of the proposed diagnostic test, such as the AST, and the gold standard in the same large group of patients. With the liver biopsy, the difficulties include a small but nonzero risk of mortality or serious morbidity as well as discomfort and expense. The gold standard itself may be an imperfect indicator of the disease being studied. This imperfection (inaccuracy) will decrease the observed sensitivity and specificity of the diagnostic test if the gold standard’s inaccuracy is independent of the diagnostic test’s result. However, inaccuracies in the gold standard may have opposite effects under other circumstances. Consider again our example of the AST test for hepatic fibrosis. If the pathologist who interpreted the liver biopsy was aware of the AST test when interpreting the histopathologic specimen, in
56
Handbook of Non-Invasive Methods and the Skin, Second Edition
equivocal cases he or she may have overdiagnosed hepatic fibrosis in the presence of an abnormal AST, and underdiagnosed the disease when the AST was normal, therefore increasing artificially the measured sensitivity and specificity of the AST test. A second type of bias would occur if, for instance, mild asymptomatic hepatitis caused both an increase in AST levels and a systematic overdiagnosis of hepatic fibrosis, and therefore an artifactual increase in measured sensitivity and specificity. Finally, it may be the case that the gold standard cannot be applied to all individuals subject to the diagnostic test due to logistical, cost, or ethical considerations, or due to comorbidity or risk of complications. If the result of the diagnostic test is used to determine which patients are subject to the gold standard evaluation, sensitivity and specificity estimates may be biased substantially; see Section 8.4.3. A gold standard may be unavailable, in which case measurements of validity, including sensitivity and specificity, become problematic. Nevertheless, validity may be tentatively assessed by measuring a surrogate endpoint known to be associated with the diagnostic test. In addition, test–retest reliability and correlation with other diagnostic tests for the disorder support validity.
normal. This approach has the advantage of producing simple, easy-to-understand results, yet it is clearly not as informative as the actual result reported on a continuous scale.
8.4 RECEIVER OPERATING CHARACTERISTIC (ROC) CURVES Changing the critical value of a test almost invariably changes both its sensitivity and specificity. It is useful, therefore, to understand the consequences of choosing particular critical values in more detail and to use this understanding to help make optimal choices. To address the artificial dichotomy problem, a receiver operating characteristic (ROC) curve for a diagnostic test is often developed.*
8.4.1 DEFINITION
OF THE
ROC CURVE
8.3.3 THE ARTIFICIAL DICHOTOMY
The ROC curve displays the range of sensitivities and specificities that are possible for a corresponding range of choices for the critical value. Let us assume that an AST value greater than a critical value z is labeled as a positive test result and that an AST value less than or equal to z is labeled as a negative test result. Table 8.3 shows the range of AST values found in our example, providing classifications of finer resolution than that shown in Table 8.1. If the total number of possible test results is equal to t(t = 13 in our example), then each row j defines a different cross-classification, a different 2 × 2 table, for each j = 1, …., t. The shaded cells in the fifth row of Table 8.3 (j = 5) define Table 8.1. Using the subscript j to indicate this 2 × 2 table, we thus find that a5 = 15, b5 = 3, c5 = 23 – 15 = 8, and d5 = 27 – 3 = 24, so that p5(d|D) = 15/(15 + 8) = 0.65, as noted previously. The test becomes less conservative as the critical AST value decreases, since a lower AST value is then required for the result to be reported as positive. In other words, sensitivity increases and specificity decreases as one scans the third and fifth columns from top to bottom. The ROC curve for these data is shown in Figure 8.1. This curve is a plot, for j = 1, …, t, of the observed sensitivities pj(d|D) = aj/(aj + cj) on the vertical axis against –– – 1 minus the observed specificities, 1 – pj (d|D) = pj (d|D) = bj /(bj + dj), on the horizontal axis. Scanning the rows in Table 8.3 from top to bottom corresponds to starting from the lower-left diagonal point (0, 0) in Figure 8.1 and
The terms sensitivity and specificity presume that a test result is simply normal (negative) or abnormal (positive). Typically, however, the test result is measured on a continuous scale and a critical value chosen to dichotomize the result. For instance, results of the AST test are reported initially in Système International units. Results above the critical value are labeled as abnormal, and results below,
* ROC curves were first developed in signal detection theory (Peterson et al.3). The “operating” end of this system (in this case, the true disease state) determines the signal sent to the “receiver” (the test result), together with noise; the task is then to determine what signal was actually sent. Each point on the curve describes the receiver’s criteria for distinguishing between signal and noise, and is called an operating position on the curve. A classic text on ROC techniques is that by Green and Swets;4 for medical applications, see, for instance, Metz.5
8.3.2 THE CLINICAL CONTEXT Sensitivity and specificity are used widely in part because these measures are independent of the prevalence of the disorder, as mentioned previously. However, this independence does not imply that sensitivity and specificity are constant. The diagnostic test may have different sensitivities and in different stages of the disease, or in different forms of the disease, and these differences may vary geographically. Similarly, a variety of illnesses, treatments, or medications may affect the performance of the test. The test’s validity may also depend on age, gender, socioeconomic status, ethnic background, and other demographic characteristics. Hence, when interpreting published measures of test validity, careful attention must be given to the clinical context in which such measures were determined. Test validity in a highly referred patient population may differ from that in a primary care setting; spatial and temporal factors may also affect validity. Replication of validity measures under diverse conditions is therefore quite helpful to establish the generalizability of results.
Statistical Analysis of Sensitivity, Specificity, and Predictive Value of a Diagnostic Test
57
TABLE 8.3 Data Table Used to Construct the Empirical ROC Curve Diseased Row (j) 1 2 3 4 5 6 7 8 9 10 11 12 13 Total
Critical Value of the AST Test Result, z >38 38 37 36 35 34 33 32 31 30 29 28 z 0 9 11 11 15 18 20 21 21 22 23 23 23
1.0
0.8
Disease-Free
0.2
0.4 0.6 1-specificity
0.8
1.0
FIGURE 8.1 The empirical ROC curve for the hepatic fibrosis example; from Table 8.3.
proceeding to the upper-right diagonal point (1, 1) in the plot. As indicated, a critical value of AST equal to 34 U/l shows an increase in sensitivity at no cost to specificity when compared with those values attained using a critical value of 35 U/l for this test. ROC curves help to remove the arbitrariness of clinical decision making by allowing one to investigate and control the critical value of a test to optimize decisions. ROC curves also facilitate the comparisons among
UNDER THE
ROC CURVE
Single-number summaries of ROC curves are useful for judging the validity of the test in question and also for making statistical comparisons of two or more ROC curves for competing diagnostic tests. Computing the area under an ROC curve is one way to reduce it to a single quantity. The area under an ROC curve (AUC), 0.8, for instance, can be interpreted as the probability (80%) that a randomly selected case from the diseased population will have a response to the diagnostic test worse (i.e., more abnormal, more positive) than that of a randomly selected individual from the disease-free population. For a completely uninformative test, AUC = 0.5; for a perfectly accurate test, AUC = 1.0. Estimates of the AUC can be obtained by parametric and nonparametric methods. A parametric method to compute the AUC (Dorfman and Alf7) assumes that the test results for the disease-free population are distributed as standard Gaussian (normal) with mean 0 and variance 1, and that the test results for the diseased population are distributed also as Gaussian with mean m and variance s2. Under these assumptions it has been shown that
58
Handbook of Non-Invasive Methods and the Skin, Second Edition
⎛ a AUC1 = Φ ⎜ ⎝ 1 + b2 with a =
⎞ ⎟, ⎠
μ 1 and b = σ σ
where Φ(·) is the standard Gaussian distribution function. If one makes the additional simplifying assumption that the points of the empirical ROC curve lie on the smooth curve defined by the Gaussian distribution functions, as they nearly do in our example, estimates of the parameters μ and σ2 can be obtained by routine methods; otherwise, iterative maximization routines are required. The simplifying assumption in our case yields estimates of the mean and variance of the diseased group, which are simply the sample mean and sample variance of the distribution of test responses in the diseased group after these responses have been centered and scaled by the mean and the variance of the responses in the disease-free group. Under that simplifying assumption, our example yields estimates of 2.028 and 0.971 and give an estimate AUC1 = Φ(1.455) = 0.9272. However, the maximum likelihood approach is to be preferred when the empirical ROC curve is not smooth. Maximum likelihood estimates, obtained from the program ROCFIT, courtesy of C. Metz, are 2.283 and 0.929 and give an estimate AUCML = Φ(1.443) = 0.9255. One can use the trapezoidal rule to obtain a simple nonparametric estimate of the AUC. If the ROC curve has been evaluated at t points, then the AUC can be decomposed into t – 1 trapezoidal strips (t = 13 in our example, 12 trapezoids). The area of each trapezoid is the product of the length of its base and its average height; the AUC is the sum of these areas. That is, for j = 1, …, t, the –– horizontal coordinates are xj = 1 – pj (d|D), i.e., the observed proportions of false positives at the jth level of the AST test, and the vertical coordinates are yj = pj(d|D), i.e., the observed proportions of true positives at the jth level of the AST test. The area under the ROC curve by the trapezoidal approximation is thus t –1
AUC2 =
∑(x j=1
j +1
– xj )
y j +1 + y j 2
The trapezoidal approximation to the area under the ROC curve shown in Figure 8.1 is AUC2 = 0.9147. A trapezoidal approximation such as AUC2 is smaller than the area under any smooth concave curve connecting the observed points, such as that assumed for AUC1 and AUCML. Also, the nonparametric estimate AUC2 is more sensitive to changes in the ordinates xj. It should be noted that the trapezoidal approximation reflects the fact that the results have been binned: there are ties among diseased and disease-free individuals at
identical AST values. Without such binning, the ROC curve has a staircase appearance, taking discrete jumps connected by horizontal and vertical line segments throughout. The nonparametric estimate of the AUC is thus underestimated by the binned trapezoidal approximation (see, for instance, Zweig and Campbell6). The magnitude of the underestimation depends, of course, on the number of ties and is an issue worthy of consideration when the results of a diagnostic test have been lumped into categories at a coarser resolution (e.g., definitely abnormal, probably abnormal, equivocal, probably normal, definitely normal) than those reported originally. Hanley and McNeil8 note the equivalence of the distribution of AUC2 to that of the Wilcoxon–Mann–Whitney statistic9: both measure the probability of correctly ranking a pair of disease-free and diseased patients among all such pairs in the study population. This equivalence is important, for it enables one to estimate the variance of AUC2 explicitly. Specifically, suppose that there are nD diseased and nD– disease-free patients. Let the symbol a = AUC2 denote the area under the ROC curve as approximated by the trapezoidal rule. Denote by q1 the estimated probability that two randomly selected and truly diseased patients are ranked below one randomly selected and truly disease-free patient. In addition, let q2 denote the estimated probability that one randomly selected and truly diseased patient is ranked below two randomly selected and truly disease-free patients. Then, q1 = a/(2 – a), q2 = 2a2/(1 + a), and the estimated variance of AUC2 is var [ AUC2 ] = var [ A ] =
(
1 nD nD
)
(
)
⋅ ⎡⎣ A (1 – A ) + ( nD – 1) q1 – A 2 + ( nD – 1) q2 – A 2 ⎤⎦
In our example, A = 0.9147, as stated previously. Application of the preceding approach to our example yields q1 = 0.8428 and q2 = 0.8740. Hence, the estimate of the area under the ROC curve has an estimated standard error of SE [ A ] = var [ AUC2 ] = 0.0019 = 0.0436 . A test of the null hypothesis that the true AUC is only 0.5 yields a z-score of z = (0.9147 – 0.5)/0.0436 = 9.512 using a standard Gaussian reference distribution and clearly rejects this null hypothesis.
8.4.3 CORRECTION OF THE ROC CURVE VERIFICATION BIAS
FOR
Note that ROC curve analyses may need to be corrected for verification bias in certain cases. Verification bias arises when the subjects used to assess the properties of the test have been selected in a nonrandom manner (Begg and Greenes,10 Gray et al.,11 Begg12). Proportions
Statistical Analysis of Sensitivity, Specificity, and Predictive Value of a Diagnostic Test
TABLE 8.4 The Cross-Classification of the Larger Population of Subjects from which the Verified Sample Shown in Table 8.1 was Drawn Hepatic Fibrosis? Elevated AST Level? Yes No Total
Yes
No
Total
15 16 31
3 48 51
18 64 32
of subjects selected for test validation from a larger population of subjects may differ in some systematic, nonrandom manner across the levels of test results. In general, if the selection bias favors inclusion of more subjects with high abnormal test results, then the reported sensitivity of the test is inflated artificially. Conversely, if selection bias favors inclusion of more subjects with low normal test results, then, in general, the reported specificity of the test would be inflated artificially. If selection bias favors higher proportions of subjects on both extremes of the test results, then both sensitivity and specificity are generally inflated. Verification bias can be corrected if the sampling fractions across the levels of test results are available. In our example, suppose that one knows the total number of available subjects at each level of the AST test results, only a fraction of whom have been selected for test verification. Such a possibility is shown in Table 8.4, giving the cross-classifications of a larger fictional population of subjects from which Table 8.1 was constructed. Table 8.4 is Table 8.1 with the second-row entries increased by a factor of 2. The subjects comprising Table 8.1 are a nonrandom sample from Table 8.4: those testing positive are sure to be included in the verification sample, whereas subjects with negative test results have only a 50:50 inclusion probability. Sensitivity has been inflated artificially. (Note, however, that predictive value remains unchanged; both positive and negative predictive values are not altered when a single row is multiplied by a constant.) Correction of the ROC curve analysis for verification bias employs Bayes’ theorem, as follows. Let the symbol s denote a selection indicator for positive and negative subjects, so that p(s|d) = 18/18 = 1.00 for those testing positive — sure – inclusion — and p(s|d) = 32/64 = 0.50 for those testing negative. By definition, one does not know the true disease status for those subjects not selected for verification. One knows only that fraction testing positive out of the total number of subjects, i.e., p(d) = 18/82 or 22%, and, of the selected subjects testing positive, what fraction were found to be truly diseased, i.e., p(D|d, s) = 15/18 or 83%, as given previously in Table 8.1. Similar calculations apply for those testing negative and selected. One makes
59
the additional assumption, plausible under the null hypothesis of no association, that true disease state and selection are conditionally independent given the test result. In other words, assume that P(D, s|d) = P(D|d) · P(s|d), implying P(D|d) = P(D|d, s), and similarly for – P(D|d, s). Then, by Bayes’ theorem, the observed sensitivity corrected for verification bias is
p * (d D) =
p ( D d, s ) ⋅ p ( d )
(
) ()
p ( D d, s ) ⋅ p ( d ) + p D d, s ⋅ p d
⎛ 15 ⎞ ⎛ 18 ⎞ ⎜⎝ ⎟⎠ ⋅ ⎜⎝ ⎟⎠ 18 82 = ⎛ 15 ⎞ ⎛ 18 ⎞ ⎛ 8 ⎞ ⎛ 64 ⎞ ⎜⎝ ⎟⎠ ⋅ ⎜⎝ ⎟⎠ + ⎜⎝ ⎟⎠ ⋅ ⎜⎝ ⎟⎠ 18 82 32 82 = 0.484 matching the sensitivity indicated in the complete data in Table 8.4. Similar calculations yield a corrected specificity of 0.941, also matching the specificity indicated in Table 8.4. Table 8.5 is a separate example of verification bias at a higher resolution, at a finer level of detail. Table 8.3 has been extended in Table 8.5 to include two additional columns that give the total numbers of subjects, twice the number shown in Table 8.1, and fractions selected at each level of the test results. (Results given in Table 8.1 are again indicated by shading.) The fractions of selected subjects with high AST scores are consistently greater than the fractions of selected subjects with mid-level and low AST scores: the ROC curve analysis is thus again susceptible to verification bias. To see such biases in finer detail, we need a little more notation. Denote by qk(·) that fraction of patients out of the total who have AST result at level k, and let qk(·|·, s) denote that fraction of patients giving positive or negative test results out of those included. For instance, at k = 4, qk(d) = 7/100, qk(D|D, s) – – – = 4/(4 + 1), q4(D|d, s) = 1/(4 + 1), and, again, q4(d) = 7/100. (This latter fraction is equal to q4(d): the fourth row in Table 8.5 represents the same fraction of available subjects regardless of their disease states. For each k, the – fractions qk(d) and qk(d) are identical.) Observed sensi* tivities p j (d|D) corrected for verification bias are thus j
∑ q ( D d, s ) ⋅ q ( d) k
p (d D) = ∗ j
k
k =1 t
∑ q ( D d, s ) ⋅ q ( d) k
k =1
k
,
j = 1,..., t
60
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 8.5 Augmented Data Table Used to Correct the ROC Curve for Verification Bias Verification Sample Critical Value of the AST Test Result, z
Row (j) 1 2 3 4 5 6 7 8 9 10 11 12 13 Total
Diseased Subjects With AST = z
>38 38 37 36 35 34 33 32 31 30 29 28 z
9 2 0 4 3 2 1 0 1 1 0 0 0 23
0 9 11 11 15 18 20 21 21 22 23 23 23
0 0 2 1 0 1 3 2 1 4 9 2 2 27
Total Patient Population
Subjects With AST ≤ z 27 27 27 25 24 24 23 20 18 17 13 4 2
Available
Fraction Selected
9 2 2 6 4 4 6 4 5 14 28 7 9 100
1.00 1.00 1.00 0.83 0.75 0.75 0.67 0.50 0.40 0.36 0.32 0.29 0.22 0.50
35 and 34 U/l demonstrate inflated sensitivities of the uncorrected results at the cost of deflated specificities.
1.0
0.8
8.4.4 REGRESSION METHODS
34 34 35
Sensitivity
Subjects With AST = z
0.6 35
0.4
Corrected Uncorrected
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1-specificity
FIGURE 8.2 The corrected and uncorrected ROC curves for the hepatic fibrosis example; from Table 8.5.
yielding corrected vertical coordinates for the ROC curve. Similar weighted partial sums over disease-free patients given observed negative test results and the inclusion of these patients in the verified sample yield corrected hori–– zontal coordinates, p*j (d|D), j = 1, …, t. Figure 8.2 shows the corrected and uncorrected ROC curves deriving from Table 8.3 and Table 8.5. Although the corrected ROC curve appears better overall, the indicated AST values of
In addition to the ROC curve methods discussed here, there are several multivariable methods that should be mentioned, including ordinal regression methods (see McCullagh13) and logistic regression methods (see, for instance, Hosmer and Lemeshow14 and Hunink and Begg15). The purpose of these methods is to calculate a single number that will predict the presence or absence of the disorder better than the diagnostic test alone. In order to do so, this single number is a function of the test result as well as other factors that are correlated with disease state, such as gender, age, and other intrinsic and extrinsic effects mentioned in Section 8.3.2. The multivariable function that produces this single number is termed a prediction rule. The prediction rule can be subjected to ROC analysis, and its AUC compared with that of the original diagnostic test alone. For instance, if gender, age, etc., are correlated with the outcome of the diagnostic test, then logistic regression methods allow for the inclusion of covariates x in a multivariable model that improves predictive performance. Estimated coefficients from the fit of such a model are combined with covariate values linearly to yield a prediction rule with superior performance, i.e., higher values of p(D|d, x) than those attainable through ROC analyses that do not accommodate such factors.
Statistical Analysis of Sensitivity, Specificity, and Predictive Value of a Diagnostic Test
8.5 RECOMMENDATIONS We conclude the chapter with a brief discussion of additional elements of test evaluation and our specific recommendations for future reports.
8.5.1 CONSIDERATIONS OTHER TEST EVALUATION
THAN
VALIDITY
IN
Several considerations pertain to the assessment of tests beyond quantification of test validity. If the reasons for the observed inaccuracies are discovered, then it may be possible to enhance validity or to identify settings in which validity is greatest and poorest. It is important to assess both the reliability and validity of the test, since poor reliability may be an important source of inaccurate individual test results. It is also important to determine test performance outside of the research setting and under the conditions used by others. Issues related to associated costs and health risks must also be considered. Results of cost–benefit and cost-effectiveness analyses may be crucial to the test’s adoption in settings other than one’s own. The test itself may have an impact on the clinical setting in which it is used; determination of appropriate circumstances for the test must be made. Devising a better test may not be worthwhile if test results will not affect significant therapeutic decisions. Finally, the consequences of inaccurate test results must also be considered. If a diagnostic test is less costly yet less accurate, then the consequences of errors that may follow (e.g., suffering, disability, and death) may outweigh the savings in monetary cost. In some circumstances, a formal decision analysis that takes into account all of the preceding issues may be worthwhile.
8.5.2 SPECIFIC RECOMMENDATIONS REPORTS
FOR
FUTURE
The following 12 criteria may be useful in evaluation of future study reports that claim to validate diagnostic tests (from Weinstock16): 1. The test and other potential predictors of the disorder, the disorder to be diagnosed, and the gold standard for diagnosis of the disorder are defined with sufficient clarity and detail that an independent replication of the study may be conducted. 2. The population for which the test was validated is described. This includes the spectrum of disorder among those affected, the diagnoses among those not affected, demographic and medical data, selection criteria for the test and evaluation with the gold standard, and any relevant referral patterns.
61
3. Statistical methods are described, applied appropriately, and cited in the literature clearly. 4. The test is interpreted blindly with respect to the gold standard diagnosis and also to the proposed test. In some circumstances, it may be reasonable to assess whether attempts to blind the observers were successful. 5. The reliabilities of the test and of the gold standard are estimated and reported. 6. The sensitivity, specificity, and predictive values of the test are calculated, and where appropriate, the dependence of these characteristics on other medical or demographic factors is estimated. 7. The ROC curve for the test is presented and the area under the ROC curve is also calculated and reported, if the test results are reported on an ordinal or interval scale. 8. The relation of the test to other predictors of the disorder, including results of other tests, is assessed. The incremental value of the test is determined and prediction rules are considered. 9. If a prediction rule is suggested, it is defined precisely and its validity is evaluated and reported. If feasible, its validity is determined in a group not used to derive the prediction rule, or other statistical techniques are used to estimate its validity in such groups. The derivation of the prediction rule is described clearly, including variables considered and inclusion criteria. 10. Consideration is given to the generalizability and possible sources of bias. 11. Consideration is given to mechanisms that may account for the observed inaccuracies. 12. Consideration is given to the impact of the test in practice, i.e., effect on treatment decisions, consequences of inaccuracies, testing complications, costs, and benefits. As diagnostic testing becomes more sophisticated, more costly, and scrutinized more intensively, proper attention to quantification of validity becomes crucial to test adoption, dissemination, and appropriate use.
REFERENCES 1. O’Conner, G.T., Olmstead, E.M., Zug, K., Baughman, R.D., Beck, J.R., Dunn, J.L., and Lewandowski, J.F., Detection of hepatotoxicity associated with methotrexate therapy for psoriasis, Arch. Dermatol., 125, 1209, 1989. 2. Breslow, N., Biostatistics and Bayes (with comments), Stat. Sci., 86, 557, 1990.
62
Handbook of Non-Invasive Methods and the Skin, Second Edition
3. Peterson, W.W., Birdsall, T.G., and Fox, W.C., The theory of signal detection, Trans. IRE Prof. Group Inf. Theory, PGIT-4, 171, 1954. 4. Green, D.M. and Swets, J.A., Signal Detection Theory and Psychophysics, revised edition, Krieger, Huntington, NY, 1974. 5. Metz, C.E., Basic principles of ROC analysis, Semin. Nucl. Med., 8, 283, 1978. 6. Zweig, M.H. and Campbell, G., Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine, Clin. Chem., 39, 561, 1993. 7. Dorfman, D.D. and Alf, E., Maximum likelihood estimation of parameters of signal detection theory and determination of confidence intervals: rating method data, J. Math. Psychol., 6, 487, 1969. 8. Hanley, J.A. and McNeil, B.J., The meaning and use of the area under a receiver operating characteristic (ROC) curve, Radiology, 143, 29, 1982. 9. Colton, T., Statistics in Medicine, Little, Brown & Company, Boston, MA, 1974.
10. Begg, C.B. and Greenes, R.A., Assessment of diagnostic tests when disease verification is subject to selection bias, Biometrics, 39, 206, 1983. 11. Gray, R., Begg, C.B., and Greenes, R.A., Construction of receiver operating characteristic curves when disease verification is subject to selection bias, Med. Dec. Making, 4, 151, 1984. 12. Begg, C.B., Biases in the assessment of diagnostic tests, Stat. Med., 6, 411, 1987. 13. McCullagh, P., Regression models for ordinal data (with discussion), J. R. Stat. Soc. Ser. B, 42, 109, 1980. 14. Hosmer, D.W. and Lemeshow, S., Applied Logistic Regression, John Wiley & Sons, New York, 1989. 15. Hunink, M.G. and Begg, C.B., Diamond’s correction method: a real gem or just cubic zirconium, Med. Dec. Making, 11, 201, 1991. 16. Weinstock, M.A., Validation of a diagnostic test, Arch. Dermatol., 125, 1260, 1989.
9 Sample Size Calculation Claus Bay and Susanne Møller Mathematical Statistical Department, LEO Pharma, Ballerup, Denmark
CONTENTS 9.1 9.2 9.3
Introduction..............................................................................................................................................................63 Sample Size and Power ...........................................................................................................................................64 t-Test for Continuous Data ......................................................................................................................................64 9.3.1 Example with Independent Data .................................................................................................................64 9.3.2 Example with Paired Data...........................................................................................................................65 9.4 Chi-Square Test for Dichotomous Data ..................................................................................................................65 9.4.1 Example with Independent Data .................................................................................................................65 9.5 Discussion and Recommendations ..........................................................................................................................65 References .........................................................................................................................................................................66
9.1 INTRODUCTION Planning a scientific experiment consists of deciding on a number of equally important components: main purpose, design, variables to measure, methods of measurements, and hypotheses to test. The statistical planning is an integral part of this process. The statistical methods and tests of the specified hypotheses should be considered at an early stage in the process. An important design feature is the number of observations on a particular response variable that the experiment should produce. The consequence of not considering this in relation to the questions the experimenter wants the study to answer may very well be that the size of the study is inadequate and therefore turns out to be inconclusive. The sample size must be motivated by statistical considerations concerning difference to detect and power of statistical test (e.g., as stated in the EEC guidelines for Good Clinical Practice for Trials on Medical Products [GCP]1 in Section 9.4.3). The idea underlying the calculation of sample size of an experiment is to be able to detect, with a suitably high probability, an important difference between experimental units (e.g., treatment groups in a clinical trial), if such a difference exists. Thus, the experimenter should decide on the size of the difference he would not like to overlook (often called the minimal clinical relevant difference) and how sure he would like to be on the decision made from the study, i.e., the probability of finding a difference if it really exists (= the power) and the probability of finding a difference
when no true difference exists (= level of significance). Also, it should be decided how to measure the difference. When a number of variables are measured, it is preferable to name one variable or derived parameter to be of primary interest and consider the rest of secondary importance. The sample size is then determined based on this primary variable. If this procedure is considered infeasible, the sample size can be determined for all those variables considered of primary interest, and the study is dimensioned according to the largest of the calculated sample sizes, in this way ensuring at least the specified power in respect of all variables. The following sections will give a simple introduction to how sample size calculations can be done for most experiments and examples to illustrate the methods that will be given. Only two-sided tests are considered, since one-sided tests are rarely used in practice. It is emphasized that the formulas are only approximate, but for the vast majority of cases the gain in accuracy by applying the exact methods is negligible and need only to be considered for small sample sizes of less than about 20 for continuous data. It is beyond the scope of this chapter to go into a detailed discussion of the mathematical statistical theory from which the following formulas are derived. The interested reader is referred to Desu and Raghavarao2 for a thorough introduction to sample size methodology. Altman3 gives an excellent introduction to statistical methods in general for medical research, and in Cox and Hinkley4 the more theoretical aspects of test theory can be found. The special case of bioequivalence studies in general is dealt with in Chow and Liu.5 63
64
Handbook of Non-Invasive Methods and the Skin, Second Edition
9.2 SAMPLE SIZE AND POWER To be able to understand the calculation of sample size, it is important to understand the concept of power of a statistical test. Consider the situation where the means μ1 and μ2 of two normal distributed statistics are compared by a U-test. The test statistic |U| is the numerical value of the difference between the means divided by its standard error, distributed as N(0, 1). The conventional null hypothesis is then H0, μ1 – μ2 = 0, against the alternative hypothesis HA, μ1 – μ2 ≠ 0. The p value of the test is the probability of having observed the actual data (or more extreme data) when H0 is true, i.e., true means are equal, P = P(|U| > u1–α/2| μ1 = μ2), where u1–α/2 is the 1 – α/2 fractile of the standard normal distribution function (e.g., a = 0.05, u1–α/2 = 1.96, and P(|U| > 1.96|μ1 = μ2) = 0.05). The cutoff point for statistical significance, the significance level, denoted by a, is equal to the probability of rejecting the null hypothesis when it is true, i.e., the probability of obtaining a false positive result. This value is also referred to as a type I error. From the specification of the alternative hypothesis it appears that a whole range of values of μ1 and μ2 are contained in the set defined by HA. For every pair of values, the probability of accepting H0 when it is false is β = P(|U| < u1–α/2| μ1 – μ2 ≠ 0), i.e., the probability of a false negative result can be calculated. This value is called a type II error. The power function f(α, δ) = P(|U| > u1–α/2| μ1 – μ2 = δ) is monotonously increasing with increasing value of δ. Note that the power function is equal to α for δ = 0. This follows from f(α, 0) = P(|U| > u1–α/2| H0) = α. For a specific alternative H1 defined by μ1 – μ2 = δ0, the type II error, β, is equal to 1 – f(α,δ0). The quantity 1 – β is called the power of the test. When β is high, the power 1 – β is low, and it is unlikely that a true difference will be detected, i.e., yielding a statistically significant result of the statistical test, as the probability of failing to reject H0 is high even when H0 is false. This is obviously not desirable for the experimenter, so the aim is to have high power. Let us now examine the general situation where the test statistic X is normal distributed as N(0, Σ 20 ) under the null hypothesis, H0, and normal distributed as N(δ, Σ12 ) under the alternative hypothesis, H1, where δ < 0 or δ > 0. Often Σ 20 and Σ12 will depend on N: Σ 20 = σ 20 /N and Σ12 = σ12 /N. With the significance level a and power 1 – β, Lachin6 shows that the sample size in this case is obtained by ⎡ u1–α /2 σ 0 + u1–β σ1 ⎤ N ( α, β, δ ) = ⎢ ⎥ δ ⎣ ⎦
2
TABLE 9.1 Commonly Used u-Fractiles Significance Level, α
u1-α/2
Type II, Error, β
u1–β
1% 5% 10%
2.58 1.96 1.65
20% 10% 5%
0.84 1.28 1.65
In the following it will be shown how this simple expression can be used in the case of Student’s t-test for equality of means of normal distributed variables with unknown variance and for chi-square tests for proportions. The corresponding u-values for commonly used values of a and b appear in Table 9.1 (e.g., α = 5%, u1–α/2 = 1.96; β = 20%, u1–β = 0.84). With these and the above formula, the sample size can be determined for most experiments.
9.3 t-TEST FOR CONTINUOUS DATA Let (xi)i=1,…,N and (yi)i=1,…,N be sets of mutually independent observations from normal distributions with means m1 and m2, respectively, and variance σ2. The usual t-test statistic is T = ( x – y ) s 2 N where x and y are the averages of the xis and yis, respectively, and s is the pooled estimate of the standard deviation σ from the two samples. The difference x – y is normal distributed as N(0, 2σ2/N) under H0 and normal distributed as N(δ0, 2σ2/N) under H1. In this case the sample size of each group necessary to have power 1 – β to detect a difference equal to or greater than δ0 is ⎡ ( u1–α /2 + u1–β ) s ⎤ N = 2⎢ ⎥ δ0 ⎢⎣ ⎥⎦
2
The variances are equal under H0 and H1, and σ0 and σ1 are substituted by the estimate s. Because the true variance is unknown and replaced by an estimate of the variance, the formula is only approximate since T is t distributed. The exact value of N is obtained by substituting the u-fractiles with the corresponding t-fractiles. N is then found by iteration.2 For paired data, s is the estimate of the intrasubject standard deviation.
9.3.1 EXAMPLE
WITH INDEPENDENT
DATA
Consider a clinical study in psoriatic patients. Two topical treatments are compared in a parallel group design. The response variable is the percentage reduction in psoriatic area and severity index (PASI; Frederikson and Petterson7) from start of treatment to end of treatment. The betweenpatient standard deviation of the percentage change in PASI is assumed to be 35%-points. The clinician wants
Sample Size Calculation
65
with 80% probability (power) to detect a true difference between treatments of 10%-points. The sample size calculation gives 2
⎡ (1.96 + 0.84 ) 35 ⎤ N = 2⎢ ⎥ = 192 patients in eaach group 10 ⎦ ⎣
N= ⎡ u1–α /2 2 π 0 (1 – π 0 ) + u1–β π1 (1 – π1 ) + π 2 (1 – π 2 ) ⎤ ⎥ ⎢ π1 – π 2 ⎥⎦ ⎢⎣
2
Extensive tabulations of the sample size using the exact distribution can be found in Fleiss.8
9.3.2 EXAMPLE
WITH
PAIRED DATA
Consider an experiment in psoriatic patients where two topical treatments are compared in a right/left design. The response variable is the change in skin thickness measured by ultrasound scanning from start of treatment to end of treatment. The intrapatient standard deviation of the change in skin thickness is assumed to be 0.25 mm. The clinician wants with 80% probability to detect a true difference between treatments of 0.1 mm. The total sample size is then 2
⎡ (1.96 + 0.84 ) 0.25 ⎤ N = 2⎢ ⎥ = 98 patients 0.10 ⎣ ⎦
9.4 CHI-SQUARE TEST FOR DICHOTOMOUS DATA Let (xi)i=1,…,N and (yi)i=1,…,N be sets of mutually independent observations from binomial distributions with means (probabilities) π1 and π2, respectively. These parameters are estimated by the observed frequencies p1 and p2. The usual chi-square test statistic is used to test the null hypothesis that π1 = π2 = π0. The general formula for sample size can be applied by noting that the square root of a chi-square distributed variable with one degree of freedom is normally distributed. In this case the standard error of the difference p1 – p2 depends on the values of the proportions. The standard deviations of p1 – p2 under H0 and H1 are estimated by σ 0 = 2 π 0 (1 – π 0 ) where π 0 = ( π1 + π 2 ) 2 σ1 = π1 (1 – π1 ) + π 2 (1 – π 2 ) Using these expressions, the sample size for each group necessary to have power 1 – β against the alternative given by (π1, π2) is
9.4.1 EXAMPLE
WITH INDEPENDENT
DATA
Consider a clinical study in psoriatic patients. Two topical treatments are compared in a parallel group design. The response variable is the proportion of patients who achieve a marked improved or cleared status at the end of treatment according to the investigator’s overall assessment of treatment response. It is assumed that the average proportion of response according to this criteria is about 60%. The clinician wants with 80% probability to detect a true difference between treatments of 20%-points corresponding to π1 = 0.5 and π2 = 0.7. The sample size calculation gives N =
⎡ 1.96 ⎢ ⎣
2 0.6 (1 – 0.6 ) + 0.84 0.7 (1 – 0.7 ) + 0.5 (1 – 0.5 ) ⎤ 0.7 – 0.5
⎥ ⎦
2
= 93
patients in each group.
9.5 DISCUSSION AND RECOMMENDATIONS We have seen from the previous examples that the calculation of sample size is technically very simple and can be done with the use of a pocket calculator. The difficulties lie in specifying the different elements of the calculations. Many statisticians have experienced that making an experimenter decide on the power and the difference to detect is very difficult — thus the emphasis in the previous section to explain the concept of power of a statistical test. The fixing of power at 80 or 90% is, of course, in many cases rather arbitrary, but it is recommended to carry out experiments with at least 80% power of the statistical test, the reason being that the power can be thought of as the receiver’s (experimenter’s) risk of getting a defective item (inconclusive result). The decision about the difference to detect will depend on the purpose of the experiment. Suppose a novel drug to treat patients with a disease where no effective treatments are available, then even a small effect may be of clinical relevance. Conversely, in patients suffering from
66
Handbook of Non-Invasive Methods and the Skin, Second Edition
a disease with several efficacious treatments, a new drug may only be of interest if it shows a much better effect than placebo in a placebo-controlled trial. In any case, the difference that one can detect with a realistic number of observations depends, as we have seen, on the standard deviation of the response variable. How should the standard deviation be determined? The ideal situation is that it is well established from documented experiments with the same response variable, associated measuring technique, and comparable subject populations. However, this is often not the case. One solution is to perform a small pilot study to obtain an initial estimate of the standard deviation. If this is infeasible, it is the author’s experience that, as a rule of thumb, the biological between-individual coefficient of variation (CV) of many parameters is about 40%. Another problem that often arises is that in published data only the interindividual standard deviation is reported. This is not of much help in an experiment where differences in responses are measured within individuals, since the variance of the difference between related (paired) observations is equal to 2σ2 – 2τ, where σ is the interindividual standard deviation and τ is the (unknown) covariance. Often the covariance will be at least half the size of the interindividual variance, so in many cases, an upper limit for the relevant intrasubject variance σ 2int ra is s2, the between-individuals variance. Thus, the estimate of
the interindividual variance can, in many cases, be used as an estimate of the unknown intraindividual variance.
REFERENCES 1. CPMP Working Party on Efficacy of Medical Products, Good Clinical Practice for Trials on Medical Products in the European Community, Note for Guidance, Commission of the European Communities, 1990. 2. Desu, M.M. and Raghavarao, D., Sample Size Methodology, Academic Press, San Diego, 1990. 3. Altman, D.G., Practical Statistics for Medical Research, Chapman & Hall, London, 1991. 4. Cox, D.R. and Hinkley, D.V., Theoretical Statistics, Chapman & Hall, London, 1974. 5. Chow, S.-C. and Liu, J.-P., Design and Analysis of Bioavailability and Bioequivalence Studies, Marcel Dekker, New York, 1992. 6. Lachin, J.M., Introduction to sample size determination and power analysis for clinical trials, Controlled Clin. Trials, 2, 93, 1981. 7. Frederikson, T. and Petterson, U., Severe psoriasis: oral therapy with a new retinoid, Dermatologica, 157, 238, 1978. 8. Fleiss, J.L., Statistical Methods for Rates and Proportions, 2nd ed., John Wiley & Sons, New York, 1981.
of a Quality 10 Implementation Management System in a Contract Laboratory Working with Non-Invasive Methods Klaus-Peter Wilhelm and Jutta Hofmann proDERM Institute for Applied Dermatological Research, Schenefeld/Hamburg, Germany
CONTENTS 10.1 Introduction..............................................................................................................................................................67 10.2 What Is Quality?......................................................................................................................................................67 10.3 Quality Management and the Principles of ISO 9001:2000 ..................................................................................68 10.4 Quality in the Skin Bioengineering Laboratory .....................................................................................................69 10.5 Performance Assessment in the Skin Bioengineering Laboratory .........................................................................72 References .........................................................................................................................................................................72
10.1 INTRODUCTION
10.2 WHAT IS QUALITY?
The concept of quality has changed extensively during recent decades. The traditional thinking of quality management, which comes from the manufacturing industry, was primarily result oriented and focused on testing the output. More recent quality concepts have taken into account that “quality begins in the head” and that early analysis, planning, and arrangement of the development and production process helps to prevent subsequent errors.1,2 Prevention rather than cure is the adage, and it is applicable to all industries, including the service sector. In a climate of increasingly worldwide competition, quality management assumingly will play a key role to implement innovations rapidly and economically and to respond to the requirements of customers with appropriate stateof-the-art products and services.3 This chapter intends to illustrate the implementation of a quality management system in a dermatological contract research institute. It deals with the idea of what quality is and outlines briefly the process-based ISO 9001:2000 quality management philosophy. Additionally, the factors influencing quality in a laboratory are described and their significance highlighted. The last section provides a concept about how performance of business may be measured and evaluated using the International Organization for Standardization (ISO) standard.
It appears to be a simple question, but the answer is not as simple as expected. In the everyday context, like beauty, everybody may have their own idea of what quality is. At its simplest, quality can be defined by attributes related to the product or by its fitness for use. These categories of quality have measurable characteristics; their rating leads to the evaluation of the products’ value, and subsequently to the conclusion that one product is better or worse than another product in terms of conformity with the demands at stake. The argument not to buy might be that, for example, a car is not fast enough or that one test for mad cow disease is more sensitive than another. The complexity of what quality is nowadays is presented in the standardized ISO definition: quality is the “degree to which a set of inherent characteristics fulfills requirements.”4 Generally speaking, quality occurs when all those features of a product or service that are required by the customer are met. What does this imply for a skin bioengineering laboratory? Often the sponsor will not explicitly mention basic requirements, but silently assume their fulfillment.3 For example, a sponsor will expect as a matter of fact that a clinical study will fulfill the good clinical practice standard as well as applicable laws and statutory regulations based on these laws in the country in which the 67
68
Handbook of Non-Invasive Methods and the Skin, Second Edition
investigation takes place or that sound scientific practices will be applied. Accessory requirements such as timelines or format of report or presentation of results may be less clear. However, these characteristics offer the sponsor the opportunity to satisfy its need for individuality and the laboratory the opportunity to make its service stand out from those of the competitors.3 For a good reputation, quality is fundamental, and those who have it will find all doors open to them. If the sponsor returns and the report does not, it can, in our experience, safely be assumed that quality in terms of customer satisfaction has been achieved.
10.3 QUALITY MANAGEMENT AND THE PRINCIPLES OF ISO 9001:2000 Changing the interpretation of quality has changed the concepts of quality management as well. In the beginning, in the 1930s, there was the concept of quality control. Products were manufactured in the production line, and at the end of the line people inspected the finished products whether they complied with distinct requirements or not. If products were defective, they were thrown away or reworked. If they adhered to the specification, they were shipped to the customer. This strategy made no contribution at all to avoid or eliminate the origins of the fault. The next more effective step in the 1950s was the quality assurance approach. “Quality built in” was the motto; in-process quality control, project management, and independent audits were part of the concept. Error rates could be lowered; quality in terms of conformance with quality requirements was ensured. Competition and the “financial lid” have prompted quality initiatives extending beyond those traditional goals. Providing better quality at equivalent cost or equivalent quality at lower cost is one aim of today’s quality management. A graphical overview of the history of quality management is given in Figure 10.1. How can this purpose be achieved? As already indicated, the perception of quality depends on the customer
and on the product or service delivered. Additionally, the company’s activities to create these products or services have to be taken into account. Hence, a system to manage quality cannot be objective in the sense that every time everybody can identically reproduce the system of a company in any given situation. Each quality management system is an individual solution defined and recorded in the handbook. However, the philosophy is that the requirements for quality management systems are generic, that requirements for what a company must do to manage quality can be formulated.5 Those essential features are clarified in the ISO 9000 series of standards, which were first published in 1987, 40 years after the foundation of ISO. In December 2000, the revised ISO 9001:2000 standard was issued. Currently ISO 9000 is the most widely used series of standards for quality management worldwide.3 The standard is written without bias to manufacturing or service, but emphasizes acknowledged key factors for the management of quality.5 Key factors and some ground rules for the implementation of the ISO approach are shown in Table 10.1. Hence, quality is a management task; process management is regarded as a complement to line management and project management. In addition to the top-down approach, competent and motivated staff who accept this initiative and put it from the bottom up into practice are needed. What does managing by process mean? One of the keys to successful process management is the separation of the functional view of the organization (What has to be done?) from the operational view (How it is done?).3 A process should be stable despite organization and technology changes, whereas a procedure may be regarded as unstable, subject to change through continuous improvement. A process is generally characterized by utilizing resources to transform inputs to outputs.4 In the view that the activities of the company are a series of cooperating processes that are dependent upon each other in order to achieve quality, inputs of processes frequently are outputs from upstream processes. 3 Within a company the
Quality assurance
Quality control Verification that the requirements of quality are fulfilled Control of the final product
Quality management
System to ensure conformance with quality requirements
Corporate responsibility, design and management of the whole creation of value chain
Control at critical points during the process, modern project management and auditing to guarantee quality
Management responsibility and involvement of staff at all levels in planning and control, continual improvement
1950
FIGURE 10.1 Quality management. Brief overview of the history.
2000
Time
Implementation of a Quality Management System in a Contract Laboratory Working with Non-Invasive Methods 69
TABLE 10.1 Management of Quality: Key Factors and Ground Rules for the Implementation of the ISO Approach Key Factor
Ground Rules
Management responsibility
Demonstrate that executive management takes the responsibility for the adequacy of the quality system Develop clear beliefs and objectives and show commitment Encourage effective participation of personnel and trust
Resource management
Plan the workload of the personnel, set up role-based training and ongoing mentoring, and use reward and recognition schemes for motivation Utilize a standardized set of equipment, well-thought-out scheduling of its maintenance, quality control, and use Establish a functional facility design and set up a schedule for carrying out studies Estimate costs and use budget planning
Process approach
Define the activities necessary to obtain the desired result Manage activities and related resources as a process View and manage the series of cooperating processes as a system Keep in mind simple day-to-day management
Customer focus
Understand current and future customer needs Fulfill customer requirements and make efforts to exceed customer expectations
Performance measurement
Establish a fair way to set objectives by considering factors such as training, administrative duties, and time frames in evaluations Define pass/fail criteria or rating on organizational objectives
Evaluation and improvement of all appropriate areas of the business
Base effective decisions on the analysis of data and information
processes can be divided at least into two categories: core processes and support processes. Core processes, often called value chain processes, directly contribute to the manufacturing of the product or the service, whereas support processes ensure that the core processes can run smoothly. Typical support processes are the human resource process, the document and data control process, or the finance process. Vital for a successful process management is to clearly state appropriate objectives that can be accomplished and that allow the drawing of reliable conclusions about the current quality situation in the company.3 The companywide quality policy is one quality management measure; in addition, individual process objectives can be specified. This approach of ISO 9001:2000 is consistent with the well-known plan–do–check–act cycle, or Deming cycle, as it is often called with respect to the contribution of Edwards Deming, an American statistician, to Japan’s quality improvement efforts after World War II.6 Deming taught about problem solving and teamwork, concepts that were new to statistical quality control of his time. He recommended that business processes should be placed under a feedback loop in order to identify and change those parts of the process that needed to be improved. The diagram in Figure 10.2 illustrates the continuous process to tackle improvement in the context of business process management.4,5
Generally speaking, ISO 9001:2000 certification means “we know what we want, we say what we do, we do what we say, we prove it, and improve it.”
10.4 QUALITY IN THE SKIN BIOENGINEERING LABORATORY Skin bioengineering studies, like other clinical studies, have to respond to quality requirements of different parties. There are the rights and interests of the panelists, the interests of the sponsor and those of the consumer, and last but not least the interests of the research institute. In our view, long-term sustained success for both the sponsor and the laboratory can only be created when the legitimate interests of all these parties are respected. In order to accomplish this goal, a clear organizational structure and comprehensive quality procedures are a must. And of course, the system has to be written down, readily accessible, and user-friendly. Our approach is based on a hierarchical set of different documents, as shown in Figure 10.3. On top there is our quality policy. As a management tool this policy is basically our outward statement of commitment to respect the sponsor, the different involved parties, and our employees, and outlines briefly what we do to provide a service of high quality. It finds its practical and measurable expression, for example, in our quality
70
Handbook of Non-Invasive Methods and the Skin, Second Edition
Decide on changes and take action to continually improve process performance.
PLAN
ACT
CHECK
Monitor and measure the processes and products against quality requirements quality policy, and objectives. Report the results to decision maker.
Establish objectives and processes necessary to deliver results in accordance with customer's quality requirements, the company’s quality policy and individual process objectives.
DO
Implement the processes.
FIGURE 10.2 Plan–do–check–act cycle. Business processes should be analyzed and measured in order to identify sources of variations that cause deviations from quality requirements and to tackle improvement.
Intention and direction of the company as expressed by top management Description of the quality management system (general requirements and rules) Document that states how an activity or procedure is to be done (general instruction, not study specific)
Document that gives specific instruction for individual work steps
Quality policy
Quality manual (= QMH part A)
Standard operating procedure (SOP) (= QMH part B)
Working instruction (AA)
FIGURE 10.3 Default documents. A hierarchical set of documents is used to fully describe the quality management system. The level of detail increases from top to bottom.
objectives, which are newly defined every year after our management review. The second level is our so-called quality management handbook (QMH). It describes our organizational structure, the series of our cooperating processes, and what has to be done to handle the different factors influencing the quality of our work properly. Generally speaking, our internal standards are defined, and all elements demanded by the ISO standard and the requirements of good clinical practice are addressed.
The third and biggest part is the standard operating procedures (SOPs), also called QMH Part B. The SOPs describe how we do what we do, and therefore permit the reconstruction of events that lead to the generation of a set of data. SOPs are the further translation of guidelines and legal requirements, scientific standards, and internal standards into the real-life situation of the laboratory. SOPs are required to ensure standardization of procedures. As general instructions for organizational-, method-, and equipment-related procedures, they are not study specific.
Implementation of a Quality Management System in a Contract Laboratory Working with Non-Invasive Methods 71
The fourth level is the working instructions. These are specific instructions for individual work steps; for example, they describe how to operate our data bank to generate and print out appendices for our reports. To write down what to do and how to do it alone is clearly not enough. All documents should be accessible to staff at all times for reference and (self-)training. Auditors, internal or external, will look for evidence of training in individuals’ training records and in the knowledge of the procedures demonstrated during the interview and review of data and documentation. In our laboratory the documents are accessible via the intranet. Authorized paper copies are handed out for the equipment procedures only, and for training purposes, so-called information exemplars are circulated as paper copies. The document management system also includes regular and timely routine reviews and provision for premature review and revision when required. A security system is in place to prevent unauthorized modification and circulation of the documents, as well as to recall and archive out-of-date versions. Even with the most effectively run and comprehensive documented quality management system, of course, the final implementation is dependent on the knowledge and commitment of the personnel themselves. During our quality system audits, staff at all levels are interviewed and given the opportunity to demonstrate these qualities. It is important to emphasize that an exact recollection of precise wording is not expected, but rather an impression of understanding the quality philosophy and principles, and the ability to access detailed instructions rapidly. Furthermore, it is vital to have not only well-trained staff, but also motivated staff. This in turn demands strategies to support motivation. How important the social and technical environment of the workplace is for the personnel is often underestimated. High levels of absenteeism due to sickness, as well as high levels of staff turnover, indicate problems in the laboratory. If one keeps in mind that replacement of personnel often leads to queries and raises the potential that mistakes occur, it becomes clear how important the human factor is. Yearly interviews with every staff member are a crucial tool in our quality management system. The actual situation is discussed, and individual objectives and training programs for the next period are fixed. Finally, it is important to note that staff who recognize weak points must have the permission to define and implement improvement measures. In this context, for example, involving staff in the process of SOP writing and review can improve the daily routine at work and the acceptance of SOPs themselves. The laboratory equipment evidently directly affects the integrity of the measurement; its proper function is of prime consideration. Hence, provisions must be made for appropriate preventive maintenance. But how can reliable data and the proper function be ensured? In this context
the contribution and advantage of standards become visible. For example, the gold standard for all measurements of length was for a long time the primary meter in Paris. Today there are more modern physical methods to define the length, but the primary meter in Paris is responsible that all over the world we have the same understanding of how long a meter really is. There is a strict system to ensure this. If we are talking about numbers that our bioinstrumentation for skin provides, we often lack this kind of gold standard, but fortunately enough we do have other methods to validate new instrumentation and increase the trust in our measurements. Changing instruments requires careful measures. For example, the absolute values of skin hydration determined by capacitance measurements using different devices are not the same, even though they are based on the same measuring principles. A study comparing both the Corneometer 820 and the new version, the Corneometer 825, showed us that the data obtained with the new device were significantly lower than those of the old one.7 This led to some confusion of the user of the Corneometer, and it really took some time until it was clarified, although it was a very simple effect. In order to overcome these restrictions, sophisticated calibration methods can be used that should cover all crucial instrumental functions influencing the measurement. For the Corneometer, for example, three controls are vital: the check for equivalence of the measuring heads, for the spring’s force, and for the triggering mechanism.7 But there is not one but many laboratories, and the question is: How can the values of these different laboratories be compared? In this case, external quality assurance measures should be established. For the skin bioengineering laboratory the possibility exists to participate in interlaboratory studies. How important such comparisons are can be illustrated by the following example.8 Four laboratories measured the redness of their volunteers with Chromameters, and because they had different volunteers, there was some differences between the laboratories, as expected. However, there was one laboratory in the round-robin study that totally fell out of range. What was wrong? The Chromameter of this laboratory was not functioning right; it needed to be repaired. But the laboratory did not know about this defect until it participated in the interlaboratory comparison. Now it has implemented procedures for checking the instrument and hopefully will get earlier signs of a malfunctioning of the instrument. Nevertheless, participation in interlaboratory studies continues to be very important. Another question arises: If you have two different instruments that are built for the same objective,9 are they identical or equivalent? If there is an obvious and significant correlation between the measurements of two instruments, this correlation provides a first hint and a good overview. But in order to really check whether the
72
Handbook of Non-Invasive Methods and the Skin, Second Edition
results of two instruments are “identical,” equivalence testing with a similar statistical approach as used in bioavailability/pharmacokinetic studies should be conducted, to provide more definite answers. In the context of the skin bioengineering laboratory materials, for example, calibration standards for the pH meter, water for dilutions, or sodium dodecyl sulfate, which is often used as positive control in patch tests, also have to be considered.10 The influence that these materials can have on the quality is not always obvious. For example, a standard with poor performance can never produce a reliable result; a chemical with the expiry date long gone can be of poor quality and its normal function can no longer safely be assumed. Simple actions like comparing the new standard with the one in use or the strict consideration of expiry dates can help to prevent these defects. Additionally, a sensible selection of suppliers helps to minimize risks. In essence, the quality management system must encompass, coordinate, and manage the so-called 5 M’s — man, machine, methods, materials, and marginal conditions — in order to achieve quality.
10.5 PERFORMANCE ASSESSMENT IN THE SKIN BIOENGINEERING LABORATORY The implementation of a quality management system is not an end in itself, but is closely linked to the target of business success. To address the success of our system, different quality objectives were established. One focus has been to reduce the frequency of work returned internally between different divisions, for example, data management, reporting, and quality assurance, in order to enhance efficiency. By interviewing staff at all levels, weak points such as uncertain responsibilities, imprecise distribution and authorization of documents, or unawareness of activities that are linked closely to one another were identified and the work flow reorganized. The positive impact was evident: for studies following standard protocols, for example, the rate of return could be noticeably lowered between all personnel involved. In line with the common saying that “It is much more expensive to find a new customer than it is to service a current one” and that satisfied customers will spread good word about the company, customer satisfaction is an important quality objective. Information, for example, on the number of reclamations, the frequency of repeat clients, and the growth of the customer pool, is used to
monitor if the sponsor is satisfied. These data can be gathered without direct interaction with the sponsor, with tools that include a database of customer information and with periodic reviews of these data looking for trends. Fortunately, during the last year the number of reclamations of our reports was low, the number of customer retention, with over 80%, was definitely high, and our customer pool is continuously growing. Last but not least, the sponsor is interested in a reliable relationship with the company, and he develops it with the people who work there. Hence, a pool of highly satisfied and, in return, highly loyal customers is vital for bottomline business success along with experienced and motivated staff. Using quality objectives as a strategic tool helps to assess and improve the quality situation in the laboratory, and to stay in business successfully.
REFERENCES 1. Kamiske, G.F. and Malorny, C., Total quality management: Ein bestechendes Führungsmodell mit hohen Anforderungen und großen Chancen, Z. Führung Organ., 61, 274, 1992. 2. Feigenbaum, A.V., Total Quality Control, 3rd ed., McGraw-Hill, London, 1961. 3. Pfeifer, T., Quality Management, 3rd ed., Hanser Verlag, Munich, 2002. 4. DIN EN ISO 9000, Quality Management Systems: Fundamentals and Vocabulary, 2000. 5. DIN EN ISO 9001, Quality Management Systems: Requirements, 2000. 6. Deming, W.E., Out of the Crisis, MIT Center for Advanced Engineering Study, Cambridge, MA, 1982. 7. Wilhelm, K.-P., Possible pitfalls in hydration measurements, in Skin Bioengineering Techniques and Applications in Dermatology and Cosmetology, Elsner, P., Barel, A.O., Berardesca, E., Gabard, B., and Serup, J., Eds., Basel, Karger, 1998, p. 223. 8. Fullerton, A., Lahti, A., Wilhelm, K.-P., Perrenoud, D., and Serup, J., Interlaboratory comparison and validity study of the Minolta Chromameters CR-200 and CR300, Skin Res. Technol., 2, 126, 1996. 9. Serup, J. and Agner, T., Colorimetric quantification of erythema: a comparison of two colorimeters (Large Micro Color and Minolta Chromameter CR-200) with a clinical scoring scheme and laser-Doppler flowmetry, Clin. Exp. Dermatol., 15, 267, 1990. 10. Agner, T., Serup, J., Handlos, V., and Batsberg, W., Different skin irritation abilities of different qualities of sodium lauryl sulphate, Contact Derm., 21, 184, 1989.
11 Ethical Considerations Mikkel Noerreslet The Danish University of Pharmaceutical Sciences, Copenhagen, Denmark
Gregor B.E. Jemec Department of Dermatology, Roskilde Hospital, University of Copenhagen, Roskilde, Denmark
CONTENTS 11.1 Introduction..............................................................................................................................................................73 11.2 Morality and Ethics .................................................................................................................................................73 11.3 Medical Ethics and Bioethics..................................................................................................................................73 11.4 Bioethics and Principalism......................................................................................................................................74 11.5 Ethical Guidelines in Research ...............................................................................................................................74 11.6 Informed Consent ....................................................................................................................................................74 11.7 Non-Invasive Skin Methods ....................................................................................................................................76 11.8 Conclusion ...............................................................................................................................................................76 References .........................................................................................................................................................................76
11.1 INTRODUCTION Ethical considerations within medicine have gained much attention during the last 50 years, and the importance of integrating these aspects in medical practice or research is today well recognized worldwide. By dealing with the general ethical considerations, and more specifically with the considerations related to the interpersonal relationship between the physician (the researcher) and the patient (the human subject), the aim of this chapter is to create an awareness of these aspects. This is so that such considerations are integrated into the design of a research project using non-invasive skin methods.
dards governing society — such as respecting others, avoiding doing harm, etc., as they are introduced to us by others. When interacting with others, the historical aspect of morality comes forth — in other words, we learn the correct moral behavior, just like children are taught how to behave correctly by their parents. Therefore, in addition to being a social construction, morality is also a historical construction.5 To introduce a third variable in the construction of morality, the culture in which we live will also shape morality. What is considered ethically correct behavior in one culture might not be ethically correct in another culture. These aspects, which underline the dynamic character of morality/ethics, therefore sharpen the necessity of a continuous attention to ethics.
11.2 MORALITY AND ETHICS When dealing with ethics it is clear that morality is a central concept, since morality refers to the traditions of beliefs about what is right and what is wrong human conduct. As mentioned by DeVries and Subedi,1 morality is socially constructed and can therefore fundamentally not be a personal policy or code — individuals as such do not exclusively create their morality by making their own rules. If this line of thought is accepted, the core parts of morality therefore already exist before an individual accepts it. We thus learn about moral codices and expectations through a process of interaction with others in society, and we come to understand the normative stan-
11.3 MEDICAL ETHICS AND BIOETHICS Within the area of medicine the philosophical reflections of what is right and wrong human conduct (morality) are an essential part of medical practice and research. Considering the position of medicine and its influence in society, this appears appropriate, as it relates to medicine’s interaction with every stage of human life and death. To ensure accordance with societal norms and uniformity among physicians and other health care professionals, the morality/ethics of medical practice and research have be explicated in rules or codes of conduct. 73
74
Handbook of Non-Invasive Methods and the Skin, Second Edition
The need for this practice has been recognized for approximately 2500 years. In ancient Greece, philosophers were preoccupied with ethical considerations of life and death, and hence the act of medicine. By swearing the Hippocratic Oath (400 B.C.), physicians of ancient Greece were obliged to act in conformity with the rules of the medical profession, and to current best practice for the benefit of the patient. In those days medical ethics was preoccupied with professional responsibilities and privileges — an aspect that has characterized medical ethics throughout thousands of years. With ideas conceived in the wake of the Nuremberg trials, which took place in 1947, the traditional physician centeredness was replaced by a set of principles, which focused more on the research subjects — the patient.4 These principles were known as the Nuremberg Code. With the Nuremberg Code the ethical tradition today known as bioethics entered the scene. Contrary to the traditions in medical ethics, where physicians exclusively discussed norms governing their own conduct, theologians and moral philosophers added to the ethical discussions in bioethics. This cross-disciplinary aspect ensured a number of different traditions and new viewpoints on ethical/moral issues. At first, this new approach to ethics within medicine had a hard time establishing itself, and it was not until the crises of medicine in the 1960s and 1970s that bioethics gained momentum and influence in health policy. This shift was partly brought on by a change in society’s conceptualization of medicine and its professions and technologies. Today medical ethics has largely been succeeded by bioethics, although the two are often used as synonyms, despite their distinctions.
11.4 BIOETHICS AND PRINCIPALISM In 1977, Tom Beauchamp and James Childress6 published a set of bioethical principles that later came to be known as principalism. The four principles, which were meant to provide a starting point for moral judgment and policy evaluations within health care, were respect for autonomy, nonmaleficence, beneficence, and justice. Acknowledging the fact that simple principles do not de facto contain sufficient content to decide ethical issues alone, and acknowledging the problem in deciding which principle is to rule, it has been emphasized that information regarding the issue should always be consulted before deciding.6 Among the four bioethical principles, respect for autonomy has gradually been established as the core principle in most Western countries, especially in America. In its most extreme form, the status of respect of autonomy is exemplified by the patient’s right to end his or her own life-sustaining treatment, despite physicians’ recommendations. Similarly, patients have the right to end
participation in any medical experiments/research projects. Tracing the development back in time, it is clear that the respect of autonomy is rooted in the liberal moral and political tradition of the importance of individual freedom and choice. As stated by Tom Beauchamp and LeRoy Walters,5 autonomy refers “to personal self-governance: personal rule of the self by adequate understanding while remaining free from controlling interference by others and from personal limitations that prevent choice” (p. 19). A patient is thus autonomous when he or she is free from external constrain and when he or she has sufficient information and understanding of this to make a choice. Another explanation for the status of the governing principle might also be due to the aspect that the principle is easy to codify and implement.2 However, it should always be remembered that the rise of patient autonomy as the general bioethical principle does not imply that the other principles, nonmaleficence, beneficence, and justice, are without relevance. Beneficence, for instance, might arise as the foremost principle in special situations, e.g., in an emergency situation where the patient is unable to express his or her own will/choices (unable to give informed consent) and where no surrogate decision maker is present, e.g., in a car crash. Physicians must in such situations act in accordance with the principle of beneficence and decide the “right” way to tackle an ethical situation from a professional point of view.
11.5 ETHICAL GUIDELINES IN RESEARCH When doing medical research that involves human subjects, certain guidelines must be complied with in most Western countries. Internationally, research projects are fundamentally governed by the Declaration of Helsinki, which the local ethical committees of the member states and as minimum comply with in their guidelines. The Declaration of Helsinki was adopted by the members of the World Medical Association in 1964 and has repeatedly been amended and ratified — the latest being in 2003.
11.6 INFORMED CONSENT In bioethics, informed consent forms one of the cornerstones, on a theoretical as well as on a practical level. This is also the case in the Helsinki Declaration, and is very evident in paragraph 22: In any research on human beings, each potential subject must be adequately informed of the aims, methods, sources of funding, any possible conflicts of interest, institutional affiliations of the researcher, the anticipated benefits and potential risks of the study and the discomfort it may entail. The subject should be informed of the right to abstain from participation in the study or to withdraw
Ethical Considerations
consent to participate at any time without reprisal. After ensuring that the subject has understood the information, the physician should then obtain the subject’s freely-given informed consent, preferably in writing. If the consent cannot be obtained in writing, the non-written consent must be formally documented and witnessed.7
The informed consent is usually given in writing, but it can also be given in other ways — as long as this is “formally documented and witnessed.” The patient and the physician are directed to keep a copy of the documentation for a certain period, so that it is available in case of any possible future disagreement. Signing the informed consent document is a formal manifestation by the patient that full information has been received and understood — and that the patient wishes to participate in the research. It is by no means an irrevocable acceptance of participation, and the patient can therefore decide to refrain from the research project at any time. If the patient wishes to withdraw, this decision must not directly have any negative influence on the patient’s current or future treatment. If the patient’s withdrawal has a negative effect, e.g., if the patient withdraws in the middle of a treatment program that should not be ended abruptly, the patient must be informed about the consequences of doing so, and a solution to the problem must be sought — a solution that embraces the wishes of the patient. It is therefore a clear violation of the patient’s autonomy if the physician pressures a patient who wants to withdraw from the research project. The basic requirements for the informed consent are that it must be: 1. 2. 3. 4.
Informative Understandable Voluntary Competent
That the consent is informative means that it must contain a full package of information. It is therefore not enough to present the patient with selected information, since this will limit the patient’s basis for making a free choice. For example, when informing the patient about risks related to participation in the research project, it is important that information on all known risks (side effects, etc.), fatal as well as minor, is given and placed in context. If the actual likelihood of the risks is known, such information should of course also be given. That the consent is understandable implies that the information provided by the physicians is understood by the patient. In the attempt to ensure this, the physician must address and recognize the capacities, perspectives, choices, and actions of the patient. In other words, the physician has to accept the personal values and beliefs, which function as the foundation for the behavior of the
75
patient. It is therefore clear that dialogue between the physician and the patient is of utmost importance, i.e., whether the patient has understood the information or not, and how the patient has understood the information. Structuring the process of informing the patient in an appropriate way will usually help this task — by drawing up a manuscript, listing the minimum information that should be given, the physician (or others involved in the research project) can keep track of the process and ensure that the necessary information is given to all patients participating in the project. When informing the patient orally, the physician must keep certain aspects of communication in mind. The information must be delivered using common and simple words, with as few medical and technical terms as possible — it must be understandable for the layperson. The art of setting oneself in the position of the receiver is a difficult task, but nonetheless a skill to strive for, since it will help the physician to show empathy with the patient and to get a feel of the patients situation when delivering information. Last but not least, it is important to allow the patient to ask questions in the process. This requires that the physician knows what information to give and is competent within the area of research. If the oral presentation of information is supplemented by written information, the language of the text must be clear, somewhat simple, brief, accurate, complete, and humane. Illustrations, pictures, diagrams, etc., will usually help to support the information given — as long as they are not misleading or taken out of a completely different context. To test whether the information is understood by the patient, it is advisable to conduct a pilot study. For the pilot study to be as representative as possible, it is important that the test persons share the same characteristics as the research population. So if the research is to test the transdermal water loss (TDWL) on patients with psoriasis, it is a good idea to test the delivering of information on a few persons with psoriasis. By encouraging these test persons to evaluate the information and the process of informing in an unstructured way, by letting them give descriptions in their own words, important aspects that might otherwise have been taken for granted or missed in the design of the research project can now be gathered.3 That the informed consent is voluntary of course implies that the patient gives his or her informed consent free from coercion. If the consent is not voluntary, this must be seen as an invasion of the patient’s autonomy. The physician must therefore not put pressure on the patient by, for instance, telling the patient that a refusal will have a negative influence on future treatment. If the patient is assisted by a third person, and pressured by this person, the physician must also stress that it is the patient’s choice to participate or not.
76
Handbook of Non-Invasive Methods and the Skin, Second Edition
The last requirement is related to the fact that the person giving his or her informed consent is competent to do so. In some instances the patient is not capable of understanding or dealing with the information provided. If the patient is disorientated, depressed, or otherwise mentally and psychologically unstable, the patient will not be competent to give informed consent. In such instances there are certain requirements related to the process of informed consent, which should be fulfilled. In other instances the patient can be a child or another person not recognized as a self-determining autonomous person. In these situations certain guidelines will also have to be followed. On the basis of this, it is worth noticing that obtaining informed consent is a dynamic process — not a definitive commitment by the patient to participate in the research forever. Information about the research procedures therefore do not end with the patient giving his or her informed consent. Rather, information should be given continuously before, during, and after the research procedures. It is furthermore important that informed consent is given deliberately and explicitly by the patient, and is not presumed consent resting on the physician’s presumption.
11.7 NON-INVASIVE SKIN METHODS From the medical point of view, non-invasive skin methods are often regarded as a harmless, effective, quick, and cheap way to gather information about the patient’s skin condition, especially if compared with the invasive skin methods. If it is assumed that the patients share this view, this might very well cause problems before, during, and after the research project. When dealing with non-invasive skin methods, it is therefore important that: •
•
Risks related to the use of the methods are assessed and communicated to the patients before they give their consent The procedures of use of the methods are explained to the patients before the informed consent is given and before the methods are actually used on the patients in the clinic
When dealing with risks, it should be remembered that all research on patients involves some degree of risk, even if the risk involved is only related to the patients’ transportation to and from the research site. The planning of the research must therefore correspond to these risks. Additional risks occur if substances are applied to the skin; e.g., cosmetic research involves independent risks of irritant or allergic reactions. Although the border between
voluntary use of such substances in everyday private use and use in scientific research may be blurred, the fact that it occurs in an experimental setting puts a special onus on the researcher to have considered risk explicitly. To comply with the governing bioethical principle, it is therefore important that the physician invest time and effort in understanding the patient perspective throughout every stage of the project — and question his or her own assumptions regarding the methods used. In addition to giving information on all relevant aspects of the research, the physician should dialogue with the patient, so that the patient as well as the physician is clear on what the research project and the patient’s participation is about.
11.8 CONCLUSION Ethics is a social construction (as well as a historical and cultural one). This implies that ethics is a dynamic concept that can and will change over time in accordance with the development of society. Ethical procedures will therefore, to a certain extent, reflect society’s attitudes toward what is right and wrong. Physicians planning to do research should therefore always consult their national/local ethical committee before conducting research on patients, to ensure that they comply with the current national guidelines. In addition to this, and in relation to the use of non-invasive skin methods, physicians should acknowledge that patients might not share their view of the excellence and safety of these methods, and should therefore engage in a dialogue with patients before, during, and after the research, to ensure a common understanding.
REFERENCES 1. DeVries R. and Subedi J., Eds., Bioethics and Society: Constructing the Ethical Enterprise, Prentice Hall, Englewood Cliffs, NJ, 1998. 2. Wolpe P.R., The triumph of autonomy in American bioethics: a sociological view, in DeVries R. and Subedi J., Eds., Bioethics and Society: Constructing the Ethical Enterprise, Prentice Hall, Englewood Cliffs, NJ, chap. 3, pp. 38–59. 3. Donovan J., et al., Improving design and conduct of randomised trials by embedding them in qualitative research: ProtectT (prostate testing for cancer and treatment) study, British Medical Journal, 325: 766–770, 2002. 4. Shuster E., The Nuremberg Code: Hippocratic ethics and human rights, Lancet, 351: 974–977, 1998. 5. Beauchamp T.L. and Walters L.R., Contemporary Issues in Bioethics, 5th ed., Wadsworth Publishing Company, 1999.
Ethical Considerations
6. Beauchamp T.L. and Childress J.F., Principles of Biomedical Ethics, 4th ed., Oxford University Press, New York, 1994.
77
7. World Medical Association, Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects, revised version, 2002, http://www. wma.net/e/policy/pdf/17c.pdf, downloaded July 6, 2004.
Section II Technique, Application, and Validation
Skin Surface, Epidermal Structure, and Function Clinical Photography, Surface Imaging Techniques, and Computerized Image Analysis
Aspects in Medical/Clinical 12 General Photography Nis Kentorp Department of Dermatology, Bispebjerg Hospital, Copenhagen, Denmark
CONTENTS 12.1 Ethical Guidelines.................................................................................................................................................82 12.2 Getting Consent ....................................................................................................................................................82 12.3 Keeping Track of Clinical Photos ........................................................................................................................82 12.4 Scaling...................................................................................................................................................................82 12.5 Perspective ............................................................................................................................................................82 12.6 Depth of Field.......................................................................................................................................................82 12.7 Lighting.................................................................................................................................................................83 12.8 Room Space ..........................................................................................................................................................85 12.9 Background ...........................................................................................................................................................85 12.10 Patient Positions....................................................................................................................................................86 12.11 Movement .............................................................................................................................................................87 12.12 Ring Flash 8 ........................................................................................................................................................87 12.13 Optimal Exposure .................................................................................................................................................87 12.14 Cameras, Film, and Processing ............................................................................................................................87 12.15 Specialist Photography .........................................................................................................................................87 12.16 Output and Viewing ..............................................................................................................................................88 12.17 Guidelines for Publication ....................................................................................................................................88 12.18 Standardization......................................................................................................................................................88 12.19 Conclusion ............................................................................................................................................................88 References .........................................................................................................................................................................88
Clinical photography is an essential part of medical research, diagnostics, and documentation. Knowing how to take clear and consistent pre- and postoperative photos is one of the most precise ways to document a patient’s progress, demonstrate to prospective patients who are having the same procedure what results may look like, and create a visual record of your work. High-quality clinical photos are also important for use at scientific meetings, in medical journals, and for other educational purposes. Photographs can tell us a great deal about a patient’s condition at an instant in time, and serial photographs taken over a period can tell us much more of the story, about the progress of disease or response to treatment. So why is it that the clinical photographs we see so often are a disappointment? They are too dark, too light, dark shadows, lack of sharpness, color variations, important detail obscured, untidy, indeterminate scale, etc.
Undoubtedly, fully trained clinical photographers take the best medical photographs in professional studios with the full range of lighting and equipment available. But such a facility is not always available, and certainly not at all hours of the day and night or in every location in which they see patients. The clinical photographers are not concerned with photographic tricks, such as are seen in some advertising for anything from hair tonics to plastic surgery. Playing with angles of view, makeup and clothing, perspective tricks, soft focus, and lighting can exaggerate the benefits of treatments on offer. Images can be manipulated at the processing stage, and this is both easier to do and harder to detect in digital images. There is no place in clinical recording or publication for use of such photographic manipulation to misrepresent outcomes, and in some cases it would be illegal. Original films and written 81
82
Handbook of Non-Invasive Methods and the Skin, Second Edition
records should be retained, and for digital images an audit trail should be maintained. All institutions that deal with patients should require photographers that have been specially trained. This should ensure the best way to take the type of pictures that the doctors need. In general, the photographic output depends on what the doctor has asked for — the photographer will tell the patient what is necessary in order to obtain the required pictures. It is important that the photographs demonstrate the condition as clearly as possible. In order to do this, it may be necessary to tie or pin up the patient’s hair or to remove jewelry. In some cases any makeup should be removed. Sometimes the condition the photographer has been asked to record will mean that any clothing that would appear in the picture is removed. Occasionally it is necessary to use special instruments (such as nasal retractors) to help get a clear picture. If this is required in a specific case, the photographer will explain the instruments to the patient before the photographs are taken. The photographer should be informed of any concerns or reservations there might be about the photography, and be sure that the patient is given all the information that is needed.
12.1 ETHICAL GUIDELINES The photographer must show respect for the patient, and is the basis for the professional ethical responsibility concerning race, sexual orientation, gender, religion, political observance, and national or social origin. The photographer is responsible in legal and ethical aspects, and all information and material should be treated with the greatest confidentiality.
12.2 GETTING CONSENT The patients should sign a photo consent form. Ideally, it should be done even if it is not planned to share the photos in a journal or on a website. Taking this precaution may make it easier should you change your mind later on. If the patient is easily recognized, the patient consent forms are required if the clinical photos are to be published in a journal, for educational purposes, or on a website. This form is regarded as legally binding and very important, and as such, it must be as specific as possible.
12.3 KEEPING TRACK OF CLINICAL PHOTOS For expediency and to remember the patient’s name, age, and the procedure performed, write down all relevant information before taking the photographs, along with the
date on a piece of paper. This will help you keep track and may avoid delays and misinformation later on.
12.4 SCALING The best way to standardize scale is to use a lens that is marked with reproduction ratios on the barrel. The lens is set to a specific magnification ratio, and focusing is achieved by moving the whole camera back and forward. Some lenses, such as the Nikon Micro-Nikkor lenses, are factory marked for 35-mm cameras, but for digital photography, using shorter focal lengths, standardize on either the marked distances or a certain ratio by photographing a ruler and marking the lens barrel with tape or thin paint lines. With simpler cameras, this approach might be difficult or impossible, but standardization can still be achieved by maintaining the same distance from the patient. If the camera has a zoom lens, the amount of zoom should be fixed so that the image size remains the same at that distance. Some experimentation might be needed, and cameras will differ in their ability to achieve this predictably. The addition of a linear scale gives a comparative method of assessing the size of a lesion, although caution should be applied to any attempt to take measurements from photographs: some distortion is inevitable, and it must be remembered that a single photograph is a twodimensional representation of a three-dimensional subject. In photographs for forensic or medicolegal purposes, a two-dimensional right-angle scale is recommended.
12.5 PERSPECTIVE The perception of depth that perspective gives to a clinical photograph is an important attribute. The concept of diminishing image size with distance is familiar, but perspective also affects the appearance of objects and people as viewpoint changes: come too close and features can appear distorted, but take the photograph from too great a distance and the perception of depth is lost. That is why not only the image size, but also the distance from which a photograph is taken, should be standardized. For 35-mm film, a lens of around 100-mm focal length allows distortion-free images in most clinical situations, but a shorter focal length of around 50 to 60 mm will be needed for full-length photographs, or when working in a confined space.
12.6 DEPTH OF FIELD Depth of field is the amount of a subject that appears sharp in front of, and behind, the principal plane of focus. The smaller the lens aperture, the greater the depth of field, and this is particularly critical in close-up work where
General Aspects in Medical/Clinical Photography
83
FIGURE 12.1 A platform covered with cloth in the same color as the background. Used in order to get the right mid-point angle of the lower extremities.
there is considerable depth, such as when photographing the teeth. Relative aperture size is denoted on a lens by f numbers — the higher the number, the greater the depth of field, but the smaller the amount of light passed through the lens. Generally, clinical photography needs the maximum depth of field, so the flash should be bright enough to allow it to use the smallest possible aperture. The 35mm cameras should aim for at least f/16 and up to f/32 for close-up work where depth is important. Using a faster film will improve depth of field, but at the expense of some image quality.
12.7 LIGHTING The human eyes and brain have an extraordinary ability to adjust to different lighting conditions — we work quite happily in daylight, or under tungsten and fluorescent lights. Color temperature is measured in Kelvin (K) and pr. definition is the unit of thermodynamic temperature, the fraction 1/273.16 of the thermodynamic temperature of the triple point of water, which also, in fact, is the absolute temperature scale. (The conversion from Celsius to Kelvin is as follows: K = ˚C + 273). Normal daylight will be measured to approximately 5500 Kelvin — in blue sky and in shadow more. Tungsten light is approximately 3400 Kelvin. Halogen bulbs are stable in color temperature, but conventional light bulbs will be more red/yellowish down to 2900 Kelvin in daily/monthly use. Electronic flash produces color temperature similar to daylight.
FIGURE 12.2 Photograph taken with a daylight film in tungsten lighting (examination light). This shows a strong orange/yellow cast, too much contrast, and a small depth of field due to a small f number.
Cameras and films are less adaptable, particularly in mixed light, so photographs taken with a daylight film in fluorescent lighting will have a green cast, while those taken in tungsten lighting will have a strong orange/yellow cast. Indoor lighting is not bright enough, and the quality of the light is never appropriate for clinical photography. Electronic flash is ideal for providing a bright light source that is of such short duration that it is capable of “freezing” any movement. It is powerful enough to negate the effects of normal room lighting, although if you work in a room with strong sunlight coming in through the windows, you should close your blinds or curtains. Similarly, powerful examination lights or operating lights should be dimmed while taking photographs with flash. The existing light, e.g., strong light in the examine room, can be controlled by a higher shutter speed. If an existing — tungsten — light is used in analog photography, a film for tungsten light should be used, alternatively a filter (blue) for compensation. Again, this will
84
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 12.4 Electronic flash above the camera lens is used and shows shadow, shape, and texture. FIGURE 12.3 Examination light combined with electronic flash. The nonflash light is controlled with faster shutter speed.
affect the speed of the film. In digital photography this is not a problem, but the white balance has to be set. This is a normal procedure in professional digital photography. The above issues emphasize that the most secure method to produce acceptable photos is to use a flash. Modern compact cameras have a flash built into them, usually with automatic exposure control, making it difficult to produce a badly exposed photograph. This is fine for a quick record, but the lighting is too close to the lens axis to provide sufficient modeling effect to demonstrate shape and texture. It will also produce red eye — the red reflexes you see in people’s eyes in photographs taken with compact cameras. (Red-eye reduction modes use a preflash, which causes a delay in the shutter firing, so they are not ideal for moving subjects, such as lively children.) A still photograph does not have the advantage one has in a live examination of moving oneself or the patient
to look at a subject from any angle. Shape and texture can be revealed by off-axis lighting from a portable flashgun held in the hand that is not holding the camera, or attached to a bracket. Ensure that it is pointing directly at the subject to avoid the light falling off toward the edges. Lighting should as nearly as possible replicate what we are used to seeing: light from the sun, or indoor lighting from the ceiling, throwing shadows downward. Lighting should therefore come from above, and not below, in relation to the anatomical position in which we pose the patient. Getting this wrong can have strange effects. For the sake of consistency, always use the same camera, flash, lens, film, lighting, and patient position. It is axiomatic that the only variable among photographs taken to show change over time should be in the patient. Everything else should stay the same — viewpoint, positioning, lighting, color, magnification, perspective, contrast, and background. The relative importance of these properties of a photograph might vary: color might not always be important to an orthopedic surgeon, although to a
General Aspects in Medical/Clinical Photography
85
FIGURE 12.5 Ring flash is used and provides virtually shadowless lighting, with the flash tube wrapped around the camera lens.
dermatologist it might be the most salient feature of a condition. However, the principles of standardization should apply to any set of two or more photographs taken at different times. In practice, it is extremely difficult to standardize absolutely so many variables — the photographs might be taken by different people, in different rooms, using different cameras, lenses, or films, under different lighting, and from a different distance or angle to the patient. Slight variations in the film processing or digital treatment are highly likely. Different manufacturing batches of film — or in digital photography, the manufacture of the charge coupled device (CCD) chip — will have variations in sensitivity and response to color. Clearly, standardization requires a certain amount of planning, a systematic approach, adherence to protocols, and attention to detail.
12.8 ROOM SPACE Set aside a place solely for clinical photography to maximize privacy and patient comfort. Try to have the space situated near to the exam room for convenience.
FIGURE 12.6 Professional studio flash used in a studio gives maximum control of lighting and large depth of field.
12.9 BACKGROUND Try to have a seamless background, using a rich blue and green. Backgrounds with too much color can reflect onto, or throw a cast of the complementary color into, the subject, so the background should be dimmed blue/green or neutral white/gray. If you routinely photograph patients in a clinic or study, it is worth fixing up a plain, neutral background sheet or using the more or less sterile cloth or tissue, normally in “hospital” color. Professional medical photographers often prefer to use a black background, which requires several carefully placed lights to ensure that the edges and hair of the patient are not lost in the photograph. When photographing patients in a ward, it is best to place them against a plain white, gray, or dimmed blue/green color sheet. Use sticky tape, velcro, or bulldog clips to suspend the sheet and smooth it out as evenly as possible.
86
Handbook of Non-Invasive Methods and the Skin, Second Edition
12.10 PATIENT POSITIONS The patients should use the same positions, angles, and poses for both pre- and postoperative photos. Keep image size consistent in your pre- and postoperative shots. Occasionally, it might be useful to include contextual information in a photograph, such as showing the patient sitting up in bed, or attached to equipment. Similarly, if the photograph is to show dermatitis caused by an elasticity bra strap, it might be useful to include in the photograph the item of clothing responsible. In most cases, however, it is only the clinical appearance of the patient that is interesting, so all other distractions and possible influences to judgment should be excluded from the image. Use a chair with an adjustable back to pose the seated patient and exclude any parts of the furniture from the frame. Remove jewelry and makeup as far as possible. Allow patients time to replace them in private afterwards, and provide a mirror for that purpose. Ideally, all clothing should be removed from the field of view. This clearly requires some sensitivity and tact, as few people are used to exposing themselves to a camera. The patient can remove only the parts of clothing appropriate for each picture, rather than leaving them exposed any longer than is necessary. Photographs of patients hitching up their clothing can appear less dignified, less clinical, and probably cause the patient no less discomfort than if they removed those items completely. Modesty garments can be worn if it is not necessary to show the genitalia; disposable white underwear should be available, or use small sheets held up with tape or bulldog clips. Any extraneous items of clothing appearing at the edge of the frame should be cropped out afterwards. Try and provide photographic garments (underwear and paper gowns) for patients so that they can feel comfortable before taking photos and so they can move from the exam room to the clinical photography area. Discuss the pre- and postoperative photos. Avoid having the clinical photography feature be a surprise to patients. Let them know in advance (this should be part of your initial consultation) that photos are an essential part of medical treatment, confidential, and provide a good visual of the before and after treatment and surgery. If the patient is assisted, for example, a child held by a parent, care should be taken to avoid including that person in the photograph. While photographing a child on a mother’s lap, ask the mother to sit sideways, so that she is not seen behind. Keep helping hands out of frame, or be as discreet as possible. In close-up photographs, where a hand is seen retracting eyelids, lips, etc., ask the person who is doing the retraction to wear an examination glove. The distraction of a patient’s or parent’s dirty fingernail in the field of view can ruin an otherwise excellent closeup view. Hair should be swept back from the face, using
a hair band or hair clips. Hair clips can also be used to expose lesions on the scalp to the best advantage. While photographing the lateral aspect of the head, long hair should be swept over the opposite shoulder, or held up in a bun. Much confusion can arise from photographs — particularly close-ups — in which the patient’s position, or the orientation of a body part, is ambiguous. If we are going to produce a serial record of a patient, the positioning should be consistent. The convention used by most medical photographers is, wherever possible, to photograph the patient in the anatomical position. All clinical photographs should be viewed with reference to this position — the top of the photograph should always be nearest the top of the head. This works for most views, but some parts of the anatomy are best viewed from other angles, or in different positions. The arms are best photographed in extension, and are normally photographed in a horizontal position. However, the palms should still face forward, so the back of the forearm should be photographed with the patient facing away from the photographer. Close-up views can easily be misorientated, so they should normally be accompanied by an establishing view to show their precise anatomical position. A full-length photograph of a patient is one of the most difficult to achieve satisfactorily. Fortunately, it is not often necessary or particularly useful, except to show abnormal stature, posture, or body shape. The distribution of skin lesions is better shown with separate photographs from waist up, waist down, both front and back, with the arms photographed separately. For some conditions it is preferable to show the patient weight bearing. If the patient can stand unaided, he or she should do so in the anatomical position, as described above. While photographing the standing patient, remember that from your head position the patient’s feet are a long way away compared to his or her head. This can cause both perspective distortion, with the patient appearing to have tiny feet and legs, and a dramatic falloff in exposure between the head and feet. You should position the camera level with the patient’s mid-point — about waist level. For pictures taken of the lower extremities from the knees to the feet, a platform covered with cloth in the same color as the background is useful in order to get the right mid-point angle. It is difficult to produce satisfactory photographs of patients, other than infants and small children, in the prone or supine position. Such photographs require a greater working distance than can be achieved with the patient lying in bed — even if you stand on a chair or stepladder. A short working distance means using an extreme wideangle lens, which will result in unacceptable distortion. Babies, before they can sit or stand, can be photographed on a physiotherapy mat, or several layers of blanket, covered in a white sheet on the floor. Not only is this safer (they have nowhere to fall), but from a standing
General Aspects in Medical/Clinical Photography
position you can achieve a full-length view without distortion. You will need the assistance of a parent or helper to position the child for lateral views, and to gently extend the legs to be shown in full length.
12.11 MOVEMENT If the photo requires movement, the photographer should be the one to move with the camera and not the patient. Have the patient stand on a mark (you can have an X marked with black tape) and stand still.
12.12 RING FLASH 8 There is a common misconception that all medical photographs should be taken with a ring flash. The ring flash is an important tool for photographing cavities, where the shadow of a directional flash would obscure important detail. Thus, it is useful for dental photography or for deep wounds in surgical photography. A ring flash provides virtually shadowless lighting, with the flash tube wrapped around the camera lens. This type of lighting is very flat and reduces modeling. It also causes large circular reflections, which are particularly noticeable on wet surfaces such as the eye. For this type of work, a handheld flash is preferable: a powerful light source providing modeling and a single, small reflex.
12.13 OPTIMAL EXPOSURE Photographs that are too dark or too light are a bitter disappointment, and clinical photography presents greater challenges than almost any other type of photography. Clinical detail is easily lost in washed-out pale skin or underexposed dark skin. The extremes of light and dark in specialties such as dental or operative photography can disrupt the most sophisticated light metering systems. It is really worthwhile to do some tests for your camera–flash–film combination with different subject matter. Predetermined exposures that can be manually set for any magnification ratios are more reliable than automatic exposure meters, because automatic metering can be influenced by the background to cause the area of interest to be incorrectly exposed.
12.14 CAMERAS, FILM, AND PROCESSING In choosing a camera for clinical photography, the two main choices are between digital and conventional film and between compact and single-lens reflex (SLR) cameras. Your choice will depend largely upon the budget, and just how portable you need the camera to be. One of the biggest attractions of digital cameras is their immediacy.
87
After taking a photograph you can check to see whether you have a usable image, and you can download photographs onto a computer within minutes. There are hundreds of digital cameras in the market, presenting a confusing range of choices. If you want a compact camera that fits neatly in your pocket or medical bag, you will have to make compromises in the control you have over lighting, image size, and working distance. Greater control can be achieved with an SLR, which will allow you to change lenses and attach different types of flash. However, digital cameras of this type are prohibitively expensive and are used mainly by professional photographers. The alternative is to use a conventional camera and have your slides or negatives scanned. Many processing laboratories will process and scan film as a routine service, giving you a CD of all your images, sometimes in several different resolutions. If not shooting with a digital camera, try to always use Kodak, Fuji, or another recognized brand name in film. The film should be ISO 100 to 200. Remember to always use a flash.
12.15 SPECIALIST PHOTOGRAPHY Specialist clinical photography requires more specialized equipment and techniques. For dental photography, a range of mirrors and retractors are essential. Ophthalmic photography uses sophisticated slit lamp and retinal cameras, with which investigative photographs can be taken, such as fluorescent angiograms. Endoscopic photography uses cameras designed to attach to the instrument. Most instrument manufacturers supply, and can advise on, proprietary recording systems. Orthopedic patients might present handling problems, requiring special chairs and other equipment. In dermatology, some conditions can be photographed in the invisible spectrum, using infrared, ultraviolet, or fluorescence techniques. Some conditions can be better illustrated using special photographic techniques, such as infrared and ultraviolet (direct/fluorescence) photography, transillumination, dermaphotography, or photomicrography. Please take extra care to preserve patient confidentiality when handling patient recordings of all types, and prevent patient images from being seen or used by anyone other than the appropriate professionals. Patients’ wishes and written consent must be complied with at all times. Clinical photographs, video, and other images should all be considered to be a part of the clinical case notes. Photographs should be stored and presented appropriately for their use, and images for publication should be prepared according to the instructions to authors. Digital images for publication should be sized appropriately for the final reproduction size.
88
Handbook of Non-Invasive Methods and the Skin, Second Edition
12.16 OUTPUT AND VIEWING Consideration should be given to the conditions for viewing images. Clinical photographs might be seen as prints in a patient’s case notes, as digital images on a computer screen, or, more rarely these days, as projection slides. Prints should be viewed in good lighting — preferably daylight or white light fluorescent illumination (5500 K). Ask your laboratory to print photographs on glossy, rather than matte or textured paper. Prolonged exposure to ultraviolet rays will cause fading and color change, so keep prints stored out of direct light, preferably between sheets of acid-free paper. Slides are better for group viewing and, with consistent processing in a good laboratory, can give quite consistent results. Store them in archival-quality polypropylene filing sheets to avoid chemical damage. Computer viewing and projection of digital files can present problems of both quality and consistency. Use a good-quality monitor positioned away from brightly lit windows or colored surrounds. The most important factor to control in viewing is consistency. Use a good professional laboratory for films and do not leave films in the camera for months before processing — both exposed and unexposed film deteriorate over time. Keep your films in a cool place. Store large batches of unexposed film in a refrigerator, but remove film at least an hour before use to allow it to return to room temperature. Digital files for projection should be saved as highquality JPEG (.jpg) files, but most publishers prefer TIFF (.tif) files. The file size will depend upon the reproduction size, and many publishers will state exactly the file size to submit. It is normally safe to save them at a resolution of 300 dpi (dots per inch) at the final reproduction size. Check recent issues of the journal to see what is the maximum size at which images are normally reproduced (usually defined by the column width).
12.17 GUIDELINES FOR PUBLICATION If possible, avoid cropping and cutting the image. This is because most of the photos will be altered according to the specifications of the particular organization you send them to. If you are dealing with analog images, retain the
negatives or slides for yourself. Most organizations will ask for the actual photos; keep them for your own records.
12.18 STANDARDIZATION Standard representation is essential in clinical photography, to allow objective repeatability of views for comparison between different dates, or among different patients. Many clinical photographs are taken at fixed magnifications, and the views and magnification scales are recorded for each patient visit. Lighting, background, viewpoint, perspective, film, and processing should all be highly standardized and controlled. Some specialties have standard sets of views for particular groups of patients. These include dental/maxillofacial surgery, craniofacial surgery, facial plastic surgery, and intersex clinic.
12.19 CONCLUSION Familiarity with equipment and adherence to simple protocols can make all the difference between success and failure in clinical photography. A systematic approach is essential. This should extend beyond the photography itself to the handling and storage of photographs. Considerable attention should also be paid to legal and ethical issues before undertaking any clinical photography.
REFERENCES 1. Harvard Medical School, Educational articles, Harvard Medical School, New England Regional Research Center, Boston, MA, 2003. 2. UCLH Trust and Medical School, Guideline for Clinical Photography, UCLH Trust and Medical School, January 23, 2004. 3. Wentz MG, Clinical photography simplified: developing a personal set of views, 15: 211–214, 1995. 4. Ainslie G, Reilly J, The use of linear scales in the photography of skin lesions, J Audiov Media Med, 26: 15–22, 2003. 5. Aesthetic Surgery Journal, Paul Bernstein, public education, media relations, managing editor, Paul Kushner, senior director, marketing. 6. Institute of Laryngology and Otology, 2003.
of Compact Digital Camera for 13 Use Snap Photography Ken-ichiro O’goshi Department of Dermatology, Bispebjerg Hospital, Copenhagen, Denmark
CONTENTS 13.1 Introduction and Background ..................................................................................................................................89 13.2 Advantages of Using a Compact Digital Camera...................................................................................................89 13.2.1 Immediacy....................................................................................................................................................89 13.2.2 Storage in Image Bank ................................................................................................................................89 13.2.3 Lasting Quality ............................................................................................................................................89 13.2.4 Image Analysis ............................................................................................................................................89 13.3 Choice of Camera ....................................................................................................................................................91 13.3.1 Easy Operation ............................................................................................................................................91 13.3.2 Fast Start-up, Shooting, and Playback ........................................................................................................91 13.3.3 Fine Picture..................................................................................................................................................91 13.3.4 Zoom ............................................................................................................................................................91 13.3.5 Mobility .......................................................................................................................................................92 13.3.6 Easy Connection to Image Banks and Tools ..............................................................................................92 13.4 Process of Saving Digital Images ...........................................................................................................................92 13.5 Teledermatology ......................................................................................................................................................93 References .........................................................................................................................................................................94
13.1 INTRODUCTION AND BACKGROUND A single-lens reflex camera (Figure 13.1 and Figure 13.2) operating with photographic films is available in almost every dermatology clinic. However, compact digital cameras (Figure 13.3 and Figure 13.4) are now popular and available at a reasonable price. These both offer convenience, and images may even be taken as 1-cm close-ups. There is no need to buy films and wait for development since digital images are available within seconds. Images may be accepted, deleted, saved, and printed out directly using a color printer with or without a personal Macintosh or Windows computer (PC).
13.2 ADVANTAGES OF USING A COMPACT DIGITAL CAMERA 13.2.1 IMMEDIACY You can check what you shot and discuss the diagnoses or treatment for some hard-to-cure cases with colleagues immediately or at a conference.
13.2.2 STORAGE
IN IMAGE
BANK
Recorded images are conveniently stored in a tiny memory card (Table 13.1) or on a hard disk of your PC. You can manage recorded images that you want to recall and search them easily using an appropriate software. Storage programs are image formats that are in widespread use on websites, for example, PNG (Portable Network Graphics), JPEG (Joint Photographic Experts Group), or GIF (Graphic Interchange Format).
13.2.3 LASTING QUALITY There is no change of image over time, and replicates from the original can be produced at any time.
13.2.4 IMAGE ANALYSIS Size and number of eruptions can be measured with appropriate. NIH (National Institutes of Health) Image is a public domain image processing and analysis program. It is very useful, but only for the Macintosh. NIH Image can be used to measure area, mean, centroid, perimeter, etc., 89
90
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 13.1 Example of a single-lens reflex camera for dermatological photography.
FIGURE 13.3 Example of a compact digital camera (rear view). FIGURE 13.2 Example of a compact digital camera (front view).
of user-defined regions of interest for research. It also performs automated particle analysis and provides tools
for measuring path lengths and angles. Results can be printed out, exported to text files, or copied to the clipboard. Furthermore, you can compare different skin conditions during the treatment when you open the files on the PC monitor at the same time.
Use of Compact Digital Camera for Snap Photography
91
without waiting for a long time to save the previous image. This will help in photography of moving objects and children.
13.3.3 FINE PICTURE
FIGURE 13.4 Card slots of a color printer. The SD/MS PRO slot is used for an SD Card™ (by SanDisk) and a Memory Stick™ or Memory Stick PRO™ (by SONY). The xD/SM slot is used for an xD-Picture Card™ (by Olympus and Fujifilm) and a SmartMedia™ (by Toshiba). The CF slot is used for a CompactFlash™ (by SanDisk).
13.3 CHOICE OF CAMERA The camera market is highly competitive and offers many brands. It is not necessary to choose an expensive singlelens reflex camera for daily clinical uses. Below are some points in choosing a good camera at a store.
13.3.1 EASY OPERATION The high-quality megapixel panel should be user-friendly and easy to operate, even under direct shiny sunlight.
A 2 or 3 megapixels camera has sufficient resolution for routine clinical photos. For graphic monitors, the screen resolution signifies the number of dots (pixels) on the entire screen. For example, a 640 × 480 pixel screen is capable of displaying 640 distinct dots on each of 480 lines, or about 300,000 pixels. This means it can print out 10,000 dots per square inch. A 3 CCD (charge coupled device) facility can make recorded images very fine and clear as well as support the printout. A CCD is a volatile memory whose semiconductors are connected so that the output of one serves as the input of the next. Its advantage is small size and speed to access. Adapted multipoint autofocus gives priority to the images’ composition and prevents off-center captured images where the subjects are out of focus. It automatically adjusts the lens system relative to the distance measured between the camera and several points in the images. The autofocus function operates with an ultrasound emitter and sensor, and the measured distance and sharpness are dependent on physical evenness of the object surface. If the camera has a builtin flash, a high-intensity white light-emitting diode (LED), you can capture dark and fine images with correct/appropriate luminance even when a close-up subject is in a lowlight condition.
13.3.4 ZOOM 13.3.2 FAST START-UP, SHOOTING,
AND
PLAYBACK
Within a second after turning the camera on, you should be able to start taking pictures. With a between-shot interval of around 2 seconds, you can record the image quickly
The cameras have not only optical lenses, but also digital zoom, which provides a maximum zooming up to about 10∞. On the other hand, if you want to capture the closeup eruptions, you can get as close as even 1 cm from the
TABLE 13.1 Recent Memory Cards for Recording of Images Memory Card Memory Stick
Standard support by Windows. Can handle from black and white to full color (16,777,216 colors); basically, can save noncompressed images; optionally, can compress 16- and 256-color forms
SD Card
Invented by SanDisk Co., Matsushita Co., and Toshiba Co.; not only possible to record images, but also suitable to record music for portable digital audio player
SmartMedia
Proposed by Toshiba Co.; small as a postage stamp; widespread use for PDA (personal digital assistant) or digital cameras; adapter is necessary to connect to computer
xD-Picture Card
Invented by Olympus Co. and Fujifilm Co. in 2002; smallest size in the world; adapter is necessary to connect to computer
CompactFlash
Proposed by SanDisk Co.; possible to connect directly to your laptop computer through an exclusive adapter; this connection system is much cheaper than any other media
92
Handbook of Non-Invasive Methods and the Skin, Second Edition
13.4 PROCESS OF SAVING DIGITAL IMAGES
FIGURE 13.5 The tulip mark of supermacro mode used for close-up photography. You can even capture 1-cm close-up images with the supermacro function.
object and capture the fine images with the setting of the supermacro function (Figure 13.5).
13.3.5 MOBILITY The smaller the camera is, the more frequently you can practice with it. If every young doctor has a reasonable camera, he or she can record at anytime and consult with senior doctors.
13.3.6 EASY CONNECTION TOOLS
TO IMAGE
BANKS
AND
When buying a digital camera, a CD-ROM (compact disc read-only memory) with software to save the images on a PC, manage them simply, and even edit them in various ways is often included. You can save the images in some recording ways like JPEG, TIFF (Tagged Image File Format), BMP (bit map), PICT, etc. (Table 13.2). There are some cables in the kit to connect the camera to a PC, a printer, or a TV. Furthermore, if you have a PictBridgecompatible printer, you can print pictures directly via the supplied universal serial bus (USB) cable by connecting the camera. This allows PC-less printing.
After buying your favorite camera, read the manual book and charge the battery for the camera first. Put the memory card inside the camera, turn it on to shooting position, and push the button to zoom in or out. Set the date and time with checking in the finder. For practice, capture any image you want and delete any image you do not want to save. Choose your setting to take pictures. Select resolution of image (pixels). If you save images as a part of the patient’s record, 2 to 4 megapixels is acceptable. Memory cards have limitations regarding numbers of photos depending on the pixels of individual images (Table 13.3). When taking close-up pictures, select the macro or supermacro setting (Figure 13.5). Generally, you do not have to use the full-auto shooting mode (Figure 13.6), except for dark objects, for example, nevi on the scalp surrounded by many hairs. Even when the object is dark in color, you can take pictures with full-auto mode with half-push capturing, and the camera can automatically do the lighting without flash. In the clinic, whenever you think it is worth saving the image, you must identify the patient. If possible, you can capture the patient’s label on the front page of his or her medical record before you take a picture of the eruptions. Before you file the images, first install the software included in the package. Connect the camera directly to your PC with the supplied USB cable, or insert a memory card directly into your PC or into the memory card slot of a printer that has a function to scan it. Using editorial software, choose “Scan to PC,” name each image, and select files where you want to keep the images. On the hard disk of your PC, make two directories, one for patient image and another for ID, and save the data in separate files. You can easily identify and search
TABLE 13.2 Common Image Formats Image Format BMP
Standard support to Windows; can handle colors from black and white to full color (16,777,216 colors); basically, can save noncompressed images; optionally, can compress 16- and 256-color forms
JPEG
Compression rate is around 1/10 to 1/100; good for compression of natural photos, but not good for computer graphics; patented by Forgent Co. in July 2002
PICT TIFF
Standard support to Mac OS of Apple Co.; can save both vector data and bit map data; can save full-color images and compress data Invented by Aldus Co. and Microsoft Co.; helpful for storage of one image in various formats; pixels, color numbers, encoding procedures in the same file
PDF
Invented by Adobe Systems Co.; possible to recall almost the same layout, fonts, decorative letters, and photos as an original text
Use of Compact Digital Camera for Snap Photography
TABLE 13.3 Number of Photos Stored on a Memory Card Depends on the Capacity of Memory Card and the Selected Resolution Memory Card Capacity Resolution Setting
16 MB
64 MB
512 MB
8M 6M 4M 2M WEB
8 10 16 33 150
35 45 70 143 645
284 363 568 1150 5200
Note: M = megapixel; MB = megabyte; WEB = the amount of mode needed to record the image to present on the website or attach to e-mails.
AUTO
93
you need to fix color tones of the image, you can correct with tone curve chosen from image adjustment of the tool. When you make a new presentation with Microsoft PowerPoint®, open a new file and set the background chosen from the menu format. Then, using Adobe Photoshop, crop or select all of the image (Ctrl + A buttons), select the image, copy it (Ctrl + C buttons), and paste it (Ctrl + V buttons) on the page of Microsoft PowerPoint. Write some explanations and save it with a title. It is really helpful to e-mail with images attached. It is an easy and fast way to discuss cases or studies via the Internet. As you send e-mails, open e-mail software such as Microsoft Outlook Express® or free mail software like Hotmail of Microsoft, click the attachment button, select the image, choosing from your recorded file, attach it within the limitation of its size due to the software, and send it with some comments at the same time. Commonly, the size limit is around 2 megabytes in an ordinary free e-mail software.
13.5 TELEDERMATOLOGY
FIGURE 13.6 The AUTO mark of full-auto setting.
not only any image you want to recall, without entering the finding mode, but also the patient’s medical record, xrays, and histopathology if included. If necessary, make backups, which means other copies of the same images, in the CD or digital versatile disc (DVD) or external hard disk for storage. CD is cheaper. An external hard disk is preferable regarding speed and reasonable in price as well. Using image processing software (Table 13.4) like Adobe Photoshop®, you can crop the area with your favorite marquee tool, for example, a rectangular one. If
Telemedicine is the practice of medicine at a distance using high-tech devices such as PC, digital camera, dermoscopy, microscopy, scanner, color printer, and telecommunication system. Teledermatology refers to the practice or clinical field in dermatology using telemedicine for diagnosis and treatment over long distance. This field is rapidly growing since many patients living in distant locations need immediate help and up-to-date medical service. Teledermatology has been used only since 1996.1 Telemedicine reports2–4 included collaborations on dermoscopy and dermatopathology.5–7 However, the direct relationship between doctor and patient remains very important. Telemedicine can be nothing but optional to a medical consultation.
TABLE 13.4 Common Software for Processing and Display of Digital Images Software
Company
Classifying
Adobe Systems
Photo retouch software
To process, crop, or fix the color tone of the images
Acrobat
Adobe Systems
Application software for PDF files
To process or crop the images
PowerPoint
Microsoft
Presentation software
To make presentation slides with text and images
Outlook Express
Microsoft
Messaging software
To e-mail with text and images
Lotus Notes
Lotus Development
Groupware
To e-mail with text and images
Photoshop
Using It
94
Handbook of Non-Invasive Methods and the Skin, Second Edition
REFERENCES 1. Wootton R. Telemedicine: a cautious welcome. Br Med J 313: 1375–1377, 1996. 2. Eedy DJ, et al. Teledermatology: a review. Br Med J 144: 696–707, 2001. 3. Goldberg DJ, et al. Digital photography, confidentiality, and teledermatology. Arch Dermatol 140: 477–478, 2004. 4. Oakley AMM, et al. Retrospective review of teledermatology in the Waikato, 1997–2002. Australasian J Dermatol 45: 23–28, 2004.
5. Shapiro M, et al. Comparison of skin biopsy trigger decisions in 49 patients with pigmented lesions and skin neoplasmas. Arch Dernatol 140: 525–528, 2004. 6. Collins K, et al. Patient satisfaction with teledermatology: quantitative results from a randomized controlled trial. J Telemed Telecare 10: 29–33, 2004. 7. Collins K, et al. General practitioners’ perceptions of asynchronous telemedicine in a randomized controlled trial. J Telemed Telecare 10: 94–98, 2004.
Image Analysis of 14 Computerized Clinical Photos Stacy S. Hawkins Unilever Research U.S., Edgewater, New Jersey
CONTENTS 14.1 Introduction..............................................................................................................................................................95 14.2 Advanced Two-Dimensional Facial Prototyping Methods .....................................................................................96 14.2.1 Facial Averaging Methods: Shape, Color, and Texture ..............................................................................96 14.2.2 Characterizing Appearance Differences in Age Groups .............................................................................96 14.2.3 Characterizing Product-Induced Changes ...................................................................................................96 14.2.4 Predictive Transforms ..................................................................................................................................97 14.3 Three-Dimensional Image Analysis ........................................................................................................................97 14.3.1 Three-Dimensional Scanning/Acquisition Systems ....................................................................................97 14.3.2 Cyberware® Scanning ..................................................................................................................................98 14.3.3 Objective Ratings from Facial Scans ..........................................................................................................98 14.3.3.1 Facial Scanning for Measuring Rapid Improvement in Texture .................................................98 14.3.3.2 Facial Scanning for Measuring Improvement in Facial Sagging................................................99 14.4 Discussion ................................................................................................................................................................99 References .........................................................................................................................................................................99
14.1 INTRODUCTION One of the key considerations for a clinical face care image analysis system is to provide accurate and objective tracking of improvement to skin health and condition with intrinsic aging and repair of extrinsic factors due to photodamage. Clinical photography and high-resolution three-dimensional volumetric measurements over a relatively small area (skin surface replicas followed by laser profilometry, or in vivo profilometry systems) have been used to objectively measure these attributes.1–8 Digital photography provides a two-dimensional representation, from which different attributes may be quantified, for example, the area covered by wrinkles in the crow’s-feet, the number of wrinkles, and wrinkle length and width (Figure 14.1). Clinical photography and archiving systems have been developed for teledermatology and semiautomated classification and diagnosis applications.9–11 One of the drawbacks to developing automated image analysis algorithms from two-dimensional photographs, however, is that color/pigmentation changes, shine, and a slight change in positioning or lighting conditions may alter the performance of the image acquisition and processing phases.
FIGURE 14.1 Sample clinical images of before and after product application of effective antiaging formulation. Several features have changed in the after image, including improved overall texture, fine lines, and appearance of sagging.
Objective and sensitive image analysis is therefore dependent on optimization of the image acquisitions systems used: • •
Calibrated, high-resolution photos Optimizing views and lighting for the study
95
96
Handbook of Non-Invasive Methods and the Skin, Second Edition
Further, changes in appearance of skin that occur with health, aging, and photodamage are multidimensional and may require global facial assessment. From a two-dimensional perspective image, many parameters that relate to these multidimensional appearance scales may be derived, including shape, color, texture, frequency, and severity of attribute. From three-dimensional range images, parameters that relate to these multidimensional appearance scales include absolute depth, shape, and volume of facial features.
14.2 ADVANCED TWO-DIMENSIONAL FACIAL PROTOTYPING METHODS 14.2.1 FACIAL AVERAGING METHODS: SHAPE, COLOR, AND TEXTURE
FIGURE 14.2 Method for calculating an average face to represent skin health of all subjects at any given week in the study.
Facial averaging, or morphing, has evolved as a powerful tool for characterizing healthy skin and understanding key drivers of attractive appearance.12–16 Previous research by Perrett et al.17–21 has demonstrated the potential of facial averaging and caricaturing for the development of aging, photodamage, and healthy skin measurement models. These facial prototypes could then be used to evaluate improvement from baseline in skin along multidimensional attributes with skin care formulations. The advantages of this technique over other image processing algorithms are that there are no a priori assumptions on areas or features that improve, and only the important features/sources of variation that are representative of the whole panel will appear in a facial average. Facial prototyping and modeling can therefore drive the development of more relevant objective measures in two and three dimensions. Recent advances in facial averaging have provided more accurate modeling capabilities for representing texture. Wavelet analysis has been used to study the variation of images at multiple scales (coarse to fine) and orientations. The addition of wavelets to the computation of facial averages and caricatures was described by Tiddeman et al.22 to preserve a more accurate statistical model of texture. Facial average computation for shape, color, and texture is a four-step process (Figure 14.2):22
Previous research has shown that by including the average shape, color, and texture to generate facial models, the resulting facial average will be characterized by assessors as belonging to the appropriate age and severity of photodamage groups.23
1. Semiautomatic calculation of over 200 facial features detected in the photographs 2. Calculation of the average facial shape 3. Morphing individual images into the average shape and blending together 4. Computing the average textural features (facial rhytides, mottled hyperpigmentation, pores) and superimposing these features onto the image from step 3
Morning — mild self-foaming wash, clarifying toner, mild cream moisturizer with SPF 15 Evening — mild cleansing pillow, clarifying toner, mild cream moisturizer with no SPF
14.2.2 CHARACTERIZING APPEARANCE DIFFERENCES IN AGE GROUPS Calibrated digital images of female subjects (ages 20 to 68) were captured from a front view. Facial averages were computed for subjects in different decade groups — 20s, 30s, 40s, 50s, and 60s — with 8 to 12 subjects in each age group. The difference in age groups represents a composite of both intrinsic and extrinsic aging factors (Figure 14.3). Additionally, composite facial models for 19 healthy appearance scales have been developed for attributes, including severity of pores, lines and wrinkles, healthy glow/color, and irritation.21
14.2.3 CHARACTERIZING PRODUCT-INDUCED CHANGES Twenty-eight female subjects, ages 18 to 48, with normal, healthy skin, signed informed consent forms to participate in this skin care regimen study.24 This regimen, recommended by the investigative dermatologist for all subjects with normal, healthy skin, was as follows:
Digital photographs were taken of the subjects at baseline and after 1 and 2 weeks of product application. The photographs of the subjects at each evaluation were averaged using the facial averaging technique described in
Computerized Image Analysis of Clinical Photos
97
appearance. Overall, the skin appeared brighter as well as more moisturized (Figure 14.4).
14.2.4 PREDICTIVE TRANSFORMS Once appropriate models for a particular attribute have been developed using facial averaging, varying changes in skin condition can be morphed onto an individual image. The average overall improvement for a panel of subjects after 12 weeks using an antiaging cream may be transformed to a different individual subject to show predicted improvement with an antiaging skin care regime (Figure 14.5).20
14.3 THREE-DIMENSIONAL IMAGE ANALYSIS FIGURE 14.3 Facial averages for subjects in their 20s, 30s, 40s, 50s, and 60s.
14.3.1 THREE-DIMENSIONAL SCANNING/ACQUISITION SYSTEMS Three-dimensional scanners have been commercially available for over 15 years and have been used for anthropometry, animation, and industrial design. The principle of operation for the different systems includes triangulation with laser sources, triangulation with pattern projection (structured light systems), interferometry and Moiré fringe pattern interpretation, stereo photogrammetry, phase measuring profilometry, and interpreting depth by shape-from-shading cues. With all of these systems, there exists a trade-off between resolution and total field of view; however, recent advances in digital cameras and computer graphics rendering techniques allows for detailed capture and quantification of shape, color, and texture. With tabletop systems (scanning occurs along one
FIGURE 14.4 Before and after averages of subjects with mostimproved skin condition after 2 weeks of a cleansing, toning, and moisturizing regimen.
Section 14.2.1, to evaluate how the skin changed over time. Digital pictures of each subject’s face were “morphed” together to generate one composite image representing all study participants. Over the course of the 2week study, this image evolved as subjects were reevaluated. In this way, multidimensional improvement to skin health representing all subjects was computed and translated into one image. After reviewing the average image/facial morphs, the investigative dermatologist saw the greatest change in the cheeks, including less visible pores and smoother skin appearance. The forehead also showed significantly improved evenness of skin tone and a smoother skin
FIGURE 14.5 Transform based on before and after images of subjects using an effective antiaging formulation on a different subject’s face.
98
Handbook of Non-Invasive Methods and the Skin, Second Edition
plane), there is some falloff in depth resolution for global face views.
14.3.2 C
YBERWARE®
SCANNING
The Cyberware® (Cyberware, Monterey, CA) threedimensional scanning technology has been used in the film industry for film special effects and computer gaming applications to create virtual scenes,25,26 for reconstructive and cosmetic surgery,27–29 and in the garment industry to ensure exact-fitting clothing and in designing ergonomically friendly products.30 The Cyberware three-dimensional facial scanner shines a safe, low-intensity laser (780 nm) on a subject to create a lighted profile. A high-quality video sensor captures the profile from two viewpoints. This stereophotogrammetric view then allows the computation of the three-dimensional geometry/range data on the face. The system can digitize thousands of these profiles (scan lines) in a few seconds. Simultaneously, a second video sensor acquires color information (color camera at an angle normal to the range line being captured). A self-contained light source illuminates the surface with cold white light. The color video camera is fitted with a filter that blocks infrared light, and the range sensor is fitted with a filter that blocks visible light. This enables both color digitizing and geometry digitizing to occur simultaneously. Cyberware facial scans provide a virtual photograph model of the face. For viewing and processing the threedimensional scans, a grid that represents a very coarse texture map of the skin may be used to rotate and move to the proper zoom level; resulting images may then be rendered as a three-dimensional virtual image with or without a color overlay (Figure 14.6). It was of interest to test the utility of the measurement of facial landmarks on these surface volume scans for quantifying texture improvement with cosmetic antiaging formulations.
FIGURE 14.6 Sample three-dimensional scan views using Cyberware facial scanner. Left: Preview grid. Middle: Texture data alone. Right: Texture data with color information overlaid result in the appearance of a smoother surface than when texture is viewed alone.
14.3.3 OBJECTIVE RATINGS
FROM
FACIAL SCANS
14.3.3.1 Facial Scanning for Measuring Rapid Improvement in Texture Facial scanning measurements were taken in a randomized, split-face, double-blind, paired comparison, homeuse test of an effective antiaging topical cosmetic formula containing glycolic acid vs. its vehicle.20 The subject population was comprised of 18 healthy female subjects, between the ages of 45 and 70 years, who provided informed consent. This institutional review board (IRB)approved study was conducted over an 8-week period. The following assessments were made at weeks 0, 4, and 8: •
•
•
Visual assessment of photodamage on the face (0 to 9 scale for fine lines and wrinkles on the eye and cheek, discrete hyperpigmentation, and overall photodamage) Self-assessment of photodamage on the face (0 to 4 scale forced-choice directed difference for overall appearance, softness, smoothness, dryness, healthy appearance, even skin color, elastic appearance, firm or loose/sagging skin, fine lines, and wrinkles) Facial scans
Although previous facial clinical studies have confirmed antiaging benefits associated with the glycolic acid cosmetic formulation, there were no clinically measurable differences in visual assessment vs. the vehicle cream in this study. However, several self-perception attributes were significantly different between treatments, favoring the glycolic acid cosmetic formulation: overall appearance, softer skin, healthier appearance, less noticeable wrinkles, and firmer skin after 4 weeks, and less loose/sagging skin after 8 weeks. The facial scanner measurements showed significant improvement of the glycolic acid cosmetic formulation over the vehicle by week 4. In the crow’s-feet area, the glycolic acid cosmetic formulation delivered a significant reduction in wrinkle depth (37% reduction) after 8 weeks, while the vehicle did not significantly reduce this parameter. In addition, the glycolic acid cosmetic formulation significantly reduced wrinkle length (longest line of crow’s-feet area; Figure 14.7) by week 4, and continuing at week 8 (16% reduction). The glycolic acid cosmetic formulation improved wrinkle length significantly greater than the vehicle (p < 0.05). Therefore, facial scanning provided added sensitivity over visual assessment for discriminating effects of a cosmetic antiaging ingredient on lines/wrinkles as early as 4 weeks, where there was no significant difference in visual assessments between treatments even by week 8, with a small panel size.
Computerized Image Analysis of Clinical Photos
99
and week 25 (p < 0.01), with the exception of the vehicle cream at week 11.
14.4 DISCUSSION
FIGURE 14.7 Facial scanning results (mean reduction in length of crow’s-feet) from an 8-week study using a glycolic acid formulation vs. its vehicle cream.
14.3.3.2 Facial Scanning for Measuring Improvement in Facial Sagging In a separate 25-week monadic application facial study, facial scanning measurements were taken to assess improvement to the appearance of texture/creasing on the cheek, including the nasolabial fold (crease from the tip of the nose to the corner of the mouth; see Figure 14.8 for measurement area).20 Measurements were taken at baseline and after 11 and 25 weeks following application of two cosmetic antiaging formulations (formulation 1 and 2) relative to a vehicle control. Both treatments provided significantly greater improvement in volume around the nasolabial fold than the vehicle treatment after 11 weeks. By week 25, volume around the nasolabial fold area increased in formulation 1 by 15%, formulation 2 by 11%, and vehicle by 7%. All treatments had significantly improved from baseline condition at week 11 (p < 0.1)
FIGURE 14.8 Facial scanning results (mean improvement in appearance of lines around the nasolabial area due to sagging) from a 25-week study using two different antiaging formulations vs. their vehicle cream.
Drivers of facial skin appearance scales are multidimensional, and the apparent skin condition of one attribute may be influenced by another or a combination of other attributes. Therefore, it was of interest to develop analysis and visualization techniques to more accurately characterize healthy and photoaging skin conditions. Twodimensional facial averaging or prototyping has been previously used in facial recognition, perception, beauty, and animation applications, and may be successfully applied to quantify appearance attributes. With two-dimensional facial averaging, many attributes may be accurately quantified, including shape, color, texture, frequency, and severity. Recent advances in digital camera technology and three-dimensional scanning systems provide more accurate, sensitive, and objective measurement opportunities of absolute depth and volume for characterizing skin conditions.
REFERENCES 1. Stiller, M.J., Bartolone, J., Stern, R., et al., Topical 8% glycolic acid and 8% L-lactic acid creams for the treatment of photodamaged skin. A double-blind vehiclecontrolled clinical trial, Arch Dermatol, 132, 631, 1996. 2. Friedman, P.M., Skover, G.R., Payonk, G., Kauvar, A.N., and Geronemus, R.G., 3D in-vivo optical skin imaging for topographical quantitative assessment of non-ablative laser technology, Dermatol Surg, 28(3), 199, 2002. 3. Beitner, H., Randomized, placebo-controlled, double blind study on the clinical efficacy of a cream containing 5% alpha-lipoic acid related to photoageing of facial skin, Br J Dermatol, 149(4), 841, 2003. 4. Nardin, P., Nita, D., and Mignot, J., Automation of a series of cutaneous topography measurements from silicon rubber replicas, Skin Res Technol, 8(2), 112, 2002. 5. Piche, E., Hafner, H.M., Hoffmann, J., and Junger, M., FOITS (fast optical in vivo topometry of human skin): new approaches to 3-D surface structures of human skin, Biomed Tech, 45(11), 317, 2000. 6. Fischer, T.W., Wigger-Alberti, W., and Elsner, P., Direct and non-direct measurement techniques for analysis of skin surface topography, Skin Pharmacol Appl Skin Physiol, 12(1-2), 1, 1999. 7. Hawkins, S.S., Wright, S.L., Barrows, J., Bartolone, J., and Weinkauf, R.L., Objective Methods to Evaluate Improvement in Photodamaged Facial Skin, paper presented at the 12th International Symposium on Bioengineering and Skin, Boston, June 25–27, 1998.
100
Handbook of Non-Invasive Methods and the Skin, Second Edition
8. Meyers, C.L., Hoyberg, K., Velez, S., Wright, S.L., and Hawkins, S.S., Quantification of skin texture improvement following washing with ultra-mild liquid cleansers, poster presented at the International Society for Skin Imaging, Washington, DC, March 2001. 9. Miyamoto, K., Takiwaki, H., Hillebrand, G.G., and Arase, S., Utilization of a high-resolution digital imaging system for the objective and quantitative assessment of hyperpigmented spots on the face, Skin Res Technol, 8(2), 73, 2002. 10. Perednia, D.A., Gaines, J.A., and Butruille, T.W., Comparison of the clinical informativeness of photographs and digital imaging media with multiple-choice receiver operating characteristic analysis, Arch Dermatol, 131(3), 292, 1995. 11. Vidmar, D.A., Cruess, D., Hsieh, P., Dolecek, Q., Pak, H., Gwynn, M., Maggio, K., Montemorano, A., Powers, J., Richards, D., Sperling, L., Wong, H., and Yeager, J., The effect of decreasing digital image resolution on teledermatology diagnosis, Telemed J, 5(4), 375, 1999. 12. Perrett, D.I., Burt, D.M., Penton-Voak, I.S., Lee, K.J., Rowland, D.A., and Edwards, R., Symmetry and human facial attractiveness, Evol Hum Behav, 20, 295, 1999. 13. Perrett, D.I., May, K., and Yoshikawa, S., Attractive characteristics of female faces: preference for non-average shape, Nature, 368, 239, 1994. 14. Bruce, V., Ness, H., Hancock, P.J.B., Newman, C., and Rarity, J., Four heads are better than one: combining face composites yields improvements in face likeness, J Appl Psychol, 87(5), 894, 2002. 15. Lee, K.J., and Perrett, D.I., Manipulation of colour and shape information and its consequence upon recognition and best-likeness judgments, Perception, 29(11), 1291, 2000. 16. Calder, A.J., Rowland, D., Young, A.W., Nimmo-Smith, I., Keane, J., and Perrett, D.I., Caricaturing facial expressions, Cognition, 76(2),105, 2000. 17. Perrett, D.I., Penton-Voak, I.S., Little, A.C., Tiddeman, B.P., Burt, D.M., Schmidt, N., Oxley, R., Kinloch, N., and Barrett, L., Facial attractiveness judgements reflect learning of parental age characteristics, Proc R Soc Lond B Biol Sci, 269(1494), 873, 2002.
18. Burt, D.M. and Perrett, D.I., Perceptual asymmetries in judgements of facial attractiveness, age, gender, speech and expression, Neuropsychologia, 35(5), 685, 1997. 19. Burt, D.M. and Perrett, D.I., Perception of age in adult Caucasian male faces: computer graphic manipulation of shape and colour information, Proc R Soc Lond B Biol Sci, 259(1355), 137, 1995. 20. Hawkins, S.S., Perrett, D.I., Tiddeman, B., et al., Novel approaches in texture measurement for cosmetic antiaging evaluation, in Proceedings of the 22nd IFSCC Congress, Edinburgh, Scotland, September 2002, p. 317. 21. Hawkins, S.S., Perrett, D.I., Burt, D.M., Rowland, D.A., and Murahata, R.I., Prototypes of facial attributes developed through image averaging techniques, Int J Cos Sci, 21, 159, 1999. 22. Tiddeman, B., Burt, D.M., and Perrett, D.I. Prototyping and transforming facial textures for perception research, IEEE Comput Graphics Applications, 21, 42, 2001. 23. Hawkins, S.S., Andrew, J., Weinkauf, R.L., Tiddeman, B.P., Payne, K.R., Burt, D.M., Perrett, D.I., and Murahata, R.I. Quantification of Facial Texture by Image Averaging Techniques, paper presented at the International Society for Skin Imaging, Washington, DC, March 2001. 24. Hawkins, S.S., Subramanyan, K., Liu, D., and Bryk, M., Dermatologic Ther, 17(1 Suppl), in press, 2004. 25. New Cyberware Software Makes 3D Scans Ideal for Animation, Virtual Reality, Cyberware press release, Monterey, CA, July 28, 1992. 26. Cyberware Scanner Used to Digitize Michelangelo’s David, Cyberware press release, Monterey, CA, July 14, 2000. 27. O’Grady, K.F. and Antonyshyn, O.M., Facial asymmetry: three-dimensional analysis using laser surface scanning, Plast Reconstr Surg, 104(4), 928, 1999. 28. Bush, K. and Antonyshyn, O., Three-dimensional facial anthropometry using a laser surface scanner: validation of the technique, Plast Reconstr Surg, 98(2), 226, 1996. 29. Hubbs, L., Ed. Taking a Closer Look at Post-Surgical Skin Changes, Skin Aging, October, 100, 2003. 30. The World’s First Whole Body Scanners Bring True Human Forms to Computer Graphics, Cyberware press release, Monterey, CA, May 11, 1995.
Lens — Non-Invasive Oil 15 Magnifying Immersion Examination of the Skin H. Irving Katz and Jane S. Lindholm Minnesota Clinical Study Center, Fridley, Minnesota
CONTENTS 15.1 Introduction............................................................................................................................................................101 15.2 Objective ................................................................................................................................................................102 15.3 Methodological Principals .....................................................................................................................................102 15.4 Sources of Error.....................................................................................................................................................105 15.5 Correlation with Other Methods ...........................................................................................................................106 15.6 Recommendations..................................................................................................................................................106 References .......................................................................................................................................................................106
15.1 INTRODUCTION Visualization of the skin is perhaps the most important part of the dermatologic examination. The observation of a cutaneous sign reflects a pathologic change in the skin due to a derangement within the epidermis, dermis, and/or subcutaneous tissue. Recognition of cutaneous findings allows a definitive or differential diagnosis of a dermatological disorder. Using naked-eye inspection the experienced observer detects many visible gross morphologic features, such as size, shape, color, or type of lesion. However, some subtle or ambiguous diagnostic features require further amplification of the finding. The use of skin surface magnification techniques assists an observer in visualizing such enigmatic findings. Skin surface magnification or surface microscopy is not a new method to examine changes occurring on or within the skin. Hinselmann, Goldman, Gilje, and O’Leary, along with many others since, have reported the usefulness of skin surface microscopy methods for a variety of dermatologic conditions.1–12 In addition the past decade surface microscopy has been to study and differentiate benign and malignant pigmented melanocytic neoplasms.13–19 Extremes of skin surface magnification that are available range from just slightly more than life size to hundreds of times the size of the actual image viewed, as summarized in Table 15.1 Although high-power resolution is obtainable, such applications are not within the reach of most practitioners. Therefore, only low-power methods
TABLE 15.1 Skin Surface Microscopy Techniques Low-power surface magnification (generally 10×.
19.3.3 CORRELATION
WITH
OTHER METHODS
There is some loss of fine details at the high magnifications attainable in the SEM.
19.4 SILICONE ELASTOMER REPLICATION 19.4.1 METHODOLOGICAL PRINCIPLE The negative replica — The silicone-elastomer replication has its basis in the products developed for use in clinical dentistry.3,14 There are five main requirements on the plastic used for producing the negative mold.
FIGURE 19.1 The steps in making a metal/carbon replica of surfaces. The clean surface (a) is covered with a thin carbon or metal (Cr, Pd/Pt, Pt, Au) film at an angle (often 6 to 10°) by evaporation in vacuum (b). The evaporated film is stabilized by application of a thin plastic film, e.g., formvar (c) (hatched area). The object is removed mechanically (which often disrupts the replica) or preferably by chemical dissolution (d). The (negative) replica is stabilized by a thick layer of carbon evaporated onto the replica (e) in vacuum. The plastic film is removed by the appropriate solvent for the plastic in question. The positive carbon replica is now transferred to a conventional electron microscopic grid (f) and viewed in the TEM.
19.3 PLASTIC IMPRESSION TECHNIQUE 19.3.1 METHODOLOGICAL PRINCIPLE Before introduction of the SEM, but even more recently, replicas for light microscopy have been made from plastic compositions that required a cleaned and dry, hairless area for the replication.1,6 Consequently interest has been
1. The silicone plastic should have a low viscosity to adhere closely even to the fine details of the surface 2. It should adhere well even to wet surfaces 3. After a fast, and complete, polymerization it should be released from the original specimen without leaving any material behind. 4. It should possess an elastic memory to allow a complete return to the original status even when withdrawn from undercuts. 5. The polymerization process should not produce heat, i.e., involve an exothermic reaction which may change surface properties of the object, and cause the discomfort of the subject. The positive replica — The plastic used for producing the positive replica should cure at room temperature with as little release of heat as possible to prevent deformation of the negative mold. In the SEM micrographs accompanying this chapter (Figure 19.2) molds were made from Provil-L® (Bayer Dental D-5090 Leverkusen, Germany), which is characterized as a low-viscosity, type I silicone meeting the
150
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 19.2 Replicas of a normal skin surface (volar aspect of lower arm) obtained by the silicon elastomer two-step method. (a) Skin surface after approximately 15 min of exposure to damp cloth saturated with distilled water, 230×.
FIGURE 19.2 (d) 212×.
19.3c). It is then allowed to set for 3 min before gentle removal from the (skin) surface (Figure 19.3d). The negative replica is subsequently covered with an Araldite® plastic (CIBA-Geigy) (Figure 19.3e) which cures within 3 to 5 h depending on the volume applied. Alternatively we have used a methacrylate designed for whole-mount embedding of insects, etc., which has longer curing time. The surface of the plastic, positive replica is subsequently made conductive by gold sputtering (Figure 19.3f and g).
19.4.2 SOURCES
FIGURE 19.2 (b) 503×. There are no obvious villiform projections on the surface that has been exposed to the environment.
OF
ERRORS
The negative mold — A large negative imprint of the skin surface (i.e., >1 × 1 cm) tends to bend when loosened from the original surface and this large curvature remains when the positive replica is made (Figure 19.3f). Due to high total absorption of energy in the electron beam, a larger-than-the-stub specimen tends to be unstable in the beam, i.e., be subject to drift during viewing the SEM. The positive replica — When making the positive replica the amount of accelerator may be crucial to the final results. If the curing process occurs at too fast a rate, gas bubbles will accumulate at the replica surface.
19.4.3 RECOMMENDATIONS
FIGURE 19.2 (c) Corresponding area sampled dry on the following day, 101×.
requirements of ISO 4823, type (e) 3, category A (adhesion-induced polymerization). The silicone plastic is thoroughly mixed with an equal volume of catalyst and immediately applied to the surface to be replicated (Figure
The negative replica — The mixing of silicon base and curer is a critical stage in making a replication. The two components should be thoroughly mixed but agitation should not be so vigorous as to produce air bubbles. The drawbacks of manual mixing can be virtually eliminated when using a dual vessel ejector (Bayer Cartridge delivery dispensing gun) (Figure 19.3c). When making a negative replica for SEM care should be taken to cover a surface area no greater than the SEM specimen holder (the “stub”) to avoid the unnecessary heating that follows from having a large specimen surface. When the negative replica has been removed after setting, its surface can be inspected under a light microscope at about 40× magnification to
Skin Replication for Light and Scanning Electron Microscopy
151
FIGURE 19.3 Silicon elastomer replication of an integument surface. (a) The skin surface is briefly cleaned by a quick rinse in cool tap water and (b) blotted dry. (c) The elastomer is applied and allowed to cure for about 3 min. (d) The negative mold produced is gently removed. (e) The mold is covered with a plastic to produce a positive replica. (f) After gold-sputtering the positive replica (left) and the negative mold can be inspected for surface defects by light microscopy.
artefacts. Sometimes an improvement of resolution of details by the silicone is obtain through a quick cool water rinsing which undoubtedly removes water-soluble surface material. At low magnification (e.g., 40×) no swelling is apparent from this process. Rinsing may, however, introduce artefacts in lesional skin. It is then preferable to remove loose surface material by making two or three
impressions from the same surface rather than making the lesion subject to tap water rinse. The sequentially obtained molds can be checked against each other for artifacts. The negative imprint, the mold, is not suited for direct study in the SEM because it will melt and evaporate when hit by the electron beam. However, it can be used directly for light microscopic and photographic observations at
152
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 19.3 (g) Even large objects can be faithfully replicated. Left: fingertip of digit V of the author. (h) A stub machined to have a groove with undercut will prevent drift in large objects during SEM study.
low and moderate magnifications.8 If this is the object, rather than a study in the SEM, large areas can be replicated, e.g., 2 × 2 cm (Figure 19.3g). The positive replica — The making of a successful positive replica involves a good choice of plastic. We have used Araldite® (Ciba-Geigy) in a 1:1 mixture with accelerator. The manufacturer’s advice on mixing proportions for plastic and accelerator should be tested for each batch of plastic as it may vary during aging of these materials. It is our experience that it is not unusual that the amount of accelerator should be somewhat reduced to get a curing rate that does not produce heat and solvent gas bubbles. However, it is easy to get an incomplete curing that results in a sticky surface which deforms on removal from the negative mold. An alternative way of reducing the risk of gas bubbles at the interface between the negative mold and the plastic is to moisten the surface of the silicon imprint with the solvent of the plastic (e.g., acetone for Araldite®) immediately before pouring the plastic onto the negative template. It is advised that positive casts are inspected for the presence of gas bubbles in the surface structures under a preparation microscope after gold sputtering. If bubbles are present they usually attain a size that allows them to be seen at a magnification of 40. As an
alternative to Araldite® we have used a methacrylate designed for embedding of large objects such as insects, e.g., a beetle. This methacrylate, which takes more than 24 h to cure even in thin sheets, tends to be very brittle. It reproduces the surface details well in our experience. Pfister and Neukirchner10 used a polystyrol granulate dissolved in toluol for the positive replica. The non-cured plastic has a syrup-like consistency. To avoid air bubbles in small crevices of the negative replica the authors “moisten” it with the solvent, toluol, before applying the plastic. The hardening time of this polystyrol plastic is comparatively long, approximately 24 h. The authors claim that magnifications up to 5000× are attainable with this technique. Gold sputtering — The plastic material of the positive replica is an insolator. Gold sputtering of the surface provides a conductive film that distributes charges to ground potential but also contributes as a heat sink. The sputtering should be performed so as to get a continuous contact between the replica surface and the specimen stub. This is most easily achieved if the stub surface is cleared of the specimen at small point. When large objects are used, e.g., a replica of a fingertip with a nail, it is advantageous if the stub can be molded into the positive plastic replica. This can be achieved by making a groove with undercut (Figure 19.3h) with a milling cutter. Alternatively a cavity with undercut in the stub surface can be made using a dental drill. Through these means drift is virtually completely eliminated.
19.5 PRESENT STATUS OF REPLICATION TECHNIQUES IN DERMATOLOGY Most areas of the human integument in health1,2,6,12,15 have been described using replication techniques. In addition pathological conditions, including lesions of psoriasis,13 superficial actinic porokerastosis, as well as more unusual conditions like Gorlin’s syndrome,6 have been documented. It is noteworthy that data presented in the literature on topographic data collected by replication (and corresponding) techniques on skin and its appendices in general merely have a descriptive character and provide little, if any, functional interpretation of the findings. Cosmetic industries have long utilized SEM studies of the effect of cosmetic formulations on the skin surface and the integument appendices, but details of this information have not been publicly available and cannot be scientifically evaluated. It is obvious that topographical methods of investigating the skin surface represent an interesting and potentially fruitful area of dermatological research. Combination with morphometric systems, image analysis systems, or other physical measurement systems5 will allow quantitative analysis of changes in the surface structures as a result of the progress of a disease or a
Skin Replication for Light and Scanning Electron Microscopy
treatment of a disease. In addition to such applications a more extensive use of the excellent replication materials presently available will no doubt increase our knowledge of the dynamics of skin function in health and disease.
ACKNOWLEDGMENT I am indebted to Mr. Eva Lundewall of the Swedish branch of Phillips Industrial Electronics AB for producing the SEM micrographs and to Ms. Margareta Andersson for photography and lay-out of illustrations. The golden rule is that there are no golden rules. George Bernhard Shaw
REFERENCES 1. Chinn, H.D. and Dobson, R.L., The topographic anatomy of human skin, Arch. Dermatol., 89, 155, 1964. 2. Forslind, B., Clinical applications of scanning electron microscopy and X-ray microanalysis in dermatology, Scanning Electron Microsc., 1, 183, 1984. 3. Jokstad, A. and Mjör, L.A., Assessment of marginal degradation of restorations on impressions, Acta Odontol. Scand., 49, 15, 1991. 4. Marks, R. and Dawber, R.P.R., Skin surface biopsy: an improved replica for the examination of the horny layer, Br. J. Derm., 84, 117, 1971. 5. Marks, R. and Pearse, A.D., Surfometry. A method of evaluating the internal structure of the stratum corneum, Br. J. Dermatol., 92, 651, 1975.
153
6. Nayler, J.R., Applications of the skin surface replica technique in dermatology, J. Audiovis. Media Med., 21, 1984. 7. Pameijer, C.H., Replica techniques for scanning electron microscopy — a review, Scanning Electron Microsc., II, 831, 1978. 8. Pameijer, C.H., Replication techniques with new dental impression materials in combination with different negative impression materials, Scanning Electron Microsc., II, 571, 1979. 9. Pfefferkorn, G. and Boyde, A., Review of replica techniques for scanning electron microscopy, Scanning Electron Microsc., 1, 75, 322, 1974. 10. Pfister, T.C. and Neukirchner, A., Raster-elektronenmikroskopischer Untersuchungen am kranken Nagel mittels Abdruck-Verfahren, Fortschr. Med., 98, 1465, 1980. 11. Sampson, J., A method for replicating dry or moist surfaces for examination by light microscopy, Nature, 191, 932, 1961. 12. Tring, F.C. and Murgatroyd, L.B., Surface microtopography of normal human skin, Arch. Dermatol., 109, 223, 1974. 13. Tring, F.C. and Murgatroyd, L.B., Psoriasis — changes in surface microtopography, Arch. Dermatol., 111, 476, 1975. 14. Walsh, T.F., Waimsley, A.D., and Carrotte, P.V., Scanning electron microscopic investigation of changes in the dentogingival area during experimental gingivitis, J. Clin. Periodontol., 18, 20, 1991. 15. Wagner, G. and Goltz, R.W., Human cutaneous topography. A new photographic technique: observation on normal skin, Cutis, 23, 830, 1979. 16. Wolf, J., Die innere Struktur der Zellen des Stratum desquamans der Menschlichen Epidermis, Z. Mikr. Anat. Forsch., 46, 170, 1939.
Surface Replica Image Analysis of 20 Skin Furrows and Wrinkles Pierre Corcuff and Jean-Luc Lévêque Laboratoires de Recherche de L’OREAL, Aulnay Sous Bois, France
CONTENTS 20.1 Introduction............................................................................................................................................................155 20.2 Objective ................................................................................................................................................................155 20.3 Basic Methodology................................................................................................................................................156 20.3.1 Skin Surface Replica .................................................................................................................................156 20.3.2 Shadowing Method ....................................................................................................................................156 20.3.3 Image Analysis ..........................................................................................................................................156 20.3.3.1 Image Analyzer ..........................................................................................................................156 20.3.3.2 Analysis of the Image ................................................................................................................157 20.3.3.3 Measurement Parameters ...........................................................................................................157 20.3.3.4 Automation .................................................................................................................................158 20.3.4 Choice of Lighting Angle..........................................................................................................................159 20.4 Sources of Error.....................................................................................................................................................159 20.4.1 Replica Artifacts ........................................................................................................................................159 20.4.2 Analysis of the Image................................................................................................................................160 20.5 Correlation with Other Methods ...........................................................................................................................160 20.6 Recommendations..................................................................................................................................................161 References .......................................................................................................................................................................161
20.1 INTRODUCTION Among the various non-invasive methods described in this handbook, the description and measurement of the geometric organization of the skin relief involve the same goal, i.e., the study of underlying biological and physiological phenomena based on their impact on the surface. The skin surface “messages” result mainly from the organization of the dermis and its collagen and elastin networks, but the state of the epidermis and stratum corneum can also play a role. The problem is more complex than the simple measurement of an excretion (sebum, sweat, transepidermal water loss), as the structure is three-dimensional and needs a minimum of geometric parameters; it is therefore difficult to make a simple description and interpretation. A method developed in the 1970s to measure the roughness of metallic surfaces — mechanical profilometry — has since been adapted by several authors1–3 to study cutaneous topography. Profilometry is still widely used today, as the replacement of the physical probe by a laser beam has
overcome two major obstacles: contact between the instrument and the surface, and the time required for measurement. Given the early drawbacks of this method, we proposed in 1981 a technique based on image analysis of a skin replica, which was rapid and well adapted to the anisotropy of the skin relief, but which was prohibitively expensive.4 With the fall in the cost of electronic components and computers, image analysis has now become more accessible, and an increasing number of teams have adopted this method, which, as we shall see later in this chapter, requires impeccable technique and discipline.
20.2 OBJECTIVE At the beginning of the 1980s, the only method available was thus mechanical profilometry. It required a supple negative replica to be made, followed by a counter-replica made of hard resin (which resisted deformation by the probe). The main criticisms were the accumulation of artifacts by successive replication, the directional traces 155
156
Handbook of Non-Invasive Methods and the Skin, Second Edition
that were poorly adapted to the anisotropy of the skin surface, and the fact that the roughness parameters were difficult to interpret in terms of skin relief. Image analysis was used by few teams at that time. We discovered that it was capable of scanning the first replica in several directions and that the geometric parameters provided by this technique (orientation, number, and depth) were easy to interpret. Finally, contact between the instrument and the study surface was avoided. The objective was to describe the organization of the primary lines according to Hashimoto’s classification5 by using a minimum of parameters. When this objective had been reached, the method was extended to the study of wrinkles of the crow’s-foot area. In order to obtain objective and reliable results, we opted for an entirely automated method, which disallowed all interactive intervention.
Positive cast
Negative replica
20.3 BASIC METHODOLOGY The principle of image analysis can be likened to an air passenger watching the shadows crossing the Alps as the sun traverses the winter sky. Each mountain has a dark side and a bright side; some valleys are illuminated in the morning, others in the afternoon. The brightness increases the contrast between the areas of shadow and light, just like topologic details.
20.3.1 SKIN SURFACE REPLICA Applied to skin relief, this method carried the following constraints. The mold had to be white (as snow), matte, and opaque. Among the brands studied at the time, the silicon resin SILFLO from Flexico (England) fit the bill.6 It guaranteed reliable reproduction, the absence of further deformation, and documented artifacts.7 The fact that the samples could be stored for about 2 years was an added advantage. The negative replica is obtained in the following way. Two or three drops of Bayer catalyzer are added to 1 g of SILFLO resin and rapidly mixed in a cup with a spatula. The paste is then immediately applied to the study surface, which has first been delimited with an adhesive paper ring. The resin hardens within 2 or 3 min and the replica is removed gently by lifting from the tongue of the ring.
20.3.2 SHADOWING METHOD The cutaneous topography is such that the observer is obliged to shadow negative replicas rather than positive replicas, as is clearly shown in Figure 20.1. The sun in the above analogy is replaced by an optical fiber system providing illumination at a precisely defined angle relative to the plane of observation. For evident technical reasons, it is simpler to rotate the sample than the lighting system
FIGURE 20.1 Shadows generated at the surface of negative and positive replicas by grazing lighting. On the negative imprint the width of the shadow is large enough to allow a correct estimation of the peak height.
to simulate the movement of the sun. The negative replica with its ring is inserted into a metallic device covered with nonreflective black felt. This ensures that the sample is perfectly flat; it also means that the sample can be held perfectly horizontal and centered with regard to the light source and video camera.
20.3.3 IMAGE ANALYSIS 20.3.3.1 Image Analyzer The basic principle of the image analyzer is the segmentation of an image according to shades of gray, and this is perfect for selecting areas of shadow created by the oblique light impinging on the sample (skin furrows). In 1979, we selected the Quantimet 720 manufactured by Cambridge Instruments, as it had a number of technical features particularly adapted to this application and could be automated. We now use the Quantimet 970, which still has the essential elements of its predecessor; in this way, we can compare the results obtained in 1981 with those obtained in 1993. One important feature of this generation of system is the image format: 720 lines, 880 points per line, 605,000 pixels. To avoid introducing a bias into the measurements during sample rotation, the field of measurement must be circular, and this brings us down to 300,000 pixels (which is still a considerable amount). By way of comparison, a standard image formed of 512 × 512 pixels only provides a circular field of measurement of 180,000 pixels. Table 20.1 gives comparative data for the various measurement methods we have used. With the
Skin Surface Replica Image Analysis of Furrows and Wrinkles
TABLE 20.1 Comparison of Methods Used for Three-Dimensional Analysis of Skin Surface Replica According to the Size of the Explored Surface, the Number of Points, Depth Resolution, and Recording Time
Instrument Quantimet 970 Standard I.A. Mechanical Profilometer Optical (laser) Profilometer Confocal microscope (TSM)
(a)
(b)
Surface Area (mm2) 100 50 25 50 1
Picture Depth Points Resolution Time (no.) (μm) (min) 300,000 180,000 62,500 250,000 250,000
8 8 1 3 1
5 2 90 10 10
(c)
(d)
FIGURE 20.2 The selection of shadows by the image analyzer. In this series of pictures, white features correspond to binary images of shadows segmented at threshold level 45 (A) the principal orientation of furrows is perpendicular to the lighting; (B) binary image resulting from the vertical opening of A. (C) the principal orientation of furrows is not perpendicular to the lighting; (D) the vertical opening almost suppressed the nonperpendicular shadows.
Quantimet, one can analyze 1 cm2 of skin with a resolution of 10 mm in the horizontal plane and 8 mm in the vertical plane. One question that arises is whether more weight should be given to the field of measurement or to the resolution. Our experience shows that the area measured should be greater than 0.5 cm2 and that a resolution of 10 mm is adequate to study the primary lines. The most important factors are the sampling procedure and the reproducibility of the measurements. Obviously, the study
157
of secondary furrows needs those higher resolutions provided by profilometry. But software should be capable of analyzing separately the various classes of skin furrows. 20.3.3.2 Analysis of the Image The selection of the shadows is the most important phase of the operation. The 6-bit analog-to-digital (A/D) converter provides a scale of 64 shades of gray (0 = black, 63 = white). Figure 20.2A and C show the results obtained with a threshold at level 45, i.e., selection of all the levels between 0 and 45. The position of the light source relative to the video scan determines the significance of the measurement parameters: the light comes from the right of the screen. Subsequent operations consist of extracting only the shadows formed by peaks perpendicular to the light source, i.e., in a vertical direction. To do so, the binary image is opened by 15 pixels (erosion + dilation) with a linear, vertical structuring element (Figure 20.2B and D). This mathematical morphologic operation8 has two consequences: to eliminate small events (noise, round objects, bubbles) and to emphasize the anisotropy of the skin furrows (suppression of oblique and horizontal shadows, joining of vertical shadow segments). 20.3.3.3 Measurement Parameters The two parameters measured in binary images are known as field parameters. They correspond, in fact, to the very first stereological parameters available on image analyzers when the latter were incapable of identifying objects. The area fraction AA is the percentage surface area occupied by shadows in the field of measurement. The intercept I is the horizontal projection of these shadows. Given the particular arrangement of the replica–lighting–camera ensemble, the I lines correspond exactly to the crests of the skin furrows, and thus represent the total length of furrows in the field of analysis (Figure 20.3). On the basis of these two stereological parameters and the angle of incident light a, I can be used to estimate the number of lines per square centimeter (or per linear centimeter), while AA, I, and a can be used to estimate their mean depth. These parameters are measured at each step of the sample rotation (9˚ steps through 360˚, giving 41 series of measurements). If the values of I are plotted in polar coordinates according to the angle of rotation, one obtains the rose of directions illustrated in Figure 20.4. This graph contains local maxima, with a 180˚ symmetry, which corresponds to the main orientations of the primary lines. The orientation of each network of parallel furrows can then be deduced relative to the reference axis (the tongue of the sample); the density of lines and their mean depth can also be deduced for each main axis of furrows. All these
158
Handbook of Non-Invasive Methods and the Skin, Second Edition
Lighting
k1 = 2.10 −4 ⋅ π ⋅ n ⋅ p k1 =
Axis 1
1 − k12
⎛ k + k2 ⎞ If k1 > 1 then c = ln ⎜ 1 ⎝ k1 − k2 ⎟⎠ AA
I Axis 2
⎛k ⎞ If k1 < 1 then c = π − 2 arctan ⎜ 1 ⎟ ⎝ k2 ⎠ 2 ⋅ k1 c + π π ⋅ k2
Finally, E = L
α
FIGURE 20.3 Parameters selected by the image analyzer. AA is the fraction area of shadows, I is the intercept line drawing the crest line of negative furrows, α is the lighting angle.
If E1 and E2 are the coefficients for each of the two axes mutually forming an angle b, CDSS = E1 [1 + (E2 – 1) sinb]. The importance of this parameter has been shown in studies of chronological aging,9,10 the effect of ultraviolet irradiation,11 and the skin deformation process.12 20.3.3.4 Automation
Microrelief: 4427
2
Full automation of the image analysis procedure has been the subject of a previous publication.13 A new “robot” has recently been designed to increase the analytical capacity without augmenting the time required for a full cycle. Eighty replicas instead of 40 can be analyzed in an 8-hour period. Figure 20.5 shows details of the automated
a
1
b
c
FIGURE 20.4 A rose of direction. Arrow 1 gives the orientation of the first axis of furrows and arrow 2 the second axis.
operations, i.e., the analysis of the orientations and parameters of each network, are carried out with a computer program written in Pascal. A supplementary parameter has been forwarded to characterize the relief by means of a single value. It is known as the coefficient of developed skin surface (CDSS), which represents the tissue reserve or deformation reservoir of the skin. On the basis of the mathematical cycloid arch model,4 a coefficient E is calculated for each furrow axis. A simplified formula is provided below, which is valid when the field of analysis is close to 1 cm2 (±10%). Consider n the number of furrows per centimeter squared and p their mean depth in microns determined for a main axis:
d
e
f
FIGURE 20.5 The “robot”. (a) Objective lens, (b) optical fiber lighting, (c) the circular tray accepts 80 samples, (d) a replica in position for analysis, (e) magnetic arm, (f) autofocus motor.
Skin Surface Replica Image Analysis of Furrows and Wrinkles
The lighting angle plays a crucial role in determining the nature of the shadows that will be analyzed: the lower the light source, the greater the detail. In this way, the lighting acts as a high-pass filter. In the case of a fine and regular microrelief, such as that observed in a child, an angle of 17 or 20˚ will enable the observer to measure a large number of shallow furrows (Figure 20.6A), while an angle of 38 or 45˚ will select the few deeper furrows. In this type of topography, line density is the most sensitive parameter. In the case of aged skin (Figure 20.6B), from which the fine lines have disappeared, the lighting angle has little influence on the number of furrows, but more influence on the mean depth. With a very low angled light source, two risks arise. If the sample is not perfectly flat, large areas of shadow can be generated, but this artifact is generally easy to identify in the measurements. Very high peaks can mask lower peaks located in their shade, and this bias is, on the contrary, difficult to detect. A compromise solution has been based on the value of the coefficient E1 as a function of the lighting angle (Figure 20.7). This coefficient passes through a maximum that depends on the surface topography. For the skin microrelief, the maximum is from 17 to 26˚; we use the higher value to avoid the above-mentioned risks. The optimum value when studying crow’sfeet is 38˚.
20.4 SOURCES OF ERROR 20.4.1 REPLICA ARTIFACTS The artifacts created during the production of the skin surface replica arise from the preparation of the volunteer and the technician’s experience. It is essential that the subject remain immobile during the polymerization phase. To this end, the room should be calm, with dim lighting and a temperature of 20˚C; in addition, the subject should be comfortable and given time to relax and adapt to the surroundings. These considerations are particularly important when taking replicas of the crow’s-foot area; in this case, the volunteer is placed in the lateral decubitus position, eyes closed and face relaxed.
30
μm
20
50
10
25
Depth
LIGHTING ANGLE
Young skin
0
0 17 20
26
32
38
45
Incident light angle (degrees) μm 150
Aged skin
30
125
20
100
10
75
0
Depth
OF
Line density per cm of skin
20.3.4 CHOICE
40
Line density per cm of skin
apparatus. A circular tray presents each sample, which is translated by a magnetic arm for positioning under the camera. The replica is then raised by the autofocus motor of the Quantimet and rotated by a stepwise motor. At the end of the analysis, the replica is lowered, placed in its socket, and the next sample is presented for analysis. The time previously required to control the position of each sample is thus saved.
159
50 38 17 20 26 32 Incident light angle (degrees)
45
FIGURE 20.6 Density of lines and mean depth of furrows plotted vs. the lighting angle: (A) young skin replica of the volar forearm, (B) aged skin replica of the volar forearm.
The conditions necessary for obtaining negative replicas must be followed to the letter. The main sources of artifacts are as follows. Bubbles of sweat can form holes at the intersection of primary lines if the room is too warm or the subject stressed or emotional (Figure 20.8). Areas lacking any relief are due to inadequate mixing of the resin with the hardener (Figure 20.9). One of the most frequent causes of artifacts is polymerization of the resin before it is applied to the skin. To ensure that application has been carried out correctly, the underside of the replica should be inspected: if the surface is smooth and shiny, the application is correct; if the surface is irregular, matte, and wrinkled, the resin had already started to harden before application (Figure 20.10).
160
Handbook of Non-Invasive Methods and the Skin, Second Edition
E1 1.25
1.20 c 1.15 a
b
b
FIGURE 20.9 Replica artefacts: (a) an “academic” replica, (b) flat areas (arrows) obtained with a nonhomogenous mixture.
1.10 a 1.05
α°
1.00 17 20
26
32
38
45
Incident light angle (degrees)
FIGURE 20.7 Evolution of the cycloid arch coefficient E1 (see text) according to the lighting angle: (a) young skin of Figure 20.6A, (b) aged skin of Figure 20.6B, (c) crow’s-feet wrinkles.
a
b
FIGURE 20.10 Replica artefacts: (a) the smooth verso of an “academic” replica, (b) aspect of a replica performed during the polymerization of the resin.
Interfering lights and changes in the ambient lighting influence the reproducibility of the measurements, and it is best to work in total darkness. Incorrect focusing is also a source of error, and autofocusing is thus a major advantage. As a rule, all these types of error can be overcome by regular analysis of a standard replica during a series of measurements.
FIGURE 20.8 SEM picture of a skin surface replica showing droplets at the intersect of primary furrows.
20.4.2 ANALYSIS
OF THE IMAGE
A frequent source of error is that the replica is not perfectly flat. Analysis through 360˚ permits this type of error to be identified simply, since the values for each 180˚ segment should be symmetrical. Normally, the rigid metal sample support enables most flatness errors to be corrected. All other errors are due to the image analyzer. The most frequent are time shifts and instability of the light source or electronic circuits of the camera. The uniformity of the incident light is also a critical factor: in general, there is a light gradient that has to be corrected by the image analyzer (background substraction).
20.5 CORRELATION WITH OTHER METHODS There have been few publications comparing mechanical profilometry and image analysis by shadowing. In 1985, we reported a comparison of the effects of antiwrinkle product on the face, using mechanical profilometry for the wrinkles of the forehead and image analysis for the crow’s-foot wrinkles.14 The percentage reductions were 27% for the profile area and 30% for the CDSS. The depth of the deeper lines (>50 mm) fell by 23% in the profilometric method and 16% in the image analysis technique. Schmidt et al.15 compared the two methods on the same replicas taken from the crow’s-foot area. They found a good correlation between the depth given by image analysis and the height of the peaks given by profilometry,
Skin Surface Replica Image Analysis of Furrows and Wrinkles
and also between the CDSS and peak surfaces. The best correlation was obtained with peak heights between 50 and 100 mm, i.e., the optimal domain in image analysis with a lighting angle of 38˚. Hayashi et al.16 recently demonstrated that the fractions of shadow corrected for the lighting angle (RWA parameter) and the maximum depth of the wrinkles (V) were independent of the lighting angle. V was identical to the height given by a micrometer. Schrader and Bielfeldt17 compared data from image analysis, mechanical profilometry, Corneometer®, and methylene blue staining of the skin during cosmetic treatments. Linear regression study led to weak but significant correlations between the methods. Some authors have preferred a profilometric approach to a threshold method in image analysis: the optical profilometry traces curves corresponding to gray levels along a scanning line in the video image. The profiles are analyzed with the same parameters as those used to analyze a mechanical trace, and good correlations between the two types of profile have been reported.18 Grove and Grove,19 also using optical profilometry by image analysis, showed that there was a good correlation between the roughness parameters (Ra, Rz) and Daniell’s visual classification during treatment of the crow’s-foot area with retinoids.
20.6 RECOMMENDATIONS The study of the skin microtopography or facial lines by means of image analysis, whatever the technique used, is more a problem of sampling and reproducibility than one of sensitivity. It is thus best to use approaches that analyze as large an area as possible, and a large number of samples. The time required for analysis is consequently important, and automation clearly has a role to play. In the special case of crow’s-foot wrinkles, the number of significant events (lines) is relatively small. Studies of changes in these lines during treatment must thus be based on strictly identical areas of analysis (before and after treatment). The image analyzer is a powerful tool in this setting, since it enables the study areas to be superimposed. As we have seen, the production of the replica is a crucial step that requires the utmost care. This is an essential consideration because, despite the fact that the replica method permits samples to be studied some time after their collection, it is essential that the replicas truly reflect the state of the skin at the time they are made. Replicas are durable and easy to store, and are also readily transportable. It is thus astonishing that so few multicenter studies comparing the different available methods have been published. A consensus on the effects of antiaging treatments could be arrived at by this approach, which would finally convince the international scientific community of their efficacy.
161
REFERENCES 1. Cook, T.H., Profilometry of skin. A useful tool for the substantiation of cosmetic efficacy, J. Soc. Cosmet. Chem., 31, 339, 1980. 2. Makki, S., Mignot, J., and Zahouani, H., Statistical analysis and three dimensional representation of human skin surface, J. Soc. Cosmet. Chem., 35, 311, 1984. 3. Hoppe, U. and Sauermann, G., Quantitative analysis of the skin surface by means of digital signal processing, J. Soc. Cosmet. Chem., 36, 105, 1985. 4. Corcuff, P., de Rigal, J., and Lévêque, J.L., Image analysis of the cutaneous microrelief, Bioeng. Skin Newslett., 4, 16, 1982. 5. Hashimoto, K., New methods for surface ultrastructure. Comparative studies of scanning electron microscopy, transmission electron microscopy and replica method, Int. J. Dermatol., 13, 357, 1974. 6. Makki, S., Barbenel, J.C., and Agache, P., A quantitative method for the assessment of the microtopography of human skin, Acta Derm. Venereol., 59, 285, 1979. 7. Gordon, K.D., Pitting and bubbling artefacts in surface replicas made with silicone elastomers, J. Microsc., 134, 183, 1984. 8. Serra, J., in Image Analysis and Mathematical Morphology, Vol. 2, Theoretical Advances, Serra, J., Ed., Academic Press, London, 1988. 9. Corcuff, P., de Rigal, J., Makki, S., Lévêque, J.L., and Agache, P., Skin relief and aging, J. Soc. Cosmet. Chem., 34, 177, 1983. 10. Corcuff, P., Lévêque, J.L., Grove, G.L., and Kligman, A.M., The impact of aging on the microrelief of periorbital and leg skin, J. Soc. Cosmet. Chem., 82, 145, 1987. 11. Corcuff, P., François, A.M., Lévêque, J.L., and Porte, G., Microrelief changes in chronically sun-exposed human skin, Photodermatology, 5, 92, 1988. 12. Corcuff, P., de Lacharrière, O., and Lévêque, J.L., Extension induced changes in the microrelief of the human volar forearm: variation with age, J. Gerontol. Med. Sci., 46, 223, 1991. 13. Corcuff, P., Chatenay, F., and Lévêque, J.L., A fully automated system to study skin surface patterns, Int. J. Cosmet. Sci., 6, 167, 1984. 14. Corcuff, P., Chatenay, F., and Brun, A., Evaluation of anti-wrinkle effects on humans, Int. J. Cosmet. Sci., 7, 117, 1985. 15. Schmidt, C., Camus, C., Candiu, H., Her, C., Soudant, E., and Bazin, R., Correlation d’une technique d’analyse d’images et d’une méthode profilométrique dans l’étude des rides de la patte d’oie, Int. J. Int. Sci., 9, 21, 1987. 16. Hayashi, S., Matsuki, T., Matsue, K., Arai, S., Fukuda, Y., and Yoneya, T., Changes in facial wrinkles by aging and application of cosmetics, in Proceedings of the IFSCC Congress, Yokohama, Japan, 1992, p. 733. 17. Schrader, K. and Bielfeldt, S., Comparative studies of skin roughness measurements by image analysis and several in vivo skin testing methods, J. Soc. Cosmet. Chem., 42, 385, 1991.
162
Handbook of Non-Invasive Methods and the Skin, Second Edition
18. Marshall, R.J. and Marks, R., Assessment of skin surface by scanning densitometry of macrophotographs, Clin. Exp. Dermatol., 8, 121, 1983.
19. Grove, G. and Grove, M.J., Effects of topical retinoids on photoaged skin as measured by optical profilometry, in Methods in Enzymology, Packer, L., Ed., Academic Press, New York, 1990, p. 360.
Method for Skin Surface 21 Stylus Contour Measurement Johannes Gassmueller BioSkin Institut für Dermatologische Forschung und Entwicklung GmbH, Hamburg, Germany
Andrei Kecskés Schering AG, Dermatologie/Humanpharmakologie, Berlin, Germany
Peter Jahn Schering AG, Diagnostika Koordination, Berlin, Germany
CONTENTS 21.1 Introduction............................................................................................................................................................163 21.2 Skin Surface Measurement....................................................................................................................................164 21.3 Methodological Principle ......................................................................................................................................164 21.3.1 The Replica................................................................................................................................................164 21.3.2 The Stylus Method ....................................................................................................................................164 21.3.2.1 Technical Equipment..................................................................................................................164 21.3.2.2 Profile Characterization..............................................................................................................165 21.3.2.3 A Recent Development: The Touchless Acoustic Stylus ..........................................................165 21.4 Sources of Error.....................................................................................................................................................166 21.4.1 The Test Area.............................................................................................................................................166 21.4.2 Taking the Replica.....................................................................................................................................166 21.4.3 Handling the Equipment............................................................................................................................166 21.5 Correlation with Other Methods ...........................................................................................................................166 21.5.1 Laser Profilometry .....................................................................................................................................166 21.5.2 Image Analysis ..........................................................................................................................................167 21.6 Recommendations..................................................................................................................................................167 Acknowledgment.............................................................................................................................................................168 References .......................................................................................................................................................................168
21.1 INTRODUCTION Measuring surface texture requires a precise understanding of what is meant by surface. A surface is the boundary between two media. In medical terms, the surface of the skin is the boundary between the individual and the physical environment with all its diverse influences on the function and structure of the skin. In his handbook for surface texture analysis, Mummery1 gives a very plastic description of what we have to deal with: “When specifying a surface profile, the examiner must be aware, that there are as many surface profiles as there are landscapes. The Himalayas and the Black Forest in Germany are both
mountain ranges. This is where the similarity between them ends. The two mountain ranges differ not only in magnitude (height of the peaks) but also in their form (the shape of the peaks and valleys). Describing a surface is as complex as describing a mountain range.” The landscape of normal or impaired skin is determined by the cutaneous architecture. The arrangement and interlocking of adjacent keratinocytes, the epidermal rete ridges, the dermal papillary structure, and the cutaneous appendages all contribute to the surface form. Both internal and external processes such as aging, dehydration, hydration (cosmetic products), or atrophy 163
164
Handbook of Non-Invasive Methods and the Skin, Second Edition
(corticosteroids) continuously remodel the cutaneous landscape. Profilometry allows the objective measurement of the above-mentioned effects on skin surface and is of particular interest for the evaluation of the pharmaceutical or cosmetic action on the skin.
21.2 SKIN SURFACE MEASUREMENT The ideal approach for measuring the skin surface would be a touchless in vivo method. However, up to now no reliable and reproducible method has been reported that approaches the standard of indirect methods that rely on a negative replica of the skin. The main reason for the failure of direct measurements is the relative movement of the anisotropic skin surface while scanning the profile (pulsation of minor arteries, moving, trembling of subjects). Therefore, measuring the profile or surface texture of a replica remains the most common approach to describe surface characteristics related to the skin’s surface geometry.
21.3 METHODOLOGICAL PRINCIPLE In our first attempts in 1980 to measure the influence of topical corticosteroids on skin surface texture, we adapted the stylus method. Established instruments for profilometry, like the Hommel Tester (Hommelwerke GmbH, Schwenningen, Germany), were originally developed to measure tool traces that occur when working on a metallic surface. Later the method was modified for use in almost all areas of microgeometry. Accepted standardized parameters like roughness, waviness, total profile, spacing, and others serve to quantify surface texture. Therefore, surface texture analysis has grown into a field of its own with increasing importance not only for metalwork, but also for many other research and production branches. This is especially true for experimental dermatology.
FIGURE 21.1 Measuring set-up using a column-mounted linear traverse unit.
21.3.2 THE STYLUS METHOD A diamond stylus is traversed across the replica surface. An electrical signal equivalent to the vertical displacement of the stylus is amplified and converted into a digital signal. The digital information is analyzed by computer according to selected parameters for roughness (Figure 21.1). 21.3.2.1 Technical Equipment
Several materials have been used to obtain reliable impressions of the skin’s surface. After having tested a number of materials, Makki et al.2 found silicone rubber, a dental impression material, to be most suitable for this purpose. After evaluating several materials ourselves, we found Provil® L C.D.* silicone rubber impression material to be the material of choice. Since the Hommel Tester automatically calculates an inverse profile of the measured negative impression, a second impression of the negative primary cast is not necessary. This eliminates possible errors introduced by secondary positive casting.
Pick-up: The pick-up has two functions: it supports the stylus and acts as a transducer converting the vertical movement of the stylus on the surface to an electrical signal. For dermatological purposes, the pick-up consists of a diamond stylus that is cone shaped with a 5-mm tip radius and a 60˚ tip angle. To protect the surface of the replica (to avoid “ploughing” through the ridges), a sliding Teflon** skid (Figure 21.2) with a round-shaped tip is mounted beside the stylus. This results in a more even and constant scanning of the surface. Furthermore, the use of a skid eliminates waviness due to mechanical filtering of the profile. Transducer: The most common transducer used for dermatological purposes is a half-bridge inductive transducer. An inductive transducer is well suited for texture surface measurement because of its linearity and insensitivity to the surroundings (ambient temperature and humidity). Its compact size allows for packaging in a small housing. The resolution makes the measurement of displacements as small as 0.001 mm possible.
* Registered trademark of Bayer Dental Werke AG, Leverkusen, Germany.
** Registered trademark of E.I. du Pont de Nemours and Company, Inc., Wilmington, Delaware.
21.3.1 THE REPLICA
Stylus Method for Skin Surface Contour Measurement
165
Ra =
FIGURE 21.2 Special pick-up for use with a flexible replica.
Traverse unit: The traverse unit furnishes the relative movement between the replica and the pickup. The replica (Figure 21.1) is moved under the stationary pick-up mounted on a flexible pick-up holder. Computer: Instrument control, data analysis of the digitized signal, and data output are all computer controlled. Instruments are easily kept up-to-date by updating the software as standards change. 21.3.2.2 Profile Characterization Ra, average roughness: Ra can be called the grandfather of all roughness parameters and is still young enough to do a good job. Numeric values in micrometers (μm) are obtained. It is commonly employed because of the ease of calculation when using simple analog devices. Although several other standard parameters have been applied to quantify the profile of the skin, we still prefer the mean roughness value Ra. The different parameters for roughness are discussed in detail by Cook.3 From Figure 21.3 it can be seen that the quantity Ra is the average distance from the profile to the mean line over the traverse length of assessment. Ra is determined by the formula
x
FIGURE 21.3 Mean roughness value Ra.
y dx
(21.1)
0
lm is the traverse (scan) length and |y| is the absolute value of the location of the profile relative to the mean profile height (x-axis in Figure 21.3). Ra is a standardized roughness parameter according to relevant German and international industrial standards. Filtering: A profile filter can be compared to a sieve. If a pile of rocks and stones is put through a sieve, it will be separated into two piles. One pile consists of rocks unable to pass through the sieve, while the other is gravel able to pass through. The sieve hole size defines what is called rock and what is called gravel.1 The filtering of surface profiles follows the same rules. A filter with a defined cutoff length divides roughness (gravel) from waviness (stones). The cutoff length is analogous to the hole size of the sieve. Filtering does not change the original profile, but the way of looking at it. Statistical analysis: In addition to the average roughness Ra, the surface can be described by descriptive statistics such as variance, skewness, kurtosis, autocovariance, and autocorrelation functions, as well as Fourier analysis.1 In order to go a step further and perform a three-dimensional analysis, it is necessary to take a number of parallel measurements. This may be of interest for specific questions concerning nonhomogenic surfaces like the skin. In future mathematical models describing regularities and irregularities will probably play an important role in the description of three-dimensional images.
A very recent development that promises to revolutionize the standard stylus principle is the new touchless acoustic pick-up with a resolution of 10 nm, reflecting a precise image of the measured surface without mechanical alteration (NANOSWING, Hommelwerke GmbH, Schwenningen, Germany). The new pick-up is fully compatible with the standard equipment.4
Ra lm
∫
lm
21.3.2.3 A Recent Development: The Touchless Acoustic Stylus
y
lm = traverse length
1 lm
Principle: By means of an electronic sensor a standardized diamond tip is moved over the surface at a constant elevation. The vertical movement exactly corresponds to the surface profile. Function: In order to maintain the diamond tip at a constant elevation, the distance to the surface has to be measured continuously. To achieve this,
166
Handbook of Non-Invasive Methods and the Skin, Second Edition
oscillations with very small but constant amplitude are induced in the tip itself. The very low contact pressure of only about 0.2 mN does not result in a measurable impression on the elastic replica material. The power necessary to maintain the oscillation is measured by a precise electronic device completely integrated into the pickup. The closer the oscillating diamond tip gets to the surface, the more power is consumed as a result of the air friction between the tip and the surface. The power consumption of the oscillation serves as the input value for an electronic regulating circuit. The output is connected to an electric positioner for the diamond tip. Since the power consumption of the oscillation serves as a measure of the distance between the tip and the surface, maintenance of constant power results in a constant distance to the surface.
21.4 SOURCES OF ERROR When employing profilometry in experimental dermatology there are a number of pitfalls that can lead to wrong conclusions: (1) an adequate test site must be chosen, (2) the replica must be made, and (3) the equipment must be operated properly.
with an inner marking that leaves an impression on the hardened replica is necessary for orientation.
21.4.3 HANDLING
THE
EQUIPMENT
For comparable results the replica always has to be scanned in the same direction. The direction should be determined by the course of the tension lines. It is advisable to measure the replicas within a comparable time interval. The weight of the pick-up on the connecting surface always has to be the same (between 1 and 2 g) and should be controlled with a balance before the start of every measurement series. Precise measurements require care and a stable environment. The regular use of a calibrated roughness standard ensures proper instrument function. Vibration is one of the main sources of measurement error and should be avoided as far as possible. In most buildings walking through the room while measuring causes distinct vibration. Protection from air currents is a matter of course (open windows). With the standard stylus repeated measurements on the same replica should not be performed along the same line, but have to be carried out at a sufficient distance from the previous measurement.
21.5 CORRELATION WITH OTHER METHODS
21.4.1 THE TEST AREA
21.5.1 LASER PROFILOMETRY
A suitable test site for measuring influences of topical treatments should have an even surface with a regular structure and few or no hair follicles. The volar side of the forearm meets most of the requirements. Subjects with too many hair follicles, scars, tattoos, visible dehydration (detergents), extensive sun exposure, and pathological skin conditions (scaling) have to be excluded except under special circumstances.
Laser profilometry5 is a computer-assisted structural analysis of the skin surface that uses laser beams for touchless measurements with a very high resolution. A three-dimensional profile of the surface can be stored digitally. Different parameters of roughness can be determined. Additional mathematical and statistical procedures, such as Fourier transformation and autocorrelation function, complete the analysis. The touchless laser beam is said to allow more precise imaging of the peaks and valleys of a replica independent of its elastic properties (no bending of the peaks through contact) than the conventional stylus method. However, one must be aware of the limitations of the method that are caused by the geometrical and optical properties of the object of interest. One aspect is the critical inclination (10 to 15˚) of a profile, where the laser system tends to overestimate the real depth of the profile. Furthermore, a high optical contrast (dark to bright areas) of the scanned surface may lead to a misinterpretation of the profile. Under unfavorable conditions a mistake of several hundred percent may result because the laser registers the optical contrast as well as the geometrical profile. The profile features produced by the optical contrast are indiscernable from the geometrical profile. Further errors well known to the experienced investigator result from the entry of the beam into the replica, porous surfaces, or structures with optic imaging properties.
21.4.2 TAKING
THE
REPLICA
For precise casting the replica material should be of low viscosity, fast hardening, and elastic without shrinking. The material should not produce heat while hardening. Large residues should not remain on the skin after removing the replica. The most commonly used material is silicone rubber impression material for dental purposes. With Provil® L C.D. all the above demands are covered. Other materials are listed by Cook.3 Provil L C.D. is a two-component material supplied in a ready-to-use cartridge, eliminating the necessity of further mixing. If a two-component material is used that has to be mixed by hand, it takes time and practice to mix the two components rapidly enough but at the same time without air bubbles. If the substance is already too hard, the replica will not be sufficient. For most purposes a mold
Stylus Method for Skin Surface Contour Measurement
Reg-papier RP 50
Homelwerke
Reg-papier RP 50
167
Reg-papier RP 50
Homelwerke
Reg-papier RP 50
FIGURE 21.4 Skin replica before topical therapy with a corticosteroid ointment of medium strength.
FIGURE 21.4 (Continued). Skin replica 3 weeks after topical therapy with a corticosteroid ointment of medium strength.
These inevitable but not always predictable errors should be kept in mind when referring to the very high resolution of laser profilometry. In addition, to take advantage of the highest possible resolution over large areas of the replica surface, laser profilometry is very time-consuming (requiring hours per replica). It should also be mentioned that the equipment for the laser method is considerably more expensive than that for the stylus method (both standard or acoustic pick-up). Despite the technical advantages of the laser method compared to the conventional stylus method, it remains to be seen whether this method is actually superior in practice and equal to the new touchless acoustic principle applied to the stylus method.
21.6 RECOMMENDATIONS
21.5.2 IMAGE ANALYSIS The basic principle underlying image analysis is the measurement of shadows generated by incident lighting at the surface of a replica. Main target parameters are the number and mean depth of wrinkles. A major problem with this method is uneven lighting that may result from unlevel replicas that are true to the skin structure.
A well-standardized procedure to assess the influence of different corticosteroids on the epidermal macropattern is the Duhring chamber test6 in combination with profilometry. The loss of the detailed structure of the epidermis can be quantified (Figure 21.4). No residual cream should be present on the test site. In general, it is advisable to stop treatment 24 hours before any measurement. Skin replicas made of silicone rubber impression material have the disadvantage that their elastic properties may lead to incorrect measurement of the “mountain peaks.” On the other hand, the material allows for repeated measurements by taking several replicas from the same area without substantial alteration of the skin surface. Cyanoacrylate replicas give a stable imprint of the skin surface but alter the test site by marked stripping when removing the replica. One should keep an eye on the hydration of the skin because it influences the profile considerably. This is especially important when testing substances for their influence on aging skin. A reduced roughness may be due to temporarily enhanced hydration and not to reduced depth of the wrinkles because of changes in the elastic fibers.
168
Handbook of Non-Invasive Methods and the Skin, Second Edition
Whichever roughness parameter is chosen by the experienced examiner, he or she must be aware of the method’s limitations. It must always be kept in mind that while new technologies and parameters for skin surface texture analysis may reveal statistically significant differences between two skin conditions, the most important authority to validate significant data is clinical efficacy.
ACKNOWLEDGMENT We thank B. Hughes for help with the English translation.
REFERENCES 1. Mummery, L., Surface Texture Analysis: The Handbook, Hommelwerke GmbH, VS-Muehlhausen, Germany, 1992.
2. Makki, S., Barbenel, J.C., and Agache, P., A quantitative method for the assessment of the microphotography of human skin, Acta Derm. Venereol., 59, 285, 1979. 3. Cook, T.H., Profilometry of skin: a useful tool for the substantiation of cosmetic efficacy, J. Soc. Cosmet. Chem., 31, 339, 1980. 4. Personal communication from Volk, R., Hommelwerke GmbH, Schwenningen, Germany. 5. Saur, R., Schramm, U., Steinhoff, R., and Wolff, H.H., Strukturanalyse der Hautoberfläche durch computergestützte Laser-Profilometrie, Hautarzt, 42, 499, 1991. 6. Frosch, P.J., Kligman, A.M., and Wendt, H., The Duhring chamber test for assaying corticosteroid atrophy in humans, in Percutaneous Absorption of Steroids, Mauvais-Jarvis, P., Vickers, C.F.H., and Wepierre, J., Eds., Academic Press, London, 1980, p. 185.
22 Laser Profilometry Jan Efsen Novo Nordisk A/S, Bagsvaerd, Denmark
Steen Christiansen Technical University of Denmark, Institute of Manufacturing Engineering, Lyngby, Denmark
Hans Nørgaard Hansen Copenhagen, Denmark
Jens Keiding LEO Pharma, Ballerup, Denmark
CONTENTS 22.1 Introduction............................................................................................................................................................169 22.2 Object.....................................................................................................................................................................170 22.3 Methodological Principle ......................................................................................................................................170 22.3.1 Preparation of Object: Making a Replica .................................................................................................170 22.3.2 Control of the Optical Profilometer ..........................................................................................................170 22.3.3 Calibration Methods ..................................................................................................................................171 22.3.3.1 Optical Sensor ............................................................................................................................171 22.3.3.2 Air-Bearing Table .......................................................................................................................171 22.3.3.3 Software......................................................................................................................................172 22.3.3.4 Standard Roughness Specimens.................................................................................................172 22.3.3.5 Frequency Response Analysis....................................................................................................172 22.3.4 Summary of Calibration Methods.............................................................................................................172 22.3.5 Comparison of the Optical Profilometer with Mechanical Stylus Instruments .......................................172 22.3.5.1 Measuring and Data Collection with the Profilometer..............................................................173 22.3.6 Characterization of Surfaces with Stratified Structure .............................................................................173 22.3.6.1 Preprocessing the Parameter Calculations .................................................................................173 22.3.6.2 Algorithm ...................................................................................................................................174 22.3.6.3 Other Parameters ........................................................................................................................175 22.3.6.4 Using Three-Dimensional Parameters for Characterizing Skin Replicas .................................175 22.3.6.5 What Has Quantitative Analysis of Skin Structure Been Used For?........................................177 22.4 Recommendations..................................................................................................................................................177 References .......................................................................................................................................................................177
22.1 INTRODUCTION The skin is a most versatile organ. It functions simultaneously as a protective first-line defense for the body and an area of chemical communication between the body and the external world. The inside part of the skin has been studied by a series of techniques. Histological, biophysical, and biochemical
studies have led to a detailed understanding of the internal structure of the skin, i.e., organization of cell layers, fibers, and chemical constituents in different parts of the skin. More systematic studies of the outside part of the skin have been carried on since the 1930s. In the beginning the technique was used merely to describe features on the surface of the skin or a skin replica that could be observed in a microscope. 169
170
Handbook of Non-Invasive Methods and the Skin, Second Edition
More quantitative studies of skin structure started about 30 years ago and have included both two-dimensional and, more recently, three-dimensional studies. These methods can be subdivided into two categories:
Normally skin replicas, i.e., impressions that are used as reproductions of the skin relief, are used as study objects in profilometric studies. This chapter will give a short introduction to profilometry with an emphasis on the methodology of optical profilometry. The performance of a commercially available instrument is discussed, as is the skin structure data that can be obtained using this technique and how the dermatologist can make use of these data.
UBM Laserbeam
ire
ct
io n
X-direction
Air bearing table
Yd
1. Structure mapping based on analysis of an image of the surface. These systems include image analysis of videoscans, as in the Magiscan image analysis system,1 and the fully automated system developed by Corcuff et al.2 in which the shadows cast by incident light on a replica may be used to determine primary and secondary direction and depth of furrows and also determine skin surface area. 2. Topographical mapping of the structure based on scanning of surface height (two- or threedimensional scanning). This category includes profilometers — mechanical or optical.
Optical sensor (fixed)
FIGURE 22.1 Optical profilometer system. 12
14
Analogue output 13
4 2 1
3
8 5
10
22.2 OBJECT
6 7
The purpose of this chapter is to present the operation of an optical profilometer and how it may be calibrated. We also introduce the use of software for a three-dimensional surface description and give a practical example of the use of this system.
9
22.3 METHODOLOGICAL PRINCIPLE
7 6
Sensor 11
FIGURE 22.2 Autofocus principle.
must be carefully controlled in order to avoid air bubbles getting stuck in the replica.
22.3.1 PREPARATION OF OBJECT: MAKING A REPLICA The making of a replica is in principle very easy. In our routine we first attach an adhesive ring to the skin area to be studied. A silicone rubber (Silflo® Flexico) is mixed with a catalyst, and the mixture is distributed as a thin layer covering the central opening of the ring and is allowed to harden for 3 to 5 minutes. The ring with the replica attached is gently removed, and finally the replica is cut into an appropriate size with a cutting device. In principle, this method leads to an accurate replica of the skin surface structure.3 In practice, however, there are often problems. The skin under study may need pretreatment to remove scales, hair, etc. Due to the anatomy, it may be difficult to obtain a replica of an appropriate structure and size. The process of mixing rubber and catalyst
22.3.2 CONTROL
OF THE
OPTICAL PROFILOMETER
The optical profilometer (Microfocus 1080) used in this investigation was provided by UBM Messtechnik (Ettlingen, Germany). The profilometer consists of four main parts: an optical sensor, an air-bearing table, a control unit, and a computer. The optical sensor is fixed, and the object whose surface is to be measured is placed on the airbearing table. By means of the control unit the table can be translated in two perpendicular directions, an X-direction and a Y-direction, as illustrated in Figure 22.1. It is possible to move the table 150 mm in both directions. Operation of this system is based upon the autofocus principle. An illustration of the system is shown in Figure 22.2 and described in UBM Messtechnik.6 Infrared light
Laser Profilometry
(wavelength of 780 nm) emitted from a laser diode (1) is focused onto a small spot (diameter of 1 mm) by a system of lenses (8 and 9). The light reflected from the object surface (11) is directed back into the sensor and is imaged as a pair of spots onto an arrangement of photodiodes (5). This is done in such a manner that both diodes are illuminated equally only when the objective lens (9) is precisely in its focal distance from the surface. If the distance to the object changes, the focus point is shifted too, and the illumination of the photodiodes becomes unequal. This unequal illumination of the photodiodes generates a focus error signal. A control circuit monitors the error signal and moves the objective lens accordingly. This is the autofocus principle. The movement of the lens is accomplished by a coil (6) and magnet (7) arrangement. The vertical movement of the objective lens is registered by a light barrier measurement system (10) and corresponds to variations in the height of the object surface. The standoff distance between object surface and optical sensor is approximately 2 mm. The optical sensor has two vertical measurement ranges: ±50 and ±500 μm. It is possible to obtain a vertical resolution of approximately 6 nm in the range ±50 μm and 60 nm in the range ±500 μm. The surface of the object may be viewed during the measurements through a window (4) in the sensor by using a microscope (13) and a CCD camera (14). All operations of the system are controlled by a computer.
22.3.3 CALIBRATION METHODS Calibration and control of the optical profilometer is necessary to obtain reliable results. Our calibration procedure was divided into the following five parts: • • • • •
Control of optical sensor Control of air-bearing table Control of software Calibration against standardized roughness specimens Determination of the frequency response of the optical profilometer
The tests were carried out at the Technical University of Denmark. The tests are described in detail in Efsen and Hansen.4 22.3.3.1 Optical Sensor The linearity of the sensor was tested with a sine bar. This is a metal bar mounted on two gauge blocks to obtain a well-defined angle (Figure 22.3). The sine bar was measured with the optical profilometer and the linearity determined from these results. The test was carried out for vertical measurement ranges ±50 and ±500 μm.
171
Optical sensor
Optical sensor
Sine bar Gauge block
Gauge block
Air bearing table
FIGURE 22.3 Sine bar.
Optical sensor
Optical sensor
Max. angle
Roundness standard specimen
FIGURE 22.4 Roundness standard specimen.
The ability of the optical sensor to measure inclined surfaces was tested with a roundness standard specimen. This is a glass ball with an almost perfect roundness. This specimen was measured with the optical profilometer as indicated in Figure 22.4, and the maximum allowable slope of the surface determined. A supplementary test of the sensor covered an investigation of which surfaces the sensor was able to detect. Furthermore, the control circuit of the sensor was tested by changing control parameters in the software. 22.3.3.2 Air-Bearing Table The flatness of the movement of the air-bearing table was tested with an optical flat. The optical flat was measured in two areas: 1 × 1 mm2 and 100 × 100 mm2. The small area represents a typical three-dimensional area for investigation in mechanical engineering, and the large area covers almost the entire possible movement of the airbearing table. The positioning accuracy and repeatability in both the X- and the Y-axis direction was tested with laser interferometry. Furthermore, the perpendicularity of the two axes was tested with an angle plate.
172
Handbook of Non-Invasive Methods and the Skin, Second Edition
Calibration standard specimen ISO 5436 type A 6
Profile height (μm)
TABLE 22.1 Suitable Calibration Methods
Overshoot
4 2
Object
0
Optical sensor
−2 −4
Air-bearing table
−6 Overshoot
−8 −10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Profile length (mm)
FIGURE 22.5 Measuring a calibration standard specimen (DS/ISO 5436 type A) with the optical profilometer.
Software
Standard roughness specimens
Property
Test Specimen
Linearity Maximum allowable slope Flatness Positioning accuracy of axis Repeatability Perpendicularity of axis Parameter calculation Filter test
Sine bar Roundness standard specimen Optical flat Laserinterferometry
Background noise
Static amplificationa
Frequency response analysis a
Ability to measure oscillating surfaces
Laserinterferometry Angle plate Theoretic sine profile Signal analyzer, DIN 4777 Optical flat
Calibration standard specimen (DS/ISO 5436, type A) Determination of frequency response
Overshoot
22.3.3.5 Frequency Response Analysis FIGURE 22.6 PTB parameter specimen.
22.3.3.3 Software The software was tested with a theoretic sine profile. The results were compared to already existing and tested programs and calculations performed after the standard DS/ISO 4287/1.7 The filter characteristics were determined by means of a signal analyzer (Brüel & Kjær, type 2032). A filter test as described in DIN 47778 was carried out. 22.3.3.4 Standard Roughness Specimens The optical profilometer was calibrated with standard roughness specimens designed for mechanical stylus instruments. An optical flat was used to determine the background noise in the system during a measurement. A single grooved calibration standard specimen (DS/ISO 5436, type A9) was used to determine the static amplification of the profilometer (Figure 22.5). These measurements revealed overshoot of the sensor when measuring steep edges. A parameter specimen developed by Physikalisch Technische Bundesanstalt (PTB) (Germany)5 was used to test the ability to measure real surfaces (Figure 22.6). The PTB parameter specimens were not found suitable for calibration of the optical profilometer.
By the term frequency response we mean the steady-state response of the optical sensor to a sinusoidal input. This analysis reveals the ability of the sensor to detect vertically oscillating surfaces. When measuring surfaces with lateral waviness, the frequency response of the sensor determines which wavelength components it is possible to detect. The analysis showed that the vertical range of a surface influences the response of the optical sensor.
22.3.4 SUMMARY
OF
CALIBRATION METHODS
The tests found suitable for calibration of the optical profilometer are listed in Table 22.1.
22.3.5 COMPARISON OF THE OPTICAL PROFILOMETER WITH MECHANICAL STYLUS INSTRUMENTS The performance of the optical profilometer was compared with that of mechanical stylus instruments (Rank Taylor Hobson Talysurf 5 and Rank Taylor Hobson Surtronic 3P). The comparison consisted of measuring a wide range of processed surfaces, i.e., turned, milled, and lapped surfaces. Also, the well-defined surface of a PTB parameter specimen was measured. The standardized parameters calculated from profiles obtained by the optical profilometer were on all surfaces studied higher than the
Parameter values (optical profilometer)
Laser Profilometry
173
Optical profilometer versus mechanical stylus instrument
Optical system, f 1 mm
100 Mechanical system, r > 2 mm
10 1 0.1 0.01 0.001 0.001
0.01 1 10 0.1 Parameter values (mechanical stylus) Rt (μm)
100
FIGURE 22.8 Comparison between optical stylus (laser beam) and a typical mechanical stylus.
Ra (μm)
FIGURE 22.7 Measuring industrial surfaces with both an optical profilometer and a mechanical stylus instrument.
corresponding parameters from a mechanical stylus instrument. This is illustrated in Figure 22.7, where the Ra and Rt values are plotted. Values obtained from the mechanical stylus instrument are plotted along the X-axis, and values obtained from the optical profilometer are plotted along the Y-axis. If both instruments would have given the same parameter values, all the points should have been placed on the marked line. It is obvious that parameter values obtained with the optical profilometer are higher than those obtained with the mechanical stylus instrument. This can be explained by the fact that the 1-mm laser spot detects more and deeper valleys in the surface than a mechanical stylus instrument (radius of typically >2 mm), as shown in Figure 22.8. Furthermore, a mechanical stylus works like a mechanical filter (Figure 22.9). Finally, steep edges represent a problem, as illustrated by the overshoot phenomenon shown in Figure 22.5. On a system calibrated as described, it is possible to obtain a repeatability of 5 to 10% when measuring skin replicas. 22.3.5.1 Measuring and Data Collection with the Profilometer The operation of the profilometer is controlled by an extensive software that is an integrated part of the equipment. It makes automatic scanning of a series of objects possible. Scanning takes place with the object positioned on the air-bearing table. Several objects may be scanned in one measuring sequence. In a setup table it is possible to predefine up to 99 positions on the table, and to each position a specific scan area may be defined. The information on position and area may be saved in the memory of the computer. When used together with information on speed and intensity of scanning, it allows the user to set up automatic scan procedures for all these operations. It is possible also to automize the data analysis. An extended list of standard two-dimensional roughness parameters
FIGURE 22.9 The stylus working as a mechanical filter.
and a few three-dimensional parameters is included with the software. Power spectra and autocorrelation analysis in two and three dimensions and two-dimensional material ratio curves are some of the other parameters included. Software routines for three-dimensional material ratio parameters are, however, not included. In the following we will show how such parameters can be calculated and how they may be used for description of skin surface structure.
22.3.6 CHARACTERIZATION OF SURFACES STRATIFIED STRUCTURE
WITH
This is a very common problem in engineering surfaces, since different layers in a surface have different functional properties. If the top part of a surface does not look like the valley part, the surface cannot be characterized with commonly used parameters, such as Ra and Rz, as these parameters cannot distinguish up and down in the profile. This is also a problem in skin structure studies, as these parameters cannot distinguish a groove from a ridge. Therefore, other parameters that can distinguish stratified structures are needed. 22.3.6.1 Preprocessing the Parameter Calculations A complete description of the skin surface is very complex since the replica contains many waves caused by the specific test conditions, such as the location where the
174
Handbook of Non-Invasive Methods and the Skin, Second Edition
replica has been taken and the replica preparation method. Grooves, ridges, and holes must be classified as microstructure and must be separated from the general form (the macroform), e.g., the form of the leg, the arm, or the face, before a reliable characterization of the microstructure alone can be made. Generally the true measured geometrical form — known as the raw set of data — may be classified into three categories: • • •
Microstructure, e.g., small holes, marks, and cracks Macrostructure, e.g., a cylinder, plane, or conus An arbitrary level, depending on the vertical range in which the measurement has taken place
It is difficult to distinguish between these categories since no exact limit between micro and macro has been defined yet. Different methods, e.g., fast Fourier transformation and digital cutoff filtering, have been tried,10 and one of the most suitable ways to remove the macroform seems to be subtraction of the Nth-order least squares polynomial fit from the raw set of data, as shown in this formula: R(x, y) = Z(x, y) – F(x, y) where R(x, y) is the residual surface, Z(x, y) is the original datapoint, and F(x, y) is the least squares Nth-order polynomial fit. Different methods for definition of a reference plane have also been examined, e.g., the arithmetic mean plane and the least squares plane. The least squares plane method has been evaluated to be the best since this method is more robust than the arithmetic method. All parameters should be calculated with respect to this reference plane. 22.3.6.2 Algorithm When the form element has been removed and the reference plane is defined, the parameters can be computed after the following precept:11 1. The material ratio curve is computed as follows: a. The surface is truncated into 4096 levels. b. The relative volume of material in each level is calculated, leading to the material distribution. c. The material distribution is accumulated over the 4096 levels leading to the material ratio curve, which shows how much material is found above a specified surface level (this is illustrated in Figure 22.10 and Figure 22.11).
Heightdistribution
Y
l 3%
10%
FIGURE 22.10 Illustration of height distribution curve. Y
Material ratio curve
l 0%
100%
FIGURE 22.11 Calculation of the three-dimensional material ratio curve.
2. The 40% secant is moved along the materialbearing curve until the 40% secant with least slope is found. These points are marked with A and B. 3. A line is projected through the A and B to the intersection with 0 and 100% material ratio. These points are labeled C and D, as shown in Figure 22.12. The vertical distance between C and D is the core surface depth (Sk). Sr1 and Sr2 are the material ratios at the top and the bottom above and below the core surface. 4. The areas A1 and A2 are computed. Spk is defined as the height in the triangle that has area A1 and baseline Sr1. The Svk is defined as the height in the triangle that has area A2 and baseline Sr2. Sk — the core surface depth — measures the height of the core material portion. It depicts the flat-test part of the material ratio curve, i.e., the region with the greatest increase in material. A small Sk value generally indicates that the skin surface is very smooth, since the material volume is very large. This is seen, e.g., on the skin surface of a baby where the skin is virtually devoid of wrinkles. Spk — the reduced surface peak height — denotes the height of the surface peak projecting beyond the core surface. A low Spk value indicates that the surface does not include many ridges and may be free of extreme peaks. Svk — the reduced surface valley depth — denotes the proportion of surface valleys extending into the material below the core surface. It provides useful information on grooves and holes in the surface. Generally it can be said
Laser Profilometry
175
B
A 40% y
40%
γ Spk
β
Sk
α
sekant
Svk 0
Sr1
Peak area
50 Bearing area
Sr2
100
Valley area
FIGURE 22.12 Calculation of the three-dimensional material ratio curve parameters.
that large Svk values indicate that the surface contains a large amount of wrinkles. Basically, this set of parameters divides the surface into three categories: top, core, valleys. One parameter is related to each category, and therefore, the total set of parameters is required for a suitable characterization of surfaces with stratified structure. 22.3.6.3 Other Parameters Other useful parameters for characterization of skin surfaces can be the arithmetic mean value Sa, the difference between maximum and minimum height St and skewness Ssk. Digitally the Sa is calculated as follows: 1 Sa = MN
MN
∑Z
i
i=1
where M is the number of points per profile, N is the number of profiles, and Zi is the numerical distance from least squares mean plane to the ith residual surface height. Sa is a mean value and gives an overall impression of the roughness properties of the object. St is the maximum vertical distance between highest peak and deepest valley. This parameter gives information about extreme conditions, e.g., extremely deep holes. Ssk is the skewness of the material distribution curve. This parameter can be used to describe the shape of the surface height distribution. It is given by the formula 1 Ssk = MNSq3
MN
∑Z i=1
3 i
where M is the number of points per profile, N is the number of profiles, Zi is the distance from least squares mean plane to the ith residual surface height, and Sq is the root mean square deviation of the surface. For an asymmetric distribution of the surface heights, the skewness may be negative if the distribution has a longer tail at the lower side of the mean plane, which means that the surface consists primarily of valleys and holes. This is in contrast to positive skewness, which indicates a surface including many peaks and ridges. If skewness is zero, no primary trend can be seen. These three parameters have to be handled with care and conclusions based on these parameters alone must be avoided. The definition of Sa includes the numerical values, which means that this parameter cannot see the difference between up and down. St is very sensitive for single-point values. One extreme value that may be caused by a vibration during the measurement can result in misleading conclusions. It can be seen from the formula for skewness, Ssk, that this parameter is dimensionless, since it is normalized by Sq. Therefore, skewness can never give information on absolute properties of a surface; only relative properties can be characterized. 22.3.6.4 Using Three-Dimensional Parameters for Characterizing Skin Replicas Sixteen skin replicas were investigated. An area of 5.6 × 5.6 mm2 was measured with the optical profilometer and analyzed by means of the three-dimensional parameters described above. In this investigation only the parameters Sa, St, Svk, and Sk were calculated.
176
Handbook of Non-Invasive Methods and the Skin, Second Edition
PI/DTH
MM KH-1
REPLIKA JE/HNH
24. 05. 93 LEKH1LM
150 mm
−200 mm 200 mm
5.60 mm 20 p/mm
z
5.60 mm: 20 p/mm
y
0°
x
LEO
FIGURE 22.13 Example of a profile plot from category I.
PI/DTH
MM 04-I
REPLIKA JE/HNH
24. 05. 93 LEO41LM
150 mm
−200 mm 200 mm
5.60 mm 20 p/mm
5.60 mm: 20 p/mm
z 0°
y
x
LEO
FIGURE 22.14 Example of a profile plot from category IIb.
The surface plots obtained from the optical profilometer were divided into two categories by means of visual inspection: Category I: Characteristic waviness with or without craters Category II: No characteristic waviness but possibly craters
Category II could be further subdivided into these two subcategories: Category IIa: No characteristic waviness but craters Category IIb: No characteristic waviness and no craters
Laser Profilometry
177
22.4 RECOMMENDATIONS TABLE 22.2 Categorization of Skin Replica Based on ThreeDimensional Surface Parameters
Category
Description
Suitable 3D Parameters (μm)
Value
I
Waviness, with or without craters
Sa St Sk Svk
Sa ≈ 30 210 < St < 300 80 < Sk < 120 42 < Svk < 49
IIa
No waviness, with craters
Sa St Sk Svk
Sa ≈ 20 230 < St < 280 60 < Sk < 100 30 < Svk < 40
IIb
No waviness, without craters
Sa St Sk Svk
Sa ≈ 20 170 < St < 195 50 < Sk < 70 28 < Svk < 33
Figure 22.13 shows an example of a profile plot from category I, and Figure 22.14 shows a replica from category IIb. Table 22.2 shows the combination of parameters and parameter values found in this investigation. In this material it was possible based on a calculation of the parameters Sa, St, Svk, and Sk to place each replica into one of these three categories. All four parameter values were needed to place the replica in the correct category. 22.3.6.5 What Has Quantitative Analysis of Skin Structure Been Used For? Image analysis studies have been used by Corcuff and Lévêque12 to describe development of skin structure as a function of age. They have shown that the regularity of the skin relief decreases with age, associated with the disappearance of secondary lines. They also showed that the aging process is a combination of physical age and external factors acting on the skin surface. Grove et al.,13 among others, have used optical profilometry in clinical studies on the effect of retinoid treatment on wrinkles and have shown that roughness parameters Ra and Rz correlate with clinical gradings based on clinical ratings.13 Zahouni et al.14 have used profilometric studies to calculate volume and area of leg ulcers and have used the data for evaluation of wound-healing treatments. Image analysis of melanoma has recently been shown to be of help in the clinical diagnosis of pigmented lesions,15 and a profilometric study has shown that it is possible to distinguish between melanomas and nevocellular nevi from a profilometric analysis of replicas taken from these areas.
In this chapter emphasis has been put on controlling and standardizing the operation of the optical scanner. However, it is equally important to have a standardized method of developing the material for study — the replica. The process of producing the replica and the quality of the replica should be controlled. Sometimes the skin may need pretreatment prior to taking the replica, in order to remove scales or hair or clean the skin to remove residuals from skin surface treatments. With respect to the use of the profilometer, it is necessary to specify the size of the area measured, the speed of the scanner, and the scanning intensity (i.e., points/millimeter). In addition to these specifications, which are included in the standard menu prior to starting the measurement, a series of hardware specifications are available. The actual set values should also be specified in order to make reproducible measurements and for lab-to-lab comparisons. In the test of the profilometer a test of the software was included. Only a small part of the extensive software was tested, however. An important part is the use of filtering methods in the determination of roughness parameters. If a digital filtering algorithm is used for removal of form error instead of a polynomial fit, it is important to select the correct filter algorithm. A phase-correct filter (M-filter), not the common RC2 filter (which is not phase correct), should be used. A wrong choice of filter may introduce peaks where no such peaks can be identified in reality. Also, the cutoff wavelength should be carefully chosen. In general, the cutoff wavelength should be 0.2* the evaluation length.
REFERENCES 1. Grove, G.L. and Grove, M.J., Objective methods for assessing skin surface topography noninvasively, in Cutaneous Investigation in Health and Disease, Lévêque, J.-L., Ed., Marcel Dekker, New York, 1989, chap. 1. 2. Corcuff, P., Chatenay, F., and Lévêque, J.L., A fully automated system to study skin surface patterns, Int. J. Cosmet. Sci., 6, 167, 1984. 3. Cook, T.H., Profilometry of skin: a useful tool for the substantiation of cosmetic efficacy, J. Soc. Cosmet. Chem., 31, 339, 1980. 4. Efsen, J. and Hansen, H.N., Optisk ruhedsmåling, Ed. MM.93.34, Institute of Manufacturing Engineering, Technical University of Denmark, Copenhagen, Denmark, 1993. 5. Deutscher Kalibrierdienst, Physikalisch Technische Bundesanstalt, Calibration of Stylus Instruments, Guideline DKD-R4-2, Braunschweig, Germany, 1991.
178
Handbook of Non-Invasive Methods and the Skin, Second Edition
6. UBM Messtechnik, Microfocus Measuring System Manual, Ettlingen, Germany, 1992. 7. DS/ISO 4287/1, Surface Roughness. Terminology, Part 1, Surface and Its Parametres, 1st ed., Dansk Standardiseringsråd, København, 1986. 8. DIN 4777, Teil 1, Oberflächenmesstechnik. Profilfilter zur Anwendung in elektrischen Tastschnittgeräten, Phasen-korrekte Filter, Berlin, 1988. 9. DS/ISO 5436, Calibration Specimens: Stylus Instruments: Types, Calibration and Use of Specimens, 1st ed., Dansk Standardiseringsråd, København, 1987. 10. Stout, K.J., Sullivan, P.J., Dong, W.P., Manisah, E., Lou, N., Mathia, T., and Zahouani, H., The Development Methods for Characterisation of Roughness in 3 Dimensions, Vols. 1 and 2, BCR report, EC contract 3374/1/0/170/90/2, Centre for Metrology, University of Birmingham, Birmingham and L’Ecole Centrale de Lyon, Lyon, France, 1993. 11. Christiansen, S., Function-Related 3-Dimensional Definition of Surface Microtopography, Ph.D thesis, Technical University of Denmark, Institute of Manufacturing Engineering.
12. Corcuff, P. and Lévêque, J.-L., Age-related changes in skin microrelief measured by image analysis, in Aging Skin, Lévêque, J.-L. and Agache, P.G., Eds., Marcel Dekker, New York, 1993, chap. 13. 13. Grove, G.L., Grove, M.J., Leyden, J.J., Lufrano, L., Schwab, B., Perry, B.H., and Thorne, G., Skin replica analysis of photodamaged skin after therapy with tretinoin emollient cream, J. Am. Acad. Dermatol., 25(2), 231, 1991. 14. Zahouni, H., Assoul, M., Janod, P., and Mignot, J., Theoretical and experimental study of wound healing: application to leg ulcers, Med. Biol. Eng. Comput., 30, 234, 1992. 15. Busche, H., Connemann, B.J., Kreusch, J., and Wolff, H.H., Surface Topography in the Diagnosis of Malignant Melanoma, poster session, 3rd Conference of the International Society for Ultrasound and the Skin, Elsinore, Denmark, 1993.
Evaluation of Skin 23 Three-Dimensional Surface: Micro- and Macrorelief Jean Mignot Laboratoire de Métrologie des Interfaces Techniques, Institut Universitaire de Technologie, Besançon, France
CONTENTS 23.1 Introduction............................................................................................................................................................179 23.2 Aim of the Study ...................................................................................................................................................180 23.3 Improvements in the Apparatus.............................................................................................................................180 23.3.1 Focusing System........................................................................................................................................180 23.3.2 Triangulation System.................................................................................................................................181 23.4 Quantification of Surfaces: Microrelief ................................................................................................................182 23.4.1 Parameters Obtained from the Extension of Classical and Standard Roughness Parameters to Three Dimensions...............................................................................................................182 23.4.2 Statistical Analysis.....................................................................................................................................183 23.4.3 Textural Analysis of the Skin Surface.......................................................................................................184 23.4.3.1 Directional Quantification of Furrows .......................................................................................185 23.5 Quantification of Surfaces: Macrorelief................................................................................................................187 23.5.1 Quantification of Wrinkles ........................................................................................................................187 23.5.2 Quantification of Wounds..........................................................................................................................188 23.5.2.1 Performance and Results............................................................................................................189 23.5.2.2 Theoretical and Experimental Evolution of Healing.................................................................191 23.6 Conclusion .............................................................................................................................................................191 References .......................................................................................................................................................................192
23.1 INTRODUCTION Skin surface topography has been a matter of interest for dermatologists for 50 years. The first studies were carried out by Cummins and Midlo,1 and these were followed by many others by Wolf,2 Tring and Murgatroyd,3 and Marks and Saylan.4 However, all these studies have been bidimensional, which means that the surface itself has not been studied, but one or several profiles of the surface in one or several directions have been. This simplification was necessary because of the capacities of the equipment available: no measurement device was able to analyze a sufficient number of data related to the surface analyzed within a short time. Until recent years, these profilometric studies were made with equipment designed for measuring functional surfaces of parts of machines in the mechanical industry. The measuring part of these types of apparatus is a
mechanical sensor or stylus in direct contact with the surface to be analyzed.5,6 The detector of this sensor (Figure 23.1) is usually a thin diamond tip (radius of curvature from 5 to 10 μm) displaced at a constant speed on the surface (0.5 to 1 mm/sec) according to a direction determined by the operator.
Analog signal v = f(z)
Fixed coil
z Surface
Fixed coil x
Axis
Skid Profilometer tip
FIGURE 23.1 Principle of stylus profilometer. 179
180
Handbook of Non-Invasive Methods and the Skin, Second Edition
This type of device has several main drawbacks: •
•
•
It must not be in direct contact with the skin, because the pressure of the spheric head creates a major deformation of the skin surface.7 This is why soft replicas of the surface are made,8,9 from which solid replicas are obtained, using in most cases a polymerized material.10 The scanning speed is limited in two-dimensional profilometry systems because of the mechanical contact with the analyzed surface, and this results in slow data acquisition. The results obtained with two-dimensional analysis (or roughness parameters) show the topography of the surface in one direction only. A top view examination of the skin surface (Figure 23.2), or the study of the changes of the various frequential components of the surface (two-dimensional Fourier transform) in all directions, are sufficient to show the anisotropy of this surface (Figure 23.3).
These main drawbacks have two consequences: • •
Traditional surface measurement devices are not adequate. Because of its anisotropy, the analysis of the skin surface with profilometers is not reliable.
23.2 AIM OF THE STUDY Because of the disadvantages of profilometry, it seems necessary to develop new devices that quickly acquire data and analyze skin topography, in three dimensions, and with a satisfactory speed. Nowadays, only noncontact sensors have the capacity to meet these requirements. Furthermore, the anisotropy of skin surface demands a threedimensional treatment of data. μm 220 200 180 160 140 120 100 80 60 40 20 0
2.5 mm
2.5 mm
FIGURE 23.2 Cutaneous surface, top view.
FIGURE 23.3 Two-dimensional Fourier transform of the cutaneous surface in a 25-year-old male subject.
23.3 IMPROVEMENTS IN THE APPARATUS The study of the topography of a surface is usually carried out by scanning the surface with parallel profiles.11 In the case of the scanning of a surface of 512 ↔ 512 measurement points separated by a distance of 10 μm, data acquisition would take more than 2 hours if a mechanical sensor were used under standard speed conditions (0.5 mm/sec).12 This length of time is unacceptable for assembly line monitoring. The recent development of laser diodes and their miniaturization has given rise to the production of several devices based on two principles: focusing of the laser beam and optical triangulation.
23.3.1 FOCUSING SYSTEM The principle of this system, used in industry, is explained in Figure 23.4: the beam from the laser diode (λ ≈650 to 800 nm) is focused on a point of the surface to be analyzed. This surface is displaced under the sensor at a constant speed, and any variation in height of each measured point enlarges the beam. This unfocusing induces a decrease of the energy received by the detector, which sends a signal to the linear motor that positions the optical system, thus regulating the convergence, and refocuses the beam. At each point of variation in the relief of the surface, a refocusing of the beam is necessary. The measurements, therefore, of the displacement of the optical system are equivalent to the variations of the relief. This system is of interest because there is no contact with the analyzed surface; rapid measurements are theoretically possible. However, the mechanical displacement of the optical system, at each refocusing, requires time, and this waste of time diminishes the benefit that the principle of the system might bring. It does, however, have a very good vertical definition that reaches 0.1 μm.
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
181
Laser diode
Laser diode
Pre-amplifier
Prism
P.S. detector
P.S. detector
Pre-amplifier
Detector Collimator O
O
O
O
Lens L1
Focusing lens
Lens L2
Lens L2
Surface Surface
FIGURE 23.5 Optical profilometer: triangulation principle.
FIGURE 23.4 Optical focus profilometer.
For the measurement of skin relief made on silicon rubber negative replicas,13 average measurement speeds of 0.5 mm/sec in profilometry (two dimensions) and 0.3 mm/sec in three dimensions have been obtained. This system has another disadvantage: if it meets sudden changes in height on the relief, it has problems finding a new focusing point. These changes are often due to defects (bubbles) in the silicon rubber used to make the replicas.
6 μm
23.3.2 TRIANGULATION SYSTEM This kind of system, used in research laboratories only, is based on the classical principle of optical triangulation (Figure 23.5) utilized in noncoherent light. From a point of light projected on the surface being analyzed, a corresponding image is obtained on the surface of the detector: any variation of the image of the light on the detector corresponds to a change in the height. By using a laser diode (λ = 850 nm) and photodetectors with semiconductors (position-sensing detector [PSD]), the position of the image obtained on the surface of the PSD can be measured to the micrometer, thus giving a real sensitivity of about 3 μm. Figure 26.6 shows the image obtained with a roughness standard of a maximum amplitude of 6 μm, showing the present limitations of this type of system. Figure 26.7 and Figure 26.8 compare the results obtained with two sensors of different types: one with mechanical contact, the other with a triangulation system. The results show the validity of the latter system, even for the measurement of very short displacements. The main advantage of the triangulation sensor is its speed, which is limited only by the speed of precision translators. An average speed of 10 mm/sec and accelerations of 8 mm/sec2 when starting each measurement line can be reached. With this sensor, the acquisition time of an image of 5.12 × 5.12 mm with one measurement point every 10 μm is only 5 minutes.
2.5 mm 2.5 mm
FIGURE 23.6 Roughness standard (6 μm amplitude).
μm 180 160 140 120 100 80 60 40 20 0
FIGURE 23.7 Mechanical detector.
182
Handbook of Non-Invasive Methods and the Skin, Second Edition
Since the height z = f(x, y) is known, the quantification of the surface is obtained using different methods.
μm 160
23.4.1 PARAMETERS OBTAINED FROM THE EXTENSION OF CLASSICAL AND STANDARD ROUGHNESS PARAMETERS TO THREE DIMENSIONS
140 120
Since standardization of three-dimensional measurement of surfaces has not yet been established, extending the known two-dimensional parameters is the obvious procedure to follow. A tri-dimensional parameter can correspond to a bi-dimensional parameter; for example, one of the most common two-dimensional parameters is Ra (arithmetic mean deviation of a profile), which gives the arithmetic mean of the absolute values of the profile departure, within a sampling length, i.e.:
100 80 60 40 20 0
FIGURE 23.8 Optical detector.
Ra =
Another point of interest regarding this sensor is its wide measurement capability. It can measure a few tens of micrometers as well as variations in levels of up to 8 mm. Examples of the measurements of wounds showing wide surfaces and deep holes are given below.
1 N
N
∑ z(i)
where N is the number of experimental points. The same definition can be extended to surfaces: 1 SR a = N1N 2
23.4 QUANTIFICATION OF SURFACES: MICRORELIEF
N1
N2
∑ ∑ z ( i,j) i=1
μm 60 40 20 0 1
2
3 0
0 −10 −20 −30 −40 −50 −60 −70 −80 −90 μm
FIGURE 23.9 Skin profile and Abbott curve.
20
(23.2)
j=1
A similar extension can be obtained for all the parameters quantifying heights. The bearing curve of a profile is of special interest, as it is a fundamental parameter in the study of friction and wear problems. The graph in Figure 23.9 illustrates the
The image of a surface analyzed by a sensor, whatever its principle, is represented by the law z = f(x, y), known for a discrete number of points (usually 512 × 512 or 256 × 256 points), with a vertical range between 12 and 14 bits.
0
(23.1)
i=1
40
4 60
80
5 100
6
mm
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
183
−58.2 μm −107 μm z
y x
FIGURE 23.11 Autocovariance function of a sine surface. 25.86%
44.31%
29.83%
FIGURE 23.10 Surface of plateaus at different levels.
relationship between the values of the profile bearing length ratio to the profile section level. It shows the projected length at a determined height. By applying this principle to studies on the surface of the skin, the surface of plateaus, and consequently their average size (Figure 23.10), are obtained. z
23.4.2 STATISTICAL ANALYSIS Another way to study the skin surface is by using statistical analysis that quantifies its anisotropy. It is possible to associate the autocovariance C(α, β) to the height z(x, y) of each point of the surface:
C ( α, β ) =
1 N1N 2
N1
N2
∑ ∑ z ( x,y ) z ( x+iΔx,y+jΔy) i=0
j= 0
with α = iΔx
x
FIGURE 23.12 Autocovariance C(α, β) of a cutaneous surface in a 9-year-old child.
ied wavelengths that constitute the surface. Figure 23.13A and B illustrate the spectral density of the skin surfaces of two subjects of different ages. When G(kx, ky) is known, it is possible to find the anisotropy of surfaces.14 The spectral moments described as
β = jΔy
mpq =
(23.3) This value indicates the correlation between any point of the surface and another point distant from iΔx and jΔy, with Ox and Oy being the experimental steps in the reference directions x and y, and i and j the integers of any value. The function C(α, β) shows possible periodicities; examples are provided showing the results obtained from a strictly regular surface (Figure 23.11) or from the skin surface (Figure 23.12). Using C(α, β), the bi-dimensional Fourier transform G ( kx,ky ) =
∑ ∑ C (α, β) exp– 2iπ (αkx + βky) α
β
(23.4) is the spectral density of the surface: it represents the stored energy of any part of the surface and shows the amplitude distribution of the different components of var-
y
∑ ∑ kx ky G ( kx,ky) p
kx
p
(23.5)
ky
and particularly the moments of second order m20, m02, and (zero order) m00, give the anisotropy coefficient γ2 (according to Longuet-Higgins14):
γ2 =
⎧ m2 min = 1 ⎡⎢( m20 + m 02 )+ ( m20 + m 02 )2 + 4 m112 ⎤⎥ m 2 min ⎪ 2⎣ ⎦ with ⎨ 1⎡ 2 2 ⎤ m 2 max m 2 max = ⎢( m 20 + m 02 )+ ( m 20 + m 02 ) + 4 m11 ⎥ ⎪⎩ 2⎣ ⎦
(23.6) The main directions, i.e., the two directions on the surface in which the variations of m2 are extreme, are obtained from the values of m2 min and m2 max. The examples in Figure 23.14 illustrate the correlation between the visual aspect of the surface and the value of its coefficient γ2 of anisotropy. In the case of a purely isotropic surface, γ2 → 1, and in that of a purely anisotropic surface, γ2 → 0. The
184
Handbook of Non-Invasive Methods and the Skin, Second Edition
z
y z
x
y x (a)
z y z
x
y
FIGURE 23.13 (A) Spectral density G(kx, ky) function in a 9year-old male subject; (B) spectral density G(kx, ky) in an 86year-old male subject.
values indicated in Figure 23.14 show that the influence of aging is well described by variations of this parameter.
23.4.3 TEXTURAL ANALYSIS
OF THE
SKIN SURFACE
The particular structure of the skin, made up of plateaus crossed by valleys, can be quantified by special operators applied to the image in its entirety. The techniques developed for image analysis make the quantification of furrows possible. Using a method proposed by Peuker and Douglas,15 cutaneous furrows can be described by using techniques initially developed for earth relief study.
x (b)
FIGURE 23.14 (A) The skin surface of a 9-year-old male subject; (B) the skin surface of an 86-year-old male subject.
Any closed area, including a part of a furrow, shows particular variations of the local relief z = f(x, y), characterized by two valleys (Figure 23.15). Thus, the height variations (zi – zM) will show four changes of sign for any point M belonging to a cutaneous furrow, and the deepest furrows will be distinguished from the smallest by the value of the amplitude of the difference. Since any point selected belongs to a furrow (Figure 23.16), the density of furrows can be calculated, as well as their vertical distribution compared to the vertical distribution of the whole surface.
0
μm 70 0
Zi-Z M 0
0.5
1
mm
FIGURE 23.15 Detection of particular points (furrows): analysis of the height variations along a closed contour.
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
185
FIGURE 23.16 Recognition of the points in furrows.
The ability to separate points on furrows from points belonging to the whole surface obviously induces greater sensitivity. For example, antiwrinkle cosmetic products are developed to act mainly on deep furrows. An investigation of the whole surface would therefore drown the particular information needed in unnecessary experimental points. Furthermore, a selective study of the furrows of medium or large size (secondary or main furrows) can be obtained using discriminating thresholds. Another method used in classifying cutaneous furrows into two families is the Fourier analysis or frequential analysis of the surface. The Fourier transform F(u, v) of the surface z(x, y) is given by discrete variables x = k and y′ = k′: F ( nΔu, mΔv) =
∑ ∑ z ( k,k′ ) exp – 2iπ ( nk+mk′ ) N k
k′
(23.7) where u = nDu and v = mDv in the frequency space. One of the interesting features of this breakdown is that the entire surface can be reconstructed from only one part of the initial spectrum, i.e., after performing a lowpass filtering. The initial spectrum is composed of two parts: one with long wavelengths only (characteristic of plateaus separated by main furrows), the other with short wavelengths (characteristic of secondary furrows crossing the plateaus partially or completely), i.e.: k1
Z ( nΔu,mΔv) =
k2
∑ ∑ z ( k,k′ ) exp – 2iπ ( nk+mk′ ) N+ 0
0
∞
∞
∑ ∑ z ( k,k′ ) exp – 2iπ ( nk+mk′ ) N k1
k2
(23.8) A first approximation of the values of limits k1, k2 of wavelengths is given by calculating the average distance
between plateaus when the cutaneous surface is cut by its mean plane. With this value, an approximate surface is reconstructed using the inverse transform on the part of the spectrum showing the low frequencies only: Z LF ( k,k ′ ) =
∑ ∑ Z ( nΔu,mΔv) exp + 2iπ ( nk+mk′ ) n
m
(23.9) In Figure 23.17, such a reconstruction applied to a skin profile can be seen: (a) the initial profile, (b) the corresponding total spectrum, (c) the profile reconstructed on the wavelength band [k1, ∞], and (d) the low-frequency part obtained from determining the mean distance Sm between plateaus (the limit k1 = distance parameter Sm). The minima of profile (c) show the position of the main furrows. The main furrows, detected by one or the other of the above-mentioned methods, are then subtracted from all the detected furrows, leaving only secondary furrows. Both families of furrows can then be quantified separately (Figure 23.18), resulting in a good description of the skin surface relief. 23.4.3.1 Directional Quantification of Furrows Several methods of determining the linear forms of an image are available: one of them is the method proposed by Groch,16 which first scans the image under study in two perpendicular directions so that the forms can be detected as precisely as possible, and then uses thinning and compression algorithms (Lu and Wang,17 Holt et al.18) to reduce the calculation time. Such a method, used by Awajan et al.19 in the detection of cutaneous furrows, provides good general results but poor directional sensitivity. As proposed by Rosenfeld et al.21 and Duda and Hart,22 application of the Hough20 transform to the extraction of linear forms is a more precise method. Hough uses a change of space: two points on a straight line of the
186
Handbook of Non-Invasive Methods and the Skin, Second Edition
29.61 15.62 1.63 −12.36 −26.34 −40.33 0
1
2
3.1
4.1
5.1 (mm)
6.1
7.2
8.2
9.2
10.2
84
98
112
126
140
6.1
7.2
8.2
9.2
10.2
84
98
112
126
140
(z)
(a) 65 52 39 26 13 0 0
14
28
42
56
70 (f ) (b)
27.83 16.51 5.19 −6.14 −17.46 −20.78 0
1
2
3.1
4.1
5.1 (mm)
(z)
(c) 65 52 39 26 13 0 0
14
28
42
56
70 (f ) (d)
FIGURE 23.17 (a) Initial profile, (b) Fourier spectrum, (c) reconstructed profile from low frequencies, (d) Fourier spectrum of low frequencies.
image correspond to a point (r, υ) in the parameter space. In the space of parameters (r, υ), the resulting curve corresponds to each straight line intersecting a point xi yi of the image space: r = f ( θ, x i y i ) = x i cos θ + y i sin θ
(23.10)
Therefore, a sine curve in the parameter space corresponds to each point (xi yi) of the image space; the alignment of points on the image space will be expressed by a group of sine curves intersecting at the same point in the space of the parameters (Figure 23.19). Several preliminary operations are necessary before applying this transform to the investigation of cutaneous furrows. The reference image must be simplified so that
the points at the bottom of the furrows appear. As there are furrows of varied widths, it is necessary to develop a thinning technique for these linear forms so that an image can be obtained, which reproduces the original distribution of furrows and is composed of two different levels only: the first one classifying all the points belonging to furrows (level 1) and the second one presenting all the external points (level 0). The analysis of furrows is conducted as follows: • • •
Determination of cutaneous furrows Thinning of furrows Quantification
Furrows are localized by applying a local operator, which identifies a furrow on a point of the image. Several
(μm)
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
187
then the points likely to be eliminated will be
69.94 55.95 41.97 27.98 13.99 0
2 ≤ A (P ) ≤ 7 0
1
2
3
4
5.1
6.1
7.1
8.1
5.1 6.1 (mm)
7.1
8.1
9.2
The condition A(P) > 2 is necessary to retain the extremities of the skeleton, and the condition A(P) 7 states that the current point belongs to the edge. All the points detected by the first scanning, thus belonging to the edges of the furrows, are analyzed by a second scanning, and those satisfying the following relation are eliminated:
10.2
(a) 69.94 55.95 41.97 27.98 13.99 0 0
1
2
3
4
9.2
(23.12)
10.2
| ΣAi – ΣBi | = 3
(b)
with Ai and Bi defined in Figure 23.21. Examples of such a configuration are shown in Figure 23.22. The skeleton obtained after the application of this equation is not always of unit thickness, especially at the intersection of several furrows. To obtain a skeleton with a unit thickness, the point under discussion is eliminated as soon as one of the configurations shown in Figure 23.23 is found. The result of the application of these different operators, shown in Figure 23.24, is compared to the original cutaneous surface. From this result, the Hough transform can be seen to express the directional distribution of furrows (density and directions).
FIGURE 23.18 (a) Main furrows, (b) secondary furrows. O
X
Y r
300
2 points
200
3 points
100
23.5 QUANTIFICATION OF SURFACES: MACRORELIEF
0 −100 −200 0°
45°
90°
135°
180°
Up to this point, we have studied only cutaneous microrelief. Two other sorts of skin surface characteristics, wrinkles and wounds, can be studied. They demand other methods of measurement.
θ
FIGURE 23.19 Line detection using the Hough transform.
operators can be used, particularly those proposed by Awajan,23 Cocquerez,24 Groch,16 Jimenez and Navalon,26 and Montavert.27 Awajan uses a specific operator covering eight neighboring points, P1 to P8 (Figure 23.20): if A ( P ) =
∑ Pi
23.5.1 QUANTIFICATION
OF
WRINKLES
The quantification of wrinkles is possible by using the silicon rubber replicas described earlier, and a sensor with a sufficient vertical range, under the condition that a suitable method be utilized. Usually the aim of this kind of study is to test the efficiency of an antiaging product, and the investigation is therefore based on the comparison of
(23.11)
i
P5
P6
P7
1
0
1
0
0
1
1
0
1
1
0
1
1
1
1
P4
P
P8
0
1
0
0
1
1
0
1
1
1
1
1
1
1
1
P3
P2
P1
0
1
0
1
0
1
1
0
1
1
0
1
1
0
1
FIGURE 23.20 (a) Neighboring pixels, (b) masks used in the thinning algorithm (point elimination).
188
A1
Handbook of Non-Invasive Methods and the Skin, Second Edition
A2
A3
P B1
B2
B3
A1
A2
B1
B1
P
A3
B2
B2
B3
P
B3
A1
B2
B1
A2
B3
P
A1
A3
A2
A3
FIGURE 23.21 Configurations for pixel elimination.
1
1
1
−
1
1
0
−
1
0
0
−
−
1
−
0
1
1
0
1
1
0
1
1
0
0
0
0
0
−
0
−
1
−
1
1
1
1
−
1
−
0
−
0
0
0
0
0
1
1
0
1
1
0
1
1
0
−
1
−
−
0
0
1
−
0
1
1
−
1
1
1
FIGURE 23.22 Examples of application.
0
0
0
1
1
1
0
0
0
0
1
1
1
0
1
1
0
0
1
0
0
1
0
0
1
1
0
0
FIGURE 23.23 Masks used to obtain skeletons with a width of one unit.
To meet the first condition, replicas are made at t0 and t1 on similar parts of the body, the size of each replica being larger than the surface to be studied. Then the replica is scanned by a three-dimensional apparatus, either mechanical or optical, the scanned area also being larger than the wrinkle(s) to be examined. Both analyzed surfaces (Figure 23.25) are visualized: an area is selected on one of them and reproduced on the other one, either by a visual position control or using the intercorrelation of both samples, which consists of bringing to a maximum the product of correlation: C ( α, n ) =
∑ ∑ z ( x ,y ) z ( x + α, y + α ) 1
i
i
i
i
where z1 is the height of a point on the first image with coordinates xi yi and z2 is the height of a point on the second image. This product reaches a maximum when both images are correlated perfectly. Once selected and carefully positioned, the areas are quantified as follows. Each profile is obtained by cutting the analyzed surface by a plane perpendicular to the mean plane of the surface. The real volume of the wrinkle is then the sum of all the elementary volumes. An elementary volume is located between two parallel neighboring planes, the surface of the valley, and the two successive straight lines connecting the extreme points of each profile comprising the upper surface (Figure 23.26). The maximum depth h, on average, of the whole surface is calculated:
23.5.2 QUANTIFICATION
two types of results: the evolution of a wrinkle without treatment (the reference) and the evolution of another wrinkle after treatment. The results of the analysis depend on the different evolutions of both surfaces during the treatment (times t0 and t1). To be reliable, this comparison study must be conducted on similar surfaces (same size and necessarily similar position) and be quantified by representative parameters.
2
(23.13)
1 h= N
FIGURE 23.24 Skeleton of the area shown in Figure 23.16.
i
α
N
∑h
i
(23.14)
i=1
OF
WOUNDS
The geometric characteristics of large wounds can be measured with one optical system because of its vertical range capacity and its wide precision translators. A typical example is a leg ulcer, which can present differences in levels of several millimeters on a surface of a few square centimeters. The healing progress of such wounds is monitored by making silicon rubber replicas of the surface at various times. It is useful to make such surface replicas for two main reasons: •
Keeping surface prints is important because they can be used later to check measurements and even to try new methods. These replicas can be stored over a long time.
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
Perimeter: 12.1 mm Surface: 8.07 mm2 Mean height: 123 μm Mean depth: 352 μm Volume: 992 mm2 × μm
189
Perimeter: 12.1 mm Surface: 8.13 mm2 Mean height: 76.5 μm Mean depth: 213 μm Volume: 622 mm2 × μm
FIGURE 23.25 Measurement of topographical parameters made on the same area, before and after treatment.
the wound is absolutely painless. It can reproduce the tiniest details: amplitudes of about a micrometer can be detected, which is sufficient for this kind of investigation. The white color of this material is accepted by the optical triangulation captor. In addition, it takes only 3 to 5 minutes to make a replica, even a large one. hi
23.5.2.1 Performance and Results To analyze the surface of a wound (Figure 23.27), a scan of this surface is made line by line. The measuring step along a line and the distance between two neighboring lines are practically multiples of 10 μm. The result of this scanning is a table of N × L points, since the scanned area is generally rectangular. The acquisition time depends on the size of the surface and the motor used (step-by-step motors); the step motor can be of 10, 50, or 100 μm, producing measuring speeds of up to several tens of millimeters per second, the maximum frequency being 6000 steps per second. The maximum scale of the vertical measuring range is 8 mm; it can be increased by using a large-scale system.28,29
FIGURE 23.26 Evaluation of the volume of a part of a furrow.
•
Only one measuring system is used, even if the prints come from other sources, thus allowing multicentric studies. Risk of error is reduced with just one system only.
Materials like SILFLO®* have been widely tested in dental surgery and in dermatology. This material is completely safe and causes no wound reaction; application to * Registered trademark of Flexico Developments Ltd.
7.5 mm
42 mm
FIGURE 23.27 Ulcer replica scanning.
56 mm
190
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 23.28 Determination of the boundary of the ulcer.
Leg
axis
FIGURE 23.29 Elimination of the local form of the leg.
Perimeter: 639 mm Surface: 297 cm2 Mean height: 817 μm Volume: 24265 mm3
FIGURE 23.30 Geometric parameters of a leg ulcer.
Once the surface z(x, y) is analyzed, the geometrical parameters are obtained in the following manner: flattening the surface by the least squared plane and determining the edge of the wound, following it from a top view of the surface, with heights appearing as false colors (Figure 23.28). The perimeter of the boundary and the surface of the wound are the first parameters to be used, directly plotting the edges of the wound with transparent paper.30 This method needs the operator’s judgment, and it can be replaced either by following up the contour on a screen with a mouse or by an automatic determination of neighboring points.31 Three-dimensional representation of wounds has been tried by photostereogrammetry,32 but this method demands heavier apparatus and is not as accurate as the direct measurement of a relief by triangulation. The exact plot of coordinates x, y, z of any point on the surface plays a part in the evaluation of the volume of the wound, which is a valuable parameter for the study of the healing process. However, the volume is significant only after elimination of the local form of the leg. As shown in Figure 23.29, this elimination is made by scanning the surface of the wound parallel to the longitudinal axis of the leg, and at a distance always greater than the real size of the wound. With this method it is possible to obtain, at the beginning of each profile, part of the healthy skin, which will be the reference. For each profile, a segment of the straight line joins the reference parts and gives the general direction of the leg where there is no wound. The segments as a whole form a surface that reproduces the general shape of the leg without any wound; thus, the real wound that will be analyzed is represented by the volume between the surface measured (included inside the boundary) and the “initial” surface of the leg. Figure 23.30A shows the initial stage of an ulcer and Figure 23.30B the evolution of its boundary during the time of treatment, and the corresponding values of
Perimeter: 516 mm Surface: 185 cm2 Mean height: 658 μm Volume: 14239 mm3
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
191
A1 B1
C1
A1 t
C2
t + dt
A2
r1 B1
r2 B3 B4
A3 r3 A4 r4
FIGURE 23.31 Contour of the ulcer at moments t and t + dt.
parameters: perimeter, surface, and volume. Taking into account the volume of the wound avoids imprecise results in which the periphery and surface are likely to evolve in an unknown direction, whereas the volume decreases in accordance with the treatment duration.
FIGURE 23.32 Theoretical ulcerous leg evolution as a function of time.
23.5.2.2 Theoretical and Experimental Evolution of Healing In view of this ability to trace the boundary of the wound, interesting future developments can be considered, such as theoretical research on its evolution, and thus the prognosis of its evolution. Interesting theoretical research was carried out by Amiez33 using geometric criteria combined with the results of a clinical investigation. She studied the displacement of the contour of an ulcer over a specific period. If C1 and C2 are two contours at two very close moments t and t + dt (Figure 23.31), contour C2 is deduced from C1 by successive progressions defined by quantities r1, r2, r3 … of points A1, A2, A3 … of the contour at t. Any element such as Ai Ai + 1 scans the elementary surface defined by the quadrilateral Ai, Ai + 1, Bi + 1, Bi. With this method, the whole boundary can be split into several quadrilaterals. According to the clinical observations made by Agache,34 the change of the contour depends on its local curvature: an arc of the contour with a large curvature radius will change more quickly than an element with a smaller radius, so that the surface dSi scanned by the element di + 1 is given by dSi = Kdt di + 1, where K is a constant. Amiez links di + 1 to quantities ri, ri + 1, for the N points of the boundary, thus developing a system with N equations, with N unknown. Therefore, the resolution of the system is possible, and with the knowledge of the boundary at an instant t, knowledge of the contour at a further instant t + dt is produced. This new contour is then the basis for a new calculation, giving the solution at the next stage. The successive healing stages of the contour of the wound can be foreseen with this iterative method. Figure 23.32 gives an example of the theoretical evolution of the real boundary of an ulcer applying the above method; Figure 23.33 shows the results of the theoretical calculation as well as the experimental results (broken lines) applied to a real boundary.
FIGURE 23.33 Ulcerous leg contour evolution. Theoretical contour, continuous lines; experimental contour, broken lines.
23.6 CONCLUSION The study of the microrelief of the surface of the skin is of interest to dermatologists, surgeons, and manufacturers of cosmetic products. The skin relief connected to the dermis and epidermis is dependent on numerous parameters that play a role in its evolution. Therefore, precise knowledge of skin relief can provide important information regarding the effects that certain illnesses, aging, and radiation have on human beings. The capabilities of skin microrelief study have been greatly improved because of the existence of three dimensions in topographical studies. Because of the use of both the real space z = f(x, y) and the frequency (or wavelength) space, a large field of application has been opened up, where furrows can be separated into two families: one family that is connected to the dermis and one that is connected only to the stratum corneum. Parameters specific to the cutaneous surface can take the place of traditional two-dimensional parameters, which were borrowed from the methods used in measuring industrial surfaces. The transfer to three dimensions opens up the field of study of texture, because the direction of the furrows can be detected and quantified, showing in particular the effects of the local mechanical deformation and aging. The macrorelief of the skin is composed of furrows of great amplitude, wrinkles, and wounds.
192
Handbook of Non-Invasive Methods and the Skin, Second Edition
The study of wrinkles can be performed with much greater precision in three dimensions because the isolated detection of one and the same area at different stages of a treatment becomes possible. The quantification of geometric parameters such as surface, volume, and depth of the same part of a wrinkle offers a reliable way to test the efficiency of antiwrinkle products. Wounds evolve, like wrinkles, according to different parameters. The healing process can be measured in an objective and quantitative manner by following the evolution of the geometric parameters: perimeter, surface, and volume of the wound. These are detected by making replicas that are then utilized for measuring the relief. Analyzed areas of large size demand the use of noncontact sensors, which are the only devices capable of working in three dimensions at satisfactory speeds. The samples discussed throughout this chapter show the interest in and necessity for three-dimensional measurements of the geometric characteristics associated with the micro- and macrorelief of the skin, as well as all the potential possibilities of texture analysis.
REFERENCES 1. Cummins, H. and Midlo, C., Palmar and plantar epidermal ridge configuration in European-Americans, Am. J. Phys. Anthropol., 9, 471, 1926. 2. Wolf, J., Das oberflächenrelief der menschlichen Haut [Skin surface relief in man], Z. Mikr. Anat. Forsch., 47, 351, 1940. 3. Tring, F.C. and Murgatroyd, L.B., Surface microtopography of normal skin, Arch. Dermatol., 109, 223, 1974. 4. Marks, R. and Saylan, T., The surface structure of the stratum corneum, Acta Derm. Venereol., 52, 119, 1972. 5. Thomas, T.R., Recent advances in the measurement and analysis of surface microgeometry, Wear, 33, 205, 1975. 6. Snaith, B., Edmonds, M.J., and Probert, S.D., Use of a profilometer for surface mapping, Precision Eng., 141, 87, 1981. 7. Xie, Y., in Quantification de la topographie des surfaces, no. 246, Thesis, University of Besançon, France, February 1992, chap 1. 8. Sampson, J., A method of replicating dry or moist surfaces for examination by light microscopy, Nature, 191, 932, 1961. 9. Sarkany, I., Method for studying the microtopography of the skin, Br. J. Dermatol., 74, 254, 1962. 10. Cook, T.H., Craft, T.J., Brunelle, R.L., Norris, F., and Griffin, W.A., Quantification of the skin’s topography by skin profilometry, Int. J. Cosmet. Sci., 4, 195, 1982. 11. Williamson, J.P., The microtopography of surfaces, Proc. Inst. Mech. Eng., 182, 21, 1968. 12. British Standard 1134, 1972, revised 1988. 13. Makki, S., Barbenel, J.C., and Agache, P., A quantitative method for the assessment of microtopography of human skin, Acta Derm. Venereol., 59, 285, 1979.
14. Longuet-Higgins, M.S., The statistical analysis of a random moving surface, Philos. Trans. R. Soc. London Ser. A, 249, 966, 321, 1957. 15. Peuker, T.K. and Douglas, D.H., Detection of surface points by local parallel processing of discrete terrain evaluation data, Comput. Graphics Imaging Process., 4, 373, 1975. 16. Groch, W.D., Extraction of line shaped objects from aerial images using a special operator to analyse the profiles of functions, Comput. Graphics Imaging Process., 18, 347, 1982. 17. Lu, H.E. and Wang, P.S.P., An improved fast parallel thinning algorithm for digital pattern, in IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, June 19–23, 1985, p. 364. 18. Holt, C.M., Stuart, A., Cunt, M., and Perrot, R.H., An improved parallel thinning algorithm, Commun. ACM, 30, 156, 1987. 19. Awajan, A., Rondot, D., and Mignot, J., Quick method of measuring the furrows distribution on skin surface replicas, Med. Biol. Eng. Comput., 27, 379, 1989. 20. Hough, P.V.C., Method and Means for Recognizing Complex Patterns, U.S. Patent 3,069,654, December 18, 1962. 21. Rosenfeld, A., Thurston, M., and Lee, Y.H., Edge and curve detection for visual scene analysis, IEEE Trans. Comput., C2D, 562, 1971. 22. Duda, R.O. and Hart, P.E., Use of the Hough transformation to detect lines and curves in pictures, Commun. Assoc. Comput. Mech., 15, 11, 1972. 23. Awajan, A., Detection et analyse des structures linéaires d’une image. Applications biomédicales et industrielles, no. 78, Thesis, University of Besançon, France, 1988. 24. Cocquerez, J.P., Analyse d’images aériennes: extraction de primitives rectilignes et antiparallèles, Ph.D. thesis, Paris Sud (Orsay) University, France, 1984. 25. Cocquerez, J.P. and Devars, J., Détection de contours dans les images aériennes: nouveaux opérateurs, Traitement Signal, 2, 45, 1985. 26. Jimenez, J. and Navalon, J., A thinning algorithm based on contours, Comput. Vision Graphics Imaging Process., 99, 186, 1987. 27. Montavert, A., Obtention d’une ligne médiane par connexion de l’axe médian, in 5th Congress of AFCETINRIA, Grenoble, France, November 27–29, 1985, p. 777. 28. Chuard, M., Mignot, J., Nardin, P., and Rondot, D., Range expansion and automation of a classical profilometer, J. Manuf. Syst., 6, 223, 1987. 29. Zahidi, M., Assoul, M., Bellaton, B., and Mignot, J., A fast 2D/3D optical profilometer for wide range topographical measurement, Wear, in press. 30. Carrel, A. and Hartmann, A., Cicatrisation of wounds. The relation between the size of a wound and its rate of cicatrisation, J. Exp. Med., 24, 429, 1916. 31. Zahouani, H., Assoul, M., Janod, P., and Mignot, J., Theoretical and experimental study of wound healing: application to leg ulcers, Med. Biol. Eng. Comput., 30, 234, 1992.
Three-Dimensional Evaluation of Skin Surface: Micro- and Macrorelief
32. Eriksson, G., Eklund, A.E., Torlegard, K., and Dauphin, E., Evaluation of leg wear treatment with stereophotogrammetry, Br. J. Dermatol., 101, 123, 1979.
193
33. Amiez, G., Cicatrisation des ulcères, paper presented at the National Meeting of Numerical Analysis, Port Barcarès, France, 1988. 34. Agache.
Morphological Tree of the 24 The Cutaneous Network of Lines H. Zahouani Laboratoire de Tribologie et Dynamique des Systèmes, Ecully, France
Ph. Humbert Laboratoire de Biologie et d’Ingénierie Cutanée, Besançon, France
CONTENTS 24.1 Introduction............................................................................................................................................................195 24.2 Material and Method .............................................................................................................................................196 24.3 Fourier Transform of Skin Lines Network and Frequency Range.......................................................................196 24.3.1 Spatial Frequency Range of Lines Network .............................................................................................197 24.3.2 Fourier Spectrum Representation ..............................................................................................................197 24.3.3 Skin Lines Orientation in the Fourier Space ............................................................................................198 24.3.4 Spectral Rose of Skin Lines Anisotropy...................................................................................................198 24.4 Lines Identification by Anisotropic Spectral Filtering .........................................................................................198 24.4.1 Directional Extraction of the Wrinkle.......................................................................................................199 24.4.2 Directional Extraction of Lines Family ....................................................................................................200 24.5 Determination of the Three-Dimensional Tree Skin Spectrum............................................................................200 24.5.1 Statistical Analysis of Skin Lines Morphology ........................................................................................200 24.5.2 Morphological Tree of Skin Network of Lines ........................................................................................201 24.6 Conclusion .............................................................................................................................................................203 References .......................................................................................................................................................................203
24.1 INTRODUCTION The skin surface shows a specific topography depending on the anatomical site, age and sex. In general, the skin morphology presents a deterministic network of lines, who by its organization expresses all the multidirectional tensions of elastic fibers and the collagen beams. Microlines, primary lines, fine wrinkles and wrinkles represent, in fact, the special organization of collagen bundles and elastic fibers in the superficial dermis, and there is a relationship between the morphology of skin lines and elastic network [1,2]. Different functions can be attributed to the lines network. The first function is the retention and drainage canals of the sebum and sweat. They collect preferentially and retain for a long time the substances applied to the skin: they are thus preferential sites of percutaneous absorption.
This reservoir function allows the applied topical products to be stored on the skin surface and then eventually to diffuse in its different layers. The second function is mechanic; during aging the depth, width, density, and orientation of skin lines change. Some lines become more marked; they evolve progressively in marked anisotropy connected to the decrease of the elasticity of the collagen fibers. The resistance of the skin to traction prevails in a direction, different according to areas from the body, which follows the principal lines (Langer Lines). Langer Lines were established on the corpse starting from the oval form taken by a wound carried out with a round punch, and they correspond to the large axis of the oval. An excised skin narrows more and presents a minimum of extensibility in this direction. It thus acts of an anisotropy of the spontaneous tension of the skin, with distinguishing well from an additional tension of this one induced by a stretching of muscular or visceral origin. Hashimoto [3] gave a precise classification of the line network scales: the 195
196
Handbook of Non-Invasive Methods and the Skin, Second Edition
primary lines are clearly marked and are between 20 and 100 μm deep. The secondary lines are more discrete and correspond to a depth of 5–40 μm from the diagonals to the primary lines. Tertiary and quaternary lines cannot be seen visually. The tertiary lines correspond to the corneocyte border (about 0.5 μm) and the quaternary lines correspond to each corneocyte morphology (about 0.05 μm). In this work we have developed a new approach that is able to classify the skin network of lines with respect to the depth, width, and orientation. The method involves last microscope or optical fringe projection measurements [4,5] and image processing. Fourier transform and associated image processing techniques can also assist automated identification and processing of the features of interest. The global anisotropy between different skin line components has been represented in a polar histogram in the form of a spectral rose of the frequencies. The multiscale spectrum of the lines network can be represented as a three-dimensional tree that shows the hierarchic organization of the skin morphology vs. the scale of lines, wavelength, and orientation: f (ρ, λ, θ). This morphological spectrum is highly useful in finding the correlation between the scale of lines and aging or cosmetic treatment.
24.2 MATERIAL AND METHOD The last 30 years have seen a great step forward in the field of three-dimensional microscopy, especially in the analysis of the surface topography. The methodology was greatly improved when its principle was developed so as to make it possible to reproduce the surface topography in a three-dimensional space. This technological achievement allows scientists to reconstruct three-dimensional images from a nanometric scale and to reconstruct threedimensional images from a nanometric scale to a macroscopic relief. The vertical and lateral resolution and the maximum vertical range of the principal three-dimensional profilers used to assess the skin relief are summarized in the Table 24.1. Depending on the range of skin relief and zone, we analyze in this work the skin relief by laser defocusing system or fringe projection method [4,5]. The defocusing laser microscope uses an optical measuring head, which is a servomotor operating as a transmitter-receiver. A laser diode (transmitter) sends out a beam of light to the surface and the reflected signal is sent back to a set of four photodiodes (receiver). A servomotor drives a lens in order to get a maximal reflected light intensity on the photodiodes. The diameter of the spot on a flat surface is 1 μm; this diameter varies when it is transmitted to a rough surface. The diameter variation of the spot in the course of the measurement triggers an automatic localization control by vertical shifting of the lens. The lens is part of a vertical movement system, which records the localization height
TABLE 24.1 Principle Three-Dimensional Profilers Measurement System Tactile profiler Fringe projection Laser triangulation Laser defocusing microscope Confocal microscopy
Vertical Resolution 0.1 5–10 1 0.3
μm μm μm μm
0.003 μm
Lateral Resolution 2–5 20–50 50 1
Vertical Range of Measurement
μm μm μm μm
6000 μm 1000 μm 10 mm 1000 μm
1 μm
500 μm
that forms the topographical signal. The fringe projection method involves an optical measuring procedure, which uses a combination of gray code and phase-shift techniques to generate three-dimensional information. It is possible to record absolute spatial coordinates of all object points in the image area recorded with a high degree of accuracy in less than 1 sec. The measuring system consists of a projection unit and a CCD camera, which are fixed at the “triangulation angle.” With the gray code method, gratings with a rectangular brightness distribution and differing number of lines are projected consecutively. The number of lines is doubled with each projection run, which unambiguously defines the stripe order for each image point. When using the phase-shift technique, only one grating with sinus-shaped intensity distribution is projected multiple times with varying phase positions.
24.3 FOURIER TRANSFORM OF SKIN LINES NETWORK AND FREQUENCY RANGE Fourier analysis is a powerful tool used to condense information represented in the spatial (or time) into the frequency domain, and to enhance the information to detail individual frequency or wavelength values. It reveals the absolute and relative contributions of different wavelength components to the mean square height of surfaces. In addition, it is more convenient to define and to separate the so-called roughness and waviness components by twodimensional spectral analysis. In Fourier analysis the basis function is the exponential function, although Fourier series are frequently written in alternative form using the trigonometric functions sine and cosine. In this work the skin topography components have been analysed by two-dimensional Fourier transform. The determination of the real and imaginary parts of Fourier transform permits the computation of the amplitude, direction, and wavelength of each spatial frequency [6,7]. The complex Fourier transform computed from scanned data is given by the relation
The Morphological Tree of the Cutaneous Network of Lines
0
N/2
197
N–1
–N/2 –N/2
0
f (x, y) → F(vu, vv)
N/2
High frequencies
0
N/2
High frequencies
0
Low frequencies N/2
N–1
(a)
(b)
FIGURE 24.1 Spatial frequencies repartition.
F ( vu , vv ) =
1 1 N M
N −1 M −1
∑∑ f (x , y ) j
vlu =
k
J =0 k =0
(24.1)
1 1 1 = = = 0.1953 cycles mm –1 N Δ x M Δ y 512 × 0.01 (24.4)
exp − 2 πi ( vu , x j , vv , yk ) equivalent to a wavelength of λl = 1/0.1953 = 5.12 mm which represents the spectrum of surface topography defined in the finite bandwidth by the low and high spatial frequencies limit, respectively:
vlu =
u v cycles mm −1 , vlv = NΔ x MΔ y
vhu =
1 1 cycles mm −1 , vhv = 2Δ y 2Δ x
(
(
)
and
) (24.2)
(Δx, Δy are the sampling steps) the subscripts u and v indicate the wavelength and orientation of a sine wave in the original data, where N and M are the number of data points along x and y axes.
24.3.1 SPATIAL FREQUENCY RANGE NETWORK
OF
LINES
The bandwidth of the frequency range is given by the relation (24.3). For example, if N = M = 512 as the choice for the maximum sample size. If the sample intervals are fixed at Δx, = Δy = 10 μm, this sample size makes the sample dimension 5.12 mm × 5.12 mm. Hence the wavelength of the longest component of skin furrows that can be identified in the sample is approximately 5.12 mm. The frequency range of the analysis can be readily calculated from the expression (3). The high frequency limit
vhu =
1 1 1 = = = 50 cycles mm –1 2 Δ x 2 Δ y 2 × 0.01
24.3.2 FOURIER SPECTRUM REPRESENTATION Two-dimensional data recorded containing N × M points with an origin located at j = k = 0 will give rise to N × M array or spectral coefficients under the two-dimensional Fourier transformation. The spectral coefficients will form the pattern of the skin relief. The origin will be located in a position that corresponds to the origin of the data array. The frequency of these spectral coefficients increases along directions that correspond to the positive directions of the x and y axes of the array up to the Nyquist frequencies (N/2 and M/2). The coefficients are then repeated in inverse order. The representation of the Figure 24.1a has no physical significance. To overcome this, the quadrants of the spectrum can be rearranged as illustrated in Figure 24.1b. This operation moves the origin to the center of the array, and the pattern formed by the rearranged quadrants is that which would be generated by an optical Fourier transform system or diffraction (Fraunhoffer) type analysis [8]. In practice the exchange of quadrants is effected by a mathematical operation rather than by physical rearrangement of the transform coefficients. Gonzalez and Wintz (1977) show that multiplying f (xj, yk) by (–1)j+K is equivalent to shifting the origin to (N/2, M/2) as required. So the low frequencies of skin relief spectrum F(νx, νy), are located in the center of the spectrum, by using the frequential translation property [6,7]: TF[exp(–2πi(vu0x + vv0y)f(x,y)] = F(vu – vu0, vv – vv0) (24.5) for
(24.3) vu 0 = vv0 = equivalent to a wavelength of λh = 1/0.02 = 0.02 mm or 20 μm. Similarly, the low frequency limit νlu is given by
we have
N 2
(24.6)
198
Handbook of Non-Invasive Methods and the Skin, Second Edition
μm
60.4 51.3 42.3 33.2 24.2 15.1 6.0 –3.0 –12.1 mm –21.1 –30.2 –39.2 –48.3 –57.4 –66.4 –75.5 –84.5
3
2
1
0
0
1
2 mm 2D View of skin lines topography
3 Spectrum
FIGURE 24.2 Example of Fourier spectrum skin lines.
N N⎞ x y ⎛ TF ⎡( -1) ( -1) Z ( x,y ) ⎤ = F ⎜ vu − , vv – ⎟ ⎣ ⎦ ⎝ 2 2⎠ (24.7) Figure 24.2 shows the Fourier spectrum of the volar forearm skin lines.
24.3.3 SKIN LINES ORIENTATION SPACE
IN THE
FOURIER
The complex coefficients F(νu, νv) of a two-dimensional spectrum can be used to evaluate the amplitude and power spectra of the topography. The magnitude of these coefficients has the normal interpretation. The frequency domain coordinates, however, have special significance. The subscript (u, v) indicate the wavelength and orientation of a sine wave in the original data. The axes u and v are perpendicular in Fourier space, so the frequency of a particular coefficient in the transform is given by
(
1 2 2 v
v = vu2 + v
)
(24.8)
whiles the orientation of the wave, relative to the u axis, is given by ⎛ u⎞ θ = tan −1 ⎜ ⎟ ⎝ v⎠
24.3.4 SPECTRAL ROSE
OF
(24.9)
SKIN LINES ANISOTROPY
The amplitude of skin relief spectrum is given by: A(νu, νv) = 2|F(νu, νv)*F(νu, νv)|
(24.10)
where F(νu, νv)* is the complex conjugate of F(νu, νv). The amplitude spectrum can be expressed with polar coordinates A(ρ,θ). The various topographical components with respect to the wave direction identified in the polar spectrum can be directionally represented in the polar diagram in the form of a spectral rose [8,9]: RS = (θ), (0 ≤ θ ≤ π)
(24.11)
generated with the computation of the direction of each spatial frequency, identified with respect to the x axis in the original sample [6]. The geometrical form of the spectral rose represents the global anisotropy and permits the identification of the direction of each line’s family. Figure 24.3 shows the spectral roses with respect to the age of the skin morphology.
24.4 LINES IDENTIFICATION BY ANISOTROPIC SPECTRAL FILTERING Traditionally the anisotropy of surface topography has been studied by different anisotropy indexes, defined initially by L. Higgins [11] and adapted to surface roughness by Nayak [12]. The major part of the anisotropy indexes have been defined in the sense of the variation of the surface gradient vector. The method developed in this work quantified the anisotropy of skin relief by computing the direction of each component in the spectrum. The basic idea of this method is to introduce the notion of the anisotropy between form, waviness, and microrelief using the direction as an indicator parameter of each component, which allows exact extraction of the spatial frequencies of each morphological family. These two methods of anisotropic extraction have been developed to extract the low or high frequencies vs. the anisotropy of the relief morphology of skin.
The Morphological Tree of the Cutaneous Network of Lines
μm
5
21 years 4
mm
3
2
33 years 1
0 0
1
2
3
4
μm
5
62 51 40 29 18 7 –4 –15 –26 –37 –48 –59 –70 –81 –92 –103 –114
5
4
3 mm
62 51 39 27 16 4 –7 –19 –30 –42 –53 –65 –77 –88 –100 –111 –123
199
2
1
0
mm
μm
54 years
4
3
Spectral rose
2
85 years
1 Skin Morphology 0
1
2
mm
3
4
μm
5
122 100 78 56 34 12 –10 –32 –54 –76 –98 –120 –141 –163 –185 –207 –229
2
3
4
5
3
4
5
5
4
mm
5
0
1
mm
mm
101 85 68 52 36 19 3 –13 –30 –46 –62 –79 –95 –111 –127 –144 –160
0
3
2
1
0
0
1
2 mm
FIGURE 24.3 Spectral roses with respect to the age of skin lines morphology.
5 4.5 4 3.5 3 2.5 2 1.5 1 0.5
1 1.5 0 0.5 mm 0
2 2.5
3 3.5
4 4.5
4.5 4 3.5 3 2.5
5 4.5 4 3.5 3 2.5 2 1.5
5
1 0.5
1 1.5 0 0.5 mm 0
Original morphology
2 2.5
3 3.5
4 4.5
5
2 1.5
4.5
1 0.5 0 0 mm
0.5
1
1.5
2
2.5
3
3.5
4
The micro relief
Extraction of the wrinkle
FIGURE 24.4 Extraction of the wrinkle.
24.4.1 DIRECTIONAL EXTRACTION
OF THE
WRINKLE
Two fundamental sampling templates are introduced to extract the specific morphology in the frequency domain [6]. The first template is the parallel line template, which involves the computation of the total power observed along a series of superimposed parallel lines across the Fourier spectrum. If we consider the case of the wrinkles frequencies oriented in the x or y axis of the spectrum. The extraction of the frequencies can be realized by the
inverse Fourier transform of the frequencies oriented in x or y direction as following: + vx
Z ( x, y )Wrinkles = TF −1
∑ ⎡⎣ F ( x , v )⎤⎦ x
y
θ⊥ u or // u
v = − vx
(24.12) Figure 24.4 shows the separation of the principal wrinkle and the furrows oriented differently.
200
Handbook of Non-Invasive Methods and the Skin, Second Edition
Following Langer’s investigations of mechanical properties of the integument, recent in vivo studies confirmed and assessed the anisotropy of the skin network. The anisotropy is generally characterized by different mathematical or imaging analysis methods. These approaches consider the anisotropy as a nonhomogenous directional distribution D(θ) of skin lines, represented in an x,y plane without any information about volumetric anisotropy. Unfortunately this method doesn’t represent the reality of the skin network distribution, which can be represented by three anisotropic families:
90°
0°
(a)
FIGURE 24.5a Directional sampling.
24.4.2 DIRECTIONAL EXTRACTION
OF
LINES FAMILY
The second method is the radial sampling template; the wedge sampling filter integrates spectral energy in a number of given angular directions (Figure 24.5a):
Z(x,y) = αF(micro-relief, θ1) + βG(Langer-Lines, θ2) + γ H(wrinkles, θ3) (24.14) α, β, and γ are the amplitudes skin relief, which depend on sex, age, zone, and pathology. θ1, θ2, θ3 are dependent on the depth scale and wavelength of skin morphology.
θ j + Δθ
Z ( x, y )texture ( x, y ) Δθ = TF
−1
∑ ⎡⎣ F ( v , v )⎤⎦ x
y
θj
(24.13)
24.5.1 STATISTICAL ANALYSIS MORPHOLOGY
As shown in Figure 24.5b, this Fourier decomposition is adapted to the separation of the multidirectional volarforearm skin lines.
μm
2 mm
Δθ = 0° → 90° Skin morphology top view
0 0
1
mm
2
3
3
2 μm
60.4 51.3 42.3 33.2 24.2 15.1 6.0 –3.0 –12.1 –21.1 –30.2 –39.2 –48.3 –57.4 –66.4 –75.5 –84.5
3
μm
2
79.4 70.6 61.9 53.1 44.4 35.6 26.8 18.1 9.3 0.6 –8.2 –16.9 –25.7 –34.4 –43.2 –51.9 –60.7
mm
30.2 26.0 21.9 17.8 13.6 9.5 5.3 1.2 –2.9 –7.1 –11.2 –15.4 –19.5 –23.7 –27.8 –31.9 –36.1
θ ≈ 0°
1
1
0
0
1
mm
2
3
θ ≈ 90° 1
Δθ = 90° → 180° 0
0
1
mm
2
μm
3
55.1 46.8 38.5 30.3 22.0 13.7 5.4 –2.9 –11.1 –19.4 –27.7 –36.0 –44.2 –52.5 –60.8 –69.1 –77.4
3
2
1
0 0
1
0
1
mm
2
3
3
2
1
0
2 mm
(b)
FIGURE 24.5b Directional spectral extraction of skin lines.
SKIN LINES
The nature of the skin relief shows that each local motif of lines network can be represented by three parameters:
3
mm
μm
35.3 30.1 24.9 19.8 14.6 9.4 4.3 –0.9 –6.1 –11.2 –16.4 –21.5 –26.7 –31.9 –37.0 –42.2 –47.4
OF
mm
270°
24.5 DETERMINATION OF THE THREEDIMENSIONAL TREE SKIN SPECTRUM
Frequencies of lines familiy
mm
Directional Sampling 180°
3
The Morphological Tree of the Cutaneous Network of Lines
3.0%
6%
2.5%
5%
2.0%
4%
1.5%
3%
1.0%
2%
0.5%
1%
0.0%
0%
201
Morphological Rose
90
110
70 50
130
ρ
0
170 λs
24
116
30
150
720
10
180
0%
λsmean = 186 μm
Zmean = 33.49 μm
1%
2%
3%
0
FIGURE 24.6 Statistical distribution of lines morphology. 45.5
120
90
120
60
90
90
120
60
60
36.7 28.0
150
150
30
150
30
30
19.2 10.5 1.7 –7.0
0% 0.25 0.5 0.75
–1.2 → 14.4 μm
14.4 → 29.9 μm
90
90
120
–15.8 –24.5
0
0% 0.25 0.5 0.75
120
60
150
0% 0.25 0.5 0.75
0
29.9 → 45.5 μm 90
60
150
30
0
120 30
60
150
30
–33.2 –42.0
0% 0.25 0.5 0.75
–50.7
–47.8 → –32.3 μm
–59.5
90 120
0
0% 0.25 0.5 0.75
0
–32.3 → –16.7 μm 120
60
90
0
0% 0.25 0.5 0.75
–16.7 → –1.2 μm 90
120
60
60
–68.2 –77.0
150
150
30
150
30
30
–85.7 μm
–94.5
0% 0.25 0.5 0.75
0
0% 0.25 0.5 0.75
–78.9 → –63.4 μm
–94.5 → –78.9 μm
0
0% 0.25 0.5 0.75
0
–63.4 → –47.8 μm
FIGURE 24.7 Anisotropy of furrows vs. the depth of the relief.
Z(x,y) = f(ρ,λ,θ)
(24.15)
ρ the amplitude, λ the wavelength, and θ the direction. The local elements ρij, λij, and θij are quantified after the spectral sampling of different furrows family of the skin relief. The density of distribution of each parameter ρij, λij, and θij can be quantified as shown in Figure 24.6.
24.5.2 MORPHOLOGICAL TREE OF LINES [13–15]
OF
SKIN NETWORK
The morphological rose of skin lines network is generated by computing the density of lines oriented in a direction θ identified with respect to the x axis in the original sample. The geometrical form of the morphological rose
represents the local and global lines anisotropy and permits to quantify the phase of surface topography. The knowledge of the local heights of each line’s motif ρ permits the representation of the morphological rose vs. the line depth: RM(θ) = f(Δρ)
(24.16)
This original representation of the anisotropy vs. the level of lines family can be used to study the volumetric aspect of anisotropic skin network. Figure 24.7 shows an example of the anisotropy of lines vs. the depth of the relief after the spectral sampling of the lines in different directions.
202
Handbook of Non-Invasive Methods and the Skin, Second Edition
ρ
Sarah tree 3D view
? ?
? ?
μm
? ?
3D view
45.5 28.0 10.5 –7.0 –24.5 –42.0 –59.5 –77.0 –94.5
? ?
50.8 33.9 17.0 0.1 –16.8 –33.8 –50.7 –67.6 –84.5
θij, λij
? ? ?
θ = 0°, ϕ = 15°
90
180
180
0
270
Zij(x, y)
μm
→ (ρij, λij, θij)
270
FIGURE 24.8 Morphological tree of skin lines.
μm
93 75 57 39 21 3 –15 –33 –51 –69 –87 –105 –123 –141 –159 –177 –195
4
25 years 3
110
90
70 50
mm
130
2
150
30
170
1
10
180
0%
1%
2% 3%
0
0 0
1
2 mm
3
4
88 63 37 12 –13 –39 –64 180 –90 –115 μm 270
90 0
84 years 3 110
mm
μm
98 85 71 58 45 31 18 5 –8 –22 –35 –48 –62 –75 –88 –102 –115
130
2
50
10
170 180
0
70
30
150
1
90
0
1
2 mm
0% 1% 2% 3% 4% 5% 6% 7% 8% 9%
3
0
85 56 27 –1 –30 –58 –87 –116 –144 μm
180
90
270
0
(a)
FIGURE 24.9a Assessment of skin aging.
This method has been extended to the totality of the skin lines and represented in a three-dimensional tree. Figure 24.8 represents the tree of a volar forearm network of lines of a young women, from the summit of the plateau to the deepest valley of lines in the form of a threedimensional tree. The z axis of the tree represents the
height of the furrows; the density of lines orientations is represented for each scale of the skin relief family. An illustration of this method applied to the assessment of the aging and cosmetic application is given in Figure 24.9a, b, and c [13–16].
The Morphological Tree of the Cutaneous Network of Lines
203
–60 –40 –20 0 20 40 60 Z μm
100 Before
2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0
–60 –40 –20 0 20 40 60 Z μm
2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 mm
80 60 Altitude μm
40 0 –20 –40 –60 –80 –100 0
90
180
0
1 2 3 Density %
4
0
1 2 3 Density %
4
100
6 months after treatment
80
2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 mm
20
60 Altitude μm
40 20 0 –20 –40 –60 –80 –100 0
90
180
(b)
FIGURE 24.9b Reorganization of the secondary lines under the effect of vitamin C [16].
24.6 CONCLUSION Since the 1980s, the study of the aging process or the effect of a cosmetic by the analysis of the morphology of the skin lines has used quantifying techniques developed for mechanical engineering. As mentioned above, the morphology of the cutaneous relief is not to be compared to the surface of an iron sheet. The skin physiology seems mainly mechanical, merely following the dimensions of a contracted and elastic medium dermis. The main lines are formed in the superficial dermis where they remain visible after separation from the epidermis, whereas the epidermis loses every trace of them. However, the secondary lines marked on the plateau of the relief can be of epidermic origin. For this reason an original analysis technique was developed that is well adapted to the physiological phenomena. This new approach is similar to a mathematical microscope that can reveal the skin at different magnifications and shows the results with a morphological tree. This novel method displays the hierarchy of the many families of the skin lines, taking into account their depth, width,
and direction. With this new technique, it is now possible to quantify objectively and accurately the morphological changes of the lines during the aging process or after application of a dermocosmetic. The precision of the analysis evidences the influence of a cosmetic on the branches of the tree by distinguishing the smoothing, restructuring, or thinning of the lines and wrinkles. The current technological advances will spread the use of this technique in skin pathologies, probably supplemented usefully by videomicroscopy.
REFERENCES 1. C.E. Pierard, C. Franchioment, C.H.M. Lapierre, “Le vieillissement, son expression au niveau de la microanatomie et des propriétés physiques de la peau.” Int. J. Cosmet. 2:209, 1980. 2. C. Plewing, “Regional differences of cell sizes in the human stratum corinum. Effect of sex and age.” J. Invest. Dermatol., 73:67, 1979. 3. L. Hashimoto, “New methods for surface ultrastructure.” Int. J. Dermatol. 13:357–381, 1974.
204
Handbook of Non-Invasive Methods and the Skin, Second Edition
1 Mean depth = 387 μm
Before
30 28
26 24
22 20
18 16
14 12
10 8 6
4 2 4 2 0 0 mm
6
Altitude mm
0.5
22 20 18 16 14 12 10 8
0 –0.5 –1
180
270
–1.5 0
Mean depth = 182 μm
1
2 3 4 5 Density %
24 22 20 18 16 14 12 10 8
After 28 26 24
22 20
Altitude mm
0.5
18 16 14 12
10 8 6 4 2 0
0 –0.5 –1
6 4 2
180
0
270
–1.5 0 1 2 3 4 5 6 Density %
(c)
FIGURE 24.9c Wrinkle treatment. 4. S. Jaspers, H. Hopermann, U. Hoppe, R. Lunderstadt, J. Ennen, “Rapid in vivo measurement of topography of human skin by active image triangulation using a digital micromirror device.” Skin Res. Technol., 15:195–207, 1999. 5. K.J. Stout, P.J. Sullivan, W.P. Dong, E. Mainsah, N. Luo, T.G. Mathia, H. Zahouani. “The development of methods for the characterisation of roughness in three dimensions.” Commission of the European Communities, Brussels, (ISBN 7044 13132), 1993. 6. H. Zahouani, “Spectral and 3D motifs identification of anisotropic topographical components. Analysis and filtering of anisotropic patterns by morphological rose approach.” Int. J. Mach. Tools Manuf., 38:615–623, 1998. 7. I. Sherrington and E.H. Smith, “Fourier models of the surface topography of engineering components.” Surface Topography, Vol. N° 1, 1988, pp. 11–25. 8. G. George Lendaris, Gordon L. Stanley, “Diffraction Pattern Sampling for Automatic Pattern Recognition,” Proc. of the IEEE, Vol. 58 N° 2, February 1970, pp. 198–216. 9. H. Zahouani, PhD thesis, Besançon, 1989. 10. H. Zahouani, V. Jardret, T.G. Mathia, “Morphological characterisation of rough material,” Surface Modification Technologies, Vol. 3, Edited by T.S. Sudarshan and M. Jeandin, The Institute of Materials, Vol. 3, pp. 135–147, 1995.
11. M.S.L. Higgins, “Statistical properties of an isotropic random surface,” Philosophical Transactions of the Royal Society, Vol. 250, Series A, 1957, pp. 157–174. 12. P.R. Nayak, “Some aspects of surface roughness measurement,” Wear, Vol. 26, 1973, pp. 165–174. 13. H. Zahouani, R. Vargiolu, Ph. Humbert, “3D Morphological tree representation of the skin relief. A new approach of skin imaging characterisation,” International Federation of the Societies of Cosmetic Chemists. Paper N° 30, pp. 69–80. Cannes, France, September 14–18, 1998. 14. H. Zahouani, S-H. Lee, R. Vargiolu, “The Multi-Scale Mathematical Microscopy of Surface Roughness. Incidence in Tribology.” Lubrication at the Frontier. Elsevier Science B.V., pp 379–390, 1999. 15. H. Zahouani, M. Assoul, R. Vargiolu, T. Mathia, “The morphological tree transform of surface motifs. Incidence in tribology.” Int. J. Mach. Tools Manufact., 41: 1961–1979, 2001. 16. Ph. Humbert, M. Haftek, P. Creidi, C. Lapiere, A. Richard, A. Rougier, H. Zahouani. “Topical ascorbic acid on photoaged skin. Clinical, topographical and ultrastructural evaluation: double-blind study vs. placebo.” Exp. Dermatol., 12:237–244, 2003.
of Methodologies for 25 Comparison Evaluation of Skin Surface Contour and Wrinkles: Advantages and Limitations Motoji Takahashi and Motoki Oguri Shiseido Research Center, Yokohama-shi, Japan
CONTENTS 25.1 Introduction............................................................................................................................................................205 25.2 General Description of the Instruments to Measure Skin Surface Morphology .................................................206 25.2.1 Skin Surface Replicas and Image Analysis ..............................................................................................206 25.2.1.1 Skin Surface Contour Measurement1 .........................................................................................206 25.2.1.2 Wrinkle Measurement by the Shadowing Method2...................................................................207 25.2.2 Optical Profilometry Using Skin Surface Replicas and Three-Dimensional Analysis ............................208 25.2.3 In Vivo Measurement (PRIMOS7 and FOITS8) ........................................................................................208 25.2.4 Parameters for Skin Surface Contour in Three-Dimensional Analysis....................................................208 25.2.5 Parameters for Wrinkles in Three-Dimensional Analysis ........................................................................209 25.3 Comparison of Image Analysis and Optical Profilometry....................................................................................209 25.3.1 Skin Surface Contour ................................................................................................................................209 25.3.2 Wrinkles.....................................................................................................................................................209 25.4 Comparison of Replica and In Vivo Measurement ...............................................................................................211 25.5 Advantages and Limitations of Each Method ......................................................................................................211 25.6 Conclusion .............................................................................................................................................................212 References .......................................................................................................................................................................212
25.1 INTRODUCTION In the past, different methods and instruments have been developed to measure skin surface contour or wrinkles. The most common technique is the replica method, where a silicon imprint of the skin is measured with mechanical tactile or optical devices or by image analysis instead of the living skin. Recently, in vivo measurement without taking replicas, which means direct measurement on the skin of the living person, has been developed. Image analysis of a replica has been widely used because of the low cost of instruments and simple manipulation, though it can scarcely measure the exact peak height of skin surface
topography. Optical profilometory using replicas can get precisely three-dimensional data in sufficient speed to compare with tactile data. On the other hand, in vivo measurement has an advantage to escape from experimental artifacts when taking a replica and making possible repeated measurement at the same area on the skin. However, this method has a disadvantage that the results are not a little affected by pulsatory motion or by body movement. This chapter describes a comparison of image analysis of skin surface replica, optical profilometry using a replica, and the in vivo method from the standpoints of accuracy, reproducibility, handling, and so on.
205
206
Handbook of Non-Invasive Methods and the Skin, Second Edition
25.2 GENERAL DESCRIPTION OF THE INSTRUMENTS TO MEASURE SKIN SURFACE MORPHOLOGY 25.2.1 SKIN SURFACE REPLICAS AND IMAGE ANALYSIS 25.2.1.1 Skin Surface Contour Measurement
1
In measurement of skin furrows using negative skin replicas and image analysis, the replicas are illuminated from three directions with 120° between each pair of directions in a light-tight box to reduce light other than that from the controlled light source. In skin surface contour measurement it is necessary to illuminate the replica evenly by light because the pattern of skin furrows running in every direction is very crucial, unlike in wrinkle measurement. The images of the replicas are taken by digital camera. The gray level (levels of brightness) at each pixel is obtained, and the binary images are prepared by classifying each pixel with a gray level above and below a certain level (slice level) as black or white. The binary image is reduced and linearized to extract the characteristics of skin furrows. The thinned image is obtained by substituting a broken string of black dots in the horizontal direction for its central point, and the linearized image is obtained by substituting fine lines in the thinned image for vertical, horizontal, and 45° diagonal lines (Figure 25.1). Various parameters are defined as shown in
TABLE 25.1 Parameters Used for Image Analysis of Skin Surface Replica1 Index KSD
RMAX
VC1
VC2
AVSD16
ALL
KOSU
LEN
NUM WD (a)
(c)
(b)
(d)
FIGURE 25.1 Extraction of characteristics of skin surface pattern from negative replica image. (a) Camera image of skin relief; (b) binary image; (c) thinned image; (d) linearized image.1
Definition Standard deviation of gray level value at each pixel in camera image Ratio between the maximum and minimum values of average run length of black dots in four directions (vertical, horizontal and 45° diagonal), each in binary image Variation coefficient of number of black dots of each 13 × 13 mesh composing binary image Variation coefficient of number of white dots of each of 13 × 13 mesh composing binary image Standard deviation of number of black dots in each of 4 × 4 meshes composing binary image Total number of black dots in binary image Number of meshes with a proportion of black dots above 45% in binary image divided into 13 × 13 meshes Average length of a straight line between the points of intersection in linearized image Number of straight lines in linearized image Ratio of dots between binary and thinned image
Meaning Average of skin roughness Anisotropy of skin furrows
Anisotropy of skin furrows
Anisotropy of skin ridges
Size of hair follicles
Proportion of hair follicles and skin furrows Number of larger hair follicles
Average length of skin furrows
Number of skin furrows Average width of skin furrows
Table 25.1 and calculated for these images to extract their geometric characteristics. Using KSD (average of skin roughness: standard deviation of gray level at each pixel in the image) and VC1 (anisotropy of skin furrows: variation coefficient of number of black dots in each 13 × 13 mesh composing the binary image), age-related changes in the skin surface contour at cheek sites for 295 Japanese females ranging from 3 to 65 years of age were examined.1 The typical images for each generation are shown in Figure 25.2. These parameters can describe the change in skin surface pattern with aging (Figure 25.3 and Figure 25.4); VC1, especially, could detect subtle age-related changes in skin surface pattern. VC1 and KSD can also detect the
Comparison of Methodologies for Evaluation of Skin Surface Contour and Wrinkles: Advantages and Limitations 207
9Y
15Y
25Y
33Y
42Y
55Y
62Y
FIGURE 25.2 Change in skin surface pattern at cheek site with aging.1
8.0
0.6 n.s.
0.5
n.s. ∗∗
∗∗
6.5
∗∗
∗∗
0.3
∗∗: p < 0.01 (compared with 3–9Y) Mean ± S.E.
0.2 ∗∗: p < 0.01 (compared with 3–9Y)
6.0 5.5
0
10
20
30 40 Age (years)
50
60
∗∗
∗∗
0.4
7.0
VC 1
KSD
7.5
n.s.
∗∗
∗∗
Mean ± S.E.
n.s.
0.1 70
0
0
10
20
30 40 Age (years)
50
60
70
FIGURE 25.3 Change in surface roughness at cheek site with aging.1
FIGURE 25.4 Change in anisotropy of skin surface furrows at cheek site with aging.1
change in skin surface contour caused by skin hydration cream applied to a cheek with dry skin. After daily treatment with cream for 4 weeks, KSD increased from 8.83 to 11.0 and VC1 decreased from 0.45 to 0.39, as shown in Figure 25.5. Both parameters are very useful in the efficacy test of cosmetics to examine skin surface contour changes.
perfectly flat. The peak height and area of shadows created by the oblique light impinging on the replica are analyzed to examine the characteristics of wrinkles. The most important operation is the extracting of shadows recognized as wrinkles, because the peak heights of shadows depend on the threshold of gray level to be used. Recently, the calibration method using a standard scale for wrinkle measurement was proposed for obtaining exact peak height.3 However, the shadowing method has a few intrinsic disadvantages: small wrinkles are hidden by adjacent large ones, and the peak height of wrinkles changes depending on sample (replica) tilt. Many parameters are calculated in wrinkle analysis; the following ones are generally used in the shadowing method: the fraction of
25.2.1.2
Wrinkle Measurement by the Shadowing Method2
A negative replica is illuminated by oblique light at a precisely defined angle relative to the plane of observation in the wrinkle measurement. The replica must be kept
208
Handbook of Non-Invasive Methods and the Skin, Second Edition
based on a touch-free fringe projection, and PRIMOS, based on active triangulation in conjunction with the phase-shift technique. Both make it possible to identify and readjust previously recorded skin areas at video rate, and can measure skin surface configuration of the human body in three dimensions. Because of avoiding the need to make replicas, they are free from unavoidable loss of information connected with the replica technique.
25.2.4 PARAMETERS FOR SKIN SURFACE CONTOUR IN THREE-DIMENSIONAL ANALYSIS
Before VC 1: 0.45 KSD: 8.83
There are many two-dimensional parameters to be used in the field of surface metrology, such as Ra (roughness average), Rz (10-point height), Ry (maxium height of the profile), and Sm (mean spacing of profile irregularities).9 In order to describe the skin surface contour, these parameters are extended to surfaces, i.e., three dimensions, and defined as SRa, SRz, SRy, and SSm, respectively. 1 l f ( x ) dx : Arithmetical mean deviation l 0 of profile. f(x) is the height of the assessed profile.
Ra = After VC 1: 0.39 KSD: 11.0 (42, female, cheek)
1 l 2 l1 f ( x, y ) dxdy : Ra extended l 1l 2 0 0 to three dimensions. f(x,y) is the height of the assessed profile.
SRa (1) =
FIGURE 25.5 Change in skin surface microtopography after application of hydration cream for 4 weeks.1
shadows in a defined area (WA), the average peak height of all wrinkles (WH), and the highest peak in wrinkles (HP).
25.2.2 OPTICAL PROFILOMETRY USING SKIN SURFACE REPLICAS AND THREE-DIMENSIONAL ANALYSIS Currently, instead of a mechanical sensor or stylus in direct contact with the surface to be analyzed (mechanical profilometry), a noncontact optical sensor is frequently used (optical profilometry), whose principles are focusing of the laser beam4 (confocal scanning laser microscopic method) and optical triangulation (slit projection method5 or pattern projection method6). These methods have an advantage in that they can quickly acquire data and analyze skin topography in three dimensions.
25.2.3 IN VIVO MEASUREMENT (PRIMOS7 FOITS8)
∫
∫ ∫
Ra 0 + Ra 45 + Ra 90 + Ra 135 :Ra0,Ra45, Ra90, 4 and Ra135 are defined as Ra in 0, 45, 90, and 135° directional lines, respectively (Figure 25.6).
SRa (2 ) =
90° 135° 45°
0°
AND
Nowadays, nontouching systems for a rapid in vivo measurement of skin topography are developed and commercially available. They are FOITS, whose technique is
FIGURE 25.6 Ra, Rz, Ry, and Sm are calculated in four different directions (0, 45, 90, and 135°) for extension to three dimension. Three-dimensional parameters (SRa, SRz, SRy, and SSm) are obtained by averaging the values in four directions, respectively.
Comparison of Methodologies for Evaluation of Skin Surface Contour and Wrinkles: Advantages and Limitations 209
5
Rz =
5
∑
ypi +
i =1
∑y
vi
i =1
: Average of the height of 5 the five highest peaks (ypi) plus the five deepest valleys (yvi) over the evaluation length.
Rz 0 + Rz 45 + Rz 90 + Rz135 :Rz0, Rz45, Rz90, and 4 Rz135 are defined as Rz in 0, 45, 90, and 135° directional lines, respectively. Ry = (Maximum of |ypi |) + (maximum of |yvi |): the maximum peak to lowest valley vertical distance over the evaluation length. SRz =
Ry 0 + Ry 45 + Ry 90 + Ry 135 :Ry 0 , Ry 45 , Ry 90 , 4 and Ry135 are defined as Ry in 0, 45, 90, and 135° directional lines, respectively.
TABLE 25.2 Correlation Coefficient with Two-Dimensional (VC1 and KSD) and Three-Dimensional (SRa, SRz, SRy, and SSm) Parameters VC1 VC1 KSD SRa(1) SRa(2) SRz SRy SSm
KSD
SRa(1) SRa(2)
SRz
SRy
SSm
1 –0.499 –0.215 –0.209 –0.254 –0.318 0.207 –0.499 1 0.651 0.647 0.746 0.773 –0.018 –0.215 0.651 1 0.964 0.892 0.904 0.324 –0.209 0.647 0.964 1 0.929 0.936 0.275 –0.254 0.746 0.892 0.929 1 0.961 0.240 –0.318 0.773 0.904 0.936 0.961 1 0.089 0.207 –0.018 0.324 0.275 0.240 0.089 1
Note: N = 58.
SRy =
1 Sm = N
N
∑ Si : The mean spacing between peaks. i =1
A peak must cross above the mean line and then back below it. Si is the width of each peak. Sm 0 + Sm 45 + Sm 90 + Sm 135 : Sm0,Sm45, 4 Sm90, and Sm135 are defined as Sm in 0, 45, 90, and 135° directional lines, respectively.
SSm =
25.2.5 PARAMETERS FOR WRINKLES DIMENSIONAL ANALYSIS
IN
THREE-
The total volume of all wrinkles in a defined area (WV) and the profile length ratio (Lr)5 are used for quantification of wrinkles other than WA, WH, and HP, described in Section 25.2.1.2. Usually the aim of this kind of study is to test the efficacy of antiaging products, and the investigation is therefore based on the comparison of two types of results: the evolution of a wrinkle before treatment (the reference) and the evolution of another wrinkle after treatment. To be reliable, this comparison study must be conducted on the same skin surface (same size and same position).
25.3 COMPARISON OF IMAGE ANALYSIS AND OPTICAL PROFILOMETRY 25.3.1 SKIN SURFACE CONTOUR Correlation coefficients with the parameters used in image analysis (KSD and VC1) and those in optical profilometry (SRa, SRz, SRy, and SSm) are shown in Table 25.2.
Because three-dimensional parameters obtained by profilometry are originally defined as those showing surface roughness of the industrial materials, KSD, which is the skin surface roughness parameter, correlates well with SRa, SRz, and SRy. On the other hand, VC1, which shows anisotropy of skin furrows, does not correlate with these three-dimensional parameters. From this result KSD obtained by image analysis seems to be a good substitutive parameter for SRa, SRz, and SRy in three-dimensional analysis. Considering that VC1 can detect subtle agerelated changes in anisotropy of skin furrows and has no relationship with these three-dimensional parameters generally used in industry, VC1 is a unique parameter to describe skin surface contour.
25.3.2 WRINKLES The relationship between image analysis and optical profilometry in WA and WH is shown in Figure 25.7. There is a good correspondence between them in WA, but not in WH, though the correlation coefficient is a little higher. As WH is an average peak height of all wrinkles, the discrepancy between image analysis and profilometry depends on the difference of methodology. Typical cases of discrepancy between them in peak height of wrinkles are shown in Figure 25.8 and Figure 25.9. Figure 25.8 shows an example that a peak height obtained by image analysis is lower than that obtained by optical profilomtry, as the average height of roughness profile following the analyzed wrinkle is higher than the reference line. On the other hand, Figure 25.9 is an opposite case, where the peak height is higher in image analysis than in profilometry. It is considered that two-dimensional parameters like WA (the fraction of wrinkles in a defined area) have a good correspondence of image analysis to profilometry, but not three-dimensional parameters like WH.
210
Handbook of Non-Invasive Methods and the Skin, Second Edition
WH (mm)
30
0.3
Image analysis
Image analysis
WA (%)
20
0.2
0.1
10
r = 0.731, n = 31
r = 0.639, n = 31 0
20 10 Optical profilometory (a)
0
30
(%)
0
0
0.1 0.2 Optical profilometory (b)
0.3 (mm)
FIGURE 25.7 Relationship between image analysis by the shadowing method and optical profilimetry in wrinkle parameters WA (fraction of wrinkle area) and WH (average peak height of wrinkles).
10
10
600
300
500 200
8
100
100 0
7 0
2
4
6
8
0 −100
10 6
−100 −200
μm
200
mm 5
−300
Y-direction (mm)
μm
300
9
9
8 0
2
4
6
8
10
7
mm 6
−200 −300
5
4 4 3
Y-direction (mm)
400
25° 3
25° 2
2 1 1 0 500 0 −500 μm
FIGURE 25.8 Example showing that peak height of wrinkles obtained by image analysis is lower than that by profilometry.
0 500 0 −500 μm
FIGURE 25.9 Example showing that peak height of wrinkles obtained by image analysis is higher than that by profilometry.
Comparison of Methodologies for Evaluation of Skin Surface Contour and Wrinkles: Advantages and Limitations 211
Optical profilometry by confocal scanning laser microscope (HD 100D, Laser tech. Corp., Japan )
In vivo measurement by PRIMOS (reversed)
FIGURE 25.10 Comparison of in vivo measurement and optical profilometer with replica from the same skin area on the ventral side of the forearm. The result of the in vivo measurement is reversed compared with that by the profilometer. The replica method with optical profilometer shows finer and sharper skin surface pattern. The measuring area is 10 × 10 mm. OpenGL10 representation mode.
25.4 COMPARISON OF REPLICA AND IN VIVO MEASUREMENT According to a previous paper,7 the difference between in vivo and replica measurement is noticeable. For example, the two-dimensional roughness parameter Rq (root mean square) of the in vivo results measured by PRIMOS was 10 to 30% higher than the corresponding value measured by the replica obtained from the same part of the human body.7 Authors also reported that the measured texture of replica at the back of the hand was blurred, whereas the in vivo result showed finer and sharper structures. From these results they concluded that in vivo measurement is superior to the replica method. We also measured the skin surface contour of the ventral forearm and crow’s-feet at exactly the same area using replica and in vivo (PRIMOS) methods. Results are shown in Figure 25.10 with OpenGL
representation mode and somewhat different from the previous result.7 In our result profile, the in vivo measurement has more noises than that of the replica method, especially in forearm measurement, where the texture is finer and shallower than eye wrinkles. It is considered that pulsatory motion or body movement still affects the result of in vivo measurement.
25.5 ADVANTAGES AND LIMITATIONS OF EACH METHOD Advantages and disadvantages of image analysis of skin replicas, three-dimensional analysis of replicas with optical profilometry, and in vivo measurement are summarized in Table 25.3. In the replica technique the impression materials must penetrate into all ends of furrows and
TABLE 25.3 Comparison of Methodologies for Evaluation of Skin Surface Configuration Method
Instruments
Advantages
Disadvantages
Image analysis of replica
Combination of light source and digital camera
Optical profilometry
Optical focus profilometer Confocal scanning laser microscope Triangulation optical profilometer FOITS PRIMOS
Low cost of instrument Simple manipulation Spatially analyze orientation of skin texture High accuracy Fast analysis Reproducible Not necessary to take replica Fast analysis Repeated measurement at the same area
Artifact when making replica Not obtain exact peak height Lose lower peaks behind higher peak Artifact when making replica
In vivo measurement
Affected by body movement
212
Handbook of Non-Invasive Methods and the Skin, Second Edition
wrinkles to reproduce all details of skin topography and a short hardening time is demanded. Therefore, the skill of taking a replica influences experimental results. On the other hand, in in vivo measurement making a replica is not necessary, but data acquisition must be completed in a very short time to avoid the affect of self-movement of the living human body.
25.6 CONCLUSION The methodology for measurement of skin surface topography has been developed ranging from image analysis of replicas to the in vivo technique. Each method has advantages and limitations. Among them, in vivo measurement is quite promising, because it avoids the skin hydration effect induced by replica impression materials and the morphological changes in skin surface induced by hydration. However, this technique is slightly influenced by pulsatory motion or self-movement of the human body. Further improvement of the instrument is needed. Technical development for collecting data within 1/10 of a second would be a pressing matter to precisely analyze skin surface contour in direct measurement without taking a replica.
REFERENCES 1. Takahashi, M., Image analysis of skin surface contour, Acta Derm. Venereol. (Stockh.), Suppl.185, 9–14, 1994.
2. Hayashi, S., Matsuki, T., Matsue, K., Arai, S., Fukuda, Y., and Yoneya, T., Changes of facial wrinkles by aging, sunlight exposure and application of cosmetics, J. Soc. Cosmet. Chem. Japan, 27, 355, 1993. 3. Tamura, K., Okano, Y., Okamura, H., and Masaki, H., Quantitative method of wrinkle depth using standard scale, J. Soc. Cosmet. Chem. Japan, 35, 50, 2001. 4. Mignot, J., Three-dimensional evaluation of skin surface: micro- and macrorelief, in Handbook of Noninvasive Methods and the Skin, Serup, J. and Jemec, G.B.E., Eds., CRC Press, Boca Raton, FL, 1995, p. 107. 5. Takasu, E., Umeya, Y., and Horii, I., Development of wrinkle analyzing system using three-dimensional curved shape measurement and the changes of Japanese women’s facial wrinkles with aging, J. Soc. Cosmet. Chem. Japan, 29, 394, 1996. 6. Jaspers, S., Hoperman, H., Sauermann, G., Hoppe, U., Lunderstadt, R., and Ennen, J., Rapid in vivo measurement of the topography of human skin by active image triangulation using a digital micromirror device, Skin Res. Technol., 5, 195, 1999. 7. Hof, C. and Hopermann, H., Comparison of Replicaand In Vivo-Measurement of the Microtopography of Human Skin, research report of the Institute of Automation of the University of the Federal Armed Forces, Hamburg, Germany, 2000. 8. Rohr, M., Brandt, M., and Schrader, A., Skin surface: claim support by FOITS, SOFW J., 126, 2, 2000. 9. ISO 4287:1997, Surface texture: Profile method (Surface Metrology Guide, http://www.predev.com/smg/ parameters.htm). 10. OpenGL — The Industry’s Foundation for High Performance Graphics. http://www.opengl.org/.
Skin Surface Friction
Studies on Skin: 26 Tribological Measurement of the Coefficient of Friction* Raja K. Sivamani, Gabriel Wu, and Howard I. Maibach Department of Dermatology, School of Medicine, University of California, San Francisco, California
Norm V. Gitis Center for Tribology, Inc., Campbell, California
CONTENTS 26.1 Introduction............................................................................................................................................................215 26.1.1 Experimental Designs................................................................................................................................216 26.1.2 Hydration ...................................................................................................................................................217 26.1.3 Lubricants/Emollients/Moisturizers...........................................................................................................217 26.1.4 Probes.........................................................................................................................................................217 26.1.5 Normal Load..............................................................................................................................................217 26.2 Skin Friction Coefficient Values............................................................................................................................218 26.2.1 Hydration ...................................................................................................................................................218 26.2.2 Lubricants/Emollients/Moisturizers...........................................................................................................220 26.2.2.1 Talcum Powder...........................................................................................................................220 26.2.2.2 Lubricant Oils.............................................................................................................................220 26.2.2.3 Emollients and Moisturizers ......................................................................................................220 26.2.3 Probes.........................................................................................................................................................221 26.2.4 Anatomic Region, Age, Gender, and Race ...............................................................................................222 26.3 Conclusion .............................................................................................................................................................222 References .......................................................................................................................................................................223
26.1 INTRODUCTION Mechanically, friction allows us to keep from slipping as we step out of the shower, hold Styrofoam cups of coffee in the morning, or turn the steering wheel in our cars. Because the skin is a surface itself, it is convenient to analyze and describe it in terms of a surface phenomenon like friction; friction studies on skin provide valuable insight into how the skin interacts with other surfaces. It also provides information about the skin under various conditions (e.g., age and gender) and under various chemical treatments (e.g., lotions and moisturizers). Studying the friction of skin supplements other mechanical tests. Friction studies can be conducted with
noninvasive methods and give a measure of the skin’s health — skin hydration, for example: Naylor1 showed that moistened skin has an elevated friction response, and El-Shimi2 demonstrated that drier skin has a lowered friction response. Friction provides a quantitative measurement to assess skin. The friction parameter generally measured is the coefficient of friction. To measure the friction coefficient, a surface is brought into contact with another and moved relative to it. When the two surfaces are brought into contact, the perpendicular force is defined as the normal force (N). The friction force (F) is the force that opposes relative movement between the two surfaces. From Amonton’s law, the coefficient of friction (μ) is
* Modified with permission from Skin Research and Technology, 9, 227–234, 2003.
215
216
Handbook of Non-Invasive Methods and the Skin, Second Edition
defined as the ratio of the friction force to the normal force: μ = F/N The friction coefficient can be measured in two ways: the static friction coefficient (μs) and the dynamic or kinetic friction coefficient (μk). The static friction coefficient is defined as the ratio of the force required to initiate relative movement and the normal force between the surfaces; the dynamic or kinetic friction coefficient is defined as the ratio of the friction force to the normal force when the two surfaces are moving relative to each other. For simplicity, much of the research has focused on the dynamic friction coefficients wherein the two surfaces move at a relative constant velocity. Most of the friction studies on skin have dealt with the dynamic friction coefficient, and the subscript k is usually dropped. This overview references the dynamic coefficient of friction unless otherwise noted. According to Amonton’s law, the dynamic friction coefficient remains unchanged regardless of the probe velocity or applied normal load in making the measurement. Amonton’s law holds true in the case of solids with limited elastic properties. Although Naylor1 concluded Amonton’s law was true, later studies by El-Shimi,2 Comaish and Bottoms,3 and Koudine et al.4 have found that skin deviates from Amonton’s law, since the friction coefficient increased when the normal load was decreased. El-Shimi2 and Comaish and Bottoms3 reasoned that the rise in friction coefficient resulted from the viscoelastic nature of the skin, allowing for a nonlinear deformation of the skin with increasing load.
26.1.1 EXPERIMENTAL DESIGNS Various experimental designs have been devised to measure the friction on skin. They focus on measuring friction by pressing a probe onto the skin with a known normal force, and then detecting the skin’s frictional resistance to movement of the probe. The designs fall into two categories: 1. A probe moved across the skin in a linear fashion 2. A rotating probe in contact with the skin surface In the linear designs, the probe movement is accomplished in several ways. Comaish and Bottoms3 utilized one of the simplest linear designs: they moved the probe across the skin by attaching it to a pan of weights by means of a pulley. Weights are placed in the pan such that the probe slides over the skin at a constant velocity. This allows for the calculation of the dynamic friction coefficient by dividing the total weight in the pan by the normal load on the probe.
FIGURE 26.1 UMT test setup for volar forearm. The forearm is strapped into place with a holder to immobilize the forearm. The copper probe is then brought down to carry out the friction and electrical measurement. (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
More sophisticated linear designs followed the simple design used by Comaish and Bottoms,3 providing motorized unidirectional movement of the probe or the use of a reciprocating motor to move the probe back and forth. In both designs the motorization affords greater control in maintaining the velocity of the probe. Strain gauges measure the friction force as the probe moves along the skin surface. Figure 26.1 shows a biomedical tribometer friction measurement device where the normal load and the probe speed are computer controlled. The second design category measures friction with a rotating wheel pressed onto the surface of the skin with a known normal force. Highley et al.5 measured the frictional resistance by determining the angular recoil of the instrument as the wheel contacted the skin. They measured this angular recoil by recording the proportion of light that hit a dual-element photocell. An electrical signal was then generated in proportion to the frictional resistance. Comaish et al.6 developed a portable, handheld device (Newcastle friction meter) that relied on a torsion spring to measure the skin’s frictional resistance. The devices are surveyed in Table 26.1. An important part of designing a friction measurement apparatus is choosing the probe size, shape, and material.
Tribological Studies on Skin: Measurement of the Coefficient of Friction
217
TABLE 26.1 μ) of Untreated Normal Probe and Apparatus Used to Measure the Dynamic Friction Coefficient (μ Skin In Vivo Author Naylor1 El-Shimi2 Comaish and Bottoms3 Koudine et al.4 Highley et al.5 Prall7 Cua et al.8 Johnson et al.9 Asserin et al.10 Elsner et al.11 Sivamani et al.17 Sivamani et al.24
Probe Size and Shape
Probe Material
Motion of Test Apparatus
Maintenance of Normal Load
8-mm-diameter sphere 12-mm-diameter hemisphere 15-mm-diameter annular ring Hemisphere lens
Polyethylene Stainless steel (rough), stainless steel (smooth) Teflon, nylon, polyethylene, wool Glass
Linear, reciprocating Rotational
Static weights Static weights
Linear
Static weights
Linear
Disc Disc 15-mm-diameter disc 8-mm (radius of curvature) lens 3-mm-diameter sphere 15-mm-diameter disc 10-mm-diameter sphere
Nylon Glass Teflon Glass
Rotational Rotational Rotational Linear, reciprocating
Static weights; balance beam Spring load Spring load Spring load Static weights
Ruby Teflon Stainless steel
Linear Rotational Linear
Copper
Linear
13-mm-diameter cylinder
Because friction is an interaction between two surfaces, the probe geometry and material will affect the values calculated for the friction coefficient of the other surface. Several shapes and material have been used, as outlined in Table 26.1. Also, results will be more accurate when the probe’s normal force is maintained at a constant value or continuously monitored; previous methods used to maintain the normal force include spring mechanisms or static weights to weigh down the probe (Table 26.1). These parameters are revisited critically later in this article. Much effort has been spent in understanding how skin friction changes with differing biological conditions and upon the application of various products to the skin surface. These studies have been of interest to various industries that manufacture products meant as skin topical agents because friction measurements can provide clues regarding the effectiveness of their products.
26.1.2 HYDRATION Hydration is a complex phenomenon influenced by intrinsic (i.e., age, anatomical site) and extrinsic (i.e., ambient humidity, chemical exposure) factors. These factors can affect the mechanical properties of the skin, and research has been performed to correlate hydration levels with the skin’s friction coefficient.24 Hydration studies have investigated how increases and decreases in skin hydration correlated with the friction coefficient. In past studies, researchers generally induced increases in skin hydration through water exposure. However, decreases in skin
Balloon; static weights Spring load Computer servofeedback control Computer servofeedback control
hydration were not experimentally induced and dehydration studies were performed between subjects with normal skin and subjects that had clinically dry skin.2,12
26.1.3 LUBRICANTS/EMOLLIENTS/MOISTURIZERS Much of the reviewed research has been devoted to ascertaining how the application of certain ingredients influences the skin surface, of interest to the cosmetic/moisturizer and lubricant industries. The studies focused on the effects of talcum powder,2,3 oils,2,3,5,14 and skin creams/moisturizers.7,14,17,24 Hills et al.15 analyzed how changes in the friction coefficient, following emollient application, differed with temperature.
26.1.4 PROBES As mentioned earlier, the probe geometry and material influence the measured value of the friction coefficient because friction is a probe–skin interaction phenomenon. Few studies have examined probe effects: El-Shimi2 studied probe roughness and Comaish and Bottoms3 probe roughness and material.
26.1.5 NORMAL LOAD Friction measurements can offer quantitative insight into changes on the skin surface, and the UMT offers technical advances over existing friction measurements. The control of the probe speed and the real-time monitoring of the normal load allow for real-time calculation of the friction
218
Handbook of Non-Invasive Methods and the Skin, Second Edition
Friction coefficient vs. normal load (50 g sensor)
TABLE 26.2 Reported Values of the Dynamic Friction μ) for Untreated Normal Skin Coefficient (μ In Vivo
1.2
0.8 μ
Author 0.4
Naylor1 El-Shimi2
0.0 0
5
10
15
20 25 30 Load (grams )
35
40
45
50
FIGURE 26.2 Friction coefficient vs. normal load. The friction coefficient increased as the normal load was decreased, suggesting that the skin does not follow Amonton’s law. The probe was moved at 5 mm min–1 (n = 4). (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
coefficient. As seen in Figure 26.2, the control of the load is important because the friction coefficient does not adhere to Amonton’s law. Wolfram18 theoretically deduced that the friction coefficient would relate to the normal load as follows: μ ∝ N–1/3 where N is the applied normal load to the skin. Sivamani et al.17 found that the friction coefficient related to the normal load as follows: μ ∝ N–0.32 and Koudine et al.4 found the dependence on the applied normal load to be μ ∝ N–0.28
26.2 SKIN FRICTION COEFFICIENT VALUES Friction is an important characteristic of skin because it allows us to execute many of our daily activities. In addition, friction studies offer insight into how skin and the skin surface change across age, gender, race, anatomical site, and chemical applications. This can provide better information about expected skin variations in the population and why certain topical applications are effective. Comparative studies are particularly useful in following how the skin mechanically changes under different conditions. Previous studies have reported various values for the skin’s friction coefficient. Dynamic friction coefficient measurements (Table 26.2) fall in the range 0.12 to 0.7; however, most fall in a narrower range of 0.2 to 0.5 (Figure 26.3). Besides natural variations in skin, the wide range
Comaish and Bottoms3
Koudine et al.4 Highley et al.5 Prall7 Cua et al.8
Johnson et al.9 Asserin et al.10 Elsner et al.11
m 0.5–0.6 0.2–0.4 (stainless steel, rough) 0.3–0.6 (stainless steel, smooth) 0.2 (Teflon) 0.45 (nylon) 0.3 (Polyethylene) 0.4 (wool) 0.24 (dorsal forearm) 0.64 (volar forearm) 0.2–0.3 0.4 0.34 (forehead) 0.26 (volar forearm) 0.21 (palm) 0.12 (abdomen) 0.25 (upper back) 0.3–0.4 0.7 0.48 (forearm) 0.66 (vulva)
in results may be due to differences in probe movement, geometry, controlled monitoring of the normal force, and material chosen to make the friction measurement. In designing the friction measurement apparatus, the two types of probe movement utilized were rotational and linear (Table 26.1). As a result, the linear probe constantly moves over untested skin and the rotational probe spins over tested skin. This can lead to discrepancies in reported values for the skin friction coefficient. Another important source of variation may be the ability to control the normal force while the probe is testing the skin surface. The skin friction instruments are designed to measure the frictional resistance of the skin, and it is assumed that the normal force is constant. During a test the normal force may not stay constant as a result of an uneven skin surface, inaccurate spring, or a nonuniform distribution of static weights placed above the probe head. Therefore, the assumption of a constant normal force may be incorrect and can lead to variation in the calculated friction coefficient. A third source for variation is the choice of the probe material. Because friction is a surface phenomenon between two materials, the choice of the probe influences the numerical value obtained for the friction coefficient.
26.2.1 HYDRATION Hydration studies revealed that drier skin had lowered friction while hydrated skin had an increased amount of
Tribological Studies on Skin: Measurement of the Coefficient of Friction
219
Ranges in the dynamic coefficient of friction measurements Naylor (1) El-shimi (2) Comaish bottoms (3) Koudine et al (4) Highley er al (5) Prall (7) Cua et al (8) Johnson et al (9) Asserin et al (10) Elsner et al (11) Sivamani et al (17) 0
0 .1
0 .2
0 .3 0 .4 0 .5 μ (dynamic coefficient of friction)
0 .6
0 .7
0 .8
FIGURE 26.3 Outline of the ranges in the dynamic coefficient of friction. These ranges reflect measurement of untreated normal skin friction in vivo. (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 227–234, 2003.)
friction (Table 26.3). However, the skin response is more complex, because very wet skin also has a lowered friction coefficient, much like the characteristics of dry skin.16 Most studies focus on an intermediate zone of hydration where the skin has been moistened without an appreciable slippery layer of water on the skin. Results in Table 26.3 show that the increases in friction were varied, and this possibly results from the various probes used. Although the addition of water increases the friction coefficient, this effect lasts for a period of minutes before the skin returns to its normal state.2,5,14,17 The water has an effect of softening the skin, and this in turn allows for a greater contact area between the probe and the skin. Also, water results
in adhesive forces between the water and the probe. Thus, there is more frictional resistance between the skin and the probe, resulting in a higher friction coefficient.18 Since the water evaporates in minutes, the skin returns to its normal state in the same time frame. For dry skin, the skin becomes less supple and the probe does not achieve as much contact area; this allows the probe to glide more easily over the skin surface. This results in a lowered friction coefficient, as seen in the isopropyl study17 and in prior studies involving subjects with clinically dry skin.2,12 The agreement between the experimentally induced dry skin and clinical dry skin is expected.18
TABLE 26.3 μ) with Increasing Comparative Studies of the Changes in Dynamic Friction Coefficient (μ Hydration (Hydration) and Decreasing Hydration (Dryness) Author Naylor1 El-Shimi2 Comaish and Bottoms3 Highley et al.5 Prall7 Johnson et al.9 Lodén et al.12 Nacht et al.14 Sivamani et al.17 a
Probe Material Polyethylene Stainless steel (rough), stainless steel (smooth) Wool, Teflon Nylon Glass Glass Stainless steel Teflon Stainless steel
% Increase Due to Hydration μmoist – μnormal)/μ μnormal} × 100 {(μ
% Decrease Due to Dryness μnormal – μdry)/μ μnormal} × 100 {(μ
80 100–200 (stainless steel rough)
— 28 (stainless steel, rough), 41 (stainless steel, smooth) — — —
40a (wool), 400a (Teflon) 500 200 100–233 — 45 55 (in vitro)
33 (hand), 41 (back), 14 (arm) — 10 (in vivo)
Comaish and Bottoms studied the change in the static friction coefficient in their hydration study.
220
Handbook of Non-Invasive Methods and the Skin, Second Edition
% Change in friction ceofficient
Pertrolatum Heavy mineral oil
Glycerin Baseline
50 40 30 20 10 0 −10 −20 −30
∗Time
−1
0
1
2 3 Time after application (hrs)
4
5
6
= −1 is immediately prior to application; Time = 0 is immediately after application
FIGURE 26.4 Effect of lubricant cosmetic ingredient on skin friction coefficient. Amount applied of each material: approximately 2 mg/cm2 (mean of five subjects, but p value was not published). Time = –1 is immediately prior to application; time = 0 is immediately after application. (Adapted from Nacht, S. et al., J. Soc. Cosmet. Chem., 32, 55–65, 1981.)
The studies on lubricants, emollients, and moisturizers are important for cosmetics and products developed to make the skin look and feel healthier. The literature reports that the important qualitative characteristics in skin topical agents are skin smoothness, greasiness, and moisturization.17,19 Previous research has tried to describe these subjective, qualitative descriptions in a quantitative fashion by correlating them against the friction coefficient. Prall7 tried to find a quantitative correlation for skin smoothness but was unable to make a direct correlation to the friction coefficient until he added skin topography and hardness into the analysis. Nacht et al.14 found a linear correlation between perceived greasiness and the friction coefficient (Figure 26.4). 26.2.2.1 Talcum Powder El-Shimi2 and Comaish and Bottoms3 showed that the friction coefficient decreased after the application of powder. El-Shimi2 found the friction coefficient to decrease by 50% after application; Comaish and Bottoms,3 in analyzing the static friction coefficient, observed an insignificant change for a wool probe and a 30% decrease in friction with a polyethylene probe. However, they also found that wetting the talc powder caused an increase in the measured friction. 26.2.2.2 Lubricant Oils A lowering in the friction coefficient is the initial effect after the application of oils and oil-based lubricants.2,5,14 Nacht et al.14 and Highley et al.5 also showed that after the initial decrease in friction, the oils eventually raised the
% Change in friction coefficient
26.2.2 LUBRICANTS/EMOLLIENTS/MOISTURIZERS
120 100 80 60 40 20 0 −20 −40
0 (Not greasy)
A B
C
D
1
2
E
F 3
4
5 (Very greasy)
Mean score of greasiness
FIGURE 26.5 Correlation between changes in the friction coefficient and the sensory perception of greasiness. A to F represent different creams that were applied to the skin. The reported percent change in the friction coefficient is immediately after application, and the greasiness scores were subjective evaluations. (Adapted from Nacht, S. et al., J. Soc. Cosmet. Chem., 32, 55–65, 1981.)
skin’s friction coefficient. The results of the lubricant cosmetic studies by Nacht et al.14 are shown in Figure 26.5. 26.2.2.3 Emollients and Moisturizers Prall7 and Nacht et al.14 found that the friction coefficient rises with the addition of emollients and creams in a similar fashion to water. However, the effects of the creams lasted for hours, while the water effects lasted for about 5 to 20 minutes.7,14,17 Sivamani et al.17 quantified the friction, greasiness, and stickiness of the skin following application of creams and treatments (Figure 26.6 to Figure 26.8). Hills et al.15 also studied emollients, but they examined how different emollients compared against one another and how changes in temperature changed the friction coefficient. At a higher temperature (45˚C), most emollients lowered the friction coefficient to a greater degree than at a lower temperature (18˚C).
Tribological Studies on Skin: Measurement of the Coefficient of Friction
2.5
Coefficient of friction
Mean 1.5
Amplitude
1
Amplitude/Mean
0.25
2
0
10
20
30 Time (sec)
40
50
60
FIGURE 26.6 Calculation of the amplitude/mean measurement. The mean refers to the mean value of the measured friction coefficient as indicated on the graph. The amplitude refers to the deviation seen during the friction coefficient measurement as indicated on the graph. Then the amplitude is divided by the mean to calculate the amplitude/mean. It has been suggested that this value represents the smoothness of the skin surface.4,17 (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
0.1 0.05
•
Amplitude/Mean
0.1
0.05
• Distal left forearm
Distal right Proximal right forearm forearm
All
FIGURE 26.7 Amplitude/mean on the untreated volar forearm. No significant differences were found for different anatomical sites between the left and right volar forearms or between distal and proximal sites on the same volar forearm. (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
When lubricant/moisturizers are applied to the skin, the skin friction is affected in three general ways:14,18 A large, immediate increase in the friction coefficient, similar to water application, that follows with a slow decrease in the friction coefficient. These agents can be interpreted to act by immediate hydration of the skin through some aqueous means to give the immediate increase in friction. In Figure 26.4, creams A, B, and C represent this type of lubricant/moisturizer.
∗
Petrolatum
Glycerin Occlusion Intervention
Untreated
FIGURE 26.8 Amplitude/mean measurements for interventions. The application of glycerin and the PVDC occlusion increased the amplitude/mean of the volar forearm. Also, the addition of glycerin raised the amplitude/mean significantly more than the PVDC occlusion. Petrolatum significantly decreased the amplitude/mean, and this is quantitative evidence of petrolatum’s greasiness (p < 0.001). (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
0.15
0
∗
0.15
0.5
0
∗
0.2
0
•
221
An initial decrease in the friction coefficient that is followed by an overall increase in the friction coefficient over time. These agents are fairly greasy products (Figure 26.5), and this greasiness causes the immediate decrease in the friction coefficient. The eventual rise in the friction coefficient is probably due to the occlusive effects of these agents.21 In other words, these products and ingredients act to prevent water loss from the skin, thereby increasing the hydration of the skin. Representations of a few ingredients that elicit this response are in Figure 26.5 and are represented as cream F in Figure 26.4. A small, immediate increase in the friction coefficient that then increases slowly with time. These agents are interpreted to act as a combination of effects seen in the previous two cases. These lubricants/moisturizers have ingredients and agents that serve to both hydrate the skin through some aqueous method and prevent water loss through some occlusive mechanism. Because of the presence of these occlusive agents, which tend to be more slippery, the immediate rise in the friction coefficient is lower than in products that fall into the first category listed above. In Figure 26.4, this is seen in creams D and E.
26.2.3 PROBES El-Shimi2 and Comaish and Bottoms3 compared probes (Table 26.2 and Table 26.3) and found that smoother probes gave higher friction coefficient measurements. ElShimi2 noted that higher friction coefficient measurements were made with a smoother stainless-steel probe than with
222
Handbook of Non-Invasive Methods and the Skin, Second Edition
a roughened stainless-steel probe. Comaish and Bottoms3 found a similar result with two types of nylon probes: a sheet probe and a knitted probe. The sheet probe (the smoother of the two) gave a higher friction coefficient measurement. El-Shimi2 postulates that the smoother probe forms more contact points with the skin and has a greater skin contact area than the rougher probe, resulting in more resistance from the skin and a larger measurement for the friction coefficient.
26.2.4 ANATOMIC REGION, AGE, GENDER, RACE
AND
Few studies address the effects of anatomic region, age, gender, or race as they pertain to the friction coefficient. To date, no significant differences have been found with regard to gender8,22,24 or race.23,24 The friction coefficient varies with anatomical site: Cua et al.8,22 found that friction coefficients varied from 0.12 on the abdomen to 0.34 on the forehead. Elsner et al.11 measured the vulvar friction coefficient at 0.66, whereas the forearm friction coefficient was 0.48. Sivamani et al.24 found that the proximal volar forearm had a higher friction coefficient than the distal volar forearm (Figure 26.9). Manuskiatti et al.23 studied skin roughness and found significant differences in skin roughness at various anatomical sites. Differences in environmental influences (e.g., sun exposure) and hydration may account for this. Elsner et al.11 showed that the more-hydrated vulvar skin had a 35% higher friction coefficient than the forearm, and this is in agreement with hydration studies that contend that skin has an increased friction coefficient under increased hydration. Age-related studies have been contradictory, where some authors found no difference8,22,24 (Figure 26.10) and others found differences.10,11 Cua et al.22 showed no differences in friction with respect to age except for friction measurements on the ankle. Elsner et al.11 also performed age-related tests and found no differences in the vulvar friction coefficient, but observed a higher forearm friction coefficient in younger subjects. They postulate that the skin on parts of the body that become exposed to sunlight can undergo photoaging, and thus forearm skin shows evidence of age-related differences while the light-protected vulvar skin does not.11 Asserin et al.10 concluded that younger subjects had a higher forearm friction coefficient than older subjects. There are few gender-related and racial friction studies. Cua et al.8,22 and Sivamani et al.24 found no significant friction differences between the genders. Manuskiatti et al.23 found no significant racial (black and white skin) differences in skin roughness and scaliness. Sivamani et al.24 found no differences in volar forearm friction among different ethnicities before and after chemical treatments (Figure 26.11 and Figure 26.12).
1 0.8 0.6 0.4 0.2 0
300 250 200 150 100 50 0
Coefficient of friction (untreated- all volunteers) ∗ ∗
Distal left forearm
Distal right arm
Proximal right arm
Electrical impedance (untreated- all volunteers) ∗ ∗
Distal left forearm
Distal right arm
Proximal right arm
FIGURE 26.9 Friction coefficient and electrical impedance. There were no significant differences between the distal left volar forearm and the distal right volar forearm. The proximal right volar forearm had a significantly higher friction coefficient and a significantly lower electrical impedance than the distal right volar forearm, and the proximal right arm friction and electrical impedance measurements were different from those of the distal right arm (p < 0.001). (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
COF between age groups (untreated skin) Young Old 1 0.8 0.6 0.4 0.2 0
∗ ∗
Distal left forearm
Distal right arm
Proximal right arm
El between age groups (untreated skin) 300 250 200 150 100 50 0
Distal left forearm
Distal right arm
Proximal right arm
FIGURE 26.10 Age-related comparisons of friction and electrical impedance. No significant differences were apparent between old and young skin on the volar forearm. Within each category, the proximal right arm friction and electrical impedance measurements were different from those of the distal right arm (p < 0.001). (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
26.3 CONCLUSION Although there have been limited studies dealing with the measurement of the skin friction coefficient, past studies
% Increase from untreated skin
Tribological Studies on Skin: Measurement of the Coefficient of Friction
Asians Caucasians
African Americans Hispanics/Latinos
400 300 200 100 0
% Decrease from untreated skin
of the test apparatus is an extremely important factor, because test design parameters can also have an influence on friction measurements. A better appreciation of the importance of the friction coefficient will become clearer as measurement methods improve and allow for greater accuracy.
REFERENCES Occlusion
Petrolatum Intervention
Glycerin
FIGURE 26.11 Coefficient of friction across ethnicity. Data represent the increase in friction when compared to untreated skin of the volar forearm. No significant differences were found between the different ethnic groups. Petrolatum and glycerin increased the friction coefficient significantly more than PVDC occlusion (p < 0.01). The increase in the friction coefficient due to petrolatum was not significantly different from the effect of glycerin. (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
Asians Caucasians
African Americans Hispanics/Latinos
0
−25
−50
−75
223
Occlusion
Petrolatum Intervention
Glycerin
FIGURE 26.12 Change in electrical impedance across ethnicity. Data represent the decrease in electrical impedance when compared to untreated skin of the volar forearm. No significant differences were found between the different ethnic groups. Glycerin lowered the electrical impedance significantly more than PVDC occlusion or petrolatum (p < 0.01). The decrease in the electrical impedance due to PVDC occlusion was not significantly different from the effect of petrolatum. (Adapted from Sivamani, R.K. et al., Skin Res. Technol., 9, 299–305, 2003.)
and our study17 show that differences in skin, because of various factors — such as age and hydration — can be correlated with the friction coefficient. Friction coefficient studies can serve as a quantitative method to investigate how skin differs on various parts of the body and how it differs between different people. It is also a useful method for tracking the changes resulting from environmental treatments, such as sunlight, and when various chemicals are applied to the skin, such as soaps lubricants and skin creams. The reviewed studies also indicate that the design
1. Naylor, P.F.D., The skin surface and friction, Br J Dermatol, 67: 239–248, 1955. 2. El-Shimi, A.F., In vivo skin friction measurements, J Soc Cosmet Chem, 28: 37–51, 1977. 3. Comaish, S., Bottoms, E., The skin and friction: deviations from Amonton’s laws, and the effects of hydration and lubrication, Br J Dermatol, 84: 37–43, 1971. 4. Koudine, A.A., Barquins, M., Anthoine, Ph., Auberst, L., Leveque, J.-L., Frictional properties of skin: proposal of a new approach, Int J Cosmet Sci, 22: 11–20, 2000. 5. Highley, D.R., Coomey, M., DenBeste, M., Wolfram, L.J., Frictional properties of skin, J Invest Dermatol, 69: 303–305, 1977. 6. Comaish, J.S., Harborow, P.R.H., Hofman, D.A., A hand-held friction meter, Br J Dermatol, 89: 33–35, 1973. 7. Prall, J.K., Instrumental evaluation of the effects of cosmetic products on skin surfaces with particular reference to smoothness, J Soc Cosmet Chem, 24: 693–707, 1973. 8. Cua, A., Wilheim, K.P., Maibach, H.I., Frictional properties of human skin: relation to age, sex and anatomical region, stratum corneum hydration and transepidermal water loss, Br J Dermatol, 123: 473–479, 1990. 9. Johnson, S.A., Gorman, D.M., Adams, M.J., Briscoe, B.J., The friction and lubrication of human stratum corneum, in Thin Films in Tribology, Dowson, D. et al., Eds., Elsevier Science Publishers, Amsterdam, 1993, pp. 663–672. 10. Asserin, J., Zahouani, H., Humbert, Ph., Couturaud, V., Mougin, D., Measurement of the friction coefficient of the human skin in vivo. Quantification of the cutaneous smoothness, Colloids Surf B Biointerfaces, 19, 1–12, 2000. 11. Elsner, P., Wilhelm, D., Maibach, H.I., Frictional properties of human forearm and vulvar skin: influence of age and correlation with transepidermal water loss and capacitance, Dermatologica, 181: 88–91, 1990. 12. Lodén, M., Olsson, H., Axéll, T., Linde, Y.W., Friction, capacitance and transepidermal water loss (TEWL) in dry atopic and normal skin, Br J Dermatol, 126: 137–141, 1992. 13. Sulzberger, M.B., Cortese, Jr., T.A., Fishman, L., Wiley, H., Studies on blisters produced by friction, J Invest Dermatol, 47: 456–465, 1966. 14. Nacht, S., Close, J., Yeung, D., Gans, E.H., Skin friction coefficient: changes induced by skin hydration and emollient application and correlation with perceived skin feel, J Soc Cosmet Chem, 32: 55–65, 1981.
224
Handbook of Non-Invasive Methods and the Skin, Second Edition
15. Hills, R.J., Unsworth, A., Ive, F.A., A comparative study of the frictional properties of emollient bath additives using porcine skin, Br J Dermatol, 130: 37–41, 1994. 16. Dawson, D., in Bioengineering of the Skin: Skin Surface Imaging and Analysis, Wilhelm, K.-P., Elsner, P., Berardesca, E., Maibach, H., Eds., CRC Press, Boca Raton, FL, 1997, pp. 159–179. 17. Sivamani, R.K., Goodman, J., Gitis, N.V., Maibach, H.I., Friction coefficient of skin in real-time, Skin Res Technol, 9: 235–239, 2003. 18. Wolfram, L.J., Friction of skin, J Soc Cosmet Chem, 34: 465–476, 1983. 19. Denda, M., in Dry Skin and Moisturizers: Chemistry and Function, Lodén, M., Maibach, H., Eds., CRC Press, Boca Raton, FL, 2000, pp. 147–153. 20. Wolfram, L.J., in Cutaneous Investigation in Health and Disease: Noninvasive Methods and Instrumentation, Leveque, J.-L., Ed., Marcel Dekker, New York, 1989, chap. 3.
21. Zhai, H., Maibach, H.I., Effects of skin occlusion of percutaneous absorption: an overview, Skin Pharmacol Appl Skin Physiol, 14: 1–10, 2001. 22. Cua, A.B., Wilhelm, K.-P., Maibach, H.I., Skin Surface lipid and skin friction: relation to age, sex, and anatomical region, Skin Pharm, 8: 246–251, 1995. 23. Manuskiatti, W., Schwindt, D.A., Maibach, H.I., Influence of age, anatomic site and race on skin roughness and scaliness, Dermatology, 196: 401–407, 1998. 24. Sivamani, R.K., Wu, G.C., Gitis, N.V., Maibach, H.I., Tribological testing of skin products: gender, age, and ethnicity on the volar forearm, Skin Res Technol, 9: 299–305, 2003. 25. Sivamani, R.K., Goodman, J., Gitis, N.V., Maibach, H.I., Coefficient of friction tribological studies in man: an overview, Skin Res Technol, 9: 227–234, 2003.
Friction Evaluation by 27 Skin Unidirectional Stress Using a Friction Tester Mariko Egawa and Motoji Takahashi Bioengineering Research Labs, Shiseido Co., Ltd., Yokohama, Japan
CONTENTS 27.1 Introduction............................................................................................................................................................225 27.2 Methodological Principle ......................................................................................................................................225 27.3 Skin Friction ..........................................................................................................................................................226 27.3.1 In Vivo Measurement of Skin Friction ......................................................................................................226 27.3.2 Relationship between Skin Friction and Other Physiological Parameters...............................................227 27.3.3 Skin Surface Friction and Sensory Evaluation .........................................................................................228 27.3.3.1 Skin Surface Friction after the Application of Emulsion..........................................................228 27.3.3.2 Skin Surface Friction and Sensory Evaluation by Experts .......................................................229 27.3.3.3 Skin Surface Friction and Sensory Evaluation (by Consumers)...............................................230 27.3.3.4 Skin Surface Friction and Fitting of Emulsion in Sensory Evaluation ....................................231 27.4 Recommendation ...................................................................................................................................................231 References .......................................................................................................................................................................231
27.1 INTRODUCTION Although sensory evaluation is important for evaluating cutaneous diseases, and efficacy of cosmetic products and drugs for external use, the differences with the examiner, physiological and physical fluctuations, reliability, and reproducibility must be taken into consideration. Since there are no convenient methods for measurement of sensory evaluation on human skin, objective measurement using instruments is necessary. Sensory properties have been considered to be related to skin friction,1–10 and various devices to measure skin friction have been developed in recent years. These devices have been categorized mainly into two types: one using a probe rotating on the skin and the other using a probe sliding on the skin. The portable friction meter11–13 is one of the many rotatingtype devices.14–21 This handheld device consists of a spring-loaded Teflon wheel that rotates at a constant speed with a constant load. Skin friction has been reported to increase with skin hydration15,16 and with application of moisturizers using this instrument.17 There are other reports using this type of instrument.13,18 The relationship between friction and capacitance or transepidermal water
loss (TEWL) in atopic dry skin and normal skin has also been reported by the oscillating method.19 On the other hand, there are a few sliding-type devices.22–26 As the sensor probe moves like the actual movement of fingers, this type should be more reliable in efficacy evaluation of cosmetics than the rotating type. But there have been few studies on the relation between tactile sensation and skin physiological parameters, including frictional properties.
27.2 METHODOLOGICAL PRINCIPLE The KES-SE Friction Tester25 (Kato Tech Co. Ltd., Kyoto, Japan), a sliding-type device, is a commercial one developed for evaluation of surface frictional property. The apparatus is shown in Figure 27.1. An arm holder is available in addition to the commercial specification of this device for measurement on the human forearm. A sensor holder including a fingerprint-type sensor was placed on a friction detector. In this condition, a fixed load (average force of the finger when touching an object) was added to the sensor. The load and measurement distance are changeable for this purpose. When the sensor moved, the coefficient of sliding friction was recorded. This device is 225
226
Handbook of Non-Invasive Methods and the Skin, Second Edition
30 mm
(b)
(a)
FIGURE 27.1 Friction measuring unit. (a) The general view of the friction measuring unit is shown. An arm holder was attached to measure skin friction on the ventral forearm in vivo. The contact probe was moved on the skin surface on the ventral forearm in the direction from the elbow to wrist. (b) The surface of the contact probe touching the skin is covered with 20 piano wires arranged in parallel and has an area of 100 mm2. The probe was moved in the direction shown by the arrow.
designed to correspond to human skin tactile with an internal second filter and can measure the kinetics of sliding friction. The low frequency below 1 Hz and direct current were eliminated by an active second filter (dumping factor = 0.6, cut frequency band = 1 Hz), based on the fact that the change of long wavelength is not felt rough to human touch. The signal passing through the internal second filter was magnified 10 times, and output was converted to voltage. This voltage was accumulated in an integrator and the result expressed numerically. The average coefficient of friction (μav) and its mean deviation (MD) are usually used with this instrument to indicate the surface friction numerically. μav has been reported to correspond to smoothness sensed by human touch and MD to asperity sense, also by human touch.24
27.3 SKIN FRICTION 27.3.1 IN VIVO MEASUREMENT
OF
SKIN FRICTION
A KES-SE friction tester equipped with an arm holder was used for measurement of skin surface friction. The following measurement conditions were used in our study: the moving speed of the sensor, 1.1 mm/sec; the shift distance, 30 mm; and load, 0.244 N. The sensor gently slid on the skin surface of the ventral forearm in the direction from elbow to wrist shown by the arrow in Figure 27.1, and the data were obtained from each measurement of the mid-zone from 5 mm of each end. The surface of
the sensor touching the skin is covered with 20 piano wires as an imitation of a fingerprint and is 100 mm2 (10 × 10 mm) in area. The parameters usually used with this device are μav and MD. μav is the average coefficient of friction for a 20mm length after cutting off 5 mm each from the starting point and the end point, where the total measurement length was 30 mm. MD indicates the mean deviation for the coefficient of friction for a measurement length of 20 mm, identical with the area used for the calculation of μav: ⎛ ⎛ L ⎞⎞ μ av = ⎜ 1 ⎜ max ⎟ ⎟ ⎝ ⎝ L max ⎠ ⎠ ⎛ ⎛ L ⎞⎞ MD = ⎜ 1 ⎜ max ⎟ ⎟ ⎝ ⎝ L max ⎠ ⎠
∫
L max
μ dL
L min
∫
L max
μ − μ av dL
L min
μ = coefficient of friction The skin surface on the ventral forearm as the test site was measured at 1 hour after washing once with soap. Then, the obtained μav and MD were measured, with the sensor load varying from 2.44 × 10–1 N to 4.89 × 101 N. The moving speed of the sensor was fixed at 1.1 mm/sec during this series of measurements. The effect of load of sensor on μav was stable in the range of load used in this study. This supports previous studies using almost the same applied load (load = 0.003 to 6 N).16,21
Skin Friction Evaluation by Unidirectional Stress Using a Friction Tester
μav, MD (x10)
Water content 0.6
60
0.4
40
0.2
20
0
Before Just after
1.5 4 Time (hr)
5.5
27.3.2 RELATIONSHIP BETWEEN SKIN FRICTION OTHER PHYSIOLOGICAL PARAMETERS
80
MD (x10)
7
Water content (a.u.)
μav
0.8
0
FIGURE 27.2 Effect of water content in the stratum corneum on μav or MD. The changes in μav and MD were measured before and after applying 2 μl/cm2 distilled water on the ventral forearm. μav and MD transiently increased after application of water, and then decreased with the decrease in water content of the stratum corneum measured with the Corneometer CM825.
Figure 27.2 shows the chronological changes of μav, MD, and water content in the stratum corneum27 measured by Corneometer CM825 (Courage+Khazaka Electronic Gmbh, Köln, Germany) after the application of water (2 μL/cm2) on the ventral forearm. μav and MD increased immediately after the application of water and later fell to almost the same level as before the application. These findings suggested that μav and MD on the ventral forearm would be influenced by water content in the stratum corneum.
AND
In total, 53 male or female Japanese volunteers in good health (ages 20 to 51 years) participated in this test. Test sites were on the ventral forearm. In order to avoid environmental influence, volunteers stayed in an air-conditioned room for 30 minutes after washing their forearms with soap. Then, the water content in the stratum corneum,27 skin friction, and viscoelastic properties (Cutometer SEM575, Courage+Khazaka Electronic Gmbh, Germany) were measured. Mechanical parameters are shown in Figure 27.3.28 Following these measurements, skin surface replicas29,30 were obtained from the same test sites. Aging of volunteers influenced neither μav nor MD, as shown in Figure 27.4. The relationship between μav or MD and other physiological parameters was examined by simple linear and stepwise multiple regression analysis, as shown in Table 27.1. A significant correlation was observed between μav and water content in the stratum corneum by simple linear regression analysis. In addition, by stepwise regression analysis it was shown that the value of μav was markedly influenced by water content in the stratum corneum and skin surface patterns, such as Ra(h), which is the arithmetic mean skin roughness value in the direction horizontal to that of the movement of the friction probe 2 9 ; KSD, which indicates average of skin roughness30; and VC1, which indicates anisotropy of skin furrows.30 By stepwise regression analysis, KSD and water content in the stratum corneum were recognized as the most influential factors on MD. Other parameters such as R0, which is a deviation length of the skin when it was
R3
R0 = e(a) R1 = e(a + b)
R0 = Uf
R2 = (e(a) − e(a + b)/e(a) = Ua/Uf Uv
Elongation (mm)
227
R3 = e((r∗a) + ((r −1)∗b)) R4 = e((a + b)∗r)
Ur
R5 = (e(a) − e(a + 0.1))/e(0.1) = Ur/Ue
e(0.1) = Ue
R6 = (e(a) − e(0.1))/e(0.1)) = Ur/Ue
Ua
R7 = (e(a) − e(a + 0.1))/e(a) = Ur/Uf
e(a + 0.1)
R8 = ((e(a)∗a∗100)/f(a) − 1)∗100 Ue Uf
f(x) = sum of e(a)
R4 R1
0 0.1
a a + 0.1
a+b Time (s)
(r∗a) + (r − 1)∗b
(a + b)∗r
FIGURE 27.3 Viscoelasticity parameter. The viscoelasticity parameter was measured using a cutometer: SEM575. The deformation of the cutaneous area was measured after mechanical suction, followed by release of pressure, repeated five times. The parameters were from R0 to R8, as generally used.
228
Handbook of Non-Invasive Methods and the Skin, Second Edition
0.04 r = 0.177 p = 0.205
MD
μav
0.8
0.4
0
0
20
40
60
Age (years)
r = 0.087 p = 0.579
0.02
0
0
20
40
60
Age (years)
FIGURE 27.4 Change in μav and MD with age. No age-related changes were found in the μav or MD values on the ventral forearm.
TABLE 27.1 Stepwise Multiple Regression Analysis of μav or MD vs. Other Physiological Parameters
μav
MD
r
R2
RMSE
p value
Water Ra(h) KSD VC1 R0
0.655 0.230 –0.051 –0.056 0.193
0.4293 0.4474 0.4673 0.4987 0.5271
0.086
1000 J/sqcm) in psoriatic patients (abstract). In 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 1990, p. 59. 45. Borroni G, Vignai G, Vignoli GP, et al. PUVA-induced viscoelastic changes in the skin of psoriatic patients. Med Biol Environ 17:663–671, 1989. 46. Hargens AR, Millard RW, Pettersson K, Johansen K. Gravitational haemodynamics and oedema prevention in the giraffe. Nature 329:59–60, 1987. 47. Wickman M, Olenius M, Malm M, Jurell G, Serup M. Alterations in skin properties during rapid and slow tissue expansion for breast reconstruction. Plast Reconstr Surg 90:945–951, 1992. 48. Serup J, unpublished observations.
Chamber Method for 66 Suction Measurement of Skin Mechanics: The Cutometer® Ken-ichiro O’goshi Department of Dermatology, Bispebjerg Hospital, Copenhagen, Denmark
CONTENTS 66.1 Introduction and Background ................................................................................................................................579 66.2 Measuring Principle...............................................................................................................................................579 66.3 Measuring Device and Practical Use ....................................................................................................................580 66.3.1 Measuring Conditions and Preconditioning of Individuals ......................................................................581 66.4 Physiological Variables and Normal Skin.............................................................................................................582 66.4.1 Age and Sex...............................................................................................................................................582 66.4.2 Study of Skin Disease ...............................................................................................................................582 66.4.3 Study of Product Efficacy .........................................................................................................................582 References .......................................................................................................................................................................582
66.1 INTRODUCTION AND BACKGROUND Elasticity of the skin is necessary for structural support and for interference with facticial stimuli from outside, for joint action, and for the function of the vessels and nerves. Elasticity is mainly controlled by the collagen fibers and the surrounding intercellular ground substance, which consists primarily of water and proteoglycans. Measurement of elasticity or, to be more precise, viscoelasticity is important in various conditions of healthy skin and skin diseases, as exemplified by skin aging and scleroderma. Test of the efficacy of treatments is another application. The Cutometer® (Courage + Khazaka Electronic GmbH, Cologne, Germany) has been introduced as a device that can measure the viscoelastic properties of the skin in vivo. It provides valuable information on physiological and pathological changes of human skin as well as the efficacy of topical treatment. The Cutometer is now well recognized as a commercial standard tool in dermatological and cosmetic research. The Cutometer operates similarly to the Dermaflex®, introduced before the Cutometer.
FIGURE 66.1 The probe of the Cutometer is small and convenient enough to measure skin areas that are difficult to reach.
66.2 MEASURING PRINCIPLE The Cutometer is a suction chamber method. Negative air pressure is applied to the skin surface through the probe aperture (Figure 66.1). The resultant elevation of the skin surface into the suction chamber is measured. Elevation 579
580
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 66.2 The head of the Cutometer has several aperture sizes (2, 4, 6, and 8 mm in diameter) to fit different skin sites and the different study requirements.
is measured by a noncontact optical system in the device that consists of a light transmitter and a light recipient. There are two glass prisms, which project the light from transmitter to recipient. The diminution of the infrared light beam, depending on the elevation of the skin, is measured.
66.3 MEASURING DEVICE AND PRACTICAL USE
measurement of the epidermal elasticity. The 4- or 6-mmdiameter apertures are for study of the outer skin layers, and the 8-mm-diameter aperture is for measurement of full-thickness elasticity. The probe is connected to the main unit with an air tube and an electric cable. The pressure can be adjusted between 50 and 500 mbar and can be built up immediately or gradually at a controlled rate, as decided. The suction time and relaxation time can be changed from 0.1 to 60 seconds, and the number of measurement cycles from 1 to 99. The main unit contains the pump and the evaluation electronics. Two measuring modes can be chosen, a stress–strain technique and a time–strain technique. With the stress–strain mode, the vacuum is increased over a selected period. Then, the deformation (in millimeters) is displayed as a function of negative pressure (in millibars). With the time–strain mode, which is mostly used to study viscoelastic properties of human skin, a selected vacuum is applied for the selected period. The skin deformation is shown as a function of time. The deformation curve is created from rapid deformation representing an elastic section, followed by a viscoelastic section, and finally a viscous section2 (Figure 66.3). The time–strain mode may be used with three cycles of 5-second traction under negative pressure of 500 mbar separated by 5-second relaxation periods. The skin deformation is plotted as a function of time. The resistance of the skin to be sucked up by negative pressure (firmness or distensibility) and its ability to return into its original position (elasticity or elastic relaxation) are displayed as
The Cutometer handheld probe, with a distinctive central suction head (Figure 66.2), can be chosen with different exchangeable aperture sizes (2, 4, 6, and 8 mm in diameter). The aperture can be chosen depending on the level in the skin affected by the disease or condition (Table 66.1). The 2-mm-diameter aperture is primarily used for
TABLE 66.1 Parameters Proposed by Agache et al.1 Parameters
Interpretation
Ue Uv Uf Ur Ua
Immediate deformation: extensibility Viscoelastic contribution: plasticity Final deformation: distensibility Immediate retraction Final retraction after removal of the vacuum
Ua/Uf
Gross elasticity of the skin, including viscous deformation Pure elasticity Biological elasticity The ratio between delayed and immediate deformation
Ur/Ue Ur/Uf Uv/Ue
–
–
FIGURE 66.3 Typical graphical registration of a strain–time curve on human skin using the Cutometer. A load (vacuum) of 500 mbar was applied for 5 seconds, followed by a 3-second relaxation period. (From Berndt, U. and Elsner, P., Hardware and measuring principle: the Cutometer®, in Bioengineering of the Skin: Skin Biomechanics, Elsner, P., Berardesca, E., Wilhelm, KP., and Maibach, HI., Eds., CRC Press, Boca Raton, FL, 2002.)
Suction Chamber Method for Measurement of Skin Mechanics: The Cutometer®
TABLE 66.2 Aperture Sizes (2, 4, 6, and 8 mm in Diameter) and Level in the Skin Folded or Elongated Aperture Size
Layer of the Skin
2 4 6 8
Epidermal (nonuniform) folding Outer skin folding Outer skin elongation Full skin (uniform) elongation
curves at the end of each measurement. From these curves, among the calculated values, the biological elasticity and viscoelastic/elastic ratio with Ur/Uf and Uv/Ue units, respectively, are calculated (Figure 66.2). The following parameters and nomenclature were proposed by Agache et al.1 (Table 66.2): Ue as the immediate deformation or skin extensibility; Uv as the delayed distension reflecting the viscoelastic contribution of the skin; Uf as the final skin deformation (skin distensibility); Ur as the immediate retraction; Ua as the final retraction after removal of the vacuum; Ua/Uf as the ratio of total retraction to total deformation, which is called gross elasticity of the skin, including viscous deformation; Ur/Ue closely resembles Ur/Uf and is also used as a measure of elastic recovery; Ur/Uf as the ratio of immediate retraction to the total deformation, which is called biological elasticity; and Uv/Ue as the ratio between delayed and immediate deformation, which indicates the relative contributions of the viscoelastic plus viscous and the elastic distension to the total deformation. All these parameters are functions of skin thickness. The Uv/Ue ratio increases with decreasing elasticity. High values of these ratios — maximum = 1 (100%) — indicate a high level of elasticity. Ideally, if the skin was an isotropic physical material, the experimental values should be standardized for skin thickness, which can be determination in vivo by ultrasound. Ratios of the above parameters can be taken with no measurement of skin thickness. Ratios are specific for a body site. Skin is as a biological material viscoelastic, and mechanics known from isotropic materials such as Young’s module, depending on the thickness of the sample, are not directly applicable to anisotropic skin.
66.3.1 MEASURING CONDITIONS AND PRECONDITIONING OF INDIVIDUALS Skin elasticity is a dynamic function. Skin contains about 75% water, and the viscous component is thus very important to control and standardize. Persons should be in a neutral water balance and not dehydrated, since this reduces the skin turgor. Skin turgor (from the Latin turgor, swelling) is the water tension of the skin. Turgor is known
581
to reflect the intradermal and general hydration state, as well as the fiber system of the skin, particularly agedrelated changes.3 Repeated stress and deformation of skin result in further distension or elongation, known as creaping or hysteresis. Following a distension, it may take up to an hour for the skin to regain the spontaneous state. Hysteresis of connective tissue is the mechanical change behind warming up in sports. Thus, measurement of skin elasticity with a suction cup may not be repeated on the same site for 1 to 2 hours due to mechanical recall. Applications and conditions that may cause swelling of the skin shall be avoided, such as sun exposure, skin irritation, and major physical activity. Medicines like diuretics and hormones may easily influence skin mechanics. Diurnal variation and orthostatic position may be significant. In the morning, water redistributes and accumulates in the legs shortly after standing up. The skin has an important junction as a water reservoir in the regulation of total body water. The zero situation is important to control. The joint position shall be carefully standardized since the spontaneous stretching of the skin when measurements start will influence the result. Serup4 described that the function in relation to joint motions is vital in evolution and is directly reflected in the Langer lines, which are oriented in directions that clearly respect free motion of rotator joints and hinge joints. The important mechanical skin functions can be followed as Serup described:4 • • • •
To provide a protective and supporting cover of the body To remain tense but allow free motion of joints To counteract gravitation To be soft and pliable to allow effective contact with physical surfaces as a basis for simple and complex sensory perceptions (touch, including stereological perception of objects, pain, heat, cold, and others)
The skin is far from being mechanically uniform or isotropic. It is very different from physical materials because of the woven structure with layers and fibers. The stratum corneum and epidermis are supposed to be relatively rigid, the papillar dermis quite soft and pliable, and the papillar dermis is tense and the mechanically strongest structure of the skin. Skin elasticity is obviously dependent on body region. The thin skin of extremities is due to the woven structure, more rigid than the thick skin of the trunk. Skin becomes more rigid toward the ground, which can be seen as an evolutionary adaptation to gravity, with the pretension of the skin preventing water accumulation in legs during daytime.
582
Handbook of Non-Invasive Methods and the Skin, Second Edition
66.4 PHYSIOLOGICAL VARIABLES AND NORMAL SKIN 66.4.1 AGE
AND
SEX
Age and sex affect the viscoelastic properties of skin as described by Cua et al.5 in 33 healthy subjects. Generally, Uv/Ue increased while Ur/Uf decreased with age. Responses were variable with respect to load applied. Variability within anatomical regions was also noted, but differences between the sexes were not statistically significant for most regions. Couturaud et al.6 followed the evolution of the biomechanical properties of the skin and the microdepressionary network (mDN) as a function of age with different probe diameters of the Cutometer. Measurements obtained with the Cutometer on two groups of average ages 30 and 60 years showed modifications of skin biomechanical properties as a function of age: decrease in elasticity and extension, while fatiguability increases, with age. This was independent of the area and probe diameter. Elsner et al.7 determined the mechanical properties of human genital skin with the use of the Cutometer. Uv/Ue and Ur/Uf were both significantly lower in vulvar than in forearm skin. Ur/Uf decreased significantly with load in vulvar, but not in forearm skin, whereas Uv/Ue increased, while Uv/Ue was not load dependent in either site. Uv/Ue remained constant with age in forearm and vulvar skin. In the vulva, but not in the forearm, Uv/Ue was significantly correlated with body height.
66.4.2 STUDY
OF
SKIN DISEASE
Studies were conducted in psoriasis,8 systemic sclerosis,9–11 Ehlers–Danlos syndrome,12 striae distensae,13 hypertrophic scarring,14 and diabetes mellitus.15 Also, changes of biomechanical properties following the application for ultraviolet light,16 laser treatment,17 reconstructive dermal substitutions,18 and liposuction19 have been reported.
66.4.3 STUDY
OF
PRODUCT EFFICACY
The effect of cosmetics such as emollients, moisturizers, chemical peelings, antiaging creams, etc., studied with the Cutometer were reported.20,21
REFERENCES 1. Agache PG, et al. Mechanical properties and Young’s modulus of human skin in vivo. Arch Dematol Res 269, 221, 1980. 2. Elsner P. Skin elasticity. In Bioengineering of the Skin: Methods and Instrumentation, Berardesca E, Elsner P, Wilhelm KP, Maibach HI, Eds. CRC Press, Boca Raton, FL, 1995.
3. Lockhard, RD, Ed. Living Anatomy, a Photographic Atlas of Muscles in Action and Surface Contours, 2nd ed. Faber & Faber, London, 1962. 4. Serup J. Mechanical properties of human skin: elasticity parameters and their relevance. In Bioengineering of the Skin: Skin Biomechanics, Elsner P, Berardesca E, Wilhelm KP, Maibach HI, Eds. CRC Press, Boca Raton, FL, 2002. 5. Cua AB, Wilhelm KP, Maibach HI. Elastic properties of human skin: age, sex, and anatomical region. Arch Dermatol Res 282(5), 283–286, 1990. 6. Couturaud V, Coutable J, Khaiat A. Skin biomechanical properties: in vivo evaluation of influence of age and body site by a non-invasive method. Skin Res Technol 1, 68, 1995. 7. Elsner P, Wilhelm D, Maibach HI. Mechanical properties of human forearm and vulvar skin. Br J Dermatol 122, 607, 1990. 8. Dobrev HP. In vivo study of skin mechanical properties in psoriasis vulgaris. Acta Derm Venereol (Stockh) 80, 263, 1999. 9. Dobrev HP. In vivo study of skin mechanical properties in patients with systemic sclerosis. J Am Acad Dermatol 40, 436, 1999. 10. Enomoto DN, et al. Quantification of cutaneous sclerosis with a skin elasticity meter in patients with generalized scleroderma. J Am Acad Dermatol 35, 381, 1996. 11. Nikkels-Tassoudji N, et al. Computerized evaluation of skin stiffening in scleroderma. Eur J Clin Invest 26, 457, 1996. 12. Henry F, et al. Mechanical properties of skin in EhlersDanlos syndrome, type I, II, and III. Pediatr Dermatol 13, 464, 1996. 13. Pierard GE, et al. Tensile properties of relaxed excited skin exhibiting striae distensae. J Med Eng Technol 23, 69, 1999. 14. Fong SS, Hung LK, Cheng JC. The Cutometer and ultrasonography in assessment of postburn hypertrophic scar: a preliminary study. Burns 23(Suppl. 1), 12, 1997. 15. Yoon HK, et al. Quantitative measurement of desquamation and skin elasticity in diabetic patients. Skin Res Technol 8, 250, 2002. 16. Habig J, et al. Einflu einmaliger UVA- und UVBBestrahlung auf Oberflächenbeschaffenheit und viskoelastische Eigenschaften der Haut in vivo. Hautarzt 47, 515, 1996. 17. Koch RJ, Cheng ET. Quantification of skin elasticity changes associated with pulsed carbon dioxide laser skin resurfacing. Arch Facial Plast Surg 1, 272, 1999. 18. van Zujilen PP, et al. Graft survival and effectiveness of dermal substitution in burns and reconctructive surgery in a one-stage grafting model. Plast Reconstr 106, 615, 2000. 19. Henry F, et al. Mechanical properties of skin and liposuction. Dermatol Surg 22, 566, 1996. 20. Fischer T, et al. Instrumentelle Methoden zur Bewertung der Sicherheit und Wirksamkeit von Kosmetika. Akt Dermatol 24, 243, 1998. 21. Greif C, et al. Beurteilung einer Körperlotion für trockene und empfindliche Haut. Kosmet Med 19, 24, 1998.
Chamber Method for 67 Suction Measurement of Skin Mechanics: The New Digital Version of the Cutometer André O. Barel,1 W. Courage,2 and Peter Clarys1 1 2
Faculty of Physical Education and Physiotherapy, Vrije Universiteit Brussel, Brussels, Belgium Courage+Khazaka Electronic GmbH, Köln, Germany
CONTENTS 67.1 Introduction............................................................................................................................................................583 67.2 Object of This Study .............................................................................................................................................584 67.3 Methodological Principle ......................................................................................................................................584 67.3.1 Description of the Measuring Probe .........................................................................................................584 67.3.2 Description of the Measuring Modes........................................................................................................584 67.3.3 Handling of the Probe ...............................................................................................................................585 67.3.4 Analysis of the Measuring Systems..........................................................................................................585 67.3.4.1 Strain–Time Curves....................................................................................................................585 67.3.4.2 Stress–Strain Curves...................................................................................................................586 67.3.4.3 Reproducibility of Skin Deformation Parameters and of the Modulus of Young ....................587 67.4 A Short Overview of the Results ..........................................................................................................................587 67.4.1 Single Stress–Time Curves........................................................................................................................587 67.4.2 Repetitive Stress–Time Curves..................................................................................................................588 67.4.3 Stress–Strain Curves..................................................................................................................................588 67.5 Factors Influencing the Measurements..................................................................................................................588 67.5.1 Influence of Load.......................................................................................................................................588 67.5.2 Influence of the Diameter of the Suction Device .....................................................................................588 67.5.3 Influence of the Orientation of the Probe .................................................................................................588 67.5.4 Influence of Pressure of Application on the Skin.....................................................................................589 67.5.5 Preconditioning of the Skin Surface .........................................................................................................589 67.6 Applications ...........................................................................................................................................................589 67.7 Conclusions............................................................................................................................................................589 References .......................................................................................................................................................................589
67.1 INTRODUCTION The mechanical properties of the human skin have been extensively studied in the past, most in vitro and less in vivo.1 Skin is a complex organ that, like many other biologicals, presents in a combined way the typical properties of elastic solids and viscous liquids.2 As a consequence, the mechanical properties of the skin are called viscoelastic. Typical properties of viscoelastic materials are nonlinear stress–strain properties with hysteresis (the stress–strain curves obtained on loading
will not be superposed on the curves obtained by unloading).2–5 Furthermore, the deformation of the skin is time dependent with a typical phenomenon of creep. The creep is characterized as an increasing deformation of the skin in function of time when a constant stress is applied on this material. The viscoelastic properties of the skin are due to the components of the skin: collagen fibers, elastin fibers, and cells impregnated in a ground substance of various proteoglycans and glycoproteins.6 583
584
Handbook of Non-Invasive Methods and the Skin, Second Edition
67.2 OBJECT OF THIS STUDY The general purpose consists of measuring noninvasively the biomechanical properties of the human skin in vivo by means of a simple, reliable, and reproducible biophysical technique. There have been in the past many investigations of the mechanical properties of human skin using various equipment that measure the deformation of the skin after application of uniaxial- or biaxial-oriented forces. In these studies, normally healthy and diseased skin situations must be considered. The influence of physiological factors such as age, sex, normal and actinic aging, diseases, and changes in biomechanical properties induced by various topical treatments can also be examined using mechanical measurements. Many experimental instruments and devices have been developed in research laboratories, but only very few instruments are commercially available: a torsion method (Torque Meter*) and three instruments using the suction method (Dermaflex*, Dermalab*, and Cutometer*). This report describes the new digital version of the Cutometer MPA 580 based on the suction method. The instrument measures the vertical deformation of the skin surface when the skin is pulled in the circular aperture of the measuring probe after application of a vacuum. The use of the older analog version of the Cutometer (SEM 575) has been described by many authors7–11 and will also be reviewed in Chapter 66 of this book.12
67.3 METHODOLOGICAL PRINCIPLE 67.3.1 DESCRIPTION
OF THE
MEASURING PROBE
Figure 67.1 shows a schematic representation of the measuring handheld probe, which is attached to the main apparatus with an electric cable and an air tube. Most technical information concerning the digital version of the Cutometer has been obtained from the manufacturer.13 A variable vacuum (ranging from 20 to 500 mbar) is applied on the skin surface through the opening of the probe.The resolution of applied pressure is equal to 1 mbar. The skin surface is pulled by vacuum in the aperture of the probe. The depth of the skin penetration is measured by an optical system that measures in function of skin penetration the diminution of light intensity of a green light beam (light-emitting diode, LED). Calibration of this optical system is carried out in the factory with a micrometer pin (from 0 to 3.0 mm). The skin adjacent to the opening of the probe is maintained in position by an external guard ring attached to the * Torque Meter® is a registered trademark of Dia-Stron Ltd., Andover, U.K.; Dermaflex® and Dermalab® are registered trademarks of Cortex Technology, Hadsund, Denmark; and Cutometer® is a registered trademark of Courage-Khazaka, Cologne, Germany.
Vacuum tube
Spring
Spring
Movable part of the probe
External part of the probe
Aperture of the probe Guard ring
FIGURE 67.1 Schematic view of the measuring probe of the Cutometer. Probe with a small aperture (2 mm diameter).
probe shield (external diameter of 25 mm) or can be fixed with a double-sided sticking ring. The measuring probe is applied vertically on the surface of the skin with a constant pressure by means of a spring (50 gm/cm2 or 2 × 103 N/m2). The skin deformation (in millimeters) can be measured by this new optical system (improved optical lenses) with a resolution equal to 2 μm. As a consequence, the deformation curves of the new instrument present a much smoother tracing than the old version, where the resolution was 10 μm. Above a penetration depth of 200 μm the accuracy is 3%; below a penetration depth of 100 μm there is no linearity in the optical system. The standard measuring probe has an aperture of 2 mm diameter (test area of about 3 mm2). Optional probes with apertures of 4, 6, and 8 mm are available for studying the mechanical properties of larger skin areas. With the larger measuring probes deeper layers of the skin (dermis and perhaps hypodermis) are deformed by suction. The digital probe of the MPA 580 Cutometer is connected to a central MPA 5 multiprobe apparatus. The MPA 5 apparatus is connected to a PC with a standard Windows software directing the Cutometer. The results can be displayed as curves and values (see later in this chapter). The Courage-Khazaka software allows storage of various data concerning the volunteer, date and time of experiment, skin area, external temperature and relative humidity, type of probe used, and mode of measuring technique. In addition, graphic display of the obtained experimental curves allows calculations of individual values.
67.3.2 DESCRIPTION
OF THE
MEASURING MODES
Essentially two different measuring techniques are available: stress–strain mode and strain–time mode. In the stress–strain mode the deformation of the skin (strain) is displayed as a function of the stress (load–vacuum).
585
600 0.6 0.5 0.4 0.3 0.2 0.1 0.0
Uv
500
Ur Uf
Ue
−10
−5
0
5 10 15 Time (sec)
20
25
Load P (mbar)
Deformation (mm)
Suction Chamber Method for Measurement of Skin Mechanics: The New Digital Version of the Cutometer
400 300 200 100
FIGURE 67.2 Graphical representation of a strain–time curve obtained for forearm skin with the SEM 575 Cutometer. Aperture, 2 mm; time of application, 10 sec relaxation time, 10 sec no pretension applied; load, 500 mbar. Skin deformation in millimeters.
In the strain–time mode the deformation of the skin is showed as a function of time (Figure 67.2). In both experimental modes the choice of vacuum (from 20 to 500 mbar), the duration of the measurements (from 0.1 to 320.0 sec), and the number of measurement cycles (from 1 to 10) can be preset. The apparatus offers a choice of four measuring modes. The four measuring modes are the result of different combinations of choices of application rates and release rates of vacuum on the skin. Mode 1: Measurement with constant negative pressure. Mode 2: Measurement with linear rising and falling of negative pressure. Mode 3: Measurement with first constant, then linear falling of negative pressure. Mode 4: Measurement with linear rising of negative pressure and the abrupt cessation of negative pressure. The strain vs. time curve (mode 1) is mostly used in the mechanical studies on human skin and applications in the field of dermatocosmetics. In measuring system 2, where the deformation of the skin is measured during a linear increase in vacuum from 0 to maximal 500 mbar, and subsequently during the linear decrease of vacuum, the resulting graphical display can be automatically replotted as a stress–strain curve (Figure 67.3).
67.3.3 HANDLING
OF THE
PROBE
The skin immediately adjacent to the opening of the measuring probe is normally held in position by the guard ring of the probe on application of the probe on the skin surface or by the use of a double-sided sticking ring in order to reduce some lateral displacement of the skin adjacent to the opening during the suction.
0 0.0
0.1
0.2 0.3 0.4 Deformation (mm)
0.5
0.6
FIGURE 67.3 Graphical representation of a stress–strain curve obtained for forearm skin with the SEM 575 Cutometer. Aperture, 2 mm; linear increase and decrease in vacuum, 50 mbar/sec total application time, 20 sec.
In order to minimize this lateral displacement when no sticking ring is used, it is possible to pretension the skin before the measurements are carried out. Pretension of the skin is carried out by applying a preliminary suction on the surface of the skin during a short time (about 0.1 sec) before the real vertical deformation measurements are executed.
67.3.4 ANALYSIS
OF THE
MEASURING SYSTEMS
67.3.4.1 Strain–Time Curves In the strain–time mode, which is mostly used in viscoelastic studies on human skin, the vacuum is applied for a period varying from 1 to 10 sec followed by a relaxation of 1 to 10 sec. Figure 67.2 shows the results of a typical experiment carried out on the volar part of the forearm with the old version of the Cutometer (SEM 575). The deformation parameters used in most studies in order to describe the different parts of the curve are those proposed by Aubert et al.5 and Escoffier et al.14 Ue is the immediate deformation–skin extensibility. Uv is the deviation, which reflects the viscoelastic contribution of the skin. Uf is equal to the total deviation of the skin. Ur is equal to the immediate recovery of the skin after removal of vacuum. Due to the slow return (creep) to the original state of the skin after application of a given load, the deformation Ur does not really reach a constant plateau value. A value of 0.1 second after application of suction and removal of vacuum is systematically taken as the time values for measuring Ue and Ur. Figure 67.4 shows the results of a typical experiment carried out on the skin with the new digital version of the Cutometer (MPA 580). The new deformation parameter Ua is equal to the total recovery of the skin. The residual deformation of the skin R is equal to Uf – Ua. The standard Courage-Khazaka software calculates
586
Handbook of Non-Invasive Methods and the Skin, Second Edition
E(mm)
Amplitude/Time 0.54 0.52 0.5 0.48 0.46 0.44 0.42 0.4 0.38 0.36 0.34 0.32 0.3 0.28 0.26 0.24 0.22 0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0
Cutometer results 1 repetition
R0 = Uf R1 = Uf − Ua R2 = Ua/Uf R3 = last max. amplitude R4 = last min. amplitude R5 = Ur/Ue R6 = Uv/Ue R7 = Ur/Uf R8 = Ua R9 = R3 − R0 Ue
Ur
Ua
F1
Uf − Ua
Uf = R0 F0 = surface F1 = surface
0
2
1
3
4
1(s)
FIGURE 67.4 Graphical representation of a strain–time curve obtained for forearm skin with the MPA 580 Cutometer. Aperture, 2 mm; time of application, 2; sec relaxation time, 2 sec. Skin deformation in millimeters.
automatically the following parameters: Ro = Uf, R1 = Uf – Ua, R2 = Ua/Uf, R5 = Ur/Ue, R6 = Uv/Ue, R7 = Ur/Uf, and R8 = Ua. In addition, the software calculates the surfaces F0 and F1. F0 is the surface between the real curve and the value corresponding to the maximal deformation Uf when going from start of suction to stop of suction. F1 is the surface between the real recovery curve and the value corresponding to the maximal recovery going from stop of suction to stop of measurement. Both surfaces reflect the viscous part (slow deformation and slow return of the skin to the original part) of the viscoelastic properties of the skin. The deformation vs. time curves obtained for the second, third, and subsequent deformation–relaxation cycles (up to 10 cycles) are similar to the first curve. But the curves are progressively shifted vertically upward as a consequence of the slow return of the skin to the original state (Figure 67.5). In this repetitive mode the total duration of one cycle is short (1-sec suction and 1-sec recovery). The software calculates on this repetitive curve the following parameters: R3 is equal to the last maximum amplitude and R4 is equal to the last minimum amplitude. In addition, the software calculates the surface F2, which is equal to the surface between the real curve and the value corresponding to the maximal deformation R3 after 10 cycles when going from start of suction to stop of the 10 cycles, the surface F3 (surface between the two
repetitive curves), and the surface F4 (surface between the minimal deformations and the time axis). All the deformation parameters, Ue, Uf, Uv, and Ur, are dependent on skin thickness. Since there are significant differences in skin thickness between women and men, and since skin thickness varies with age, the use of these extrinsic deformation parameters for comparative studies is not adequate.14 Consequently, either intrinsic skin deformation parameters are used (deformation × skin thickness) or ratios of the extrinsic parameters are considered. The ratio Ur/Ue, ratio between immediate recovery and immediate deformation, is independent of skin thickness. This ratio is considered a biologically important factor for the characterization of elasticity of the skin.14,15 The ratio Ur/Uf, which closely resembles Ur/Ue, is also used as a measure of elastic recovery.16 67.3.4.2 Stress–Strain Curves In the stress–strain mode the total deformation of the skin is measured in function of vacuum during a linear increase of vacuum, followed by a linear decrease of suction (Figure 67.3). Generally, nonlinear curves are obtained with hysteresis. The stress–strain curve with loading will not be superposed with the unloading curve. Furthermore, the values of strain in the relaxation procedure do not return to the origin, and will return to zero values only after a long period of relaxation.
Suction Chamber Method for Measurement of Skin Mechanics: The New Digital Version of the Cutometer
587
E(mm)
Amplitude/Time 0.54 0.52 0.5 0.48 0.46 0.44 0.42 0.4 0.38 0.36 0.34 0.32 0.3 0.28 0.26 0.24 0.22 0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0
Cutometer surface results by repetition
F3
F4
0
1
2
3
4
5
6
7
8
9
10 t(s)
11
12
13
14
15
16
17
18
19
20
FIGURE 67.5 Graphical representation of a repetitive strain–time curve obtained for forearm skin with the MPA 580 Cutometer. Aperture, 2 mm; 10 cycles; time of application, 1 sec; relaxation time, 1 sec. Skin deformation in millimeters.
Strictly speaking, because of the nonlinearity of the stress–strain curves for human skin, the modulus of Young is not applicable. One can always define a coefficient of elasticity that corresponds to the slope of the stress to the strain curve at a given stress value.17 In the experimental ascending curves (Figure 67.3), a linear part in the stress–strain mode is clearly present between 150 and 500 mbar. From the linear part of the curves a modulus of Young (E) can be in principle calculated.1,18 For the practical calculation of the modulus of Young, a theoretical model for the deformation of the skin in the suction aperture as proposed by Barel et al.7 and Agache et al.19 can be used, allowing the calculation of the strain and the stress. In this theoretical model one assumes that the initial flat surface of the circular test area is transformed by suction in a curved surface of a segment of a sphere. 67.3.4.3 Reproducibility of Skin Deformation Parameters and of the Modulus of Young Repetitive Uf deformation measurements without skin pretension carried out on the same person show a coefficient of variability ranging from 4 to 6%, depending on the skin sites to be examined. With skin pretension the coefficient of variability is generally lower (around 4%). Similar results were obtained by other researchers when using the suction method on different skin sites.15,16 Similar repetitive measurements carried out on
the same individual in the stress–strain mode show a coefficient of variability for the modulus of Young around 10%.
67.4 A SHORT OVERVIEW OF THE RESULTS 67.4.1 SINGLE STRESS–TIME CURVES The stress–strain and strain–time curves measured on different skin sites (forearm, face, thigh, etc.) by different researchers,7,20–25 using the suction method, are in good agreement with the results obtained by other methods (tensile, torsional, elevation, and indentation).1 It is obvious that the deformation parameters (Ue, Uf, and Ur) vary in function of the load (vacuum) and the aperture of the suction probe. With the 2-mm suction probe, typical skin deformation data (Uf), ranging from 0.1 to 0.6 mm, are recorded in function of vacuum (from 100 to 500 mbar). These values of vertical skin elevation correspond typically to deformations of the epidermis and dermis, with perhaps some contribution of the hypodermis when large deformation values are measured. The values obtained for the elasticity ratio Ur/Ue are functions of anatomical sites, age, and other physiological factors. However, typical values of the elasticity ratio ranging from 0.4 to 0.9 are recorded for the different anatomical skin sites. These elasticity recovery values are
588
Handbook of Non-Invasive Methods and the Skin, Second Edition
in good agreement with the results obtained by the torsional method.14 In addition, the software calculates (Figure 67.4) the surface F0 (surface between the real curve and the value corresponding to the maximal deformation Uf when going from start of suction to stop of suction) and the surface F1 (surface between the real recovery curve and the value corresponding to the maximal recovery going from stop of suction to stop of measurement). Both surfaces reflect the viscous part (slow deformation and slow return of the skin to the original part of the viscoelastic properties of the skin) and could be considered very interesting skin parameters. These parameters are probably investigated in the cosmetic industry in order to prove the efficacy of their products, but are not published in the scientific literature.26
67.4.2 REPETITIVE STRESS–TIME CURVES Courage-Khazaka proposes in the software package to carry out 10 cycles of suction of 1 sec each (Figure 67.5). Schlangen et al.27 propose to examine the results of three longer consecutive suction release curves. The software allows calculation of the repetitive deformation vs. time curves’ different surfaces (Figure 67.5). The software calculates on this repetitive curve the following parameters: R3 is equal to the last maximum amplitude and R4 is equal to the last minimum amplitude. Starting from these values, the software calculates the surface F2 (surface between the real curve and the value corresponding to the maximal deformation R3 after 10 cycles when going from start of suction to stop of the 10 cycles), the surface F3 (surface between the two repetitive curves), and the surface F4. Once again, these surface parameters are probably investigated in the cosmetic industry in order to prove the efficacy of their products, but are not published in the scientific literature.26
67.5 FACTORS INFLUENCING THE MEASUREMENTS 67.5.1 INFLUENCE
OF
LOAD
As would logically be expected, and dependent on the opening of the suction probe, all the skin deformation parameters are nonlinearly increased in function of load (vacuum). Due to limitations of the instrument, skin deformation measurements at 400 to 500 mbar are not possible with the 8-mm suction probe. In agreement with Elsner et al.,16 the elasticity ratio and other deformation ratios are independent of load for most anatomical sites (with the exception of vulvar skin). Due to the curvature of most of the stress–strain curves, and depending on the choice of the linear part in the stress–strain curves for the calculation, the values of the modulus of Young are different and, consequently, load dependent. The values of the modulus of Young obtained in this study are of the same order of magnitude as those recently reported by Agache et al.19 In agreement with the same workers,19 the modulus of Young is more or less independent of vacuum between 150 and 500 mbar.
67.5.2 INFLUENCE OF THE DIAMETER SUCTION DEVICE
OF THE
The maximal deformation Uf of the skin increases linearly in function of the diameter of aperture of the suction device. A maximal vertical deformation of 1.2 mm can be observed with a 6-mm suction probe. It is important to mention that skin deformation measurements are not possible at vacuum loads of 400 to 500 mbar with the 8-mm suction probe.
67.4.3 STRESS–STRAIN CURVES
67.5.3 INFLUENCE PROBE
From the linear part of the stress–strain curves typical values of the modulus of Young ranging from 130 to 260 kPa were computed for the different anatomical skin sites. These in vivo values obtained by the suction method are in agreement with previous data obtained by the torsional system1,28 (modulus of Young = 42 kPa) and the suction method29,30 (modulus of Young = 129 kPa). As pointed out previously by Piérard,1 the determination of the extent of the linear part of the stress–strain curves is not always obvious and is generally rather subjectively determined by the researchers. This explains the much larger variations generally observed in the reported Young’s modulus. As a consequence, depending on the values of the applied stress and the mechanical system used in the study, the modulus of Young varies in a large range from 104 to 106 N/m2.1
The vertical deformation of the skin is recorded by the optical measuring system along a well-defined direction, which is indicated on the surface of the measuring probe. As a consequence, the eventual anisotropy in the mechanical properties of the skin could be evaluated by measuring the vertical deformation of the skin under different orientations of the probe. The mechanical properties of the forearm skin were evaluated along two perpendicular directions: parallel and perpendicular to the primary lines of the skin microtopography of the skin surface on the forearm. No significant differences were observed in our laboratory for the elasticity ratio Ur/Ue and for the modulus of Young when measurements were carried out parallel and perpendicular to the primary lines (preferential directions of the skin surface pattern) in young and middle-aged individuals.31,32
OF THE
ORIENTATION
OF THE
Suction Chamber Method for Measurement of Skin Mechanics: The New Digital Version of the Cutometer
These results seem to indicate that the suction method measures in an isotropic way the vertical deformation of the skin located at the forearm. In all experiments it is recommended to maintain the probe exactly vertical to the skin surface.
67.5.4 INFLUENCE OF PRESSURE OF APPLICATION ON THE SKIN With the spring system, the handheld probe is always applied with constant pressure on the skin surface. However, in order to reduce as much as possible small variations in pressure of application, we have found it more convenient to carry out the experiments under the following experimental setup: the skin surface to be measured is always placed in a horizontal position and the suction probe is applied with the help of a stable probe holder.
67.5.5 PRECONDITIONING
OF THE
SKIN SURFACE
The skin adjacent to the aperture of the probe is immobilized by a guard ring in order to reduce as much as possible lateral displacement of the skin toward the opening of the suction device, or can be firmly immobilized by the use of a double-sided sticking ring. The use of the sticking ring provokes an alteration of the skin surface by removal of corneocytes from the horny layer. In the strain–time mode it is possible to pretension the skin by applying a suction during a short time before the real deformation measurements are carried out (pretime setting on the instrument). It has been shown in our laboratory7 that under pretension, the values of the skin deformation parameters are more reproducible and more accurate. Measurements carried out at different skin sites (forearm, forehead, and crow’s-feet) show that with pretension the elastic recovery parameter Ur/Ue is systematically higher (typically for forearm skin in young individuals, Ur/Ue values changes from 0.82 without pretension to 0.89 with pretension). This result indicates that with preconditioning the skin regains to a greater extent the initial position after deformation.
67.6 APPLICATIONS The new digital version of the suction method is well suited, thanks to the versatility of the measurements, to study in vivo the fundamental viscoelastic properties of the dermis in normal and diseased skin. In addition to this fundamental approach of the properties of the dermis, the influence of various factors such as sex, normal and actinic aging, and anatomical skin sites can be readily evaluated by this technique. The use of the analog version of the Cutometer (SEM 575) in order to investigate the influence of aging,
589
anatomical skin sites, and of sex has been described by many authors.3,15,16,21–25,32,33 Furthermore, the efficiency of various dermatocosmetic treatments (topical cosmetic applications and treatment of various deseases) can be evaluated quantitatively by the suction method.34,35 Review articles have been published,7–11 and these topics will also be reviewed in Chapter 66 of this book.12
67.7 CONCLUSIONS Due to improvements in the optical measuring system and a new analog electronic system, the resolution (2 mm) and the accuracy (3%) are improved. The deformation curves show a continuous tracing compared to the stepwise tracing of the old version. The digital version of the Cutometer MPA allows us to measure in a simple way in vivo the viscoelastic properties of the skin. Under well-controlled experimental conditions where various parameters such as load (vacuum), aperture of the suction device, position and pressure of application of the probe, time of application and relaxation, and pretension of the skin are kept constant, reproducible and accurate stress–strain and strain–time curves are obtained that give quantitative information concerning the purely elastic and viscoelastic properties of the dermis. These parameters are used to study the properties of various anatomical skin areas in normal and diseased skin situations. Finally the influence of physiological parameters such as aging, anatomical skin sites, and the efficacy of topical dermatocosmetic treatments can be quantitatively examined by this suction method.
REFERENCES 1. Piérard, G., A critical approach to in vivo mechanical testing of the skin, in Cutaneous Investigation in Health and Disease, Noninvasive Methods and Instrumentation, Lévêque, J.L., Ed., Marcel Dekker, New York, 1989, chap. 10. 2. Larrabee, W., A finite element model of skin deformation, Laryngoscope, 96, 399, 1986. 3. Daly, C.H. and Odland, G.F., Age-related changes in the mechanical properties of human skin, J Invest Dermatol, 73, 84, 1979. 4. Vogel, H.G., Age dependence of mechanical and biochemical properties of human skin, Bioeng Skin, 3, 141, 1987. 5. Aubert, L., Anthoine, P., de Rigal, J., and Lévêque, J.L., An in vivo assessment of the biomechanical properties of human skin modifications under the influence of cosmetic products, Int J Cosmet Sci, 7, 51, 1985.
590
Handbook of Non-Invasive Methods and the Skin, Second Edition
6. Silver, F.H., Siperko, L.M., and Seehra, G.P., Mechanobiology of force transduction in dermal tissue, Skin Res Technol, 9, 3, 2003. 7. Barel, A.O., Lambrecht, R., and Clarys, P., Mechanical function of the skin: state of the art, in Skin Bioengineering Techniques and Applications in Dermatology and Cosmetology. Current Problems in Dermatology, Elsner, P., Barel, A.O., Berardesca, E., Gabard, B., and Serup, J., Eds., Karger, Basel, 1998, p. 69. 8. Barel, A.O., Clarys, P., and Gabard, B., In vivo evaluation of the hydration state of the skin: measurements and methods for claim support, in Cosmetics. Controlled Efficacy Studies and Regulation, Elsner, P., Merk, H.F., and Maibach, H.I., Eds., Springer-Verlag, Berlin, 1999, p. 57. 9. Rodrigues, L., The in vivo biomechanical testing of the skin and the cosmetological efficacy claim support: a critical overview, in Cosmetics. Controlled Efficacy Studies and Regulation, Elsner, P., Merk, H.F., and Maibach, H.I., Eds., Springer-Verlag, Berlin, 1999, p. 197. 10. Berndt, U. and Elsner, P., Hardware and measuring priciple: the Cutometer, in Bioengineering of the Skin. Skin Biomechanics, Elsner, P., Berardesca, E., Wilhelm, K.P., and Maibach, H.I., Eds., CRC Press, Boca Raton, FL, 2002, p. 91. 11. Agache, P. and Varchon, D., Exploration fonctionnelle mécanique, Dans: Physiologie de la peau et explorations fonctionnelles cutanées, Agache, P., Ed., Editions Médicales Internationales, Cachan, France, 2000, p. 423. 12. O’goshi, K.I., Chapter 66, Suction Chamber Method for Measurement of Skin Mechanics: The Cutometer®, this volume. 13. Courage, W. and Khazaka, G., Hardware and software differences between the Cutometer SEM 575 and MPA 580, technical information, Courage-Khazaka Electronic GmbH, Köln, Germany, 2004. 14. Escoffier, C., de Rigal, J., Rochefort, A., Vasselet, R., Lévêque, J.L., and Agache, P., Age-related mechanical properties of human skin: an in vivo study, J Invest Dermatol, 93, 353, 1989. 15. Cua, A.B., Wilhelm, K.P., and Maibach, H.I., Elastic properties of human skin: relation to age, sex, and anatomical region, Arch Dermatol Res, 282, 283, 1990. 16. Elsner, P., Wilhelm, D., and Maibach, H.I., Mechanical properties of human forearm and vulvar skin, Br J Dermatol, 122, 607, 1990. 17. Manschot, J.F. and Brakkee, A.J., The measurement and modelling of the mechanical properties of human skin in vivo. I. The measurement. II. The Model, J Biomech, 19, 511, 1986. 18. Piérard, G.E. and Lapière, C.M., Structures et fonctions du derme et de l’hypoderme, in Précis de cosmétologie dermatologique, Pruniéras, M., Ed., Masson, Paris, 1981, chap. 2. 19. Agache, P., Varchon, D., Humbert, P., and Rochefort, A., Non-Invasive Assessment of Biaxial Young’s Modulus of Human Skin In Vivo, paper presented at the 9th International Symposium on Bioengineering and the Skin, Sendai, Japan, October 19–20, 1992.
20. Barel, A.O. and Clarys, P., Noninvasive Measurements of the Viscoelastic Properties of the Human Skin with the Suction Method, paper presented at the 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 13–16, 1990. 21. Malm, M. and Serup, J., In Vivo Skin Elasticity of Different Body Regions: The Vertical Vector, paper presented at the 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 13–16, 1990. 22. Anfossi, T., Bosio, D., and Emanuelle, G., Influence of Environment Factors on Skin Elastometric Patterns, paper presented at the 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 13–16, 1990. 23. Barbanel, J.C., A Suction Method for Determining the Direction of Langer’s Lines, paper presented at the 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 13–16, 1990. 24. Nishimura, M. and Tsuji, T., Measurements of Skin Elasticity with a New Suction Device: Relation to Age, Sex and Anatomical Regions in Normal Skin and Its Comparison with Some Diseased Skin, paper presented at the 9th International Symposium on Bioengineering and the Skin, Sendai, Japan, October 19–20, 1992. 25. Barel, A.O. and Clarys, P., In Vivo Evaluation of Skin Ageing. Relations between Visco-Elastic Properties and Skin Surface Parameters, paper presented at the 9th International Symposium on Bioengineering and the Skin, Sendai, Japan, October 19–20, 1992. 26. Courage-Khazaka, scientific information obtained from G. Khazaka, Courage-Khazaka Electronic GmbH, Köln, Germany, 2004. 27. Schlangen, L.J.M., Brokken, D., and Van Kemenade, P.M., Correlations between small aperture skin suction parameters: statistical analysis and mechanical model, Skin Res Technol, 9, 122, 2003. 28. Agache, P., Monsieur, C., Lévêque, J.L., and de Rigal, J., Mechanical properties of Young’s modulus of human skin in vivo, Arch Dermatol Res, 269, 221, 1980. 29. Hendriks, F.M., Brokken, D., Van Eemeren, J.T.W.M., Baaijens, F.P.T., and Horsten J.B.A.M., A numericalexperimental method to characterize the non-linear mechanical behaviour of human skin, Skin Res Technol, 9, 274, 2003. 30. Diridollou, S., Patat, F., Gens, F., Vaillant, L., Black, D., Lagarde, J.M., Gall, Y., and Berson, M., In vivo model of the mechanical properties of the human skin under suction, Skin Res Technol, 6, 214, 2000. 31. Van Den Eynde, A., Invloed van de leeftijd op de viscoelastische eigenschappen van de huid, B.Sc. thesis, Vrije Universiteit, Brussels, Belgium, 1990. 32. VanWonterghem, M., Mechanische eigenschappen van de menselijke huid: invloed van verschillende factoren, B.Sc. thesis, Vrije Universiteit, Brussels, Belgium, 1991. 33. Dobrev, H., Use of Cutometer to assess epidermal hydration, Skin Res Technol, 6, 239, 2000.
Suction Chamber Method for Measurement of Skin Mechanics: The New Digital Version of the Cutometer
34. Berardesca, E., Borroni, G., Borlone, R., and Rabbiosi, G., Evidence for elastic changes in aged skin revealed in an in vivo extensometric study at low loads, Bioeng Skin, 2, 261, 1986.
591
35. Sparavigna, A. and Galbiati, G., Strain-Time Curve in the Assessment of Topical Tretinoin as an Antiageing Agent, presented at the 8th International Symposium on Bioengineering and the Skin, Stresa, Italy, June 13–16, 1990.
Chamber Method for 68 Suction Measurement of Skin Mechanics: The DermaLab Gary Lee Grove, John Damia, Mary Jo Grove, and Charles Zerweck cyberDERM, Inc., Broomall, Pennsylvania
CONTENTS 68.1 Introduction............................................................................................................................................................593 68.2 Basic Description of the DermaLab Suction Cup Hardware ...............................................................................593 68.3 The DermaLab Suction Cup as a Noncomputerized Stand-Alone Device ..........................................................595 68.4 The Computerized DermaLab Suction Cup with cyberDERM Software ............................................................596 68.5 Validation Study of the Computerized DermaLab Suction Cup ..........................................................................596 68.6 Effects of Repetitive Cycles ..................................................................................................................................597 68.7 Typical Results from Studies of Human Volunteers.............................................................................................598 References .......................................................................................................................................................................599
68.1 INTRODUCTION Whether the skin feels soft, supple, compliant, firm, etc., to the touch is directly related to its mechanical properties. Probably the most widely recognized change in the mechanical properties of the skin is its age-related loss of elasticity. It is also generally appreciated that more subtle changes, such as increased skin stiffness, can provide important clinical clues for monitoring the progression of certain systemic diseases such as scleroderma. Thus, it is not surprising that over the years, a wide variety of devices have been created to noninvasively describe the mechanical properties of the skin. Despite obvious differences in design and execution, the underlying principles of all of these devices are generally much the same, i.e., load the skin surface in a standard manner and measure the resulting deformations over time. Thus, we have devices that pull, push, tug, twist, compress, wiggle, and impact the skin. In this chapter we will describe the DermaLab suction cup, which is manufactured by Cortex Technology (Hadsund, Denmark). Some aspects of an earlier version of this device have been previously covered by Serup1 and Pedersen et al.2 Since that time a number of minor improvements have been made to the basic device by Cortex Technology, and a dedicated application program has been written by cyberDERM, Inc. (Broomall, PA) that allows
the DermaLab suction cup to be computerized. Both the stand-alone and the computerized versions will be reviewed in this chapter.
68.2 BASIC DESCRIPTION OF THE DERMALAB SUCTION CUP HARDWARE The DermaLab suction cup consists of a light plastic probe (Figure 68.1) that forms a closed chamber when attached to the skin surface using double-sided sticky tape. Within the probe chamber there are two narrow beams of light that are run at different heights parallel to the skin surface and serve as elevation detectors. A computer-controlled vacuum pump is used to progressively increase the suction within the chamber. Since the time at which each of the light beams is blocked can be electronically detected, the amount of suction in kilopascals (kPa) required to lift the skin to that point can be easily determined and electronically recorded by the computer. Figure 68.2 shows a schematic diagram consisting of four panels that portray the sequence of events that occurs during a measurement procedure. When the probe is first placed on the skin, its surface will be flush across the opening of the suction chamber, as shown in Figure 68.2a. The progressive increase in suction will cause the skin to be drawn into the chamber (Figure 68.2b), and eventually 593
594
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 68.1 The lightweight plastic probe of the DermaLab suction cup.
the skin will be lifted to the point where the light beam of the lower elevation detector (level 1) will be blocked, as shown in Figure 68.2c. When this occurs, the vacuum at that point will be electronically recorded. If the pump is allowed to continue, sufficient suction will be created to additionally lift the skin to the point where the light beam of the upper elevation detector (level 2) is blocked as well, as shown in Figure 68.2d. The amount of suction required to achieve this higher level is also electronically recovered. Since the positions of the lower and upper elevation detectors are fixed by the geometry of the probe, the strain at each level is known. In the standard probe these are equivalent to the skin being stretched to 2 and 12% extensions, respectively. The suction pressure in kilopascals
To vacuum pump
Suction cup
Elevation detectors
Skin
required to lift the skin to each point provides a measure of the stress at that level. This means that the DermaLab suction cup can calculate the mechanical properties of the skin based on Hooke’s law,3 which was first formally stated by the English mathematician in 1660 as an anagram in Latin. Since no one managed to break his code, years later he translated the phrase to be “The power [sic] of any springy body is in the same proportion with the extension.” Today, Hooke’s law3 in its simplest form is generally stated that “the strain of any material is proportional to the load applied to it.” This means that if an applied tensile stress of x units will stretch a specimen of 1 unit, then a stress of 1.5x will produce an elongation of 1.5 units, a stress of 2x will produce a deformation of 2 units, and so on. The factory manual states that if certain basic assumptions are made, then the stiffness of the skin or its Young’s modulus (E) can be calculated from this stress–strain relationship as follows:
Δx = ψ ⋅ p ⋅
where x = deviation middle of surface, ψ = elasticity constant for measured object estimated from civil engineering tables, p = negative pressure, r = radius of surface measured, and s = thickness of object measured.
Elevation detectors
Skin
(c)
To vacuum pump
Elevation detectors
Skin
To vacuum pump
Suction cup
(a)
Suction cup
r4 Δp ⇒ E = 0.3125 ⋅ E ⋅ s3 Δx
(b)
FIGURE 68.2 Operating principles of the DermaLab suction cup.
Suction cup
To vacuum pump
Elevation detectors
Skin
(d)
Suction Chamber Method for Measurement of Skin Mechanics: The DermaLab
595
TABLE 68.1 Displayed Parameters When DermaLab Suction Cup Is Used as a Stand-Alone Device
ELAST RES1
RESN TIM1 TIMN
Description
Comments
The calculated elasticity modulus based on the first measurement cycle The differential vacuum necessary to elevate the skin from detection level 1 to detection level 2 on the first suction cycle
May not be appropriate for human skin measurements Very useful in describing stiffness of skin, but actual value depends upon specific probe configuration Same as RES1 Misleading due to suction curve not being linear Same as TIM1
Same as RES1 but on last suction cycle The time required for the skin to be raised to detection level 2 on the first suction cycle Same as TIM1 but on last suction cycle
In these calculations ψ is set by the manufacturer to be 0.5, r is defined by the chamber geometry as 5 mm, and s is set by the manufacturer to a standard skin thickness of 1.0 mm. We agree with Serup1 that the calculation of a Young’s modulus in the above manner is not normally considered to be appropriate for complex composite structures such as the skin. Not only is the skin highly anisotrophic and viscoelastic, but it is also composed of various layers, each of which has a different mechanical resistance. Moreover, the actual values for both (elasticity constant) and s (skin thickness) are not known and can vary considerably, depending upon the age of the patient and the anatomical region being measured. Nevertheless, we feel that this parameter must be discussed in this chapter since it is the ELAST parameter that is provided as part of the printed output of the DermaLab suction cup when used in the noncomputerized stand-alone mode.
68.3 THE DERMALAB SUCTION CUP AS A NONCOMPUTERIZED STAND-ALONE DEVICE Table 68.1 provides a listing of the various parameters that appear on the liquid crystal display of the DermaLab when used as a stand-alone instrument not interfaced to a PC. These include ELAST, which refers to the calculated elasticity modulus (E), which in our opinion should be considered to be of only limited value due to the reasons described above. We also have some reservations as to how useful the two elevation time values (TIM1 and TIMN) can be. These are the time to reach the upper detector at level 2 on the first cycle (TIM1) and the last or nth cycle (TIMN). The DermaLab suction cup does not have a vacuum tank or barometric control of the buildup of negative pressure. Instead, this device is designed with a specially constructed tube system so that the vacuum is slowly applied over a period of perhaps 30 to 60 seconds, until a plateau
is reached where the negative pressure within the chamber is maintained in equilibrium with the capacity of the pump. With stiffer skin it will take longer for this equilibrium to be achieved, which is the notion underlying the use of the elevation times TIM1 and TIMN. In the more recent version of the DermaLab suction cup, a switch has been provided to further control the flow rate of the specialized vacuum system. For most skin sites, such as the arms and legs, the switch should be set at NORMAL, which is the default setting. With skin that is very loose, such as the face and neck, it is recommended to use the REDUCED setting, which according to the factory manual reduces the rate at which the suction pressure is applied. At the time of the original paper by Serup1 the actual shape of the negative pressure curve in the chamber was not exactly known but was supposed to be nonlinear. We have been able to acquire the profile of the negative pressure curve by incorporating a digital manometer into the suction line and placing the suction cup probe on a rigid, nondeformable surface. As shown in Figure 68.3, the time required to reach equilibrium is approximately the same, 60 Suction pressure in KPa
Parameter
Normal 50 40 30 Reduced 20 10 0
0
5
10
15 20 25 30 Seconds with pump on
35
40
45
FIGURE 68.3 Negative pressure curves developed by DermaLab suction pump at either normal or reduced settings.
Handbook of Non-Invasive Methods and the Skin, Second Edition
but the level at which this occurs is considerably lower for the REDUCED setting than for the NORMAL setting. Moreover, it is quite clear that the shape of both curves is clearly nonlinear. This means that although a stiffer material will take longer to be stretched to the upper level, there is no simple relationship between that time and the stiffness of that material. In other words, a doubling in elevation time does not mean a doubling in skin stiffness. Indeed, the shape of the pressure curve clearly indicates that as time increases, the degree to which the actual stiffness is overestimated by relying on the elevation time values will increase as well. This leaves only RES1 and RESN as possible measures of stiffness that are available when the DermaLab is used as a stand-alone instrument. Both values indicate the differential vacuum necessary to elevate the skin from detection level 1 to detection level 2 in either the first or the nth and last cycle. Since these values are central to the computerized version of the DermaLab, we will deal with them in more detail in the next section.
68.4 THE COMPUTERIZED DERMALAB SUCTION CUP WITH CYBERDERM SOFTWARE As is the case with all of the DermaLab modules, the suction cup can also be configured to run in continuous mode, so that a stream of data flows via an RS-232 interface to a PC where it can be processed using specialized data acquisition programs. Actually, two data outputs are provided at a sampling rate of 20 per second, with the first being the status of the lower-level detector (ON or OFF) and the second being the negative pressure within the chamber at that point in time in kilopascals. The pressure information can be plotted in real time on a strip chart recorder so that the operator can watch the development of the suction curve in the suction chamber. By monitoring the status of the lower-level detector, the pressure developed at the point in time when the light beam of the elevation detectors is broken can easily be captured. Since the pressure pump is automatically turned off when the upper detector light beam is broken, there is no need to monitor the status of this detector, because the pressure profile will drop off immediately. Thus, it is a relatively simple matter to compute the differential vacuum necessary to elevate the skin from the lower detector beam at level 1 to the upper detector beam at level 2, which is equivalent to RES1 or RESN of the stand-alone instrument. Again, RES1 and RESN refer to values that are respectively derived from either the first or the nth and last cycle. Moreover, it is also a simple matter to provide the actual stress values in terms of pressure at both detector levels, where the strain is known due to the probe’s geometry. In the case of the standard probe, this is equiv-
alent to 2% extension at the lower level and 12% extension at the higher level. Although it would also be possible to easily extract the time required to reach the upper detector level from the time base of the pressure curve, to yield an output identical to TIM1 or TIMN of the stand-alone device, we do not think that measurements based on this parameter are easy to compare due to the nonlinear pressure curve, as previously discussed.
68.5 VALIDATION STUDY OF THE COMPUTERIZED DERMALAB SUCTION CUP We would like to emphasize that the underlying physical principle of the DermaLab suction cup is Hooke’s law.3 To determine how well the computerized device adhered to Hooke’s law,3 we used a 0.012-inch-thick latex sheet that could be stretched to different degrees before the probe was attached. By marking two reference lines 2 inches apart upon the sheet, with no tension upon it, one could achieve extensions of 25, 50, 75, and 100% by stretching these lines so that the gaps between the reference lines were 2.5, 3.0, 3.5, and 4.0 inches, respectively. The DermaLab suction cup probe could then be reattached to the latex sheet that has been stretched to various degrees. Figure 68.4 displays the results obtained for the standard probe. Note that the pressure values recorded for both the upper and lower detectors are highly correlated with the degree of stretch, which is in perfect accordance with Hooke’s law,3 with R2 being nearly 1 in both cases. It is also clear that the pressure differential between the two detectors remains constant Standard suction cup 40 35
y = 0.1665x + 20.028 R2 = 0.9986
30 25 KPa
596
20 15 10
y = 0.1452x + 3.1117 R2 = 0.987
5 0
0
10
20
30 40 50 60 70 80 Percent extension of latex sheet
90
100
FIGURE 68.4 Scatter plot of the amount of suction required to lift a reference latex sheet to block either the lower or upper level detector light beams, with the sheet stretched to various degrees. Note that results are in excellent agreement with those predicted by Hooke’s law.
Suction Chamber Method for Measurement of Skin Mechanics: The DermaLab
597
“Sceleroderma” suction cup 40
35
35
30
30
25 20 15
20
y = 0.1847x + 9.44 R2 = 0.9861
15
10
10
y = 0.1506x + 1.7517 R2 = 0.9967
5 0
y = 0.1753x + 21.175 R2 = 0.9929
25
y = 0.1554x + 11.513 R2 = 0.9936
KPa
KPa
“Facial” suction cup
40
0
10
20
30 40 50 60 70 80 Percent extension of latex sheet
90
5
100
0
0
10
20
30 40 50 60 70 80 Percent extension of latex sheet
90
100
FIGURE 68.5 The results for the scleroderma and facial probes are also in excellent agreement with those predicted by Hooke’s law.
throughout this range of extension, which justifies the use of the RES1 or RESN value as a measure of the inherent stiffness of that material, provided that we remain within the linear portion of the stress–strain curve, which is clearly true in this case. It should be pointed out that special probe configurations exist. In the scleroderma version, the upper and lower elevation detectors are moved closer to the skin surface, which allows one to measure tight skin without excessive stretching. In the facial version, the lower elevation detector is moved farther away, since the skin may be so loose and soft that it may have already moved into the chamber and blocked the lower detector before the suction is applied. As shown in Figure 68.5, regardless of the configuration, Hooke’s law3 is still obeyed. Indeed, with this type of graphic presentation it is easy to appreciate that the upper elevation detector of the scleroderma probe and the lower elevation detector of the facial probe are in approximately the same relative position above the skin surface, as shown by the fact that their plots are quite similar. Again, note that the pressure differential between the lower and upper detectors for both probes remains constant through this range of extensions, as was the case for the standard probe. Although the pressure differentials differ among the various probes, this is due to the spacing of the lower and upper elevation detectors. Indeed, rather than computing a Young’s modulus, we think it more appropriate to describe the results obtained with the DermaLab suction cup in terms of a stiffness index, which is simply delta pressure in KPa/delta distance in mm
This means that skin that is firm and taut will have a much higher stiffness index than skin that is loose and saggy. In this validation study, the material being studied with the different probe configurations was the same latex sheet. Although the positions of the lower and upper elevation detectors in the standard, scleroderma, and facial suction cup probes do differ considerably, as shown in Table 68.2, the stiffness index for the latex sheet was the same. Moreover, if we increase the thickness of the latex sheet being measured, there will be a corresponding increase in the stiffness index, as one would predict. Although the focus of this chapter is skin biomechanics, one should also realize that the DermaLab suction cup can be easily employed to measure the material properties of various elastic sheets in a meaningful manner.
68.6 EFFECTS OF REPETITIVE CYCLES With the DermaLab suction cup one can program the suction pump to do either a single on–off cycle or a series of repetitive cycles with an intervening resting time, which
TABLE 68.2 Stiffness Index of Latex Sheet Reference Standard as Measured with Three Different Suction Cup Configurations Probe Type Standard Scleroderma Facial
Mean
±
S.D.
12.07 12.44 12.18
± ± ±
0.66 1.28 0.45
598
Handbook of Non-Invasive Methods and the Skin, Second Edition
40
40 Latex
30 25 20 15 10 5
Skin
35
Level 2 Level 1 Suction pressure in KPa
Suction pressure in KPa
35
Level 2 Level 1
30 25 20 15 10 5
0
0 1
2
3 4 Number of cycles
5
1
2
3 4 Number of cycles
5
FIGURE 68.6 Effects of repetitive cycling on the amount of suction required to stretch latex or skin to block the elevation detectors at levels 1 and 2 with the DermaLab suction cup.
can be set to be from 1 to 10 seconds. The program can be set so that the device will automatically stop cycling after a predetermined number of cycles have been completed. During the validation studies we found that repetitive measurements of the latex sheet typically gave the same values for each cycle regardless of the number of cycles, as shown in Figure 68.6. In striking contrast, when measuring human volunteers we found that the suction required to lift the skin progressively decreases with each cycle, but will eventually reach an asymptote. This warming-up phenomenon is more or less obvious at different anatomical sites and in some individuals. Currently we are investigating how to best express these results and what factors are responsible for this type of behavior. Such studies can only be done with the computerized version of the device, which provides a complete data set for each and every cycle over time. With the stand-alone version only the information of the first and last cycles is provided, and these values provide only the pressure differentials (RES1 and RESN), not the individual stresses for levels 1 and 2 for each cycle.
68.7 TYPICAL RESULTS FROM STUDIES OF HUMAN VOLUNTEERS Although the probe is very lightweight (approximately 10 g), the tethered wires and pneumatic tubing, if not properly supported, will tug on the skin and alter the biomechanical properties being measured. This is extremely important when attempting to measure where the skin is lax, such as under the eye (Figure 68.7).
FIGURE 68.7 Although the probe is lightweight, it is important, especially when the skin is lax, to support the probe so that it does not tug on the skin.
Figure 68.8 summarizes the results from a small crosssectional study involving 20 healthy normal individuals, with half of them between 20 and 30 years and the other half between 50 and 60 years of age. Striking differences were found to exist at various regions of the face, especially under the eye. The facial skin of older individuals was typically less stiff than that of younger individuals in these regions, due to loss of elasticity and increased sagging. We have also found that the DermaLab suction cup provides clinically relevant data on the mechanical properties of the skin, which may help predict the severity and progression of a number of diseases, such as scleroderma. We are especially impressed with how well the DermaLab suction cup has held up under hard use in such clinical trials. A large part of its robustness stems from there not being any moving parts in the probe.
Suction Chamber Method for Measurement of Skin Mechanics: The DermaLab
REFERENCES
40 35 20–30 years
“Stiffness index”
30
50–60 years 25 20 15 10 5 0
599
Forehead
Cheek
Under eye
FIGURE 68.8 Cross-sectional survey in which the stiffness index was measured at various regions of the face of younger and older adults with a computerized DermaLab suction cup.
1. J. Serup. Hardware and measuring principles: DermaLab. In Bioengineeering of the Skin: Skin Biomechanics, P. Elsner et al, Eds. CRC Press, Boca Raton, FL, 2002 pp. 117–121. 2. L. Pedersen, B. Hansen, and G.B.E. Jemec. Mechanical properties of the skin: a comparison between two suction cup methods. Skin Res Technol 9: 111–115, 2003. 3. www.RobertHooke.com.
Measurement of 69 Twistometry Skin Elasticity Pierre G. Agache Department of Functional Dermatology, University Hospital, Besançon, France
CONTENTS 69.1 Introduction............................................................................................................................................................601 69.2 Technical and Theoretical Considerations ............................................................................................................602 69.2.1 Equipment ..................................................................................................................................................602 69.2.2 Mechanical Testing ....................................................................................................................................602 69.2.3 Mechanical Parameters..............................................................................................................................603 69.2.3.1 Elasticity Parameters ..................................................................................................................603 69.2.3.2 Viscosity Parameters ..................................................................................................................604 69.2.3.3 Skin Rheological Model.............................................................................................................604 69.2.4 Requirements for Test Validity and Correct Interpretation.......................................................................605 69.2.4.1 Geometry of the Applied System ..............................................................................................605 69.2.4.2 Absence of Pressure Perpendicular to Skin Surface .................................................................605 69.2.4.3 Attachment of Disc to the Skin .................................................................................................605 69.2.4.4 Rate of Torque Application ........................................................................................................605 69.3 Skin Mechanical Aging .........................................................................................................................................605 69.3.1 Intrinsic Aging ...........................................................................................................................................605 69.3.2 Actinic Aging.............................................................................................................................................607 69.4 Stratum Corneum and Skin Biomechanics ...........................................................................................................608 69.5 Medical Applications .............................................................................................................................................609 69.5.1 Effect of Topical Retinoic Acid.................................................................................................................609 69.5.2 Scleroderma ...............................................................................................................................................609 69.5.3 Inherited Connective Tissue Diseases .......................................................................................................610 69.6 Conclusion .............................................................................................................................................................610 Acknowledgment.............................................................................................................................................................610 References .......................................................................................................................................................................610
69.1 INTRODUCTION For the two last decades, numerous attempts have been made to non-invasively assess the human skin mechanical behavior in vivo. This was felt to be a useful way to allow a more precise follow-up of the numerous diseases or skin states characterized by an abnormal skin induration of softening and to estimate the efficacy of treatments or cosmetics. To attain this goal, devices were constructed that operate either by inducing a skin deformation and recording the resisting force or by putting a load and assessing the resulting deformation. Four directions of loads are conceivable: a vertical pressure, a vertical suction, a linear horizontal traction, and a torsion in the horizontal plane.
Mechanical stimuli perpendicular to the skin surface have the disadvantage of involving the subcutis, at least in part. This layer has wide differences in thickness or in fat content, and consequently has widespread mechanical properties. Stimuli in the horizontal plane, on the other hand, use probes stuck on the skin surface and, if the displacement is small, may be expected to involve only the epidermis and dermis. Unidirectional stresses should consider the skin mechanical anisotropy. But Langer’s lines are not easy to find, and most of them are oblique to the body or limb axis, and thereby difficult to comply with.1 Also, the sampled area is poorly delimited with such devices. All these reasons prompted some authors to use torsional devices made of a central rotating disc and a 601
602
Handbook of Non-Invasive Methods and the Skin, Second Edition
peripheral fixed ring, thus allowing a skin narrow annulus to be twisted. Vlasblom2 from Utrecht was the first to use such a device and made a theoretical study of forces and deformations implicated. Finlay3,4 from the Strathclyde group (Glasgow) extended the investigation and applied the technique to the assessment of changes in mechanical behavior with aging. Since then, the L’Oreal group (Paris), conducted by Jean-Luc Lévêque, has been using torsional devices extensively and has contributed a great deal to the development of the technique.
69.2 TECHNICAL AND THEORETICAL CONSIDERATIONS Before using or interpreting the results of a torsional test, some elements of the relationships between stress and deformation should be recalled, as well as the requirements necessary for an experiment to be mechanically valid.
69.2.1 EQUIPMENT Torsional equipment acts through a disc glued to the skin, which is rotated by a motor powered by a controlled voltage, thereby loading the peripheral skin with a torque, the value of which can be adjusted. Under this torque the skin glued under the disc moves with it, supposedly without any brake from the subcutaneous tissue. The skin around the disc is elongated in a twisting way. This is the mechanical behavior of this part of the skin that is assessed. There are two types of twists to be considered, whether there is or is not a peripheral guard ring concentrical to the disc, also glued to the skin and immobile during the disc rotation, thereby delimiting an annular area of skin submitted to elongation (Figure 69.1). When there is no guard ring,5 the skin peripheral to the disc is implicated up to an unknown distance. As it is much less extensible than the subcutis, the latter will provide for most of the twisting strain. By contrast, when there is a guard ring the shape and limits of the twisted area are known, the sliding of the skin over the subcutis is limited, and the essential part of the twisting strain takes place within the skin itself. Finlay’s4 and Jaskowski and Maceluch’s6 devices were constructed to give repeated twists at adjustable frequencies and adjustable increasing or decreasing rates. The first one was also equipped by a strain gauge, which recorded the torque generated by the skin under a constant twist and assessed force relaxation. Lévêque and De Rigal’s7 device (Figure 69.2) has a guard ring integrated to the body of the apparatus. It delivers an analogue signal directly proportional to the disc rotational angle. In a
Guard ring
Disc
Disc
A
B
FIGURE 69.1 Skin deformation produced at the periphery of a rotating disc with (A) and without (B) a guard ring.
newer version the signal is digitalized and a microprocessor both computes the main parameters and controls the measurement phases. The applied torque can be chosen between 4 and 57 mN·m, the width of the crown of skin submitted to torque is either 1, 3, or 5 mm, and the disc radius is 18 or 25 mm. The equipment is now commercially available (Dermal Torque Meter, Diastron Ltd., Andover, U.K.).
69.2.2 MECHANICAL TESTING Finlay’s4 protocol included a progressive increase in disc rotation, then a standstill to assess the skin torque relaxation, then a progressive return of the disc position to zero. He recorded both the strain and the resisting force induced in the skin during the first run and the following ones. Jaskowski and Maceluch6 used a repeated torque application up to a resonance frequency and recorded the torque
A/D conv. C M
Current supply
P D
G
Output
FIGURE 69.2 Lévêque’s twistometer diagram. (C) Rotational sensor, (M) motor, (D) disc, (G) guard ring, (P) microprocessor. Both guard ring and disc are removable.
Twistometry Measurement of Skin Elasticity
Creep
603
Recovery
L
F
Transient
θ°
Stationary
E
I α
UR
L0
UE
(a) Time 0
60s
F
FIGURE 69.3 Skin angular deformation (Θ) vs. time upon application of a constant torque. Ue: immediate deformation, Ur: immediate recovery upon torque switching off.
oscillating amplitude, the optimum frequency, and the attenuation of torque by the skin. The simple way to use torsional tests is to apply a constant torque for a couple of seconds and record the skin angular deformation (Figure 69.3). Upon torque application there is an immediate elastic deformation (Ue), followed by a creeping viscoelastic deformation (Uv). Torque suppression is associated with an immediate recovery (Ur), which is always incomplete. This is the way Lévêque and De Rigal’s8 device works. One or several runs can be programmed and preconditioned.
69.2.3 MECHANICAL PARAMETERS In vivo mechanical tests on the skin have two purposes. The first and main one is to quantitatively assess changes that are usually detectable by palpation but not measurable otherwise. The second aim is to get access to the skin intrinsic structure as far as mechanical components are concerned, and to the structure-function relationships of these components. The absolute parameters concerning the strength of skin elasticity and viscosity need to be derived from the assumptions that it is a homogeneous layer and is uniform in thickness, which is obviously wrong, but is a very rewarding approximation. 69.2.3.1 Elasticity Parameters In simple elongation tests the well-known equation of Young’s modulus (E) is a commonly used way to express the stiffness at the elastic phase: E = σ(1 – ν)/ε
(69.1)
where σ is the stress (ratio of force to the section area submitted to force), ε is the strain (elongation-to-initial sample length ratio), and ν is the ratio of relative narrowing to strain (Poisson’s ratio). In a torsional experiment the deformation is more complex because elongation is replaced by shear and is
L E
I
α
L0 (b)
FIGURE 69.4 A diagram of skin annulus deformation in torsional experiments. (a) Deformation homogeneous through thickness. (b) Deformation predominant on the skin surface. Deformation gradients are supposed linear: (L0) initial length (width of skin annulus); (L) length after elongation; (I) internal aspect of skin annulus, facing disc; (E) external aspect of skin annulus, facing guard ring, (F) force; (α) deformation angle.
rotational. The skin can be deformed through its full thickness to the same extent (Figure 69.4). This occurs when the force is high enough to act in depth while only applied on the surface, and when the skin annulus is wide enough to allow the force gradient within the skin to fully reach the deeper layers. In that case, the skin is supposedly isotropic in torsion tests and the mechanical parameter is the shear modulus μ (Lamé’s coefficient) given by the formula μ = σ(1 – ν)/α
(69.2)
where α is the deformation angle, ν Poisson’s ratio, and σ the stress as calculated by the ratio of force (torque/radius of disc) to the area submitted to torsion (skin thickness × disc perimeter). Agache et al.9 have proposed the following formula:
μ=
M 2 πe r1 r2 Θ
(69.3)
where M is the torque momentum, r1 the disc radius, r2 the guard ring inner radius, e the skin thickness, and θ the rotational angle of the disc at equilibrium. The Lamé’s coefficient is usually close to 0.4 E; therefore, it is proposed to compute E by simply dividing the above formula by 0.4. Hence,
604
Handbook of Non-Invasive Methods and the Skin, Second Edition
E=
1 M × 0.4 2 πe r1 r2 Θ
K1
(69.4) K0
η1
If only E variation is considered, the same formula gives the following equation by differentiation:10 ΔE/E = –(ΔUe/Ue + Δe/e)
(69.5)
η0
Maxwell model Kelvin voigt model
FIGURE 69.5 Bürger’s model of skin mechanical behavior, and corresponding extension (k0, k1) and viscosity (η0, η1) parameters
When the torque is low and mostly when the skin annulus is narrow, the superficial layers of the skin are more implicated than the deeper ones because the tangential force is applied on the skin surface (Figure 69.4). In that case the deformation has two components, a visible and measurable one in the superficial plane, and an invisible and immeasurable one in the vertical plane (the shear strain gradient in depth). It could be possible, from the deformation seen at the skin surface, to calculate both of these strain components and, consequently, the relevant, more complex shear coefficient and corresponding Young’s modulus equivalent. But in the present state of experimental conditions, which are submitted to large variability in the results, obtaining such precision seems unrealistic, or at least unnecessary.
This law was also validated for the recovery part of the curve, and allowed a correct determination of the immediate recovery.
69.2.3.2 Viscosity Parameters
69.2.3.3 Skin Rheological Model
The experimental protocol designed by Finlay4 includes a typical stress–relaxation step. The observed half-time of skin torque relaxation during this step is inversely related to skin viscosity for a given initial torque. Accordingly, this time would change for another torque value at the start of the relaxation step. As in an elongation test, in torsional experiments where a constant torque is applied, the deformation vs. time, as shown in Figure 69.3, can be described by the following equation (Vlasblom):
Both for better intuitive understanding of what occurs within the skin during traction and easier computing of mechanical parameters, rheological models of skin have been proposed. Wijn12 used Burger’s model for their uniaxial elongation experiments (Figure 69.5). Pichon et al. used the same model for skin torsional experiments under a constant force and using the above-quoted creep law (Equation 69.7). Accordingly the equation for the model is
U(t) = Ue + Uv (1 – e–t/τ) + At
In an experiment on six subjects with a 1-mm-wide skin annulus, they found m = 0.33 ± 0.03 and 0.335 ± 0.06 for 9 × 10–3 and 12 × 10–3 Nm torques, respectively. Increasing the width of the annulus did not change the m value, although increasing the interindividual variation. Accordingly, they proposed the following equation for the creep: ε = Uv (1 – e–t/τ) + At1/3
ε( t ) =
(69.6)
where Ue is the immediate deformation, Uv the viscous deformation (transient creep), and At a linear deformation following a longer time of torque application (stationary creep). Vlasblom showed that in the stationary creep the deformation is not proportional to time and suggested a constant term should be added to At. As this term is small compared to elastic deformation, he thought it possible to neglect it. Pichon et al.,11 using the finite differences method, undertook to determine the creep law without any influence of Ue and proposed the following equation for the creep deformation ε: (69.7)
(
)
(69.9)
The three phases of experimental strain (Ue, transient creep, and stationary creep) correspond to each of the three terms of the equation. The characteristic parameters of the springs (k0, k1) and dashpots (η0, η1) should correspond to those of special structures or arrangements within skin, e.g., the elastic resistance of elastic fibers to elongate or collagen network to deform, and the viscous resistance to displacement. All parameters of Bürger’s model can be obtained from the in vivo experiment, as follows: •
ε = ε0 Atm
σ σ σ t1 3 + 1 – e – tk1 η1 + k 0 k1 η0
(69.8)
Stress: σ = force/area submitted to stress (i.e., disc perimeter × skin thickness)
Twistometry Measurement of Skin Elasticity
•
• • •
605
Young’s modulus E (i.e., first spring strength at the casual level of first spring tension): E = k0 = σ/Ue Second spring strength: k1 = σ/Uv curvilinear (i.e., transient creep) Viscosity associated with the first spring: η0 = σ/Uv t1/3 Viscosity associated with the second spring: η1 = τk1
This model looks satisfactory, but has been used in only one published study.11 Accordingly, in the next sections the calculations relative to creep used only the second member of Equation 69.6.
69.2.4 REQUIREMENTS FOR TEST VALIDITY CORRECT INTERPRETATION
AND
Barbenel and Payne13 in a report to the International Society for Bioengineering and the Skin have presented the requirements for a torsional testing to give interpretable and reliable results. Most of them should be recalled, along with some additional warning.
T (cm. cN) A
B
330
220
A
110
B
Θ° 5
10
FIGURE 69.6 Experimental plot of Ue (Θ°) vs. applied torque (T) under 11.25 kPa pressure (A) and under 24.52 kPa pressure (B) onto skin surface. Higher pressure induced an artifactual curve where the curvilinear relationship between stress and strain was no longer visible. (From Lévêque, J-L, et al., Arch. Dermatol. Res., 269, 221–232, 1980. With permission.)
69.2.4.4 Rate of Torque Application 69.2.4.1 Geometry of the Applied System The area submitted to torque should be delimited by a guard ring in order (1) to allow the skin itself to be deformed and reduce to a minimum the skin sliding over the subcutis, and (2) to calculate the strain (i.e., deformation/initial length). The width of the twisted skin annulus should be narrow enough to prevent torsion of the subcutis. In any case, the disc diameter and inner guard ring diameter should be quoted, and also the torque value. The rotational angle θ should be small (inferior to 10˚), and the radius r of the rotating disc should be large enough to allow the displacement rθ, which is circular, to be considered as linear. 69.2.4.2 Absence of Pressure Perpendicular to Skin Surface As shown in Figure 69.6, such a pressure would change the mechanical behavior of the skin. In practice, the applied pressure, if any, should always be mentioned. 69.2.4.3 Attachment of Disc to the Skin Slipping or deformation of the attachment system should be avoided. Accordingly, the mechanical behavior of the attachment system used should be tested beforehand at the same torque as in the experiment. This was the case for Lévêque and De Rigal’s device, as tested on a steel plate,7,8 and in Finlay’s experiments.4 Anyway, the type of adhesive used should be indicated.
As skin is viscoelastic, any application of the torque at a low rate would allow both an elastic and viscous deformation to take place at the same time. A very low rate (quasi-static experiment) would reduce the viscous resistance to almost zero, and consequently assess only the elastic resistance. On the other hand, a very high deformation rate would overcome viscous resistances lower than the applied stress, and assess the elastic resistance through the measurement of immediate deformation. In any case, the rate of applied torque should be mentioned.
69.3 SKIN MECHANICAL AGING 69.3.1 INTRINSIC AGING The skin mechanical behavior over the life span, as assessed by torsional experiments in vivo, was first studied by Finlay4 in the lateral aspect of the forearm on a 4-mmwide skin annulus using a 15-mm-diameter rotating disc, inducing on the skin an 11-kPa pressure (Figure 69.6). The run consisted of progressively twisting the skin annulus (rotation of the disc 2˚ per second) up to about 20 mN·m torque, then maintaining the same twist level for 1 min before relaxing the twist down to zero at the same rate. During such a run the torque first rises up to a peak, then subsides partially while the deformation is maintained (force relaxation), then falls to a negative value; i.e., the skin would have remained deformed if it had not been forced to come back to baseline.
606
Handbook of Non-Invasive Methods and the Skin, Second Edition
The angular deformation associated with a 2 mN·m torque was felt by the author to be a good parameter of “what can be sensed by palpation.” Other chosen parameters were the ratio peak to final torque (i.e., the end of relaxation) and the time for half relaxation, indicating the relaxation intensity within 1 min, and rate, respectively. Only the angular deformation at low torque was found to significantly decrease with aging, indicating an increase in skin stiffness. Three other runs were done after the first, each one at a 1-min interval, and the same parameters pooled. Again, only the angular deformation at low torque was significantly depressed with aging. Finally, a fifth run in the reverse direction was done 1 min after the fourth run. Only the ratio peak to final torque was found significantly lower with aging (p = 0.05), indicating a decrease in force relaxation within the extended skin during this time interval. As the skin components can be grossly compared to springs and dashpots connected both in parallel and in series (Figure 69.5), the stiffer skin and less relaxing forces inside skin may denote either a weakening of springs or strengthening of dashpots (viscosity), or both. Unfortunately, the panel of subjects comprised a bulk of middle-aged adults and very few children or aged people, which may explain the small number of significant correlations found in this study. Sanders,5 using Vlasblom’s apparatus and formulas, computed the torsional modulus of the outer forearm skin in 19 healthy subjects aged 6 to 61 years. The torque was 0.83 mN·m through an 8.7-mm-diameter disc without any guard ring, so that the twisted area was kept undefined. The pressure on the skin induced by the disc was not specified. The immediate elastic deformation (Ue) was found to increase almost linearly with aging, and the delayed viscoelastic deformation (Uv) seemed to increase only beyond 40 years of age. From these data a torsional elastic modulus of 20 to 100 kPa was computed and found to decrease with aging. As the torque was very weak, only elastic fibers were supposed to be implicated. Accordingly, the decreased skin stiffness with aging was tentatively ascribed to the well-known decay of elastic fibers over the life span. However, as said previously, the absence of guard ring let the subcutis bear the major (if not the entire) part of the deformation. As this layer is loose, this explains the very low Young’s modulus observed and the increase with aging. These observed parameters were probably those of the subcutis, not of the skin. Jaskowski and Maceluch6 investigated the dynamic torsional behavior of forearm, forehead, and abdomen skin in 380 normal subjects aged 5 to 80 years. The crown of skin between disc and guarding was made to vibrate by the rotating oscillation of the disc, up to a maximum amplitude (resonance). The observed parameters were the resistance, the frequency, and the attenuation. The skin stiffness (Nm/rad) rose steadily from 20 to 60 years, then
sharply beyond 60 years, and attenuation (Nm/rad) followed the same increase. The resonance frequency and resistance slightly rose over the life span. An investigation using a torsional device similar to Findlay’s, i.e., with a guard ring, was undertaken by Lévêque’s8 and Agache’s9 groups in a series of experiments in 1980. In the two first studies on 141 subjects in the age range 3 to 89 years, they used a device made of a central disc 25 mm in diameter and a skin annulus 5 mm wide, and exerted a 12.6-kPa pressure on the skin. The protocol consisted of abruptly putting (within 15 ms) a set torque for 2 min on the volar forearm and recording the angular rotation (24 mV per degree). Two stresses were applied at some distance on the same forearm: 9 and 28.6 mN·m. Under the lower stress the immediate deformation Ue decreased by about 30% during the first two decades, kept stable until 60 years, then rose by about 10%. But the product of immediate deformation by half skinfold at the same site (as assessed by a caliper) showed a steady and significant decrease with aging. Surprisingly, at almost any age this product was lower in females than in males, indicating a higher Young’s modulus under this tensile stress, and accordingly a stiffer skin. Under the higher stress the immediate deformation Ue strongly decreased until the age of 30, then rose moderately, while the Young’s modulus, as calculated by the above-quoted formula, slightly decreased to a minimum in the 20s, then rose progressively with aging. Again, with this high torque the Young’s modulus, i.e., the skin stiffness, was greater in females in all age classes beyond 20 years. The absolute values of this parameter lie in the same range for both stresses, between 0.6 and 2.9 MPa, but they were statistically higher with the lower torque. The direction of the difference would favor a technical origin rather than a physiological one, as the higher torque was too slight to induce any damage to the skin, and anyway gave no unpleasant sensation upon application. This work was undertaken again 9 years later14 using the same torsional equipment, but owing to the newly available technique of assessing skin thickness by ultrasound high-resolution imaging (25 MHz). The measurements were done on the volar forearm set in the same position as in the previous paper, on a panel of 123 subjects. Ultrasound studies showed a thinner skin before 10 years and after 70 years of age, with nonsignificant variation in between and a thickness significantly greater (+16%) in males than in females in the whole range of ages. The skin annulus submitted to twist was reduced to 3 mm wide in order to minimize the possible implication of subcutis. Also, the applied torque values were reduced to 2.3 and 10.4 mN·m, as compared to 9 and 28.6 mN·m in the previous paper, using a central disc of 18 mm diameter instead of 25 mm. No significant change over the life span or differences between sexes were found for either immediate extension or creep. Only did the product
Twistometry Measurement of Skin Elasticity
607
1.00
Elasticity: UR/UE
0.90 0.80 0.70 0.60 0.50 5
15
25
35
45 55 Age
65
75
90
FIGURE 69.7 Skin elasticity (Ur/Ue) over the life span with 2.6 mN/m torque (dotted line) and 10.4 mN/m torque (dashed line). Bars indicate SEM. (From Escoffier C, et al., J. Invest. Dermatol., 93, 353–357, 1989. With permission.)
Ue × thickness significantly (p < 0.01) decrease by about 85% beyond 70 years of age. By contrast, the immediate recovery steadily decreased with aging and also the ratio Ur/Ue (Figure 69.7). This trend was highly significant (p < 10–4). It confirms a current observation that skin folding keeps longer as age advances. The relaxation time of the creep decreased with aging; this was only significant with the higher torque. That means a less viscous dermis with advancing age, data in accordance with the well-known decline in mucopolysaccharide content.15 The results of all above-quoted studies, as presented in Table 69.1, show an agreement for a decreased skin immediate extensibility with aging, mostly beyond 60 years, and an increase in the viscous delayed extension. As the skin thickness decreases in the 60s, the result is an increase of the elastic modulus with aging beyond this age. Only Sander’s data are discrepant with this trend. As stressed previously, the main reason seems to lie in the absence of guard ring, and this experiment was probably an assessment of the subcutis distensibility.
Other discrepancies include the force relaxation time, i.e., the ability of structural elements inside the skin to move relative to each other. As this time is reduced, as shown in Agache’s,9 Lévêque’s,16 and Escoffier’s14 experiments, their mobility would increase with aging. The reverse was observed by Finlay.4 An explanation could lie in the difference in the type of experiments. While Finlay investigated the mobility of elements during a maintained stretch (for relaxation time), the two other papers dealt with the mobility during a maintained stress (deformation time). The former would implicate the strength of an interior spring attempting to recoil against viscosity, and the latter the resistance to a deformation through associated interior spring and viscosity. Accordingly, the strength of the interior spring would diminish with aging while spring resistance to extension plus viscosity also diminishes. So the discrepancy would be only apparent, not real.
69.3.2 ACTINIC AGING A more specific study of the influence of sun exposure of skin mechanical aging was conducted by Lévêque’s group in 1988.17 This witty investigation was made on a panel of 35 professional cyclists at the end of the period of intensive training in spring and early summer in often sunny countries (Spain, Italy, France) before entering the Tour de France competition. Almost all of them wore a short-sleeved shirt, dividing the outer aspect of their arm into a sun-covered and a sun-exposed area. This was evident by the suntan limit. As UVA rays do not penetrate the skin beyond the papillary dermis, the authors used a narrower ring (1 mm wide) with an 18-mm-diameter disc, in order to restrict the shear stress to the superficial layers of the skin. A 9 mN·m torque was applied for 10 s on both exposed and covered areas on the same arm, each at 1 cm distance from the suntan limit. In exposed areas the rotation angle was reduced by 33% (p < 0.0001) and the skin
TABLE 69.1 Literature Data on Intrinsic Aging as Assessed by Torsional Tests
Author Finlay Finlay Sanders Lévêque Agache Escoffier Escoffier
Disc Diam (mm) 15 15 8.7 25 25 18 18
Torque (mN/m)
Width of Skin Annulus (mm)
2 20 0.83 9 28.6 2.3 10.4
4 4 no 5 5 3 3
Deformation with Aging Ue Range (mm)
Elastic
0.4–1.7 0.1–0.4 0.6–1.5 0.4–0.6 1.1–1.2
↓ → ↑ ↓ ↓ ↓ ↓
Viscous
↑ → ↑ ↑
Elastic Modulus Range (MPa)
0.02–0.1 0.85–2.94* 0.64–2.04* 0.16–0.36* 0.34–0.62*
Note: Ue is immediate elongation. Asterisks indicate figures not produced by the authors but computed from Equation 69.4.
608
Handbook of Non-Invasive Methods and the Skin, Second Edition
thickness increased by 22.5% (p < 0.0001). As calculated by Equation 69.5, the mean Young’s modulus moved up from 541 to 698 kPa, thus demonstrating a skin stiffening following repeated sun exposure. Also, the relative immediate recovery was reduced by 22% (p < 0.0001), denoting a substantial loss of elasticity. This work was extended by a study in three racial groups (15 blacks, 12 whites, 12 Hispanics) made on both the ventral and dorsal forearm with a 15 mN·m torque, an 18-mm-radius disc, and a free skin crown of 3 mm width.18 The torque was set abruptly, maintained for 60 s, then abruptly suppressed. While no difference between races were observed in skin thickness (as assessed by ultrasound A-scan, 15 MHz), in all subjects the dorsal forearm skin was thicker. The skin elastic modulus E on the volar forearm was the same in all groups, whereas in the dorsal forearm it was statistically lower in blacks. Also, in blacks there was no difference in E values between the two sides, while the difference was significant in other groups. Viscoelasticity was lower in the dorsal forearm, but significantly lower only in whites and Hispanics. Finally, the clinical elasticity (recovery/extensibility) was lower on the dorsal forearm except in blacks, while irrespective of groups and sites it steadily decreased with aging. In summary, the volar forearm skin had the same mechanical behavior in all races. Differences appeared on the dorsal forearm, where blacks differ by the absence of alterations related to chronic sun exposure such as an increase in Young’s modulus (i.e., stiffness) and a decrease in skin extensibility. Also a black–Hispanic–white trend was noted on the dorsal side for such alterations. The results of these two studies show that, mechanically speaking, long-lasting and repeated sun damage resembles intrinsic aging in reducing skin elasticity and increasing Young’s modulus. But the former is associated with a thicker skin, while it is the contrary for the latter. Accordingly, chronic actinic damage cannot be simply assimilated to premature aging.
69.4 STRATUM CORNEUM AND SKIN BIOMECHANICS From a mechanical standpoint, the skin is a stratified composite material whose superficial layer is the stiffest, but represents only 1/100 of the total thickness. For that reason, its role in whole skin mechanical behavior is often overlooked. Torsional experiments offered Lévêque and De Rigal7 the possibility to investigate in vivo the mechanical behavior of stratum corneum (SC). Using a 9 mN·m torque, they first studied the direct effect of pouring water on a skin crown either 1, 3, or 5 mm wide in the volar forearm of 10 volunteers. The result was a highly significant increase in the Ue parameter, 80, 40, and 15% for crown widths of 1, 3, and 5 mm, respectively. Indirect SC
hydration by 15, 30, and 120 min occlusion with a plastic sheet in six subjects caused Ue to rise progressively and significantly. The rise was also significantly steeper with a 1-mm-wide crown than with a 5-mm-wide crown. These Ue variations are inversely related to Young’s modulus ones. On the other hand, only SC mechanical properties can be modified by pouring water on the skin surface. Therefore, this experiment demonstrated that (1) SC takes an important part in the mechanical behavior of whole skin, (2) using a narrow crown of twisted skin the contribution of SC to the assessed mechanical parameters is strongly increased, and (3) the SC moisturization can be accurately assessed in vivo by torsional experiments. In the same paper, the efficacy of cosmetic formulations on SC hydration was assessed using the same technique in 13 volunteers. The significantly increased Ue found 1 h following application of either oil-in-water (O/W) or water-in-oil (W/O) creams devoid of moisturizers, or petrolatum, partially or totally subsided 1 hour later, but at a lower rate with petrolatum and with W/O emulsion. Addition of moisturizers such as 10% lactic acid or PCNa or urea had the same effect, not greater, by the first hour, and unexpectedly no sustained effect was found by the second hour following application. Only 10% glycerol had a protracted hydration effect, although less marked than with lactic acid. In a long-term study carried out on the legs of three groups of 14 volunteers, selected for showing dry skin patches, two O/W preparations containing 10% glycerol or 10% urea were applied twice daily for 3 weeks. By the first, second, and third weeks both preparations brought a significant rise in Ue, while the vehicle used as a blank did not differ from the control site. Only the effect of glycerol was still measurable (p ≤ 0.05) 1 week after cessation of treatment.7 The efficacy of other cosmetic products on SC compliance was also investigated by the L’Oreal group with the same device and experimental conditions in 10 subjects.19 The compliance parameter Ue was raised by 34 ± 0.9% 2 min following total forearm immersion in hot water (30˚C) for 3 min. Also, Ur and Ur/Ue significantly rose, by 40 ± 6 and 5%, respectively. The increased skin elasticity could be related to the rise in skin temperature together with SC hydration. In the same paper, the variations of the same parameters over 24 d of treatment are presented; each measurement was done at time intervals 24 h after the last application. All three parameters showed a steady and highly significant increase. These experiments also demonstrated the daily individual variations probably related to climatic changes (relative humidity and temperature). The dryness of facial skin was assessed on the forehead and cheeks in 55 selected subjects using both an ordinal scale (0 to 3) and torsional experiments.20 On cheeks there was a striking inverse relationship between Ue and the clinical dryness score (r = –0.66, p < 0.001).
Twistometry Measurement of Skin Elasticity
This was also significant, although less marked on the forehead (r = –0.41, p < 0.002). This body of data on the clear-cut influence of the SC hydration effect on whole skin mechanical parameters stresses the role of SC in overall skin mechanical behavior, which often had been underestimated. The conclusive evidence, i.e., modification of skin extensibility and elasticity upon SC removal, was brought again by L’Oreal’s group in the following experiment.19 After 10, 15, and 20 SC strippings with an adhesive tape stuck for 15 s under 1 kg cm–2 on forearm skin in eight volunteers, Ue, Ur, and Ur/Ue steadily and significantly rose relative to baseline. This simple experiment confirmed unequivocally that SC takes a substantial part in the skin stiffness and recovery following extension and should be accounted for as a functional as well as a structural component in any interpretation of data concerning the mechanical properties or behavior of the skin. This is the confirmation that from a mechanical standpoint, skin is a composite material made up of three layers of different thickness, stiffness, and elasticity, working in parallel.
69.5 MEDICAL APPLICATIONS 69.5.1 EFFECT
OF
TOPICAL RETINOIC ACID
Because they are supposed to take place within SC and the papillary and subpapillary dermis, the changes induced by topically applied retinoic acid (RA) might modify the mechanical behavior of the skin. Consequently, torsional devices that are stuck on the skin and put force onto the superficial layers are particularly well suited for assessing this effect. Pichon et al.11 in a group of 17 subjects evaluated the extensibility and viscosity parameters as described in Chapter 62, using both 3- and 1-mm-wide skin crowns. With the wider annulus they found a significant decrease in both types of parameters, while only extensibility parameters (k0, k1) were significantly decreased with the narrow (1-cm-wide) annulus. The difference could be ascribed to the remodeling effect of RA on both the epidermis and subpapillary dermis. The same drug was tentatively used to reduce the skin side effects of long-term systemic corticosteroid treatment in 27 kidney graft recipients.21 Each patient’s volar forearms were treated either by RA, 0.5% (or 0.025%) in O/W emulsion, or vehicle alone, at random, for 180 d. Skin thickness, as determined by an ultrasound A-scan device working at 25 MHz, was found intensively decreased at the start and significantly increased by the 60th day of treatment onward, mostly in women. Torsional tests used a 3-mm free skin crown and an 18-mm disc radius. The skin elasticity (Ur/Ue) was increased in the treated vs. control areas by a factor of 3 to 20% by the 60th day and further on, but this was only significant in women. As no experiments were done with the narrower skin crown (1
609
mm), the results are difficult to interpret in terms of location of structural change.
69.5.2 SCLERODERMA Sclerodermas are diseases characterized by induration of the lower dermis and uppermost subcutis, in either localized and well-defined plaques or bands as in morphea or diffuse as in progressive scleroderma. Increases in thickness and lower extensibility have been non-invasively documented using ultrasound echography and a suction extensometer. Kalis et al.10 confirmed these findings by using L’Oreal’s torsional device with a 3 mN·m torque applied through an 18-mm-diameter disc and a 3-mm-wide free skin annulus. The skin thickness was measured by A-scan echography. In active plaques of morphea (n = 5) the results were highly significant for both increased skin thickness (60 ± 5%) and decreased extensibility (–67 ± 11%) relative to a symmetrical healthy control area. In regressing lesions (n = 12) the skin thickness did not differ from the control area, but the extensibility was still lower, although to a lesser extent. Using Equation 69.5, the authors concluded that there was a negligible and not significant difference in the elastic modulus E in case of active plaques (ΔE = –6%), whereas in the case of regressing lesions E was significantly higher (ΔE = +49%). From these data one can infer that the excess of collagen synthesized during the sclerotic phase retains normal mechanical properties, whereas at the regressing phase the progressive dermal atrophy is associated with a stiffer collagen. Measurements in progressive systemic scleroderma were made on the volar forearm. The 11 patients were compared to 10 age-matched controls. The skin thickness was about twice that of controls (+106 ± 18%) and extensibility (Ue) was reduced by 68 ± 8%. Both these differences were highly significant. As a surprising result, E was decreased by 38%. But the skin elasticity was also reduced. Humbert et al.22 used the same device and a 15-MHz ultrasound A-scan for skin thickness measurement in a patient with morphea treated with 1,25-dihydroxyvitamin D3. After 6 months of therapy Ue rose by 211%, while skin thickness remained unchanged (1.41 and 1.35 mm, respectively). Accordingly, the Young’s modulus E decreased by 204%, a finding in accordance with clinical scoring. This result was confirmed in a series of five patients with morphea whose duration of disease had varied from 2 to 10 years.23 They received oral 1,25-(OH)-2 vit D3 (mean dose, 1.75 mg/d). In line with clinical improvement, a significant decrease in the Young’s modulus (–62%, p < 0.01) appeared after 2 to 24 months of therapy. In an open uncontrolled study Humbert et al.24 included in the same protocol 11 patients suffering from systemic scleroderma. A significant decrease in the
610
Handbook of Non-Invasive Methods and the Skin, Second Edition
Young’s modulus (–36%, p < 0.01) was also noted at the end of the study. Furthermore, these investigations demonstrated the possibility of using this device for assessing the effect of a treatment over time in a single patient.
69.5.3 INHERITED CONNECTIVE TISSUE DISEASES Torsional measurements on the volar forearm were performed on five patients with type 2 Ehlers–Danlos syndrome, using the L’Oreal device, and the results compared to those of age- and sex-matched healthy subjects (five for each patient).25 The torque was acted by a 18-mmradius disc with a free skin annulus of 3 mm width. The skin thickness was considered half of the skinfold in the same area, as measured by a caliper. While a significantly lower skinfold thickness was found in patients than in controls (1.1 ± 3 mm vs. 1.5 ± 0.2 mm, respectively), the skin extension (Ue) was significantly increased (40 ± 11 mmV vs. 25 ± 5 mmV, respectively). By combining these data, no change in the elastic modulus was found in these patients. Also, the ratio Ur/Ue (elasticity) was found to be identical to that of controls. The conclusion was that only skin thinning could account for the clinically observed extensibility in these patients. By contrast, in three cases of Marfan’s syndrome the skinfold thickness and distensibility Ue of each patient were higher than those of his or her five age- and sexmatched controls. The product Ue × thickness was significantly higher by 46 and 53% on left and right forearms, respectively, and the Young’s modulus as calculated by Equation 69.5 was decreased by 60%. On the other hand, skin elasticity was found to be slightly and nonsignificantly increased. From these data the authors suggested that Marfan’s syndrome was associated with “a true decrease in tissue stiffness.” Some years later this was confirmed by an abnormality in the fibrillin network in that disease.26
69.6 CONCLUSION Torsional experiments as used mostly with l’Oreal’s device proved a reliable and easy-to-handle tool for assessing skin mechanical behavior in current practice. As in any use of instrument, however, great care should be taken by the clinician to comply with the above-cited requirements for the validity of the experiment. Over the last 12 years this allowed significant advances to be made concerning skin physiology and in vivo mechanical properties. By contrast, only a few investigations have been made in disease, at least for two reasons. The first one is the dimension of the guard ring, which initially was 54 mm and now is reduced to 40 mm in outer diameter, thus requiring a rather large, flat area of skin, which is uncommon in pathology. The second reason is that the device, named twistometer, was a L’Oreal prototype, not
commercially available, and consequently usable by only a few groups. This is no more the case today, as the Dermal Torque Meter has come to the market. Dermatologists, whether they are interested in skin physiology or pathology, now have an easy-to-use technique to investigate one of the main skin functions, and possibly have an insight into the functioning of skin’s major components. They feel indebted to the cosmetic industry to have created the device, designed the protocol, and made the experimental studies needed to validate clinically and structurally relevant skin mechanical parameters.
ACKNOWLEDGMENT The author acknowledges Jean-Luc Lévêque for his outstanding contribution to our knowledge in skin biomechanics physiology.
REFERENCES 1. Meirson D, Goldberg LH. The influence of age and patient positioning on skin tension lines. J Dermatol Surg Oncol 19: 39–43, 1993. 2. Vlasblom DC. Skin Elasticity. Ph.D. thesis, University of Utrecht, The Netherlands, 1967. 3. Finlay B. Dynamic mechanical testing of human skin “in vivo.” J Biomech 3: 557–568, 1970. 4. Finlay B. The torsional characteristics of human skin in vivo. J Biomed Eng 6: 567–573, 1971. 5. Sanders R. Torsional elasticity of human skin in vivo. Pflügers Arch 342: 255–260, 1973. 6. Jaskowski J, Maceluch J. Nowe mozliwosci badan wlasciwosci mechanicznych skory czlowoieka. Wiad Lek 35: 1149–1155, 1982. 7. Lévêque JL, De Rigal J. In vivo measurement of the stratum corneum elasticity. Bioeng Skin 1: 13–23, 1985. 8. Lévêque JL, De Rigal J, Agache P, Monneur C. Influence of ageing on the in vivo extensibility of human skin at a low stress. Arch Dermatol Res 269: 127–135, 1980. 9. Agache P, Monneur C, Lévêque JL, De Rigal J. Mechanical properties and Young’s modulus of human skin in vivo. Arch Dermatol Res 269: 221–232, 1980. 10. Kalis B, De Rigal J, Leonard F, Lévêque JL, Riche O, Le Corre Y, De Lacharriere O. In vivo study of scleroderma by non invasive techniques. Br J Dermatol 122: 785–791, 1990. 11. Pichon E, De Rigal J, Lévêque JL. In vivo Rheological Study of the Torsional Characteristics of the Skin. Paper presented at the 8th International Symposium of Bioenginering and the Skin, Stresa, Italy, June 13–16, 1990. 12. Wijn PFF. The Alinear Viscoelastic Properties of Human Skin In Vivo for Small Deformations. Ph.D. thesis, University of Nijmegen, Holland, 1980. 13. Barbenel JC, Payne PA. In vivo mechanical testing of dermal properties. Bioeng Skin 3: 8–38, 1981.
Twistometry Measurement of Skin Elasticity
14. Escoffier C, De Rigal J, Rochefort A, Vasselet R, Lévêque JL, Agache P. Age-related mechanical properties of human skin: an in vivo study. J Invest Dermatol 93: 353–357, 1989. 15. Fleischmajer R, Perlish JS. The vascular inflammatory and fibrotic components in scleroderma skin. Monogr Pathol 24: 40–54, 1983. 16. Lévêque JL, Corcuff P, De Rigal J, Agache P. In vivo studies of the evolution of physical properties of the human skin with age. Int J Dermatol 23: 322–329, 1984. 17. Lévêque JL, Porte G, De Rigal J, Corcuff P, Francois AM, Saint Leger D. Influence of chronic sun exposure on some biophysical parameters of the human skin: an in vivo study. J Cutan Aging Cosmet Dermatol 1: 123–127, 1988/89. 18. Berardesca E, De Rigal J, Lévêque JL, Maibach HI. In vivo biophysical characterization of skin physiological differences in races. Dermatologica 182: 89–93, 1991. 19. Aubert L, Anthoine P, De Rigal J, Lévêque JL. An in vivo assessment of the biomechanical properties of human skin modifications under the influence of cosmetic products. Int J Cosmet Sci 7: 51–59, 1985. 20. Lévêque JL, Grove G, De Rigal J, Corcuff P, Kligman Am, Saint Leger D. Biophysical characterization of dry facial skin. J Soc Cosmet Chem 82: 171–177, 1987.
611
21. De Lacharriere O, Escoffier C, Gracia AM, Teillac D, Saint Leger D, Berrebi C, Debure A, Lévêque JL, Kreis H, De Prost Y. Reversal effects of topical retinoic acid on the skin of kidney transplant recipients under systemic corticotherapy. J Invest Dermatol 95: 516–522, 1990. 22. Humbert P, Dupond JL, Rochefort A, Vasselet R, Lucas A, Laurent R, Agache P. Localized scleroderma: response to 1,25-dihydroxyvitamin D3. Clin Exp Dermatol 15: 396–398, 1990. 23. Humbert P. Unpublished data. 24. Humbert P, Dupond JL, Agache P. Treatment of scleroderma with oral 1,25-dihydroxy vitamin D3. An open study. Acta Dermatol Venereol, in press. 25. Bramont C, Vasselet R, Rochefort A, Agache P. Mechanical properties of the skin in Marfan’s syndrome and Ehlers-Danlos syndrome. Bioeng Skin 4: 217–227, 1988. 26. Mollister J, Godfrey M, Sakai LY, Pyeritz RE. Immunohistologic abnormalities of the microfibrillar fiber system in the Marfan syndrome. N Engl J Med 323: 152–159, 1990.
70 Levarometry Shabtay Dikstein Unit of Cell Pharmacology, School of Pharmacy, The Hebrew University of Jerusalem, Jerusalem, Israel
Joachim Fluhr Department of Dermatology, Friedrich Schiller University, Jena, Germany
CONTENTS 70.1 70.2 70.3 70.4 70.5 70.6
Introduction............................................................................................................................................................613 The Measuring System..........................................................................................................................................613 Methodological Principle ......................................................................................................................................613 The Sensitivity and Reproducibility of the Levarometry Measurements.............................................................614 Validation ...............................................................................................................................................................614 Correlation with Other Methods ...........................................................................................................................615 70.6.1 Pull, Guard Ring, and Weight of Probe ....................................................................................................615 70.6.2 Elevations...................................................................................................................................................615 70.7 Conclusions............................................................................................................................................................615 References .......................................................................................................................................................................615
70.1 INTRODUCTION
70.3 METHODOLOGICAL PRINCIPLE
The mechanical properties of the human skin are known to change with age: the skin becomes lax and wrinkled with the loss of elasticity and turgor.1–9 A broad clinical approach to the prevention and treatment of these changes requires a non-invasive method of assessing functional consequences of structural changes in different skin compartments with age. Such a method would enable the investigator to quantify the changes in the mechanical properties of the skin with age, as well as changes due to environmental causes such as solar damage. It would also help to quantify the effects of different treatments. Levarometry is a method of evaluating skin slackness.
To measure skin slackness, we developed a device called the levarometer, 1 0 which is a modified Schade instrument11,12 with increased sensitivity. The increased sensitivity was achieved by using electronic techniques, allowing the skin to be analyzed in vivo in the low range of the stress–strain curve, shown by Daly13 to be the most sensitive parameter in differentiating between young and old skin elasticity measurements. The levarometer (Figure 70.1) operates by applying a perpendicular pull to the skin without a guard ring. A circular piece of Perspex® with a diameter of 0.5 cm is attached to the skin by double-sided adhesive tape. This tape is then attached to a counterbalanced measuring rod so that the net pressure of the system is less than 1 g/cm2. Different weights can be applied to the rod, thereby providing any desired elevating force. The elevating forces used are usually low, in the range of 5 to 40 g/cm2. The measuring rod is connected to a linear variable differential transformer (LVDT), the output of which is recorded graphically. The precision of the measuring system is 2 μm. The main problem is in ensuring that there is no friction between the measuring rod and the LVDT. The most sensitive measurement is the immediate levarometric response without measuring the creep. As a
70.2 THE MEASURING SYSTEM One of the main characteristics of wrinkled skin is that it stretches more readily than smooth skin when force is applied, i.e., wrinkled skin is more slack than smooth skin. To measure this skin slackness, one has to apply a perpendicular pull to the skin without a guard ring, thus enabling the skin to respond freely to the applied force. Levarometry works on this principle.
613
614
Handbook of Non-Invasive Methods and the Skin, Second Edition
0.25
System net pressure is less than 1 g/cm2
20–30 Years >64 Years
Electronics Different weights LVDT
Recorder
Elevation in cm
0.20 0.15 0.10 0.05
Forehead
0.00 0
Measuring area - 0.2 cm2
10
20
30
40
50
Pull g/cm2
FIGURE 70.1 Schematic view of a levarometer.
FIGURE 70.2 The dependence of levarometric measurements on age. N = 15 in each age cohort.
standard, we use a 4-g weight, since the surface area of the probe is 0.2 cm2 and the pull is 20 g/cm2 (see Section 70.5). A theoretical treatise of levarometry has been published by Lanir et al.5
When measuring the same area at different times, the variance of the measurements is less than 5.5%.14 When performing sequential measurements over the same area of 4.5 × 4.5 cm2, the deviation from the mean is 6%. This deviation is very small considering that the skin is not a homogeneous tissue.14 Topical application of a humectant (5% glycerine) on the skin did not affect the measurements. This implies that the measurements are not affected by environmental conditions or treatments that have a direct or indirect influence on the stratum corneum.14
70.5 VALIDATION Does levarometry differentiate between young and old skin? In order to answer the most interesting question for investigators in the field of skin aging, two groups, young women (20 to 30 years) and older women (65 to 80 years), were compared, with 15 healthy volunteers in each group. The measurements were performed on the forehead, and four elevating pulls were applied with 5, 10, 20, and 40 g/cm2.14 The results (Figure 70.2) show that levarometry distinguishes between the two groups. The difference was significant for all the elevation forces (p < 0.02 for 5 g/cm2 and p < 0.01 for 10, 20, and 40 g/cm2). On the basis of the forehead measurements we decided that the standard pull for all other measurements should be 20 g/cm2. This pull was chosen because it is the highest pull within the linear range. With a lower pull there is greater risk of human error, since small pulls are more difficult to handle.
Male Female
0.25 Elevation in cm
70.4 THE SENSITIVITY AND REPRODUCIBILITY OF THE LEVAROMETRY MEASUREMENTS
0.30
0.20 0.15 0.10 0.05 Forearm 0.00 0
20
40
60
80
100
Years
FIGURE 70.3 Change of standard levarometric measurement according to age for women and men. N = 22 in each age and sex cohort.
Can levarometry differentiate between male and female skin? The inner forearm skin was assessed in two groups: young subjects (20 to 30 years) and older subjects (61 years) were measured at the standard pull of 20 g/cm2.14 There were 22 women and 22 men in each age group. The inner forearm was selected as the measuring site since it is usually less exposed to natural UV, and thus extrinsic aging is less dominant as an influencing factor. The results of the forearm measurements show that levarometry differentiates between the two age groups and also between men and women (Figure 70.3). Most of the men tested in both age groups had less skin elevation than women. This difference was more marked in the older age group; in the young age group it was 40%, while in the older group it was 75% (p < 0.01). What is the difference between young and old skin by levarometry at 20 g/cm2 pull? The measurements of the female forehead skin showed that the change between the
Levarometry
615
young and old cohort is 160% (p < 0.01) (Figure 70.2). The measurement on the volar forearm skin shows a change of 60% for the young vs. old male cohort (p < 0.01), whereas in the young vs. old female cohort the change is 100% (p < 0.01) (Figure 70.3).
ring, indentometry, and ballistometry,14,18–22 suggests that levarometry without a guard ring is highly discriminating between old and young skin, and also between old male and old female skin. Levarometry is therefore suited to the study of aging skin.
70.6 CORRELATION WITH OTHER METHODS
REFERENCES
There are two research groups using similar methods to levarometry.15,16 The results are compared below.
70.6.1 PULL, GUARD RING, AND WEIGHT OF PROBE The elevating forces used by Pierard15 are 16 to 128 g/cm2, with a measuring disc diameter of 14 mm, and 64 to 512 g/cm2 for a diameter of 7 mm. Gartstein16 uses 10 g of elevating force with a measuring disc diameter of 3 mm. Pierard7 uses guard rings of different diameters. We found that if the ratio between the diameter of the measuring disc and the diameter of the guard ring is at least 6, then the ring has minimal influence on the levarometry values. Otherwise, one is measuring skin elasticity rather than skin slackness. Dikstein’s10 group did not use such a guard ring. In Pierard’s method,15 the disc is glued to the skin by cyanoacrylate, whereas Gartstein et al.16 use a vacuum. Dikstein et al. used double-sided adhesive tape instead of cyanoacrylate to avoid the difficulty of applying the cyanoacrylate only on the precise area of the measurement. It is important to note that the starting net pressure of the measuring probe on the skin should be minimal, so that it does not indent the skin before elevating it.
70.6.2 ELEVATIONS Pierard15 measured loading deformation, which is a measurement of the change 20 s after loading the force, i.e., measuring immediate levarometry with the immediate creep. Gartstein et al.16 used 0.25 s for loading time, thereby measuring immediate levarometry, allowing the same time for recovery, with a total cycle time of 0.50 s, recorded by a PC. Gartstein et al.16 also measured skin elasticity as the percentage of recovery after deformation. The reproducibility of Gartstein’s system is 4%,16 Pierard’s is lower than 8%,17 and Dikstein’s is lower than 5.5%.14
70.7 CONCLUSIONS Gartstein et al.16 found that young, undamaged skin stretches less and recovers more completely than aged or solar-damaged skin. The results from Dikstein’s group showed that young skin is less slack than aged skin (see Section 70.5). A comparison of levarometry with other mechanical methods, such as torsion, elevation with guard
1. Agache, P.G. et al., Mechanical properties and Young’s modulus of human skin in vivo, Arch Dermatol Res, 269: 221, 1980. 2. Lévêque, J.L. et al., In vivo studies of the evolution of physical properties of the human skin with age, Int J Dermatol, 23: 322, 1984. 3. Pedersen, L., B. Hansen, and G.B. Jemec, Mechanical properties of the skin: a comparison between two suction cup methods, Skin Res Technol, 9: 111, 2003. 4. Adhoute, H. et al., Influence of age and sun exposure on the biophysical properties of the human skin: an in vivo study, Photodermatol Photoimmunol Photomed, 9: 99, 1992. 5. Lanir, Y. et al., The influence of ageing on the in-vivo mechanics of the skin, Skin Pharmacol, 6: 223, 1993. 6. Oikarinen, A., Aging of the skin connective tissue: how to measure the biochemical and mechanical properties of aging dermis, Photodermatol Photoimmunol Photomed, 10: 47, 1994. 7. Gniadecka, M. et al., Skin mechanical properties present adaptation to man’s upright position. In vivo studies of young and aged individuals, Acta Derm Venereol, 74: 188, 1994. 8. Batisse, D. et al., Influence of age on the wrinkling capacities of skin, Skin Res Technol, 8: 148, 2002. 9. Aframian, V.M. and S. Dikstein, Levarometry, in Handbook of Non-Invasive Methods and the Skin, Serup, J. and Jemec, G.B.E., Eds., CRC Press, Boca Raton, FL, 1995, p. 345. 10. Dikstein, S. and A. Hartzshtark, In-vivo measurement of some elastic properties of human skin, in Bioengineering and the Skin, Marks, R. and Payne, P.A., Eds., MTA Press, England, 1979, p. 43. 11. Schade, H., Die Elasticitatsfunction des Bindegewebes und Die Intravitale Messung Ihrer Strorungen, Z Exp Pathol Ther, 11: 369, 1912. 12. Kirk, E. and S.A. Kooming, Quantitative measurements of the elastic properties of the skin and subcutaneous tissue in young and old individuals, J Gerontol, 4, 27, 1969. 13. Daly, H.C. and G.F. Ödland, Age related changes in the mechanical properties of human skin, J Invest Dermatol, 73: 84, 1979. 14. Manny, V., In-Vivo Deformation by Small Forces as a Criterion for Assessing Skin Ageing, M. Phann. thesis, Hebrew University, Jerusalem, 1989. 15. Pierard, G.E., Investigating rheological properties of skin by applying a vertical pull, Bioeng Skin Newsl, 2: 31, 1980.
616
Handbook of Non-Invasive Methods and the Skin, Second Edition
16. Warren, R. et al., Age, sunlight, and facial skin: a histologic and quantitative study, J Am Acad Dermatol, 25(Pt. 1): 751, 1991. 17. Pierard, G.E., Structure et properties mechaniques descompartiment adventitiel et reticulaire du derme, thesis for Agrege de l’Enseignement Superieur, University of Liege, Belgium, 1984. 18. Brozek, J. and W. Warren-Kinzey, Age changes in skinfold compressibility, J Gerontol, 15: 45, 1960. 19. Graham, R. and P.J.L. Holt, The influence of ageing on the in vivo elasticity of human skin, Gerontologia, 15: 121, 1969.
20. Tosti, A., F.G. Compagno, M.L. Fazzini, and S. Villardita, A ballistometer for the study of the plastoelastic properties of skin. J Invest Dermatol, 69: 315, 1977. 21. Cook, T., H. Alexander, and M.C. Cohen, An experimental method for determining the two-dimensional mechanical properties of living human skin, Med Biol Eng Comput, 15: 381, 1977. 22. Lévêque, J.-L. and P. Corcuff, In vivo studies of the evaluation of physical properties of human skin with age, Int J Dermatol, 23: 322, 1984.
71 Indentometry Shabtay Dikstein Unit of Cell Pharmacology, School of Pharmacy, The Hebrew University of Jerusalem, Jerusalem, Israel
Joachim W. Fluhr Department of Dermatology, Friedrich Schiller University, Jena, Germany
CONTENTS 71.1 Introduction............................................................................................................................................................617 71.2 Various Measuring Systems ..................................................................................................................................617 71.3 Indentometry Measurements Using Different Methods........................................................................................618 71.4 What Does Indentometry Measure?......................................................................................................................619 71.5 General Conclusions and Recommendations for Standardized Use of the Indentometry Method .....................619 References .......................................................................................................................................................................620
71.1 INTRODUCTION When asked which of various body parts are soft or hard, people generally unwittingly demonstrate the principle of indentometry. They first point to their faces, palms, etc., with the tips of their fingers or a pencil, etc. When asked what they were attempting to determine, a typical remark would be “to see how deep the pencil would go.” Industrial measurements of hardness of materials are based on a similar principle. By the same token, in order to measure skin softness, one has to apply a perpendicular force to the skin, which will then indent in response to the applied force to a degree related to its softness. The various indentometry methods are based on this principle.1
71.2 VARIOUS MEASURING SYSTEMS The first indentometer was built by Schade2 in 1912. All other indentometry measuring systems are modifications of his system. Schade’s measuring system,2 the elastometer, consists of a registering lever and a rotating drum. One end of the lever is connected to a vertical metal rod terminating in a hemispherical brass knob with an area of 50 mm2 that is applied to the skin. The rod also carries a small platform upon which a weight (ranging from 5 to 75 g) may be placed. A pen point and ink is attached to the other end of the lever, which plots the curve on paper on the rotating drum. The measuring pressure is 10 to 150 g/cm2.
The measuring system of Kirk et al.,3,4 first published in 1949, uses a similar apparatus, with the only difference being a moving lever that is traced by a celluloid point on smoked paper instead of by a pen point and ink. The standard measuring weight was 50 g, and the measuring pressure was 100 g/cm2. The measuring system of Tregear and Dirnhuber,5 first published in 1965, consists of a weighted metal rod with a measuring area of 0.1 to 5 cm2; the rod is free to slide in a vertical glass tube and can be loaded with weights ranging from 100 to 2000 g. The rod is placed on a skinfold area, the undersurface of which rests on a thin metal disc of the same diameter as the rod. The vertical position of the rod is registered by a spring-loaded dial micrometer, accurate to 2 μm. The measuring pressure is 20 to 400 g/cm2. In 1969, Robertson et al.6 first described a measuring system consisting of a measuring rod and linkage system with a low moment of inertia. The probe is a hollow aluminum rod mounted vertically on low-friction bearings. The lower, measuring end, which is applied to the skin, has a diameter of 0.5 cm. An aluminum plate at the upper end bears the weights. This probe has a simple linkage to a rotating transducer, the outer case of which rotates in an arc around a stationary inner core in response to vertical movement of the rod. The inner core contains a photosensitive element. The output of the photoresistive cell is related to the vertical displacement of the probe and is measured on a recording voltmeter. The instrument has a response time of 50 msec and can record movements 617
618
Handbook of Non-Invasive Methods and the Skin, Second Edition
of up to 1 cm with a resolution of 0.01 cm. The internal friction is approximately 1 dyn. A 50-g weight was used in all their experiments. The measuring pressure is 250 g/cm2. The measuring system of Daly et al.7,8 first published in 1974, consists of an ultrasonic transducer connected to a lead rod 4.5 mm in diameter and 20 mm long. Loads are applied to the tissue by a servo-controlled loading system. A semiconductor strain gauge bridge measures the load directly at the transducer. The load used is 5 g, and the measuring pressure is 71 g/cm2. The instrument of Pierard,9 first described in 1984, uses an aluminum rod, the force is created by a manometer, and the measurement is carried out using a comparator. The measuring pressure used ranges from 1 to 4 N per 320 mm2. The precision of the measurement is 10 μm. The measuring pressure used was 28 to 111 g/cm2. The measuring system of Dikstein et al.,10 first published in 1981, has a measuring area of 0.2 cm2 (Perspex®), connected to a light metal measuring rod that is counterbalanced so that the net pressure of the system is less than 1 g/cm2. The measuring rod can be loaded with different weights. We decided on a standard load of 2 g, since the resultant pressure is still in the linear part of the stress–strain curve. The measuring rod is connected to a linear variable differential transformer (LVDT), and the output is recorded graphically by an electronic recorder. The measuring rod is adjusted so as to just touch the skin. Once the baseline is stabilized, the weight is suddenly applied. Measuring the same area at different times using this method, the variance of the measurements is less than 6%.11 Measuring the dispersal of measurements taken immediately one after the other over an area of 4.5 × 4.5 cm2, the deviation from the mean is 10%.11 The standard measuring pressure used was 10 g/cm2.
71.3 INDENTOMETRY MEASUREMENTS USING DIFFERENT METHODS Indentometry measurements were used to assess skin softness in a variety of physiological conditions: old and young skin, male and female skin, and in various edematous conditions. The measurements of Schade2 were carried out on the leg below the knee and on the forearm below the elbow in various edematous conditions. The effect of age on the initial indentation and on elastic recovery was also studied. It was found that in edema, both the indentation and the elastic recovery decrease. With age, it is mainly the elastic recovery that changes (decreases). The measurements of Kirk et al.3,4 were carried out on the medial surface of the tibia, using a weight of 50 g. The measurements were performed in women aged 20 to 101 years and in men aged 18 to 104 years. The observa-
tions showed a definite decrease in indentation with age, and an even more marked decline in the degree of immediate resiliency (rebound — called by the author elasticity) of the skin following removal of the weight. Women’s skin in general was found to possess higher elastic properties than that of men of similar ages; the elasticity values recorded for women were usually of the same order of magnitude as those exhibited by men 10 years younger.4 Indentometry measurements were also carried out on the skin covering the lower and upper tibia. The weight used was again 50 g. The measurements were done on skin of aged (60 to 86 years) and young (18 to 22 years) volunteers. The results showed a marked difference between the two groups: the depth of indentation in the young skin was much greater than in the older age group; the immediate rebound of the tissue after removal of the weight was also much greater in the young than in the aged group.3 A force of 100 to 2000 g acting over 0.1 to 5 cm2 for up to 20 min was used by Tregear and Dirnhuber5 to compress human skin in vivo. The subjects were young white male adults. The area tested was the skin of one leg over the tibia, and the dorsal aspect of one forearm. The maximum compression produced was 0.6 mm. The time course of the compression was not exponential; there was a long-continued tail of deformation, with the compression remaining after the force had been removed. The speed of compression was increased by increasing the force. Tregear did not carry out comparative physiological studies. The measurements of Robertson et al.6 were made on the dorsum of the hand, forearm, biceps, triceps, knee, calf, and foot. Measurements were carried out on three groups of women: nonpregnant, pregnant with no clinical edema in the third trimester, and pregnant with clinical edema. In addition, measurements were performed on one pregnant woman who developed widespread edema in late pregnancy. The indentation measurements produced a similar pattern to the compressibility measurements by a modified caliper: increasing depth of indentation was accompanied by a lowered index of compressibility. Kydd et al.7 measured the anterior-medial surface of the tibia. Twenty volunteers, of both sexes, aged 8 to 86 years, participated in the study. Age did not have a dramatic effect on the response of compressive loading. In the 8 to 10 years age group, there was a tendency toward greater compression. Age was, however, shown to have its most significant effect on the recovery phase. Children aged 8 to 10 years showed almost immediate recovery, to about 97%; adults 15 to 23 years old showed 90% recovery in the same period; those aged 72 to 86 showed recovery to about 67% in 10 min, and it took 4.5 h for complete recovery.
Indentometry
The areas used for testing by Pierard9 were the volar forearm and the tibia. The effect of age and sex on these measurements was not stated. Dikstein et al.12 measured the forehead in females aged 2 to 70 years using a pressure of 10 g/cm2. In 20-year-old patients the indentation was found to be 0.043 cm, and at the age of 70 it was 0.054 cm. The elastic recovery at the age of 20 was 80.5%, compared to 65.5% at the age of 70. Thus, skin of older females has less resistance to the applied force and less elastic recovery after removing the force than does that of younger subjects. Manny,11 using the apparatus of Dikstein, measured the indentation on the forehead and the indentation of the back of the hand. The indentation of the back of the hand was measured with both open palm and closed fist. The volunteers were women aged from 20 to 30 years and over 65 years. There was no difference in the indentation between the two age groups in this experiments. However, a difference between the open-palm and closed-fist measurements could be measured, with a higher indentation of open palm. On the forehead Lanir et al.13,14 compared different age groups. These studies showed that the response to indentation loading was analyzed in reference to its glycosaminoglycan (GAG)-containing ground substance and fibers’ network microstructure. In the second study,14 these authors measured the forehead skin response in young (20 to 26 years) and old (64 to 80 years) subjects. The analysis suggests that low-load indentation and small deformation levarometry are well suited for aging studies since the skin response under these tests can be directly related to its structure and constituent properties (known to be affected by aging). However, these results suggest that levarometry is more sensitive to aging than indentometry.14 In an additional study from Hartzshtark et al.,15 topically applied pharmacological agents, which are assumed to raise the c-AMP level, decrease the low-pressure indentation value of the forehead skin of certain human volunteers. Using the same method, Robert et al.16 assessed the influence of social and work-related factors on rheological properties of aging skin. The measurements were carried out on the forehead, using a pressure of 10 g/cm2. Three different populations were studied: nuns (females), office employees (both sexes), and workers (both sexes). For the purposes of this study, subjects were defined as young with an age less than 55 years and old if they were over 55. Theses authors found that the elastic rebound (after removing the weight) decreased steadily with age. This effect was more pronounced in females. Working women lost their elasticity more rapidly than nuns, and the male workers lost their elasticity more rapidly than did the office employees. An indentometric approach was used for in vitro testing of biological material by Piruzian et al.17 They tested different passive and active mechanic load mechanisms.
619
71.4 WHAT DOES INDENTOMETRY MEASURE? The basic question is, Which skin layer is the most important for indentation measurements? In the stratum corneum, 4% lactic acid (pH 4.2) applied to the skin of four volunteers with dry skin of high pH was found to have no effect on indentation. Also, 5% glycerin, which is a strong humectant, was tried as a moisturizer on six volunteers with dry skin, without an effect on indentation.18 In an another experiment the stratum corneum was removed by stripping techniques, monitoring the surface pH to ensure that the stratum corneum was indeed removed. It was found that stripping did not affect indentation measurements.18 From those experiments it is clear that the stratum corneum does not affect indentation. In the dermis the forehead skin of three male volunteers was injected intradermally with 0.2 cm3 saline, and susequently (at least 10 d later), the same volume of saline containing 300 IU of hyaluronidase, 4 U of elastase, or 4 U of collagenase was injected. The results showed that saline decreased indentation, whereas hyaluronidase, and to a lesser degree elastase, increased indentation. Collagenase had no significant effect. The decrease in indentation caused by saline suggests, therefore, that the water moisture content in the dermis decreases indentation. Elastase, and even more so hyaluronidase, increases indentation in spite of the presence of saline. The interpretation of those experiments is that in the dermis, the state of the ground substance elastin network might be responsible for the observed changes in indentometry.18 More evidence of the effect of the ground substance on indentation was found in rat skin.5 A cross section cut through compressed rat skin showed that most of the compression occurred in the dermis. Since liquids are virtually incompressible, for the dermis to be compressed, the water must have shifted out sideways from the compressed area. The dermal fluid, the ground substance, is known to contain enough mucopolysaccharide to make it highly viscous.19 Diffusion through the dermis is greatly increased when the polysaccharides are depolymerized by hyaluronidase.20
71.5 GENERAL CONCLUSIONS AND RECOMMENDATIONS FOR STANDARDIZED USE OF THE INDENTOMETRY METHOD Indentometry is a technique for measuring skin softness. In effect, it actually measures the water status of the dermis, which is largely determined by the amount of mucopolysaccharides and their ability to bind water in the measured area. The immediate indentation measurement using indentometry was found in most studies not to discriminate sufficiently between young and old skin.4,7,12,16 The
620
Handbook of Non-Invasive Methods and the Skin, Second Edition
difference between the two age groups can be better differentiated by measuring the elastic rebound, which is the immediate rebound of the skin after unloading the force. It was found that in young skin, after unloading the force, the immediate rebound comprised a greater percentage of the initial indentation.2–4,7,11,12,16 Although some studies showed age dependence and differences between the skins of males and females with the indentometry method, it seems to us that indentometry is of most use to evaluate edematous conditions, and altered water handling of the dermis.2,6 The present data lead us to the following recommendations for a standardized use of the indentometry method: 1. In order to compare different physiologic conditions — young vs. old skin, male vs. female, and various edematous conditions — one must work within the linear area of the stress–strain curve of the skin area being measured. The standard indentation force used by Dikstein et al. is 10 g/cm2; this force is in the low linear area of the stress–strain curve of the forehead, the area on which the measurements are made.10 2. The pressure of the measuring apparatus itself should not exceed 20% of the total measuring pressure. The weight of Dikstein et al.’s apparatus creates an initial pressure of 1 g/cm2, which is 10% of the final measuring pressure. The exact initial pressure of the apparatus on the skin is not specified in most of publications. 3. During the measurement no lateral movement should occur between the skin and the apparatus. The reproducibility of the measurements should be greater than 90%.11
REFERENCES 1. Aframian, V.M. and S. Dikstein, Indentometry, in Handbook of Non-Invasive Methods and the Skin, Serup, J. and Jemec, G.B.E., Eds., CRC Press, Boca Raton, FL, 1995, p. 349. 2. Schade, H., Untersuchungen zur Organfunktion des Bindegewebes, Z. Exp. Pathol. Ther., 11: 369, 1912. 3. Kirk, E. and S.A. Kvorning, Quantitative measurements of the elastic properties of the skin and subcutaneous tissue in young and old individuals, J. Gerontol., 4: 273, 1949.
4. Kirk, J.E. and M. Chieffi, Variation with age in elasticity of skin and subcutaneous tissue in human individuals, J. Gerontol., 17: 373, 1962. 5. Tregear, R.T. and P. Dirnhuber, Viscous flow in compressed human and rat skin, J. Invest. Dermatol., 45: 119, 1965. 6. Robertson, E.G. et al., Two devices for quantifying the rate of deformation of skin and subcutaneous tissue, J. Lab. Clin. Med., 73: 594, 1969. 7. Kydd, W.L., C.H. Daly, and D. Nansen, Variation in the response to mechanical stress of human soft tissues as related to age, J. Prosthet. Dent., 32: 493, 1974. 8. Daly, C.H. and G.F. Odland, Age related changes in the mechanical properties of human skin, J. Invest. Dermatol., 73: 84, 1979. 9. Pierard, G.E., Evaluation de proprietes mecaniques dela peau pa les methodes dindentation et de compression, Dermatologica, 168: 61, 1984. 10. Dikstein, S. and A. Hartzshtark, In-vivo measurement of some elastic properties of human skin, in Bioengineering and the Skin, Marks, R. and Payne, P.A., Eds., MTA Press, England, 1981, p. 43. 11. Manny, V., In-Vivo Deformation by Small Forces as a Criterion for Assessing Skin Ageing, M. Pharm. thesis, Hebrew University, Jerusalem, 1989. 12. Dikstein, S., A. Hartzshtark, and P. Bercovici, The dependence of low-pressure indentation, slackness and surface pH on age in forehead skin of woman, J. Soc. Cosmet. Chem., 35: 221, 1984. 13. Lanir, Y. et al., In-vivo indentation of human skin, J. Biomech. Eng., 112: 63, 1990. 14. Lanir, Y. et al., Influence of ageing on the in vivo mechanics of the skin, Skin Pharmacol., 6: 223, 1993. 15. Hartzshtark, A. and S. Dikstein, The use of indentometry to study the effect of agents known to increase skin cAMP content, Experientia, 41: 378, 1985. 16. Robert, C. et al., Study of skin ageing as a function of social and professional conditions: modification of the rheological parameters measured with a noninvasive method — indentometry, Gerontology, 34: 284, 1988. 17. Piruzian, L.A. et al., [Study of biological materials based on indentometry], Izv. Akad. Nauk. SSSR Biol., 5: 769, 1990. 18. Dikstein, S. and A. Hartzshtark, What does low-pressure indentometry measure? Arztliche Kosmetologie, 13: 327, 1983. 19. Ogston, A.G. and J.E. Stanier, The physiological function of hyaluronic acid in synovial fluid, J. Physiol.,119: 244, 1953. 20. McLean, D. and C.W. Hale, Studies on diffusing factors, Biochem. J., 35: 159, 1941.
72 The Gas-Bearing Electrodynamometer C.W. Hargens Philadelphia, Pennsylvania
CONTENTS 72.1 Introduction............................................................................................................................................................621 72.2 Instrumental Application .......................................................................................................................................622 72.3 Instrumentation ......................................................................................................................................................623 72.4 Data Reduction ......................................................................................................................................................624 References .......................................................................................................................................................................625
72.1 INTRODUCTION The present chapter deals with the fundamental mechanical properties of the skin and their quantitative measurement. This is basic to an evaluation of existing dermal conditions and the efficacy of any treatment. The method is direct and not implied by inference from some other test or property, such as electrical conductivity or sonic propagation. It is the method a mechanical or materials engineer would use, the determination of stress (applied force) vs. strain (resulting deformation), to describe the characteristics influencing the performance of any structural substance. The skin is one of the most important structural substances holding the body together. We can further demonstrate that the skin’s principal strength resides in a few outer cell layers of the stratum corneum. This can be easily shown with the gas-bearing electrodynamometer (GBE) instrumentation if one examines an area where these thin cell layers are stripped away by just a few repeated applications of Scotch® tape (adhesive). A glistening layer is quickly reached, which the instrument will show has no strength of containment. In fact, such a test reveals that the mechanical modulus of the stripped area diminishes from a high value to almost nothing. This is the same phenomenon occurring in burns of the skin, which have a similar destructive effect. The term dynamometer may be somewhat foreign to the biomedical field, but technically it implies a device for measuring power, force, electrical current, or voltage. In this case, we are interested in a convenient and accurate way to measure force. A dynamometer involves the interaction of electrical and mechanical quantities, in essence a kind of transducer. Electric current in a conductor can be measured, and through its interaction with a magnetic
field, mechanical force thereby will be determined; or the reverse, measure force and derive the value of current in an electrical conductor. Usually force is determined through an arrangement of coils, magnetic fields, and a current measurement. The adjustable nature of these quantities and devices allows one great latitude in designing measuring systems to suit various purposes. Also, the precision and accuracy can be of a high order, because the calibrating forces, currents, and voltages involved can be fundamentally verified through standards. In the measurement of the skin’s mechanical properties stress–strain (force–displacement) modulus the principles just mentioned have been utilized. To further understand what is involved, it must be recognized that the skin, like other body tissues, is a viscoelastic substance; i.e., it is partly spring-like and partly viscous. Thus, the mechanical moduli one seeks to determine are significantly rate related due to the large viscous component of the modulus. In other words, the observed deformation will depend upon the speed with which a force is applied. This is why one cannot be satisfied with static measurements that were done in earlier times, for they will not explain dynamic behavior that influences our subjective perception of skin quality. The fundamental reason for there being a complex, rate-related modulus to describe skin behavior is the same as in the case of all large molecule polymers. In all of these materials there is a finite time required for a stressimposed molecular rearrangement to take place. To this phenomenon we assign the simplifying term viscosity, and it is present in soft solid substances as well as in fluids that have been studied extensively in this regard.1 Skin is composed of a well-known and complicated architecture consisting of such viscoelastic materials. As just mentioned, the main mechanical influences of this 621
622
Handbook of Non-Invasive Methods and the Skin, Second Edition
kind, as far as skin measurements are concerned, reside in the cells of the stratum corneum and the natural cements that hold the aggregate together. The GBE is sensitive enough to show that the mechanical integrity of these thin membranes is strongly influenced by small external effects such as moisture and other topically applied reagents. Returning to the matter of feasible instrumentation for skin tests, one needs equipment capable of dynamic measurement. The interesting phenomena relating to skin involve time constants on the order of 1 s. If one has sufficiently lightweight equipment, with agile components that can be quickly accelerated, these features of elasticity and viscosity can be recorded. One can apply small forces, actually a periodic forcing function such as a sine, at the 1-s rate and observe by suitable recording means the resulting deformations as a phase-displaced signal. Details of the application of these principles are explained in Section 72.2. Another point of initial consideration concerns the best direction in which to apply the test stresses in the skin to clearly enhance the desired responses. It turns out that the most sensitive indications of modulus changes in a membrane such as the stratum corneum will be obtained with the shear mode, i.e., the force vector applied directly in the plane of the surface. Techniques that indent the surface or attempt to raise a bell-shaped deformation normal to the surface with suction have been used but are far less sensitive. This is because the measurement relies upon only the smaller vector components of the principal stresses. Finally, one finds that a stress–strain characteristic taken in the shear mode gives a very adequate measurement. Plotted in real time (approximately a 1-s cycle), the diagram will be a smooth ellipse in four quadrants. It is best to keep within the linear parameters of the skin if a simple numerical modulus is to be obtained. Beyond these limits, the tissue’s response becomes very nonlinear, and analysis becomes much more complicated, involving a Fourier analysis of the individual stress and strain waveforms. The consequent interpretation of results is not worth the trouble, since the linear response involving a simple phase shift between sinusoidal stress and strain waveforms is quite adequate for the comparative evaluations one is seeking. The ellipse is of course just a display suitable for calculating the phase angle between X and Y waveforms and the modulus or dynamic spring rate of the test surface. This will be explained further in Section 72.4.
72.2 INSTRUMENTAL APPLICATION Having introduced the basic objective of the noninvasive protocol, we now present the method that has been developed over several decades and used successfully for as long.2–9 The protocol is to adhere a probe to a chosen test
FIGURE 72.1 Dynamometer attached to a facial site.
site on the skin by means of an accepted medical adhesive similar to products used in surgical procedures. The surface area contacted by the probe need be only a few square millimeters. The probe attaches in turn to the GBE, as shown in Figure 72.1. Here a facial site is being examined. The probe moves with the reciprocating action of the GBE. The amplitude of motion is usually about 1 mm, so the active area of the skin under test might be said to extend up to 10 mm beyond the attachment point, depending upon the type of skin involved and its pretension state. A more detailed view of what takes place around the probe in the plane of the skin during the stressing of the surface would seem to be the following. Stratum corneum is in itself so stiff that bodily actions are only possible because there is a certain excess of skin. That is, there are microscopic as well as macroscopic folds, creases, and ridges in its texture. As the skin moves, these function in accordion fashion to relieve the stress that would otherwise occur. Therefore, what the probe measures through the resulting strain determination, the stretch, is actually the bending of the stratum corneum within these folds, i.e., the unfolding process brought about by the stretching forces. This reality in no way diminishes the utility of the measurement, because the apparent stretchability and dynamic behavior of the skin are what one seeks to know. Regardless of the precise mechanism of dynamic response, the influence of the various skin treatments upon the measurable and subjectively sensed stiffness or dryness will become apparent from the data obtained. As the probe of the GBE applies specific forces in the plane of motion, its resulting displacement, also monitored, is controlled by the elastic and viscous moduli of the outer skin layer. It is a simple matter to display these two quantities, force and displacement, as a typical stress–strain diagram for the material.
The Gas-Bearing Electrodynamometer
A plotting device with sufficiently rapid movement, such as a storage oscilloscope, analog or digital, can display the complete elliptical diagram every second as the probe drives the skin back and forth in repeated cycles. If rapid changes in skin condition occur, as with the introduction of moisture or drying, they will appear as corresponding alterations of the diagram. For example, dryness and stiffening of the surface will appear as an elevated slope of the major axis of the ellipse; softening will cause a decrease in slope. When only a few tests are to be performed, the dynamic stress–strain plots can be measured graphically and the moduli calculated, as will be explained, from their geometry. On the other hand, in laboratories where very large testing programs continue day after day, it is best to computerize the process, taking data directly through analog-to-digital conversion of the stress and strain signals.
72.3 INSTRUMENTATION So far the GBE has been described only as to its capabilities, whereas little has been said about the mechanism itself. The probe just referred to in Section 72.2 is attached to the skin in a novel way so that the test site can be precisely preserved while the GBE is disconnected for use elsewhere. Thus, the subject can be released for other activities during the course of the test period. To facilitate this, a small plastic button serves to contact the skin, adhered by the adhesive film previously mentioned. However, the actual probe wire fits tightly into a tiny hole in the top of the button and remains thusly attached during the test. Disconnection is achieved by simply popping the probe wire out of the hole in the button. A supply of buttons and film adhesive is furnished with the instrument, along with a detailed application procedure. The GBE is visible at the top of Figure 72.1 and is mounted on a small optical bench immediately adjacent to the subject, whose head rests on the same base so as to minimize relative motion. The optical bench provides, through rack-and-pinion manual adjustments, precise but rapid positioning along the three coordinate axes. The attachment button secured to the probe wire is moved about over the subject by means of these controls until the desired test location is attained. The GBE is lowered until the tip of the probe (button) touches the skin and adheres to it. It is, of course, desirable that the subject not move for the several seconds of the test in order to avoid disturbing the recording process. This is usually not a problem, even without any particular stabilizing means for the head or other body part under study. A relaxed subject can hold quite still, even to the extent required by the rather large displacement magnification of the electronic and display systems involved. That is, 1 mm of skin displacement at the test site can correspond to 25 mm of spot movement
623
on the oscilloscope screen or plotter (a magnification of at least 25). The displacement of the probe just discussed is induced by the reciprocating armature of the GBE to which it is attached. The probe is held by a setscrew in a small plastic chuck fitted into a protruding sleeve of the armature. This chuck with probe attached can be removed as an assembly, owing to its tapered fit. This is an added convenience, as will be discovered when moving the GBE. The armature also carries a coil of fine copper wire, which moves in a strong, radial magnetic field. The field is created by a permanent magnet whose flux can be depended upon to remain constant. As the GBE’s name implies, the armature floats on a coaxial air bearing. Therefore, there is no metal-to-metal contact in the bearing, and frictional forces are reduced to those caused by the viscosity of air. This means that they are so small as to be orders of magnitude less than the skin moduli, and hence negligible. To a slightly lesser extent this is true of the hair-like leads carrying current to the coil. The force produced by the coil and magnetic field combination is in accordance with the standard electric motor equation: Force = B · l · i
(72.1)
where force is in dynes, B (magnetic flux) is in gausses, l (length of conductor) is in centimeters, and i (current) is in abamperes (1 abampere = 10 amperes). Thus, if B and l are constant, as in the dynamometer, the force is directly proportional to the current, and this fact allows one to use current to determine the force exerted by the GBE probe. Displacement (strain) is determined through another electrical device built into the GBE. This tracks the armature’s movements without exerting any force upon it. It is referred to as a linear variable differential transformer (LVDT). It consists of a primary and two secondary coils, all magnetically coupled together by a movable core. As the core moves axially with the armature through the three coils’ windings, it couples flux differentially from the primary into the two secondaries. If an electronic system is associated with the LVDT to supply current to the primary coil and rectify the combined secondaries’ output voltages (demodulate), one obtains a direct current (DC) voltage whose polarity and magnitude give an accurate record of the armature’s instantaneous position. This electronic system associated with the LVDT is referred to as a signal conditioner. The stress–strain diagram is thus created by applying a voltage corresponding to the force coil’s current to the oscilloscope’s vertical deflection system and at the same time connecting the LVDT (signal conditioner output) to the horizontal deflection system. For a viscoelastic material, whose deformation lags in time the impressed
624
Handbook of Non-Invasive Methods and the Skin, Second Edition
Stress (grams)
B/2 a
B Strain (mm)
A
FIGURE 72.2 Dynamic stress-strain diagram plotted by the GBE apparatus showing the principal measurable quantities.
sinusoidal forcing function, an ellipse is formed, as in Figure 72.2. The GBE has several important advantages for measuring the mechanical properties of soft, viscoelastic substances such as living tissue. One advantage is its ability to exert small dynamic forces and record sizeable displacements that ensue in these soft materials. Other transducers, by contrast, are very stiff, such as strain gauges and load cells, or they impose frictional elements in parallel with the test specimen. They also are usually rather bulky and inconvenient for clinical testing, or they cannot respond at high frequencies. The construction of a GBE, on the other hand, is similar to the transducer in an audio loudspeaker, but without the stiffness. For completeness it should be mentioned that a basic air-bearing dynamometer in an evolutionary form has been used at audio frequencies. However, at audio rates it operates in a slightly different fashion, because the mass of the armature, although small, and in fact negligible at the 1-Hz rate discussed earlier, would now become significant. However, it can be shown that by combining the electrical motor equation (Equation 72.1) with the corresponding generator equation involving velocity (E = Blv,) one does not need the LVDT or any other measure of displacement. Instead, it is possible to employ an electric bridge circuit to obtain precise electrical impedance analogs of the specimen’s mechanical moduli. From these electrical quantities one can calculate the actual mechanical characteristics of the material.10,11 These would combine to yield the so-called mechanical impedance. Returning to the low-frequency operation of the GBE described here, one does measure both force and displacement to produce the elliptical stress–strain diagram directly. From the diagram’s dimensions, which are calibrated in appropriate units, such as grams and millimeters, it is a simple matter to compute the moduli. The most useful modulus is the dynamic spring rate (DSR), which is the slope of the ellipse’s major axis, B/A,
as seen in Figure 72.2. As mentioned earlier, it correlates well with subjective descriptors such as dryness, softness, etc., used by trained dermatological graders.12–14 Additional information about the combined elastic and viscous parameters can be obtained by noting the openness of the elliptic loop. For example, a strictly elastic element like a steel spring would show no openness. The correct interpretation of this would be to say that there is no energy loss if the loop degenerates to a straight line; i.e., the integrated area under the curve in Figure 72.2 (the energy or work done) is the same during stretch as during return. Energy is conserved in the elastic case but not in the case involving viscous elements. There should be no mystery about the elliptical display. It is simply the Lissajous figure used to measure the phase angle between two sinusoidal waves, in this case, stress and strain. If they were shown as sine functions on two parallel time axes, one would observe the time lag or phase angle lag of the strain behind the applied stress. In our case, it is called the loss angle. In Figure 72.2 the numerical value of the loss angle is computed from the ratio of the loop opening to the total displacement. The sine of the loss angle equals this ratio: Sin θ = a/A
(72.2)
The fundamental reasons for this time lag in a viscoelastic material have been previously discussed.1 What has the loss angle to do with tissue properties in dermatological terms? It has been observed that young, healthy tissue has the least loss. This was found to be true in the case of ocular tissue when ophthalmic experiments were conducted, and it is true of skin. While speaking of the loss factor, we should interject that there is of course more to tactile sensation, self-comfort, and appearance than the dynamic shear moduli of the stratum corneum as measured by the GBE. The underlying dermal tissues are important as well, and their contribution can be conveniently measured using ballistometry. To illustrate dermal loss factor further, in a very homely analogy we are speaking about the “upholstery” of the body. That is to say, subjectively tactile sensation responds to all of the above, plus frictional effects, and there is a noticeable difference between a strictly springy surface and one with a gradual yield, even though their static stress–strain deformation coefficients may be identical. Again, note the differences between a down-filled cushion, one of foam rubber, and one packed with cotton batting. The dynamic response counts strongly.
72.4 DATA REDUCTION Graphical data reduction methods were employed when the GBE was first introduced. A Polaroid camera photographed each loop, which was later scaled and the
The Gas-Bearing Electrodynamometer
calibrated measurements tabulated. The force calibration of the GBE is straightforward and positive. First, the instrument is simply tipped up and fixed in a vertical position. The wire probe is removed from the chuck, and the dynamometer becomes a platform balance on which gram weights can be placed, a few at a time. The armature, now vertical, is balanced against gravity by means of the DC offset control of the function generator, which drives the GBE. The various balance currents introduced into the force coil of the instrument will cause a series of corresponding vertical displacements of the trace along the force axis of the oscilloscope or other plotter. The vertical amplification setting of the latter can be chosen to give a convenient scale for the force range being used in any particular experiments. Once set, of course, it should not be changed until the system is recalibrated. The displacement calibration is equally simple and positive in its overall inclusion of the system component variables. It is done with the GBE in a horizontal position and with a micrometer or other accurate scale indicating armature movement. The corresponding scale can thus be generated on the horizontal (displacement) axis of the display, again choosing an appropriate adjustment of the amplifier. When a large number of experiments must be carried out on a daily basis, a computer can be connected to the system. This will greatly facilitate recording of data, automatic calculation of DSR, and other quantities, as well as trend or statistical data summaries. The connection is made directly from the oscilloscope or recorder terminals using the signal voltages produced by the GBE system components. Each experimenter will establish his or her own protocols for equilibrating subjects’ skin when they arrive for tests. For extremely precise work, an environmental chamber has been used, although most modern laboratories have sufficiently standardized atmospheres in their general working spaces.
REFERENCES 1. Ferry, J.D., Viscoelastic Properties of Polymers, John Wiley & Sons, New York, 1961. 2. Hargens, C.W., Glaucoma and vibration tonometry, J. Franklin Inst., 207, 143, 1960.
625
3. Hargens, C.W., Viscoelastometry of biomaterials, in Proceedings of the 26th Annual Conference on Engineering in Medicine and Biology (IEEE), Minneapolis, 1973, p. 203. 4. Hargens, C.W., Instrumentation to measure the viscoelastic properties of intact human skin, in Proceedings of the 28th Annual Conference on Engineering in Medicine and Biology (IEEE), New Orleans, 1975, 17, 179. 5. Christensen, M.S., Hargens, C.W., Nacht, S., and Gans, E.H., Viscoelastic properties of intact human skin: instrumentation, hydration effects, and the contribution of the stratum corneum, J. Invest. Dermatol., 69, 282, 1977. 6. Hargens, C.W., Measurement of dynamic moduli and loss factor in viscoelastic materials using the gas bearing electrodynamometer (GBE), J. Acoust. Soc. Am., 67 (Suppl. 1), S25, 1980. 7. Hargens, C.W., The gas bearing electrodynamometer (GBE) applied to measuring mechanical changes in skin and other tissues, in Bioengineering and the Skin, Marks, R. and Payne, P.A., Eds., MTP Press, Lancaster, U.K., 1981, chap. 14. 8. Hargens, C.W., Instrumented testing of human skin in vivo, in Proceedings of the 35th Annual Conference on Engineering in Medicine and Biology (IEEE), Philadelphia, 1982, p. 12. 9. Missel, P.J., Bowman, M.J., Benzinger, M.J., and Albright, G.B., An in vitro method for skin preservation to study the influences of relative humidity and treatment on stratum corneum elasticity, Bioeng. Skin, 2, 203, 1986. 10. Hargens, C.W. and Keiper, D.A., Tonometry: challenge for electronics, in Digest of the 1961 International Conference on Medicine and Electronics (IEEE), 1961, p. 77. 11. Keiper, D.A., Dynamic mechanical properties tester for low audio and subaudio frequencies, Rev. Sci. Instrum., 33, 1181, 1962. 12. Christensen, M.S., Nacht, S., and Packman, E.W., Facial oiliness and dryness: correlation between instrumental measurements and self-assessment, J. Soc. Cosmet. Chem., 34, 241, 1983. 13. Maes, D., Short, J., Turek, B.A., and Reinstein, J.A., In vivo measuring of softness using the gas bearing electrodynamometer, Int. J. Cosmet. Sci., 5, 189, 1983. 14. Cooper, E.R., Missel, P.J., Hannon, D.P., and Albright, G.B., Mechanical properties of dry, normal and glycerol-treated skin as measured by the gas bearing electrodynamometer, J. Soc. Cosmet. Chem., 36, 335, 1985.
73 Ballistometry C.W. Hargens Philadelphia, Pennsylvania
CONTENTS 73.1 Introduction............................................................................................................................................................627 73.2 Mechanics of Ballistometry...................................................................................................................................628 73.3 Practical Ballistometry for Skin Studies ...............................................................................................................630 73.4 Data Obtained with Ballistic Measurements.........................................................................................................632 References .......................................................................................................................................................................632
73.1 INTRODUCTION The ballistometer is a neat and easy-to-use device for measuring certain properties of human skin in vivo. This statement implies the ballistometer’s usefulness, but at the same time, the expression “certain properties” is a caution. One should be very specific in all claims regarding the determination of skin properties because there are so many. One might do well to first decide which ones tell something about the subjectively sensed skin condition to be evaluated, then see which “property” is the best indicator and measure it with the correct instrument. An example of two measurements which determine entirely different skin properties concerns the role of the ballistometer in contrast to the GBE (gas-bearing electrodynamometer). The first measures certain elastic parameters below the surface, while the GBE is best for quantifying stiffness in the surface plane of the skin, the stratum corneum. Here we will speak about the ballistometer for skin studies, and it should be appreciated that this is a narrow application of the concept. Our discussion will differentiate between a strictly practical device that has served well as an experimental tool and consideration in some detail of its extended capabilities. Ballistometry, the classical use of impacting masses to measure their material properties through their interaction, is not new. Sir Isaac Newton (1642 to 1727) is credited with certain relevant philosophical propositions and experiments involving impacting bodies.1 His observations led to the conclusion that the relative velocity of two bodies after they have impacted each other is in a constant ratio to their relative velocity before impact and in the opposite direction. This constant ratio has been called the “coefficient of restitution”, usually designated by e.2
The ballistic method of investigating materials’ mechanical properties in modern times was initially applied to homogeneous, usually hard, substances such as metals.3 On the other hand, the use of the concept to examine relatively soft, viscoelastic matter is still more recent, i.e., in the middle of this century. Hollinger and Thelen4 extensively studied asphalts in the 1950s using an impacting pendulum to find the storage (elastic) and loss (viscous) moduli. The ball rebound tests of natural and synthetic rubber stocks given in handbooks show the considerable influence of temperature on polymers.3 For example, in these ballistic measurements the percent rebound of a 1.9-cm steel ball dropped 100 cm onto a natural rubber (Hevea) sample 1.9 cm thick is 45% at 20°C and 71% at 100°C. For polychloroprene (neoprene) the change was from 35 to 67% for temperature change of 20 to 100°C, respectively. Arbitrary scales have been of necessity the practice in all of these tests, and the specification of equipment details, such as the geometry of the impacting components, is essential if uniformity of results is expected. For metals the diameter of the indentation made by a small hardened steel sphere (Brinnel hardness) or the height of rebound of a small hammer (Shore scleroscope) serves as arbitrary measures of hardness. In the case of skin ballistometry the depth of penetration to particular layers of the dermis will depend to some extent upon the geometric sharpness of the impacting mass. The sharpness of course determines the instantaneous pressures exerted upon the tissue. Some standardization eventually will be helpful. The ophthalmic applications go farther back to a ballistic tonometer devised in 1930 by Vogelsang5 to assess ocular tension. The method involved photographing the rebound oscillations of a small hammer striking the cornea 627
628
Handbook of Non-Invasive Methods and the Skin, Second Edition
of the eye. Wigersma6 in 1955 apparently discovered the benefit of a lighter hammer in a method referred to as elastometry. Eventually Mamelok and Posner7 published their conviction that these measures had more to do with the mechanical properties of the cornea itself than with intraocular pressure, i.e., more in accordance with Newton’s original contention. In more recent times others have used the method to study skin properties. Here the deeper dermal structures would be influential. Tosti8 made a ballistometer for this purpose. However, one should note that the published analysis must be viewed with caution, because e is defined as an energy ratio instead of the accepted momentum ratio which Newton originally postulated. Thus, without noting this difference, one could be confused by the numerical results of the experiments. The differences result from energy being proportional to the square of the velocity, whereas momentum is mass times the first power of velocity. More specifically, the example given in the reference is of a free-falling body striking a horizontal surface. Actually a pendulum was used, and we will have more to say about pendulums. Basically it is true that kinetic energy of whatever mass system is involved should equal the starting potential energy however it is created. Then the rebound kinetic energy should equal the next potential energy peak. This reference defines e as H′/H, the ratio of rebound height, H′, to the previous starting height, H, as previously discussed. The coefficient of restitution is an important element in any analysis of the motion ensuing after the collision of two bodies. In the most elementary case it is assumed that as soon as contact begins there will be a certain period of time referred to as the “period of compression”. After maximum compression and deformation, recovery to some extent will occur during a time called the “period of restitution”; hence the name of the coefficient. These times are of course fundamentally dependent upon the period required by the materials’ molecular structures to rearrange themselves. Thus one observes the differences in behavior between those substances which are highly elastic and those that are more viscous. Experimental determination of e is traditionally done by measuring the rebound height of one object falling upon another. The expression for e is derived simply as follows:
e=
v2 – v0 v 0 – v1
(73.1)
If it can be assumed that the second body does not move, v0, their common velocity at maximum compression, equals zero, and hence we will have
e=–
v2 v1
(73.2)
In these equations v1 is the velocity of the falling mass before and v2 its velocity after impact and separation. Since these velocities are defined by the starting and final rebound heights, i.e., potential energies, one may apply the following relationships to calculate e. v1 = 2gH
and v2 = 2gh
(73.3)
Then, substituting in Equation 73.1, e=
h h or e 2 = H H
(73.4)
Thus, instrumentally one measures the height, H, from which the mass falls and the rebound height, h, to which it rises. The task of the experimenter is to find a practical way of doing this.
73.2 MECHANICS OF BALLISTOMETRY This elementary determination of the coefficient of restitution assumes that the falling mass and the surface struck are of the same material, that the surface is rigidly supported, its velocity is at all times zero, and hence the velocity of both masses at the instant of greatest compression will be equal to zero. These assumptions are not completely satisfied when ballistometry is applied to the skin. Now we must consider the details of the process further to see how valid our interpretation of the results will be. In general, when the impacting mass strikes the skin surface, several things happen. First, the elastic component of the skin begins to store some of the kinetic energy of the falling object. The subsequent release of this stored energy provides the rebound. The processes of both compression and restitution are slowed however by the viscous component whose reaction force, like a shock absorber, is proportional to velocity. It will be at a maximum at first and then decrease as indentation proceeds, energy is dissipated as viscous internal friction, and the area of contact simultaneously enlarges. The elastic component of force is not dependent upon velocity. A second deviation from simple assumptions is that all of the involved skin does not have zero velocity throughout the impact process. Instead, some parts will have been accelerated, and an exchange of momentum will occur as well as potential energy storage. This momentum initiates an acoustic wave in the skin, however rapidly attenuated, propagating in indeterminate directions. The acoustic energy will be mostly lost, although a small amount may
Ballistometry
be returned to the rebounding mass and contribute to or detract from the reaction depending upon its time phase. This is similar to a diver bouncing up and down on the end of a diving board; if he gets out of synchronism with the board, unpleasant results ensue. With skin much of the indentation energy will be lost in shear viscosity within the tissues, conversion to heat. The greatest rebound will of course occur where the ratio of elastic to viscous effect is largest. Hence the expression that the skin is “more elastic” in such cases is a fair statement and can be used to describe the observed accentuated characteristic in younger skin and for certain beneficial dermal treatments. The coefficient of restitution theoretically ranges from 0 for inelastic impact, i.e., zero after-impact velocity, to 1 for completely elastic impact (which of course never occurs). On this basis ballistometry has come to be useful for evaluating youthful and more mature skin and its response to skin treatment products. For a viscoelastic, nonlinear substance like skin, especially in compression during impact, one observes diminished times between successive rebounds. This is to be expected because the time for a mass to fall is a function of its starting height. Thus, impacts will be closer and closer in time. (See Figure 73.1 for typical rebound recordings.) One additional effect of the decreasing fall height is that the diminished impact velocity produces shallower indentation, involving essentially a different substance composed of the more superficial skin layers. This explains in part why a constant rebound ratio is not maintained. Also, it should be appreciated that, with indentation of soft substances, the area of contact increases rapidly, so that, depending upon the shape of the impacting mass, an altered stress distribution in the region occurs with depth. These last comments direct our attention to the influence of the shape of the impacting mass. This subject interested mathematicians as far back as the 19th century. Heinrich Hertz, the discoverer of wireless waves, solved the problem of impact between solid spheres and presented an extended theory of impact between a solid sphere and an elastic plate, the latter being more like our situation. These relations in mathematical form showed the dependence of the coefficient of restitution upon the material constants, Young’s modulus, density, and Poisson’s ratio.9 More recent works on this topic are the classical analyses of Timoshenko in his Theory of Elasticity in which precise relationships for stress distributions in various geometries are presented.10,11 A pendulous mass that impacts on the test surface is a better way to get control of the several ingredient parameters than by dropping weights and attempting to observe their rebounds. A compound pendulum provides a means of distributing several mass elements to increase its moment of inertia and so control the impact velocity. We can show how this is better than a simple, pivoted hammer
629
(a)
(b)
(c)
(d)
FIGURE 73.1 Sample recordings: (A and B) Reduced period with lower impact on highly elastic surface; (C) improved rebound with lower velocity impact on more viscous surface; (D) comparison of more elastic with less elastic material.
which, because of its uncontrolled impact velocity, produces only minimal results, i.e., unsatisfactory rebounds for soft, viscoelastic substances. Large rebounds are necessary to provide a high degree of discrimination among the varying viscoelastic parameters. Let us illustrate graphically the reason for controlling impact velocity when applying ballistometry to the study of viscoelastic materials. This can be shown with a graph of the force, and hence the work relationships accompanying the periods of compression and restitution. Figure 73.2 shows two plots of force acting on the indenting mass as it compresses the skin downward along the velocity axis at several speeds, one high and the other more slow. In both cases we will assume identical kinetic energy inputs to the same test site. Fundamentally it is the maximum indentation that stores the most elastic, potential, energy that will be returned as a large rebound. Identical energy input means that the total work done by the viscous and by the elastic force, i.e., the area under these curves, must be equal in both cases. It is unimportant in making our point that the exact functional relationship between these forces and the indentation of the skin cannot be known. However, the graphs show the greater apportionment of viscous energy (loss) in the highspeed case. This is of course because the viscous force is proportional to the speed, whereas the elastic force is not. Note also that the slope of the elastic force-displacement characteristic is everywhere the same as determined by the skin involved. The conclusion is that the slow-speed case stores more energy in the elastic element and to do so must mean greater indentation, as shown. Hence, the greater stored energy will occur, and the rebounds will be higher. The
630
Handbook of Non-Invasive Methods and the Skin, Second Edition
Viscous force
Viscous force
E
E O
O
XC
XC Indentation
Indentation
Elastic force
Elastic force
XC = Maximum indentation point
K
K O
O
XC
XC Indentation
Indentation
Work area (fast) OEXC + OKXC = Work area (slow) OEXC + OKXC
FIGURE 73.2 Force vs. indentation (work) diagrams of high and low velocity impact on viscous and elastic substances showing greater rebound potential for the low-velocity impact.
approximate curvature of these force characteristics is introduced to account for the varying area of contact, hence pressure, during compression and restitution. If the contact is assumed to have a conical geometry, the force will vary somewhere between a square and cubic function of the indentation.
73.3 PRACTICAL BALLISTOMETRY FOR SKIN STUDIES The qualitative measures presented in the previous section are sufficient to provide ballistometer design criteria. A practical embodiment of the ballistometer is a form of pendulum in which the pivotal angle is measured instead of mass heights. This may be done conveniently by connecting its shaft to some low-friction angle transducer such as a rotary variable differential transformer which imposes no force on the pendulum. In Reference 8 a stationary coil with its inductance varied by a moving iron core attached to the pendulum was associated with an electronic circuit to indicate angle. To directly obtain numerical values for computer entry high-resolution, digital, optical angle encoders are available. In most cases the ballistometer will operate with large angles of swing, so that some trigonometric processing will be required as
will be discussed in more detail. Still another approach is to employ an angular accelerometer whose output is integrated to give angular velocity. Programmed with a computer to deliver information precisely at the moments of impact and rebound, this arrangement can directly calculate the coefficient of restitution. Other useful functions of the computer are the establishment of a time base for release of the pendulum from its raised position and graphical display of the rebound pattern. Release of a pendulum type of moving mass will result in a certain velocity of normal impact, R dU/dt, upon the test surface, where R is the effective length of the pendulum (pivot-to-impacting point). The analysis of the pendulum’s motion must now recognize certain practical design features based upon the theory presented in the previous section. The desirability of a low-velocity, compound (physical) pendulum with a considerable moment of inertia has been discussed in terms of improved rebound. Figure 73.1 shows this improvement in measurable rebound amplitude, achieved simply by adding an adjustable counterweight to a pivoted beam assembly. Such a ballistometer is pictured in Figure 73.3 showing a bar with impacting mass and counterweight pivoted on an angle transducer.
Ballistometry
631
M1
r1
L
θ
C r2
C = Center of mass h or H
FIGURE 73.3 Ballistometer with compound pendulum and angle transducer. Impacting mass is on the right, counterweight on the left.
The angular acceleration of a compound pendulum determines its ultimate velocity. For any rotating body the angular acceleration is equal to the torque applied to it, in this case gravitational, depending upon its instantaneous angular position, divided by its moment of inertia about the pivot point. Thus the angular acceleration and hence the impact velocity can be easily adjusted by changing the positions of the impacting mass and the counterweight to alter the net gravitational torque and the moment of inertia of the system. The latter is the moment of inertia of the pivoted beam plus the contributions of the two masses. The moment of inertia of the system can be measured by timing its natural period of oscillation about the pivot and using the following relationship: 2
T I = ⎡⎢ ⎤⎥ mgL ⎣ 2π ⎦
(73.5)
I is the moment of inertia, T is the period, m is the total mass, g is gravitational acceleration, and L is the distance from the pivot to the center of mass. L can be found by balancing the beam and masses on a knife edge. The differential equation of motion for the system is d2Θ/dt2 – (mgL/I) cos Θ = 0
(73.6)
where Θ is the pivotal angle referenced to the horizontal plane of impact. Solution of the kinetic equation of motion is not absolutely necessary to the understanding, design, and use of a practical ballistometer. We can analyze the compound pendulum to see how the coefficient of restitution can be derived from energy and moment considerations. In Figure 73.4 the center of mass of the pendulum is shown at a distance L from the pivot. If the pendulum is to rotate clockwise when released, and the beam is pivoted at its midpoint, the center of mass must be to the right of
M2
FIGURE 73.4 Compound pendulum mechanics.
the pivot. Taking moments about the center of mass to find its location one obtains L=
M1 r1 – M 2 r2 M1 + M 2
(73.7)
If we can make the two masses equal, L becomes further simplified. When M1 = M2 = mg, then L=
1 (r – r ) 2 1 2
(73.8)
We will use L to calculate the effective height to which the mass of the system is raised for a given starting angle of the pendulum, Θ1, relative to the horizontal. Thus the potential energy will be obtained as PE = 2mgLsinΘ
(73.9)
This potential energy can be equated to the kinetic energy to find the velocity of impact as well as the rebound velocity, both of which are needed to find the coefficient of restitution. The kinetic energy of such a rotating system is 1/2 I (dΘ/dt)2. Solving for velocity gives v = r1 (dΘ dt ) = r1 2 mg(sin Θ) I
(73.10)
Considering this relationship for v1 where Θ1 is the starting angle of the pendulum and v2 where Θ2 is the rebound angle, the value of e may be calculated exactly as
e=
v2 = v1
sin Θ 2 sin Θ1
(73.11)
632
Handbook of Non-Invasive Methods and the Skin, Second Edition
This is the basic equation that should be used to derive the coefficient of restitution in a system that measures pendulum angles. This is in conformity with the definition of e in Equation 73.2. Equation 73.11 can be verified by showing it to be identical with Equation 73.4. This is done by substituting under the radical in Equation 73.11 the geometric expressions for sinΘ2, which is h/L, and the equivalent of sinΘ1, which is H/L. The L cancels, and one is left with the square root of h/H or Equation 73.4. We have shown how the potential energy transfer to rotational energy and back to potential is accomplished and does not alter the impact conditions. The point is made also that the angular measure is a simple way to register these processes, and the necessary trigonometric computation is not complicated. In doing this one gains control over the impact velocity which is critically important in applying ballistometry to viscoelastic materials such as the skin.
An impressive amount of data (in vivo) can be obtained with a ballistometer in a short time, particularly if the information is processed on-line via computer.12 The computer may also control a solenoid release of the ballistometer’s pendulum and start the rebound angle recording process. This may be done also with a manual electronic system. Images such as those shown in Figure 73.1 may be recorded on the screen of a storage oscilloscope whose sweep is synchronized to begin with the pendulum release. Alternatively, an X-Y plotter with good response and writing speed might be used to record raw bounce angle data. These would have to be corrected trigonometrically in accord with Equation 73.11 to obtain accurate coefficients of restitution. Experiments should be done under the same atmospherically controlled conditions as would be applied to skin studies. Subjects should be allowed to equilibrate prior to testing.
73.4 DATA OBTAINED WITH BALLISTIC MEASUREMENTS
REFERENCES
The fundamental quantity sought in ballistometrics is the coefficient of restitution. Tests on the skin as a specific extension of the measurement from its use on other materials have been reported to clearly distinguish elastic modulus differences between young and old, various body sites, as well as similar changes after pharmaceutical treatments.8,12 Ballistometry is attractive for several reasons. It is noninvasive and easy to use. No probes have to be attached to the skin. The instrument is not as expensive as for example the dynamometer. Although one cannot obtain data on the status of the stratum corneum as one does with shear measurements, it provides a practical indication of underlying tissue changes, i.e., for example, expansion from retin-A or other similar treatment. As has been pointed out, the depth of tissue responding to ballistic impact will depend upon the height to which the falling mass is raised. Thus specific skin layers will be responsive in each case. Skin structure variations between body sites and between individual subjects will of course have an effect. However, in any series of tests planned to study response to a greater or lesser depth it is advisable to select an impact velocity and not change. The moment of inertia of the pendulum, although variable, is usually not altered after the performance of the apparatus has been optimized.
1. Timoshenko, S. and Young, D.H., Engineering Mechanics, McGraw Hill, New York, 1940. 2. Loney, S.L., A Treatise on Elementary Dynamics, Cambridge University Press, New York, 1900. 3. Handbook of Chemistry and Physics, Chemical Rubber Co., Cleveland, Ohio, 1946, pp. 30, 1302, 2328. 4. Hollinger, R. and Thelen, E., Design and development of an impact tester for use with asphalt, Report R-11, National Asphalt Research Center, Franklin Institute, Philadelphia, 1956. 5. Vogelsang, K., Ueber mechanische Gewebsprüfung am Auge, Arch. f. Augenh., 108, 714, 1934. 6. Wiegersma, G., Elastometry of the eye, Am. J. Ophth., 39, 811, 1955. 7. Mamelok, A.E. and Posner, A., Measurements of corneal elasticity, Am. J. Ophth., 39, 817, 1955. 8. Tosti, A., Giovanni, C., Fazzini, M. L., and Villardita, S., A ballistometer for the study of the plasto-elastic properties of skin, J. Invest. Dermatol., 69, 315, 1977. 9. Hertz, H., J. Math. (Crelle), 92, 1881. 10. Timoshenko, S., Theory of Elasticity, McGraw Hill, New York, 1934, p. 390. 11. Lamb, H., Proc. London Math. Soc., 35, 141, 1902. 12. Fthenakis, C.G., Maes, D.H., and Smith, W.P., In vivo assessment of skin elasticity using ballistometry, J. Soc. Cosmet. Chem., 42, 211, 1991.
The Cutaneous Vasculature Skin Color and Blood Vessels
74 Colorimetry Wiete Westerhof Department of Dermatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
CONTENTS 74.1 Introduction............................................................................................................................................................635 74.2 Aim ........................................................................................................................................................................636 74.3 Methods..................................................................................................................................................................636 74.3.1 CIE Color System......................................................................................................................................636 74.3.2 Technical Details of the Colorimeters.......................................................................................................637 74.4 Sources of Error.....................................................................................................................................................639 74.4.1 Instruments.................................................................................................................................................639 74.4.2 Measurement Method ................................................................................................................................639 74.4.3 Reproducibility ..........................................................................................................................................639 74.5 Correlation with Other Methods ...........................................................................................................................639 74.6 Fields of Application .............................................................................................................................................641 74.6.1 Ultraviolet Radiation-Induced Erythema Measurement ...........................................................................641 74.6.2 Measurement of Erythema Due to Irritants and Contact Allergens .........................................................642 74.6.3 Measurement of the Blanching Effect of Corticosteroids28 ......................................................................642 74.6.4 Measurement of Skin Color1 .....................................................................................................................643 74.6.5 Measurement of Ultraviolet-Induced Pigmentation..................................................................................643 74.6.6 Measurement of Dose-Response Curves of Ultraviolet-Induced Erythema and Pigmentation52 ............644 74.6.7 Measurement of Bleaching Effect by Depigmenting Agents ...................................................................645 74.7 Conclusions............................................................................................................................................................645 Acknowledgments ...........................................................................................................................................................645 References .......................................................................................................................................................................645
74.1 INTRODUCTION Color information may be acquired and communicated in various ways. However in a scientific context there are requirements for consistency independent of time, distance, and language.1 Although the average person may be able to distinguish several thousands of colors, the description of visual observations by the use of general color terms is the least satisfactory in terms of precision. The range of names upon which people reliably agree is very limited and there is no simple way of using visual observations to describe the differences between colors.2 Color assessment based on comparison with sets of colored samples, known as color-order systems, can increase the range and reliability of color designations significantly.2 The Munsell color-order system3 is the oldest and perhaps the most widely recognized. Color comparisons may involve metamerism, that is: the actual physical composition of the colored lights might be
different but these are perceived as the same color. Therefore color-order systems are most reliable when used under fixed conditions of illumination to match colors of relatively flat surfaces that are uniformly pigmented. They are less appropriate for samples that are heterogeneous in surface structure. A particular limitation of color-order systems is that they only identify single colors and do not provide a way of specifying the nature or magnitude of color differences. The definitive method of measuring skin color uses the recording spectrophotometer adapted for reflectance readings. However these readings are not communicable, as they do not conform to an international standard. The limitations of visual observations may, in principle, be overcome by the instrumental evaluation of colors and color differences according to the system of color measurement established by the CIE (Commission Internationale de l’Eclairage).2,4–6 This system has become widely 635
636
Handbook of Non-Invasive Methods and the Skin, Second Edition
used with the availability of reliable reflectance spectrophotometric instrumentation that conforms to the recommendation of the CIE for the measurement of the color of surfaces.7–9 This method does not give information about the substances generating the color, but it is highly appropriate for color matching (e.g., grading of erythema and grading of melanin pigmentation, etc.). In fact, the use of the CIE tristimulus values is a concise and reliable way of approximating an actual normal color observer. In other words, by using these units one can imitate what is normally done by visual inspection but with greater reliability and reproducibility than a given human observer could achieve.
Yxy color space y
0.8
0.6
0.4
74.2 AIM The objective of this chapter is to present qualitative and quantitative measurement methods of color of the skin, mainly erythema and melanin pigmentation. This is done with a tristimulus computer-controlled color analyzer which measures reflected object color in an accurate and reproducible way utilizing CIELAB (CIE 1976 L*a*b*) color space values. These CIE color space parameters are proposed for the unambiguous communication of skin-color information that relates directly to visual observation of clinical importance or scientific interest.
74.3 METHODS 74.3.1 CIE COLOR SYSTEM The perceived color of objects depends on: (1) the nature of the illuminating light, (2) its modification by interaction with the object, and (3) the characteristics of the observer response. The CIE system defines these conditions as follows: (1) The relative spectral energy distributions of various illuminants, known as CIE standard illuminants, are specified and available as published tables,2,4–9 (2) the modification of an illuminant by interaction with the object is measured with a reflectance spectrophotometer having an optical configuration that conforms CIE recommendations,4 and provides a visible spectrum expressed as the fractions of incident light intensity reflected in the wavelength range 400–700 nm; (3) the nature of human color vision has been quantified for the purpose of color measurement in terms of three color-matching functions x, y, and z (Figure 74.1). Three are required because color vision has been found to be trichromatic: a single perceived color may be regarded as resulting from the effect of three separate stimuli on the visual cortex.6 Their numerical values are available as published tables and are known collectively as a CIE standard observer.2,4,5,9 They may be regarded simply as a numerical description of average human color vision. From a practical viewpoint
0.2
0
0.2
0.4
0.6
x
FIGURE 74.1 X, Y, and Z color coordinates.
their use has been made more convenient by incorporation of the tabulated values within the software provided with color-measuring reflectance spectrophotometers. Similarly, the tabulated values of the relative spectral-energy distributions of various CIE illuminants are provided within the software because they do not exist as actual physical sources of light within instrumentation. Colors are measured in terms of their tristimulus values X, Y, and Z by combining a selected table of illuminant data, the measured values of reflectance, and a selected table of color-matching functions with three summations, each having the form Σ ER x– = X At selected intervals in the wavelength range 400 to 700 nm the relative energy (E) of the chosen illuminant is multiplied by the fraction reflected (R) and the numerical value of the standard observer (x or y or z). A wavelength interval of 10 nm, which requires 31 terms in each summation, gives adequate precision for most purposes. Modern color-measurement instrumentation normally incorporates microcomputer hardware and software so that the spectral measurements and subsequent calculations are integrated so as to produce a copy of the results within a few seconds. However the tristimulus values of colors are difficult to relate to the experience of seeing them. In addition, in any study involving comparisons, contrasts, or changes, tristimulus values do not directly enable
Colorimetry
637
L∗= 100 A ΔE A’
+b∗ h0
B −a∗
+a∗
−b∗
L∗ = 0
FIGURE 74.2 Color space; CIELAB.
measurement of the difference between two colors. This concern has now been overcome by using the tristimulus values to calculate the CIE 1976 L*a*b* (CIELAB) color space values.2,4,7,9 The mathematical manipulations that convert tristimulus values to CIELAB color space values enable colors to be regarded as existing in an approximately uniform three-dimensional space in which each particular color has a unique location defined in terms of its cartesian coordinates with respect to the axes L*, a*, and b* as shown in Figure 74.2. Modern computer-controlled reflectance spectrophotometers provide for automatic calculation of CIELAB values from the spectral data they produce. The measured value of a color has been recommended by the CIE as the psychometric correlate of the visually perceived color attribute of “lightness”,4,7 to which the descriptive terms assigned might include the words “light”, “dark”, etc. In other words, L* would measure the change along a gray scale from black to white that visually varied in perceptually uniform manner. The L* scale, which ranges from 0 for a theoretical black to 100 for white, corresponds to the notion of the value attribute in the Munsell color system. The a* and b* coordinates may be conceptually related to Hering’s opponent color theory,8 which was based on the proposition that the retina of the eye contains opponent color channels that distinguish colors according to their red-vs.-green and yellow-vs.-blue attributes. In CIELAB space they are more useful when converted into polar coordinates. This facilitates the definition of a hue angle, h* = arctan (b*/a*) (Figure 74.2), which is recommended by the CIE as the psychometric correlate of the visually perceived attribute of hue (e.g., red, orange, yellow, etc.).4,8 Measured hue angles make the use of visually assigned hue terms unnecessary, although it is simple and
often convenient to relate them in a general way. The general angular position of some of the main generic hues are shown in Figure 74.1. CIE hue angle corresponds conceptually to the attribute of hue in the Munsell colororder system, but no simple relationship has been found between measured hue-angle values and Munsell hue designations.10 Colors for which both a* and b* are zero, and therefore lie on the L* axis, are termed achromatic and would be perceived as gray, white, or black. The visually perceived color attribute of “saturation”, which might be described by the use of the terms “weak”, “strong”, etc., may be measured in terms of its distance away from the L* axis in the a*b* plane. This is the length of the line C in the diagram. It is termed the CIE(1976)a*b* chroma and is calculated using coordinate geometry as C = [(a*)2 + (b*)2]1/2 It corresponds conceptually to the attribute of chroma in the Munsell system but the measured values do not relate in any simple way to Munsell designations.10 Thus the use of CIELAB coordinates enables measurement of the three attributes of a color by which it is visually distinguished. CIELAB space is not only more convenient than tristimulus values with respect to its conceptual relationship to the actual experience of seeing colors but it has the important advantage of providing a means of measuring the differences between any two colors.2,8 Their color difference (E) is calculated, using coordinate geometry, as the length of the line joining their coordinate positions: E = [(L*)2 + (a*)2 + (b*)2]1/2 In some circumstances the differences between two colors may also be considered in terms of differences in hue angle and/or chroma. Modern computer-controlled reflectance spectrophotometers used for color measurement often include the software necessary for automatic calculation of color differences from two sets of spectral data. For more details the reader is referred to the appendix of the excellent paper by Weatherall and Coombs.1
74.3.2 TECHNICAL DETAILS
OF THE
COLORIMETERS
Several colorimeters have been manufactured for medical and scientific use. Here we discuss three of the most commonly used instruments. 1. Labscan 6000® (Hunter Associates Inc., USA) scanning reflectance visible spectrophotometer has 0° illumination and 45° viewing geometry with the specular component excluded. The instrument is calibrated with a supplied white standard traceable to the National Bureau of
638
Handbook of Non-Invasive Methods and the Skin, Second Edition
Standard’s perfect white diffuser. The spectrophotometer is controlled by an IBM-XT microcomputer, which performs all color calculations from the digitized spectral data by means of a menu-driven set of programs supplied with the instrument. The skin is measured over a 30-mm diameter open circular port with a 26-mm illuminated area in the horizontal upper surface of the sensor module. Reflectance spectra over the wavelength range of 400 to 700 nm, requiring about 3 s for measurement, can be obtained. Tabulated data for CIE illuminant D65 and the CIE 1964 10′ standard observer are selected under software control and combined with the spectral data at 10-nm intervals to compute the CIELAB L*a*b* values. The latter two are then further converted to CIELAB color space and chroma. The instrumental setup is not suitable for field conditions and routine clinical applications because of its volume and weight. 2. The Minolta Chromameter CR 200® (Osaka, Japan) is a lightweight and compact tristimulus color analyzer for measuring reflected object color (Figure 74.3). Utilizing high-sensitivity silicon photocells, filtered to match CIE Standard Observer Response, the Chromameter CR 200® assures good accuracy and reproducibility (measuring error 1%). Readings, taken through the measuring head, are processed by a built-in microcomputer and then presented digitally on a liquid-crystal display. The measuring head
FIGURE 74.3 Chromameter.
also contains its own standard light source: a high-power xenon arc lamp which provides diffuse illumination from a controlled angle for vertical viewing and constant, even lighting on the object. The illumination chamber of the Minolta equipment is cylindrical, with a conical opening pointing toward the skin surface. These two parts are separated by a diffusing plate with an inner circular aperture 4 mm in diameter. The outer aperture is 11 mm, and the diameter of the estimated measuring area is 8 mm. The photo receiver is centered in the top of the chamber, and the tristimulus optical analysis unit with six silicon photocells is located directly in the probe unit (in the CR series), and in previous equipment in the main body. The meter’s precise double-beam feedback system, using six photocells, detects any slight deviations in the xenon light’s spectral distribution, and the microcomputer compensates for them, thus ensuring the utmost accuracy in measurements. The measuring area is circular and 8 mm in diameter. The Chromameter offers five measuring modes: two chromaticity-measuring modes — Yxy and L*a*b* — and three colordeviation measuring modes — ± (Yxy1 ± (L*a*b*) and Eab). The weight of the Minolta CR-200 is 1951 g. 3. The Dr. Lange Micro Color® (Dr. Bruno Lange GmbH, Dusseldorf, FRG) is computerized with a number of facilities. L, a*, and b* values are presented on a digital display. It is based on illumination with a xenon flash light. In the Lange equipment the chamber is spherical with a circular aperture, 5 mm in diameter, directed toward the skin surface. Measuring and reference photo receivers are mounted on the top of this sphere, but eccentrically aligned by 6° and 8°. Optical fibers lead the signal to the main body containing six silicon photocells for tristimulus analysis (red, green, blue) of both signal and reference light, according to DIN 5033. The Lange equipment has a short cable (600 mm) between probe and main body. The weight of the Lange equipment is 6 kg. It works at a slightly slower rate than the Minolta equipment. The technical reproducibility of the two colorimeters as given by manufacturers is the same, i.e., 0.15 E* to white. Both are calibrated against their respective white calibration tiles before use. For the practical handling of the colorimeters the reader is referred to the instruction manuals of the respective apparatus.
Colorimetry
639
74.4 SOURCES OF ERROR
12
74.4.1 INSTRUMENTS
74.4.2 MEASUREMENT METHOD The skin is not a homogeneous surface structure. It differs in primary, secondary, and tertiary skinfolds depending on the site of the body. Also the greasiness and humidity of the skin differ from site to site. This will affect the glossiness of the skin. Similarly the hairs, blood vessels, blemishes spots, and scars may determine the color aspects of the skin. The reflection from the skin is influenced by all the above factors leading to incorrect color measurement and significant site-related variations in reflectance. It is therefore imperative to carefully select the skin sites to be measured and to take the measurements in duplicate or triplicate. If erythema is to be measured it has to be borne in mind that this is a function of blood vessel size and flow. These parameters can be influenced among other things, by compression. Therefore the measuring head with the aperture must not be pressed too hard against the skin. The measuring head must also be kept in a perpendicular fashion to the skin surface. Otherwise differences in reflectance of the light source will occur.
74.4.3 REPRODUCIBILITY Another source of error lies in the method of provoking the color change. In UV-induced erythema a 3-plus visual grading with edema may be less red due to capillary compression and extracellular fluid accumulation. The time lapses of UVA and UVB induction is different. Therefore the moment of measurement is critical. If
10
8
Δa*
It is not possible to go into much technical detail within the framework of this chapter. A source of error could be the illumination. A light source comparable to the sun does not exist. Most of the artificial light sources emit incontinuous bands out of the whole light spectrum. This may give rise to metamerism. The filters used in this tristimulus chromameter are not exactly monochromatic. Especially at the periphery of the wave band of absorbtion, deviations in the filtering capacity exist that can lead to incorrect color measurement. Also the liquid or glass fiber light guide can contribute to inhomogeneous loss in light spectrum constituents. The angle of illumination of the object being measured, in our case the skin, determines the degree of specular reflection which negatively influences the color analysis of the reflected light. The lack of standardization in the production of different colorimeters will lead to incomparable results of measurement. One given instrument is usually very reproducible in measurements of the same object.5
6
4
2
29.4
47.0
75.2
117.5
UV-B dosis in mJ. cm−2
FIGURE 74.4 UV sensitivity of the skin related to body site. Diagram of mean Δ a* vs. irradiation doses given to 5 body regions of 8 volunteers: •-• right scapular back, - right lumbar back, - distal arm, - medial arm, ▫-▫ proximal arm.
follow-up measurements are required from one and the same measuring site it is necessary that care is taken to meticulously place the aperture over exactly the same place as chosen for the previous measurements. As was already explained a few milimeters away from the measuring site the characteristics of the skin may be completely different. We have devised measuring bracelets that allowed us to exactly measure the UV-irradiated sites of the skin of the lower arm for the development of erythema and pigmentation (Figure 74.4). For other sites of the body similar appliances could be devised. These devices help to increase the reproducibility of the measuring technique. In Section 74.6, Fields of Application, the methodology of inducing color changes or measuring color changes over time need to be standardized in order to be able to make chromametric measurements in a reproducible way.
74.5 CORRELATION WITH OTHER METHODS Reflectance chromameters have not been compared with spectrophotometers. However Weatherall et al.1 used a spectrophotometer, which converted the reflected wavelengths taken at 10-nm intervals into CIELAB values. Serup et al.11 evaluated two commercially available colorimeters (Minolta chromameter and Dr. Lange microcolor) and compared them with laser-Doppler flowmetry,
640
Handbook of Non-Invasive Methods and the Skin, Second Edition
which is widely used for quantification of erythema and cutaneous inflammation. Irritant reactions after application of sodium lauryl sulfate (SLS) were studied. Evaluations of technical reproducibility using white tiles and a red standard simulating erythema are presented. Repeated measurements (n = 10) were performed with the two colorimeters on their respective white calibration tiles and on a red color standard simulating moderate erythema.12 The difference between the two standard tiles was measured by the Lange equipment. The linearity of the colorimeters was examined using different color scales from the color-standard book.12 Red (erythematous scale) was registered as mainly parallel a*-axis curves both in the lower end, relevant for erythema measurement, as well as in the upper end. The Lange equipment showed a steeper increase in values in comparison with the Minolta equipment. The Lange equipment appeared more sensitive for determination of yellow, with a steeper increase in values on the b* axis. The equipment determined blue with parallel curves. However, the Lange equipment measured at a substantially lower level, i.e., about 40 U lower on the b* axis. Green was also determined at a lower level, on the a* axis, by the Lange equipment, especially at high intensities. In the three axes only a small part of the large dynamic range of the colorimeters was relevant for skin color and erythema measurement. Color measurement by colorimeters using the CIE system was useful for the characterization of erythema, confirming previous studies.3,4 In this paper dealing with light-induced erythema a more detailed evaluation of erythema was performed. Serup et al.11 concentrated on the technical aspects of color measurement. Laser-Doppler flowmetry proved useful for the characterization of erythema as shown already, with a positive correlation to clinical scoring and colorimetry (a*-axis movement toward red).15,16 The colorimetry and the flowmeter probably measure somewhat different features of cutaneous circulation, i.e., the colorimeters mainly measures the capillary accumulation of blood and the flowmeter measures the total blood flow, mainly determined by the arteriolar tone.17 Thus, the methods are, to some extent, complementary in the evaluation of erythema. The technical reproducibility of the two colorimeters was very good with low coefficients of variation, although higher for a* then for the other axes. The variation in color in the group of subjects studied was remarkably low with respect to the L value, and limited with respect to the a* value. The high coefficient of variation of b*-axis values recorded with the Dr. Lange equipment was partly attributable to the low mean values. Standard deviations might indicate a similar reproducibility of colorimeters in control skin and erythematous skin in contrast to laser-Doppler flowmetry, which showed a five times higher standard deviation in erythematous skin.
This is probably due to the small area of skin illuminated by the laser light in contrast to the colorimeters, which operate with a far larger area and thus provide a better overall assessment of the circulation. Thus, with the laser-Doppler it is of special importance that results are obtained through averaging a number of recordings of the irritant reaction studied. In contrast, a single colorimeter measurement may suffice, depending on the situation or the purpose. It is well known that laser-Doppler flowmetry, recording a dynamic situation, is sensitive to factors such as noise and talking. Colorimetry is not affected by such environmental sources of variation, making the method more suitable for routine work. The two colorimeters were essentially equally suited as tools for quantification of erythema. However they do not give identical values, and the technical part of the study concluded that they record some differences on the different color axes. Thus, a simple transformation factor cannot be applied. The CIE color system has the advantage of being internationally accepted, and equipment using this system is available. Modifications were developed by Munsell and more recently by a Swedish group, the latter system being named NCS (natural color system).18 This system was suitable for assessment of skin color during Argon-laser treatment of port wine stains. However, for the wider exchange of information it is advantageous if a uniform system is used, although experts may not find it ideal in every situation. It should not be forgotten that inflammatory responses start with vasodilatation, while edema formation takes over in advanced reactions, compressing the vasculature. Laser-Doppler flowmetry and high-frequency ultrasound measurement of skin thickness of histamine wheals, irritant reactions after application of SLS and allergic patchtest reactions show that grading of reactions relative to vasodilatation is only possible in light and moderately severe reactions.16,19,22 However, ultrasound examination also shows the development of edema within irritant reactions is less in comparison with allergic reactions. Thus, in studying irritancy, colorimetry is likely to be suitable for the grading of a relevant spectrum of reactions, which previous studies also demonstrated.13,14 In conclusion, Serup et al.,11 found that colorimetry using the CIE system is reproducible and useful for quantification of erythema, even under busy laboratory conditions. In group studies (winter) using untanned skin, the influence of melanin pigmentation appears negligible. However seasonal and regional differences in skin color need to be studied in more detail. The relevance and potential areas of application in experimental and clinical dermatology are many since this color system takes into account the actual color perception of the human eye, which spectrophotometric recording of light absorbency and reflectance does not do.
Colorimetry
74.6 FIELDS OF APPLICATION The application of reflectance chromameters in clinical practice and dermatological research is wide and still expanding. The most important applications reported in the literature are measurement of: 1. 2. 3. 4. 5. 6.
UV radiation-induced erythema erythema due to irritants and contact allergens the blanching effect of corticosteroids skin color UV-induced pigmentation dose-response curves of UV-induced erythema and pigmentation 7. the bleaching effect of depigmenting agents
641
determined with the aid of a Zeiss M4 QIII® monochromator and an EG&G 550® radiometer at a distance of 25 cm from the xenon-arc source. The UVB output was measured with an EG&G 550® radiometer fitted with a calibrated UV-enhanced silicon detector probe, on top of which a Scott UVB filter was mounted. This filter matches the erythematogenic action spectrum of normal skin at the long wave side as presented by Berger.25 The monitored UVB output of the solar simulator per second was 2.35 mW/cm2. After 24 h all irradiated sites were visually scored by two independent investigators according to the following erythema grading scale: 0 = no perceptible erythema /2 = slight or partial erythema 1 = minimal erythema with defined borders (MED) 11/2 = three arbitrarily increased steps of erythema formation without edema or vesiculation 2 (2 + erythema in the commonly used scale) 21/2 3 = erythema with induration and blisters 1
74.6.1 ULTRAVIOLET RADIATION-INDUCED ERYTHEMA MEASUREMENT The assessment of sensitivity of human skin to UV radiation is important to photochemotherapy, diagnosis of photodermatosis, photoaging, photocarcinogenesis, and photoprotection. As an example we use the method described by Westerhof et al.5 Eight healthy volunteers (four males, four females, 22 to 26 years old) with skin type II and III were irradiated on seven different regions of the body with a solar simulator. The regions were chosen following the reports of Farr and Diffey23 and Olson et al.24 On the ventral side of the left forearm, the three sites were: proximal near the elbow, distal near the wrist, and in between. On the subject’s back the four sites chosen were related to the vertebral level: upper dorsal (T3) — a spot to the left and a spot to the right of the spine, and for lumbar (L2) — a spot to the left and a spot to the right of the spine at 1 cm from the midline. Care was taken to select the spots in such a way they were without terminal hairs and without nevi or scars, etc. The test areas were then exposed to a series of four simulated solar radiation dosed increasing in delivered energy dose by a factor 22/3 of the foregoing dose. The radiation dose was chosen in such a way that one or two test spots (out of four) received a dose equal to one minimal erythema dose (MED). The solar simulator consisted of a 1000-W xenon arc in a lamphouse fitted with a quartz collimating lens (f/0.7) and a water filter with quartz windows. A dichroic (‘cold type’) mirror which reflects most of the radiation below 500 mm was placed into the beam to minimize the infrared radiation further. A liquid light guide conducted the radiation from the exit port of the lamp optics to the volunteer’s skin. An applicator that housed a quartz lens was attached to the proximal end of the light guide and produced a uniform beam of radiation, 10 mm in diameter, on the skin. The stimulated solar emission curve was
Thereafter the erythema was measured with the Chromameter II Reflectance.® The same site was measured twice on the L*a*b* mode. Of the three chromaticity values L*a*b*, the a* value is the best index for the erythema measurement. This is to be expected since a* stands for chromaticity coordinates ranging from perceived red to green in the three-dimensional color space diagram (CIE 1976). After measuring the erythematous spot, a piece of nonirradiated adjacent skin was measured for comparison. The difference between the two gives the erythematous color aspects of the skin (a*). The Chromameter was always held perpendicular to the skin surface, hardly touching it. The aperture of the measuring unit is fitted with an applicator so that the cutaneous capillaries are not compressed. Angle errors in the vertical measuring position of the instrument’s head did not show any influence on the a* and b* chromaticity coordinates and gave a slight deviation of L* (coefficient of variation 6, 8%; n = 20). The differences between duplicate readings showed a SD of 0.30 in case of L* in the range 57.0 to 67.9; a SD of 0.14 in case of a* in the range 6.5 to 20.7 and a SD of 0.21 in case of b* in the range 11.8 to 22.2. The L* value decreased when erythema developed, indicating some skin darkening but to a relatively smaller extent than the increase of a*. The b* values did not change significantly in our erythema measurements. Westerhof et al.5 tried to relate the visual ratings (0, 1/2, 1, 11/2, 2, 21/2, 3) to the Chromameter measurements.
642
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 74.5 Bracelet to make follow-up measurement with chromameter possible. The measuring head exactly fits into the holes of the bracelet, which can be applied to the lower arm in only one way as a notch should exactly overlay an ink marking made on the skin.
In Figure 74.2 the mean visual grading is plotted as a function of a*. It is easy to see that in the range of 1 to 11 there is a linear relation between visual grading and the Chromameter reading. Above a Chromameter reading of 11 a saturation of the visual grading can be seen. Westerhof et al.5 were able to confirm the findings of Olson et al.24 and Farr and Diffey.23 The sensitivity to erythema formation increases when one compares the distal part of the flexor side of the forearm with the proximal part near the elbow while keeping the energy dose of UVB radiation constant (Figure 74.5). The same is true on the subject’s back when comparing the thoracal region (T3) with the lumbar region (L2). It is also remarkable that the back is more sensitive to UV radiation than the flexor side of the forearm.
74.6.2 MEASUREMENT OF ERYTHEMA DUE TO IRRITANTS AND CONTACT ALLERGENS In dermatology and the cosmetic sciences tristimulus colorimetric measurements of erythema offer an advantage in quantifying skin reactions to allergic reactions evoked by patch tests and irritation responses. We studied the role of melanin in inflammatory reactions by provoking erythema with the potent irritant anthralin in vitiliginous skin versus normal pigmented skin.26 The locations chosen for anthralin application were the trunk, arms, and legs under the condition that normal control skin was to be as close as possible to lesional skin. These precautions were taken because of the known existence of a site-dependent inflammatory reaction to anthralin.27 Anthralin was used in concentrations of 0.1, 0.5, 1.0, and 5.0% dispensed in a cream containing 1% salicylic acid as a stabilizer; the vehiculum served as a control. The serial applications of the four concentrations of anthralin were done with
patch-test chambers. The patch-test chambers were fixed with hypoallergic cloth tape. The patch-test chambers were removed after 24 h and the irritation response was read after the next 24 h. For evaluation of the irritation response the following semiquantitative scale was used: 0, no erythema; 1, erythema (+, slightly visible erythema, ++, moderate erythema, +++, intense erythema, ++++, extreme erythema); 2, erythema with induration; and 3, erythema with induration and blisters. To avoid an optical illusion, a white paper was used to cover the test area so that the surrounding reference area was always the same. For an accurate quantitative evaluation of the anthralin-induced erythema we used the Chromameter Reflectance II® (Minolta). The control substance did not induce any inflammatory reaction. The erythematous reactions with 0.1% anthralin were weak, those with 0.5 and 1.0% tended to be stronger, and those with 5.0% showed a very intense erythema. The irritation response never reached beyond the margins of the test chamber and at the time of reading, neither induration nor blisters were observed. With the visual grading the irritation response was considered to be more intense in normal skin than in vitiliginous skin. In contrast to the visual estimations the chromameter readings indicated that the changes in redness due to anthralin were more pronounced in vitiliginous skin than in normal pigmented skin. The staining produced by anthralin appeared to be regular and of the same intensity when vitiliginous and neighboring pigmented test sites were compared. This fact was confirmed by colorimetric measurements. The relation between the degree of pigmentation and the irritation response to anthralin could only be evaluated making use of a sensitive colorimetric method. Visual estimation of the erythema can be misleading as the redness observed in the pigmented skin is known to be comprised of genuine erythema and a red component of the complex brown color. The human eye cannot discriminate these two different sources of redness. Colorimetric measurements can help to overcome this difficulty. From this study it was concluded that melanin pigment had an inhibitory effect on the anthralin-induced irritation response probably by scavenging free radicals and reactive oxygen species.
74.6.3 MEASUREMENT OF THE BLANCHING EFFECT OF CORTICOSTEROIDS28 The vasoconstrictor assay 29 (human skin-blanching assay) is a well-established method for ranking topical steroids. Cornell and Stoughton30,31 gave a rank for U.S. formulations with seven orders of potency. In Europe, the current classification for topical corticosteroids has four ranks,32,33 with group I being the most potent. In this assay the visual assessment of blanching generally gives good
Colorimetry
643
results, but under certain circumstances is not a very satisfactory method, e.g., nonoccluded tests for weak or moderately potent formulations. Many studies using quantitative instrumental methods have attempted to improve the objectivity of this test. Some have completely failed34,36 and those that succeeded were too cumbersome and time consuming for general use.37,39 The only study using a colorimetric system40 was not convincing and this method was abandoned. Queillo-Roussel et al.28 used the Minolta tristimulus colorimeter to quantify the blanching effect of topical corticosteroids in a nonoccluded vasoconstriction test. To investigate the influence of time on variations in colorimetric parameters, an initial series of measurements was performed on day one on six predetermined sites on the ventral surface of the forearm of six healthy volunteers every 2 h over a 12-h period. The colorimetric values were shown to be site related but hourly variations occurred with similar profiles for all sites. On day two, four topical corticosteroid creams, representative of their potency group, as well as a cream base were applied in a randomized double-blind manner on five predetermined sites. Visual gradings and colorimetric measurements were carried out every 2 h over the following 12-h diurnal period and were continued on day three. The colorimetric parameters L* (luminance) and a* (color hue ranging from green (–) to red (+)) gave a rank order correlated to corticosteroid potency that showed superior discrimination compared to simple visual grading. In this study L* was a more discriminative parameter than a*.
74.6.4 MEASUREMENT
OF
SKIN COLOR1
Reflectance colorimetric methods have been widely applied in studies of the chromatic characteristics of skin in the context of anthropologic studies of human population genetics.41–43 CIELAB color space parameters also have potential cosmetic applications, such as the design of prosthetic devices and the use in color matching of skin grafts in plastic surgery (e.g., after skin tumor excision). For complete color measurement it has been shown that filter colorimetry can give results that differ systematically from those based on scanning spectrophotometry.2,44 Evaluation of the ventral upper arm skin color in volunteers of various ethnic origin were measured by applying reflectance spectroscopy with CIELAB color space parameters. The volunteers classified themselves as being of European, Chinese, Indian, or Polynesian ethnicity, or of mixed race. Measurements of the visible reflectance spectrum of skin and the calculation of CIELAB values provided a practical numerical basis for quantifying the perceived color of human skin. The excess of data in a full spectrum can be reduced to a set of color space parameters that relate directly to the appearance of the color as it would be clinically observed and without
making any assumptions about the nature of the pigments involved. However the same spectral data could be used if there was a requirement beyond the simple numerical specification of appearance as in the determination of the concentration of chromophores in physiologic studies.
74.6.5 MEASUREMENT OF ULTRAVIOLET-INDUCED PIGMENTATION Tanning is a cardinal response of human skin to UV radiation. A spectrophotometric technique for measuring changes in the rate of transition between the first erythematous response and delayed pigment formation has important applications for evaluating the efficacy of sunscreens and potential pigment enhancers (e.g., psoralens with UVA and tyrosine).44,46,47 Unfortunately, the clinical usefulness of early spectrophotometers for measuring this color transition of skin has been limited by technical problems including the stability of the instrument light source, the need for external references, and the use of multiple filters to obtain the exact skin color.48,49 Furthermore, the wide spectrum of human skin colors has made comparative treatment studies with these devices virtually impossible since there is no means of recording baseline data on an individual subject for later paired comparison.50 With a handheld tristimulus colorimeter these problems have been overcome. Natural skin tones can be stored in the colorimeter memory from a subject for direct comparison to the solar exposed skin color. The capability of the tristimulus colorimeter to simultaneously evaluate the hue and saturation of skin color affords an improved opportunity to quantitate the transition from cutaneous erythema to tanning. Seitz et al.51 tested areas on the back, which were UV irradiated at the exposed sites previously estimated to produce 1, 1.5, and 2 MED for each individual. 24 h later erythema and tanning were scored by a dermatologist. The subjects received repeated UV exposures once every 48 h for 14 d. The Minolta chromameter showed subtle, continuous transition between the primary erythamous response and delayed tanning of the skin. The initial changes were below the visual threshold for detection as seen in the dermatologist’s scores and in the instrument validation responses. Linear regressions comparing the dermatologist’s scores with the meter values indicated that erythema to UVA and UVB exposure was perceived as a pure red shift in skin color, darkening in intensity with continued exposure until tanning started to appear. In contrast, tanning was perceived as an intensifying yellow hue superimposed on the existing erythema. However, since the human eye integrates all visual stimuli, any admixtures of color or other variables, e.g., skin blemishes, background skin tone, quality of lighting, etc., would partially explain the variability in the dermatologist’s erythema
644
Handbook of Non-Invasive Methods and the Skin, Second Edition
scores. These results also emphasize the difficulty in visually distinguishing between erythema and tanning.
74.6.6 MEASUREMENT OF DOSE-RESPONSE CURVES OF ULTRAVIOLET-INDUCED ERYTHEMA AND PIGMENTATION52 The assessment of sensitivity of human skin to UV radiation is important to photochemotherapy, photodermatoses, photoaging, photocarcinogenesis, and photoprotection. Today time, the most frequently used means of attempting to predict UV sensitivity, without phototesting, is classification of the persons’s skin type. This Working Classification of Sun Reactive Skin Types, as introduced by Fitzpatrick, is based on the history of an individual’s tendency to sunburn and to tan, along with some racial parameters. When more objective determination of sensitivity of the skin to UV radiation is desired, is determined experimentally. A number of investigators have sought to compare and correlate the various means of predicting and measuring UV sensitivity. Olson et al.35 reported that the MED correlated well with melanosome size, quantity, density, and distribution in various skin colors. Shono et al.36 also found a good correlation between MED and skin color, but a less clear one between the minimal melanogenic dose (MMD) and either skin color or MED. Sayre et al.37 studied the erythema action spectrum of 30 Caucasians and concluded that several factors, among which was constitutional skin color, had no significant effect on MED. Haake et al.,38 studying a population consisting of Fitzpatrick’s types I to IV, found no correlation between the UV sensitivity and color of the skin. It is inconclusive from these studies whether or not the MED correlates with skin types or skin color. Most researchers dealing with photo testing recognize that, within a so-called skin type, there is a larger interindividual variation in UV sensitivity, as assessed by MED measurements. Since a value such as the MED is merely a point on a dose-response curve, it can be expected that the full curve will yield much more information than does any arbitrarily chosen single point such as the MED. In studies of the responses of human skin to UV radiation, it is convenient to have a means of estimating the reactivity of an individual’s skin without performing photo testing. The most frequently used means at present is classification of skin into one of the six skin reactive types, based primarily on the history of previous responses of the skin to sunlight. This has proved to be a convenient method because it requires no measurements and can be done rapidly. When measurement of skin responses to UV radiation is performed, the usual method is the measurement of a MED. The MED is only an estimate of the amount of UV radiation required to produce detectable erythema and does not reveal the incremental increases in
erythema with increasing UV doses. Dose-response data for erythema would more accurately measure UV responses of human skin,23 but obtaining such data has been difficult. The availability of a sophisticated chromameter interfaced with a computer has now made possible the easy, objective measurement of erythema and pigmentation responses to UV radiation and the obtaining of doseresponse data. Westerhof et al.,52 performed comprehensive photo testing, including the determination of MED and MMD values and the measurement of dose responses for erythema and pigmentation, on a large number of volunteers and attempted to assess the value of the skin type and objectively measure skin color in predicting the skin sensitivity to UV radiation. Westerhof et al.,52 confirmed the results of Stern and Momtaz53 and those of Azizi et al.,54 who found a greater than threefold difference between the highest and the lowest MED values in each skin type, reflecting the large variation in individual MED values. The conclusion from this study52 is that the skin type does not correlate well with the MED. The reason for this may be that the skin type does not accurately estimate UV sensitivity or that the MED is not a sensitive measurement of UV responses of human skin. The lack of a close relationship of skin type with the dose-response curves for erythema and pigmentation would point to the skin type as an inadequate predictor of UV responses. With the reflectance measurement (Y) obtained with the chromameter prior to photo testing, we were able to objectively estimate skin color. The constitutional skin color did not correlate well with the skin type and neither with measured MED or MMD values. However, we found the objectively measured skin color to be the best predictor of the dose-response measurements for both erythema and for pigmentation. In lightly pigmented skin, the doseresponse curves were steep, whereas in darkly pigmented skin the curves were much flatter. From these studies52 it appears that dose-response data best measures the sensitivity of human skin to UV radiation. Because the technology to obtain dose-response information is not widely available, MED measurements will continue to be the most convenient means of measuring the response to UV radiation. However, for sophisticated photo testing in the future, such as in measuring the protection provided by sunscreens, the use of dose-response determinations might better reflect changes in UV sensitivity. The dose-response curves for erythema, which can be easily measured 24 h after irradiation of the skin, should become the standard for modern photo testing. In summary, skin color is a more valid predictor than skin type for the measurement of UV sensitivity, which is best expressed by dose-response curves. Although skin typing will continue to be used because of its convenience, one must be aware that it has severe limitations as a predictor of UV sensitivity. Perhaps if a means of easily
Colorimetry
645
estimating constitutional skin color at the site of anticipated UV exposure becomes readily available, this will prove a more valid means of predicting UV sensitivity.
74.6.7 MEASUREMENT OF BLEACHING EFFECT DEPIGMENTING AGENTS
BY
There is a need to show efficacy in the treatment of hyperpigmentary disorders such as melasma, lentigo solaris, cafe au lait spots, postinflammatory hyperpigmentation, etc. Many products have been advocated; few of them work after months of application, some have serious side effects such as depigmentation or even more severe hyperpigmentation. Duteil and Ortonne55 aimed to assess the activity of 20% azelaic acid cream in light-induced skin pigmentation in subjects. There were five test zones, all located on the middle of the back: two were treated with azelaic acid cream, two others with the vehicle, and one was left untreated. Each product was applied twice daily, 5 d a week, for 4 weeks on one zone, and for 5 weeks on the other. In the middle of the fourth week, the test zones were exposed to UVB + UVA + visible light, with a total of three times the minimal erythema dose distributed progressively over 3 consecutive days. For 7 and 10 d after the last irradiation, the induced photopigmentation was assessed by colorimetric and visual means. Compared with its vehicle, the azelaic acid cream had neither a depigmenting effect nor a preventive effect on the lightacquired skin pigmentation. Moreover, interrupting or continuing azelaic acid treatment after skin irradiation had no influence on the resulting pigmentation.
74.7 CONCLUSIONS As already stated, the complexity of the erythematogenic response in the optically complex multilayered system of the skin is difficult to approximate by a formula. Erythema indices based on such equations are bound to be oversimplifications, so that the computed mathematical interpretations might lead to inappropriate correlation with the degree of erythema. The advantage of measuring the erythema of the skin by using the CIE system of tristimulus values is that one does not need to make any assumptions about the processes underlying erythema formation. In clinical situations, e.g., when only the degree of sensitivity of the skin to UV radiation is required, one can simply estimate the degree of erythema while comparing it to normal neighboring skin. Therefore, especially in the early stages of erythema formation, it is not necessary to account for the different translucent layers of the skin, in particular the pigments in each of these layers and the degree of light scattering due to differences in turbidity. These other qualities are leveled out when comparing erythematous skin with normal skin. For scientific
investigations, however, colorimetric measurements are indispensible. Colorimeters using the CIE system allow for quantitative comparison of erythema and pigmentation in individuals and between individuals in a way that is consistent with visual judgements, but with greater reliability and reproducibility than would be possible by human observers. In this way, reliable dose response curves can be constructed. Colorimeters able to quantify reflected colors from the skin using the CIE system appear to be precise, quick, and handy in their operation.
ACKNOWLEDGMENTS This work was supported by the Dutch Pigment Cell Foundation. The constructive advise of Professor Oscar Estevez Uscanga and Dr. Henk E. Menke greatly increased the readability of this chapter.
REFERENCES 1. Weatherall IL, Coombs BD. Skin color measurements in terms of CIELAB color space values. J Invest Dermatol 99:468–473, 1992. 2. Billmeyer FW, Saltzman M. Principles of Color Technology, 2nd ed., Wiley-Interscience, New York, 1981. 3. Feather JW, Ryatt KS, Dawson JB, Cotteril JA, Barker DJ, Ellis DJ. Reflectance spectrophotometric quantification of skin color changes induced by topical corticosteroid preparations. Br J Dermatol 106:437–444, 1982. 4. Babulak SW, Rhein LD, Scala DD, Simion FA, Grove GL. Quantitation of erythema in a soap chamber test using the Minolta Chroma (Reflectance) Meter: comparison of instrumental results with visual assessments. J Soc Cosmet Chem 37:475–479, 1986. 5. Westerhof W, van Hasselt BAAM, Kammeyer A. Quantification of UV-induced erythema with a portable computer controlled chromameter, Photodermatology 3:310–314, 1986. 6. Hacham H, Freeman SE, Gange RW, Maytum DJ, Sutherland JC, Sutherland BM. Do pyridine dimer yields correlate with erythema inducation in human skin irradiated in situ with ultraviolet light (275–365m)? Photobiology 53(4):559–563, 1991. 7. Hunter RS, Harold RW. The Measurement of Appearance, 2nd ed., Wiley-Interscience, New York, 1987. 8. Marchesini R, Brambilla M, Clemente C, Maniezzo M, Sichirollo AE, Testori A, Venturoli DR, Cascinelli N. In vivo spectrophotometric evaluation of non-neoplastic skin pigmented lesions. I. Reflectance measurements. Photochem Photobiol 53(1):77–84, 1991. 9. Feather JW, Hajizadeh-Saffiar M, Leslie G, Dawson JB. A portable scanning spectrophotometer using visible wavelengths for the rapid measurements of skin pigments. Phys Med Biol 34:807–820, 1989. 10. Kollias N, Bager A. Quantitative assessment of UVinduced pigmentation and erythema. Photodermatology 5:53–60, 1988.
646
Handbook of Non-Invasive Methods and the Skin, Second Edition
11. Serup J, Agner T. Colorimetric quantification of erythema-a comparison of two colorimeters (Lange Micro Color and Minolta Chroma Meter CR-200) with a clinical scoring scheme and Laser-Doppler flowmetry. Clin Exp Dermatol 15:267–272, 1990. 12. Kollias N, Bager A. Spectroscopic characteristics of human melanin in vivo, J Invest Dermatol 85:38–42, 1985. 13. Babulak SW, Rhein LD, Scala DD, Simion A, Grove GL. Quantitation of erythema in a soap chamber test using the Minolta Chroma (reflectance) Meter: comparison of instrumental results with visual assessments. J Soc Cosmet Chem 37:475–479, 1986. 14. Westerhof W, van Hasselt BAAM, Kammeyer A. Quantification of UV-induced erythema with a portable computer controlled chromameter. Photodermatology 3:310–314, 1986. 15. Nilsson GE, Wahlberg JE. Assessment of skin irritancy in man by laser Doppler flowmetry. Contact Dermatitis 8:401–406, 1982. 16. Staberg B, Serup J. Allergic and irritant reactions evaluated by laser Doppler flowmetry. Contact Dermatitis 18:40–45, 1988. 17. Engelhart M, Kristensen JK. Evaluation of cutaneous blood flow responses by 133 Xenon washout and a laser Doppler flowmeter. J Invest Dermatol 80:12–15, 1983. 18. Malm M, Jurell G, Tonnquist G. Natural color system — a new method for evaluating skin colors during Argon laser treatment of port-wine stain. Ann Plast Surg 20:317–321, 1988. 19. Serup J. Quantification of wheal reactions with laser Doppler flowmetry. Allergy 40:223–237, 1985. 20. Staberg B, Klemp P, Serup J. Patch test responses evaluated by cutaneous blood flow measurements. Arch Dermatol 120:741–743, 1984. 21. Serup J, Staberg B, Klemp P. Quantification of cutaneous oedema in patch test reactions by measurement of skin thickness with high-frequency pulsed ultrasound. Contact Dermatitis 10:88–93, 1984. 22. Serup J, Staberg B. Ultrasound dor assessment of allergic and irritant patch test reactions. Contact Dermatitis 17:80–84, 1987. 23. Farr PM, Diffey BL. Quantitative studies on cutaneous erythema induced by ultra violet radiation. Br J Dermatol 111:673–682, 1984. 24. Olson RL, Sayre RM, Ecerett MA. Effect of anatomic location and time on ultraviolet erythema. Arch Derm 93:211–215, 1966. 25. Berger DS. The sunburn ultravioletmeter: design and performance. Photochem Photobiol 24:587–593, 1976. 26. Westerhof W, Buehre Y, Pavel S, Bos JD, Das PK, Krieg S, Siddiqui AH. Increases anthralin irritation response in vitiliginous skin. Arch Dermatol Res 281:52–56, 1989. 27. Dawson JB, Barker DJ, Ellis DJ, et al. A theoretical and experimental study of light absorption and scattering by in vivo skin. Phys Med Biol 25:695–209, 1980.
28. Queillo-Roussel C, Poncet M, Schaefer H. Quantification of skin color changes induced by topical corticosteroid preparations using Minolta Chroma Meter. Br J Dermatol 124:264–270, 1991. 29. Wan S, Parrish JA, Jeaniche KF. Quantitative evaluation of ultraviolet induces erythema. Photochem Photobiol 37:643–648, 1983. 30. Cornell RC, Stoughton RB. Correlation of the vasoconstriction assay and clinical activity in psoriasis. Arch Dermatol 121:63–7, 1985. 31. Robertson AR. The CIE 1976 color difference formulas. Color Research and Application 1977;2:7–11. 32. Polano MK, August PL. In: Polano MK (ed.). Topical Skin Therapeutics. Churchill Livingstone, Edinburgh, 1984, p. 101. 34. Fitzpatrick TB, Pathak MA, Parrish JA. Protection of human skin against the effects of the sunburn ultraviolet (290–320nm). In: Fitzpatrick TB et al. (eds). Sunlight and Man-Normal and Abnormal Photobiological Responses. University of Tokyo Press, Tokyo, 1974, p. 751. 35. Olson RL, Gaylor J, Everett MA. Skin Color, malonin and erythema. Arch Dermatol 108:541–544, 1973. 36. Shono S, Imura M, Ota M, Ono S, Toda K. Relationship of skin color, UVB-induced erythema and melanogenesis. J Invest Dermatol 84:265–267, 1985. 37. Sayre RM, Desrocher DL, Wilson CJ, Marlowe EL. Skin type, minimal erythema dose (MED), and sunlight acclimatization. J Am Acad Dermatol 5:429–443, 1981. 38. Haake N, Buhles N, Altmeyer P. Sensitivity of human skin to UV-light-practicability and limits of clinical diagnostics. Z Hautkr 62:1505–1509, 1987. 39. Berger DS. Design of a solar simulator. J Invest Dermatol 53:192–199, 1969. 40. Pathak MA, Fitzpatrick TB, Greiter F, Kraus EW. Preventive treatment of sunburn, dermatoheliosis, and skin cancer with sunprotective agents. In: FitzpatricK TB et al. (eds). Dermatology in General Medicine. McGrawHill, New York, 1987. 41. Little MA, Wolf ME. Skin and hair reflectance of women with red hair. Ann Human Biol 8(3):231–241, 1981. 42. Clarke P, Stark AE, Walsh RJ. A twin study of skin reflectance. Ann Human Biol 8:529–541, 1981. 43. Towne B, Hulse FS. Generational change in skin color variation among Habbani Yemini Jews. Human Biol 62(1):85–100, 1990. 44. Stevenson JM, Weatherall IL, Littlejohn RP, Seman DL. A comparison of two different instruments for measuring venison CIELAB values and color assessment by trained panel. N Zeal J Agr Res 34:207–211, 1991. 45. Jarnecke-Munster H. Uber die Zusammenhange der am Hauteigenstoffwechsel beteiligten Aminosauren, ins besondere Histiden und Tyrosin. Arch Dermatol Syph 180:290–293, 1940. 46. Cripps DJ. Natural and artificial photoprotection. J Invest Dermatol 76:154–157, 1981. 47. Shigeaki S, Imura M, Ota M. The relationship of skin color, spectrophotometric technique. J Invest Dermatol 84:265–267, 1985.
Colorimetry
48. Buckley WR, Grum F. Measurement of skin color, spectrophotometric technique. J Soc Cosmet Chem 15:79–85, 1964. 49. Farrington D, Imbrie JD. Comparison between visual grading and reflectance measurements of erythema produced by sunlight. Br J Dermatol 111:295–304, 1984. 50. Wasserman HP. The colour of human skin. Dermatologica 143:166–173, 1971. 51. Seitz JC, Whitmore CG. Measurement of erythema and tanning responses in human skin using a tri-stimulus colorimeter. Dermatologica 177:70–75, 1988. 52. Westerhof W, Estevez-Uscanga O, Meens J, Kammeyer A, Durocq M, Cairo I. The relation between constitutional skin color and photosensitivity estimated from UV-induced erythema and pigmentation dose response curves. J Invest Dermatol 94:812–816, 1990.
647
53. Stern RS, Momatz K. Skin typing for assessment of skin cancer risk and acute response to UVB and oral methoxalen photochemotherapy. Arch Dermatol 120: 869–872, 1984. 54. Azizi E, Lusky A, Kushelevsky AP, Shewach-Millet M. Skin type, hair color, and freckles are predictors of decreased minimal erythema ultraviolet radiations dose. J Am Acad Dermatol 19:32–38, 1988. 55. Duteil L, Ortonne JP. Colorimetric assessments of the effects of azelaic on light-induced skin pigmentation. Photodermatol Photoimmunol Photomed 9:67–71, 1992.
Color Measurement 75 Quasi-L*a*b* from Digital Images Hirotsugu Takiwaki Department of Dermatology, The University of Tokushima School of Medicine, Tokushima, Japan
CONTENTS 75.1 Introduction............................................................................................................................................................649 75.2 Objective and Methodological Principle...............................................................................................................649 75.3 Sources of Error.....................................................................................................................................................650 75.4 Correlation with Other Methods ...........................................................................................................................650 75.5 Recommendation ...................................................................................................................................................651 References .......................................................................................................................................................................651
75.1 INTRODUCTION In this era of digital images, the evolutional and rapid progress of charge coupled devices (CCD) and computers has enabled us to easily obtain and retouch high-resolution digital images. As every picture element (pixel) in digital images has color information, i.e. the brightness intensity in the red, green, and blue (RGB) channels, it is only natural that attempts have been made to quantify skin color by using digital images of the skin. In addition, unlike with reflectance instruments such as colorimeters or reflectance spectrometers, not only information about skin color itself, but also about color distribution can be obtained with digital images. However, color information obtained from digital images is inevitably influenced by the properties of CCD, the characteristics of light sources, and various other conditions under which images are taken. Therefore, if exact or “absolute” color data from digital images are needed, preliminary processing is necessary in order to convert raw data into a more reliable and standardized color scale. In this chapter, some simple methods to obtain values that approximate CIE-L*a*b* (quasi-L*a*b*) are described.
75.2 OBJECTIVE AND METHODOLOGICAL PRINCIPLE Objective: To quantify normal or lesional skin color in the L*a*b* color space by processing RGB brightness data of digital images of the skin. As described in other chapters, L* is the coordinate for the intensity of brightness, and a* and b* are the chromaticity coordinates.
Instruments: Any instruments for obtaining digital images, such as digital cameras and videomicroscopes equipped with a CCD camera, can be used. Computers and computer software for color image analysis, such as Image J (NIH freeware created by Wayne Rasband)1 and Adobe Photoshop, are also needed. Method: As the color of an object always depends on the characteristics of the illumination, digital images of the skin should always be taken under the same conditions. When images are taken with a digital camera, manual rather than automatic operation under specified conditions is recommended. If skin color is to be compared in a series of images, the distance between the camera and objects should be kept strictly constant. It is also advisable to place color chips with known color values close to the objective skin area and obtain images of the two together for color reference. The most commonly used method for estimating quasi-L*a*b* values of pixel(s) in digital images is to convert the brightness values in RGB channels to the corresponding L*, a*, and b* values. This conversion must be preceded by the preliminary formulation of regression curve equations based on an examination of the relationship between RGB and CIE-L*a*b* values of the same standard objects, such as color chips or color charts.2 In order to obtain these regression curves, it is also necessary to examine, with the aid of statistical software for multiregression analysis, the correlation between RGB brightness values of the object image and its CIE-L*a*b* values obtained with reflectance instruments. The histogram function of image analysis software can be used when the distribution and average of brightness data on a region of 649
650
Handbook of Non-Invasive Methods and the Skin, Second Edition
interest (ROI) are required for each channel (stack) of RGB. An alternative and simpler method may be to substitute RGB brightness values for the terms of XYZ in the formula of CIE-L*a*b* derivation3 as follows: Quasi-L* = 116Rg1/3 – 16 Quasi-a* = 500(Rr1/3 – Rg1/3) Quasi-b* = 200(Rg1/3 – Rb1/3) Rr,g,b = (BSr,g,b – BBr,g,b) / (BWr,g,b – BBr,g,b) where BS is the mean brightness value of the skin image, BW is that of the same size area in the image of a white standard, and BB is the background brightness. Results obtained for human and dog skin in this way show good linear correlation with CIE-L*a*b* values.3,4 Finally, an experimental technique has been developed which employs a multiband camera and special algorithm using Wiener estimation for estimating the spectral reflectance of each individual image pixel.5 This techniques is now undergoing improvements. Although not commercialized yet, it is hoped that this revolutionary method may provide not only more accurate color data but also information on optical properties of the object simply by taking its picture with a special digital camera.
the differences in the angle from the illumination source. To determine the presence of these artificial or unavoidable factors, preliminary evaluation of color images with the aid of a phantom model, such as a mannequin,2 is recommended. Images that are to be analyzed should in principle be saved in memory media or computers as TIFF files because JPEG compression may alter the fine configuration of skin surface structures or skin eruptions and therefore may influence the color values.
75.4 CORRELATION WITH OTHER METHODS When skin color is evaluated, quasi-L*a*b* values, obtained by using both multi-regression analytic method (m-L*a*b*) and “pseudo” CIE-L*a*b* formula method (p-L*a*b*), reportedly show good to excellent linear correlation with CIE-L*a*b* values measured with a MINOLTA Chromameter CR-300TM.2,3 The respective correlation coefficients between the m-L*a*b* and CIEL*a*b* values (n=187) were 0.897 (L*), 0.815 (a*) and 0.817 (b*), and those between the p-L*a*b* and CIEL*a*b* values (n=89) 0.973(L*), 0.913 (a*) and 0.837 (b*) (Figure 75.1). However, these results may depend on the instrument actually used, as the properties of the CCD
75.3 SOURCES OF ERROR Even if digital images are taken with the same camera or video equipment, the following conditions should always be kept constant in order to obtain accurate values and comparable data: 1. No direct sunlight from windows 2. No or dim external illumination 3. Same brightness intensity of flash bulb or builtin illumination 4. Same distance between camera and object 5. Same camera settings. Reference color chips, such as white standard or skin color standard, should be photographed at the same time to calibrate the illumination or check camera settings. Insufficient light diffusion due to uneven illumination may result in shiny reflections and shadows on the object surface, which in turn may cause artificial errors. If videomicroscopes are used, ROI should be located at the center of the image because darkening usually occurs in the peripheral areas of the image. In the face or extremities, anatomical curvature often causes unevenness of brightness because of variations in distance from the camera or
FIGURE 75.1 Relationship between corresponding coordinates of CIE- and quasi-L*a*b* systems. (From Takiwaki, H. et al., Skin Res. Technol. 3, 42-44, 1994. With permission.)
Quasi-L*a*b* Color Measurement from Digital Images
used cannot be expected to be identical. Therefore, it should be confirmed for each image-analytic quasiL*a*b* system that it shows good linear correlation with CIE-L*a*b* system. It should also be kept in mind that CIE-L*a*b* values of the skin depend on the instrument used.6 This phenomenon seems to result in part from the relation between the area of the measuring-head aperture and light scattering characteristics (Rayleigh scattering) of the dermis.7 Therefore, quasi-L*a*b* values thus obtained can also be expected to be influenced by the reflectance instruments that are used as reference.
75.5 RECOMMENDATION Quasi-L*a*b* evaluation from skin images appears to be the most accurate method for measuring the standardized color of very small skin areas or lesions, such as moles, ephelides, cherry angiomas and tiny pustules, that are far smaller than the area of the measuring head aperture of reflectance instruments. This method is also suitable for color evaluation of the whole area of the irregularly demarcated regions, as these areas can be selected with software for photo retouching,4 as well as for eroded or ulcerated skin lesions that are difficult to measure with a contact-type instrument.
651
REFERENCES 1. Image J – Image Processing and Analysis in Java, http://rsb.info.nih.gov/ij/ (Acc 2004-01-12) 2. Miyamoto, K., Takiwaki, H., Hillebrand, G. G., and Arase, S., Development of a digital imaging system for objective measurement of hyperpigmented spots on the face, Skin Res. Technol., 8, 227–235, 2002. 3. Takiwaki, H., Miyamoto, H., and Ahsan, K., A simple method to estimate CIE-L*a*b* values of the skin from its videomicroscopic image, Skin Res. Technol., 3, 42–44, 1997. 4. Boysen, L., Serup, J., Sorensen, P., and Kristensen, F., Use of chromametry and didital photography for objective measurement of skin color in clinically normal dogs, Am. J. Vet. Res., 63, 559–564, 2002. 5. Miyake, Y., and Yokoyama, Y., Obtaining and reproduction of accurate color images based on human perception, Proc. SPIE., 3300, 190–197, 1998. 6. Serup, J. and Agner, T., Colorimetric quantification of erythema –a comparison of two colorimeters (Lange Micro Color and Minolta Chromameter CR-200) with a clinical scoring scheme and laser-Doppler flowmetry, Clin. Exp. Dermatol., 15, 267–272, 1990. 7. Takiwaki, H., Miyaoka, Y., Skrebova, N., Kohno, H., and Arase, S., Skin reflectance-spectra and colour-value dependency on measuring-head aperture area in ordinary reflectance spectrophotometry and tristimulus colorimetry, Skin Res. Technol., 8, 94–97, 2002.
Color Calibration for 76 Practical Dermatoscopic Images Constantino Grana, Giovanni Pellacani, and Stefania Seidenari University of Modena and Reggio Emilia, Modena, Italy
CONTENTS 76.1 The Importance of Color Calibration in Dermoscopy..........................................................................................653 76.2 Instruments Characteristics and Calibration .........................................................................................................654 76.2.1 Analysis of the Video Camera’s Physical Properties................................................................................654 76.2.2 Illumination and Border Defects Correction.............................................................................................655 76.2.3 Assessment of γ .........................................................................................................................................656 76.2.4 Conversion from the Instrument’s RGB to XYZ......................................................................................657 76.2.5 Conversion from XYZ to a Known and Standard Color Space...............................................................658 76.3 An Example of Multi-Instrument Calibration.......................................................................................................659 76.4 Conclusions............................................................................................................................................................661 References .......................................................................................................................................................................663
76.1 THE IMPORTANCE OF COLOR CALIBRATION IN DERMOSCOPY Surface microscopy, which employs incident light magnification systems associated to the epiluminescence technique or polarized light, improves diagnostic accuracy with respect to simple clinical observation, especially for difficult-to-diagnose lesions [1,2]. The assessment of colors is essential for melanoma (MM) diagnosis, both for pattern analysis on dermoscopic images [3], and when employing semiquantitative methods [4–6]. Malignant lesions frequently show more than 3 colors, whereas in nevi, 3 or fewer than 3 colors are usually observed [7]. Moreover, different colors prevail in different lesion types. In order to overcome subjectivity and variability in the interpretation of dermoscopic images, several image analysis programs, also enabling color description by means of numerical data calculated from different color channels, have been recently introduced as a possible support to clinical diagnosis [8–15]. Recently, we described an image analysis method for the evaluation of colors in melanocytic lesion images based on an approach which shares some similarities with the human perception of colors [16,17]. Since color analysis has proved to be an important factor in the diagnostic process of dermatoscopic images, color degradation could negatively influence the diagnostic ability of the clinician [18]. Not much investigation
has been conducted on the interchange of dermatoscopic images and on the differences in color reproduction between different instruments or units of the same instruments. Even if this evaluation may seem insignificant, since images appear more or less similar in successive comparisons, the problem becomes apparent in automatic analysis, that is computer based color measure and characterization. Thus, the algorithms employed by different image analysis programs are strictly applicable only to pigmented skin lesions acquired with the same instrument and technique, and are not adaptable to images generated by different tools, sometimes also employing different acquisition methods [19]. For example if an instrument allows manual light intensity tuning, so that images can be adjusted by sight to appear bright enough, any comparison on the dark to light variation (for instance in the search for dark areas) is influenced by this setting [20]. If the light intensity tuning is continuously modified for every image, some difficulties could arise, leading to the question whether it is possible to use those images in an automatic framework or not. Most of the work in computerized dermatoscopy deals with the use of colors, but few studies even attempt to acquire knowledge on the color space under examination, usually only referring to an unspecified RGB color space [21,22]. Unfortunately, in the digital dermoscopy area, color standardization is very seldom taken into account, even if working in conjunction with image analysis 653
654
Handbook of Non-Invasive Methods and the Skin, Second Edition
programs. Most references are made to a generic RGB space that is implicitly assumed to be the RGB color space provided by their instrument. For instance, Vannoorenberghe et al. [21] describe a system that relies on learned probabilities to detect lesion contours directly in the RGB color space, Faziloglou et al. [22] use color histograms to detect differences between melanoma and nevus colors: to get rid of color influences they subtract the average skin color from each lesion pixel, without however considering gamma or contrast differences. In Gerger’s study [23] a completely different approach was followed, but unspecified RGB statistics were still present, disallowing any possible systematic reproduction of the published results. It is natural to assume that color calibration technologies and methods are still being studied or are as yet unknown in the dermatologic field, but this is not the case, since for example Koriki et al. [24] illustrated an on-skin lipstick color measurement system with full calibration, and in a work by Miyamoto et al. [25] a quick but effective conversion was estimated from RGB to L*a*b*, by means of a third order multiple regression analysis. The authors were also careful in referring to the obtained values as quasi-L*a*b* to stress the fact that this was an approximation of the full calibration approach. Setaro and Sparavigna [26] obtained a simple calibration by using a standard reference based on three colors and then adjusting the images comparing the measured values of the marker to the image colors by simple difference. Even this first order model is reported to produce good reproducibility. A systematic work for color measurement from video camera in dermatology was presented by Herbin et al. [27], but one of the most detailed and precise methods was provided by Haeghen et al. [28]. In this study a full calibration approach was followed and the problem of obtaining standardized images was successfully resolved. However, the final choice of sRGB color space for image exchange and handling led to the production of low contrast images and loss of details, owing to the limited area of the above mentioned color space, when applied to pigmented skin lesions acquired by digital dermoscopy. It is time for literature dealing with color images to cope with the problem of calibration, in order to generate reproducible data. In this chapter we provide a detailed and practical description of a calibration framework: since many instruments show different illumination structures, first of all illumination correction is explored; secondly an easy camera gamma estimation technique is described; thirdly, the use of specific color spaces to avoid low contrast effects due to the 8 bit quantization of color channels is described. Mention is also made to camera temperature effects over the color image reading.
76.2 INSTRUMENTS CHARACTERISTICS AND CALIBRATION The calibration techniques here described are applicable to every video camera system, and take into account problems that are very likely found in every dermatoscopic setup. For the sake of precision we will refer in detail to the specific case of an epiluminescence microscope for dermatology (FotoFinder, TeachScreen software GmbH, Bad Birnbach, Germany), that consists of a probe, comprising a CCD-chip color video camera with an integrated handle and optics for epiluminescence microscopy, a processing unit and a color monitor. Optics are set in a removable conic structure with a cylindrical transparent spacer and contact plate at the end, and with 6 bright white LEDs, positioned at the bottom of the structure, for constant illumination of the viewing area. According to the size of the lesion to be examined, 20- to 70-fold magnifications can be employed. When the probe is applied onto the skin area for imaging, a magnified color picture appears on the monitor. Focus is automatically adjusted by an electronically powered autofocus system. The S-video signal is digitized at the PAL standard resolution and digital images can be stored and processed at any time. The digitized images offer a spatial resolution of 768 × 576 pixels and 16 million colors. For the epiluminescence observation, a drop of contact medium, such as alcohol in water solution, is applied between the contact plane and the skin, enabling the recognition of subsurface structures, within and outside the lesion. The instrument is easy to handle and acquisition of the images is quick and simple. Furthermore it is one of the most widely diffused digital videomicroscopes, with over 750 pieces sold in Europe by the end of 2003.
76.2.1 ANALYSIS OF THE VIDEO CAMERA’S PHYSICAL PROPERTIES An often underrated problem is the stability of the readings taken from the digital camera. It is often assumed that, though having a signal noise component, camera readings tend to be constant over time. Unfortunately, this is not the case with many commercial camera based systems, as the FotoFinder equipment, and we assume that the same problem is present in many other devices. A first glance at the instrument’s behavior during time can be seen in Figure 76.1, in which a neutral gray patch of GretagMacbeth™ ColorChecker, precisely number 22 (xyY = 0.310,0.316,19.8), was acquired every 5 minutes from the start up of the video camera, without any change in light or room temperature. Average values taken on a selected window of the image are represented. It is clear that the camera has a so called “warm up” period during which it is practically impossible to get a stable reading of the object under examination; measurements are stable
Practical Color Calibration for Dermatoscopic Images
655
180 170 160 150
MeanR MeanG MeanB
140 130 120 110 100 0
10
20
30
40
50
60
70
80
90
100
110
120
130
Minutes
FIGURE 76.1 (See color insert following page 678.) Mean RGB values of patch 22 in successive acquisitions (5 min distance).
180 170 160 150
MeanR MeanG MeanB
140 130 120 110 100 0
10
20
30
40
50
60
70
80
90
100
110
120
130
Minutes
FIGURE 76.2 (See color insert.) Mean RGB values of patch 22 in successive acquisitions (1 min distance).
180 160 140 120 MeanR MeanG MeanB
100 80 60 40 20 0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
Minutes
FIGURE 76.3 (See color insert.) Mean RGB values of patch 22 in successive acquisitions (1 min distance with pauses every six acquisitions).
after 60 minutes. However, the frequency of acquisitions also influences the mean R, G and B values. In fact acquiring the same target every 30 seconds leads to a difference of about 10 levels per channel (Figure 76.2). This may be explained by a heating of the video camera during the acquisition due to the motorized motion of the zoom (that is reset at every acquisition) and of the auto focus. To obtain an acceptably stable reading over time, we had to insert pauses between series of acquisitions, leaving the instrument time to return to its normal conditions. Figure 76.3 illustrates a series of acquisitions of patch 22, taken at 1 minute time spans. After every 6 captures, a 10 minute pause was taken. This leads to a variation from the first to the last acquisition of less than 3 gray levels on
average, which is more than satisfactory for our aim. This procedure reproduces the routine visit of pigmented skin lesion clinic patients who usually require more than one acquisition per visit for follow up or documentation purposes.
76.2.2 ILLUMINATION CORRECTION
AND
BORDER DEFECTS
Color calibration process begins with a first step to correct the irregular illumination of the instruments. The basic assumption here is that if we are imaging a uniform reference surface, we should obtain an almost constant reading for the whole image. This is not always so because of
656
Handbook of Non-Invasive Methods and the Skin, Second Edition
the instrument’s characteristics, but we can measure the deviation from uniformity to invert it, thus obtaining a correction map for all the acquired images. A masking process must also occur to get rid of pixels that do not convey any data, such as top or bottom lines that are always black because of the frame grabber settings, or the black ring that some instruments present at lower magnification levels (20-fold for FotoFinder). The filter values are computed separately for each color channel, so if we call R(x, y) the value of the red channel, the value of the filter in that point is:
( )
FR x, y =
( )
R x, y − M R MR
where MR is the most represented value in a selected window of the image. This value can be computed from the histogram of the channel, that is a function of the channel value equal to the number of pixels with that value. MR can be obtained as the bin index with maximum value:
Gray Scale as a reference for our calibration. After computing the MRi for each step of the gray scale, and discarding those squares whose values clip at 255, we evaluate FR as: 19
( )
FR x, y = min
j ∈[1,19 ]
()
{( x, y ) : R ( x, y ) = i} ()
M R = arg max hR i i∈[ 0 ,255 ]
The window selection should be taken in a well lit area to uniform the image to the area that is best illuminated. This enables also MR to be used as an indicator of channel saturation. In fact one should be careful to avoid getting clipped values during the calibration phase, since these could disrupt the final result. In our application, since we were using the FotoFinder equipment, which has a semicircular set of photo emitting LEDs, we chose the rectangular window from (200,350) to (567,550). Moreover, this filter assessment should be carried out so that it is independent of the specific surface selected and the resulting acquisition. Thus, we used the Kodak
i R
j R
i =1
which is the median of the measured values FRi . The same process is applied to the green and blue channels, obtaining FG and FB. Each image I is then filtered for each channel using the equation:
˜I ( x, y ) = R
I R ( x, y ) 1 + FR ( x, y )
obtaining the filtered image I.˜
76.2.3 ASSESSMENT hR i = #
∑ F ( x, y ) − F ( x, y )
OF
γ
After obtaining light compensated images, the next step is to estimate the non-linear relation between the luminance factor, also known as CIE tristimulus value Y, and the digital values provided by the camera. Y is linearly related to incident light and is a standard of light measurement. Moreover, given fixed light source and geometry, it should be linearly related to the reflectance R of the surface, that can be estimated from the optical density OD as R = 10 2 −OD . Power 2 is used to scale R between 0% and 100% as it is commonly expressed. Measuring the Kodak Gray Scale with a Minolta CL-100 calibrated colorimeter, we verified that the declared optical density steps of 0.1 were indeed linearly related to Y (Figure 76.4).
100 90 80 70 60 50 40
Colorimeter Y
30 20 10 0 0
10
20
30
40
50
60
70
80
90
100
Declared reflectance
FIGURE 76.4 (See color insert.) Comparison of declared and measured reflectance of the KODAK Gray Scale.
Practical Color Calibration for Dermatoscopic Images
657
1 0,9 0,8 0,7 0,6
dr dg db
0,5 0,4 0,3 0,2 0,1 0 0
0,05
0,1
0,15
0,2
0,25
0,3
0,35
0,4
0,45
0,5
Colorimeter measured ratio Y/Yn
FIGURE 76.5 (See color insert.) Normalized RGB values of the video camera for gamma estimation.
Measuring the Kodak Gray Scale with the FotoFinder we verified the presence of gamma correction of RGB values, that is a power relation between the digital measures and the known Y values (Figure 76.5). To give an estimate of this relation we compared the normalized values of Yˆ = Y/Yn and dr = R/255, dg = R/255, db = R/255, with the following equation: y = ax γ + b and estimated the three parameters (a, b, γ) separately for the three channels. The aim was to use the N measured values di (with reference to the currently considered color channel) and the corresponding Yˆi to obtain the triplet: N
( a, b, γ ) = arg min ∑ Yˆ − ( ad a ,b , γ
i
γ i
)
+b .
i =1
The assessment was conducted by exhaustive search in a ±0.5 space around an initial starting point (1.0, 0.5, 0.5) with 0.1 steps. After finding the triplet producing the minimum value, the search was repeated at finer steps around that point, narrowing the search space progressively, up to the desired precision. If the solution was at the edge of the defined space, the search was repeated without narrowing the area, just moving the center to the best point found. Table 76.1 shows the values obtained for one of our instruments. It is interesting to note that the instrument
TABLE 76.1 Example of Estimated Gamma Parameters
R G B
a
b
1.5653 1.6465 1.8273
-0.4500 -0.4600 -0.4930
0.3275 0.3375 0.3434
presents a sort of constant offset on the Y value, probably correlated with a black level setting of the camera. The gamma value is significantly lower than typical values for video cameras, and the main difference is a different gain for each color channel, mainly enhancing the blue one. We wish to point out that these measures depend highly on the camera hardware settings, and are provided only as an example, not as “standard” figures. Following this estimation, inversion of the power relation to obtain triplets of RGB values that can be linearly transformed by matrix multiplication into XYZ triplets is simple.
76.2.4 CONVERSION FROM THE INSTRUMENT’S RGB TO XYZ Which matrix should be used to convert RGB to XYZ? Many matrices can be found in text books or on web pages, however none of these should be used. The conversion we are looking for should be assessed from the behavior of our specific instrument, which can only be observed by means of a reference object, such as the GretagMacbeth ColorChecker Color Rendition Chart (usually called the ColorChecker), a target that is often used in television broadcasting in order to evaluate the color accuracy of TV cameras. This target is produced using painted papers; therefore, it is not the ideal surface for skin reference, but its matte surfaces reduce problems regarding reflexes and the squares are large enough for quick positioning of the instrument. The ColorChecker should be handled carefully since it cannot be cleaned and the instrument’s head, used to compress the skin, could ruin the uniformity of the squares. This target was also measured with the Minolta CL-100 and resulting values were found to be quite consistent with tabulated data, supplied with the target. Every ColorChecker square was acquired and filtered with the previously estimated filter, then the average value of pixels from a central rectangle, whose sides were half the width and half the height of each image, was computed as the measured value for the patch. MX, the maximum histogram value (where X is the channel), was evaluated to reveal patches that couldn’t be correctly imaged by the
658
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 76.6 (See color insert.) Declared sRGB values of the ColorChecker and corresponding values as measured by the instrument.
instrument. Unlike other works, we did not attempt to calibrate the instrument so that it could correctly describe all the colors in the ColorChecker, on the contrary, we only chose to evaluate the settings that the “standard” clinician would use. Thus we discarded all patches resulting in MX = 0 or MX = 255. In our experiments one patch alone (the white one) couldn’t be used because of the extremely narrow range of the blue channel that saturated to 255 (Figure 76.6). Following a common procedure for optimization, we took the M valid triplets of RGB values, their corresponding XYZ ones and searched for the matrix A that gave the “best” transform: ⎡ Xi ⎤ ⎡ a11 ⎢ ⎥ ⎢ ⎢ Yi ⎥ ≅ ⎢ a21 ⎢ Z i ⎥ ⎢ a31 ⎣ ⎦ ⎣
⎛ R⎞ Θ 9 ⎜ G ⎟ = R G B RG GB BR R 2 G 2 B 2 ⎜ ⎟ ⎜⎝ B ⎟⎠
a13 ⎤ ⎡ Ri ⎤ ⎥⎢ ⎥ a23 ⎥ ⎢Gi ⎥ a333 ⎥⎦ ⎢⎣ Bi ⎥⎦
a12 a22 a32
(
for all the patches. This means that we solved the linear system given by: ⎡ ⎢R ⎢ i ⎢0 ⎢ ⎢0 ⎢⎣
... Gi
Bi
0
0
0
0
0
0
0
Ri
Gi
Bi
0
0
0
0
0
0
0
Ri
Gi
...
we used the Singular Value Decomposition (SVD) which presents the optimal solution, since it gives the minimum Euclidean difference between the known XYZ values and those assessed from RGB ones. We could have searched for other objectives, for instance to minimize the difference measured in the CIE L*a*b* color space, ΔE or CIE94, [29]. Haeghen et al. have shown experimentally that no evident benefit was obtained by doing so. Unfortunately the relation between RGB and XYZ values is not always well described by a linear transform, so to also cope with slightly more complex relations, we used a non-linear operator that included the covariance terms:
⎤ ⎡ a11 ⎤ ⎥ 0 ⎢ a12 ⎥ ⎥⎢ . ⎥ 0 ⎥ ⎢ .. ⎥ ⎥⎢ ⎥ Bi ⎥ ⎢ a32 ⎥ ⎥⎦ ⎢⎣ a33 ⎥⎦
⎡ ... ⎤ ⎢X ⎥ ⎢ i⎥ = ⎢ Yi ⎥ ⎢ ⎥ ⎢ Zi ⎥ ⎢⎣ ... ⎥⎦ In most cases, this can be solved exactly for three patches, leading to a perfect conversion for the three colors but a bad one for the others. For a resolution with more patches
)
T
This operator was described by Haeghen and provided the best accordance with the ColorChecker data. Thus, after applying this operator to the data, matrix A becomes 3 rows by 9 columns, but the optimization technique is the same.
76.2.5 CONVERSION FROM XYZ TO A KNOWN AND STANDARD COLOR SPACE At this point we have obtained an assessment of XYZ color coordinates for every measured pixel. The aim is to transform these values into another known color space enabling visualization and a simpler storage. Haeghen et al. [28] chose the sRGB color space because it is based on the phosphors used in many modern CRT-based display devices, including computer monitors. This means that an image stored in sRGB does not have to be converted before being displayed, and should look fairly realistic on a computer monitor. sRGB has a white point of 6500 K color temperature or D65, which means that the color produced by combining the full output of each color
Practical Color Calibration for Dermatoscopic Images
channel on an output device is the same as that of a blackbody at 6500 K. It has been standardized, so the meaningful exchange of images is also possible. sRGB tristimulus values have a known relationship to CIE XYZ tristimulus ones:
659
⎡ R ⎤ ⎡ 2.352 ⎢ ⎥ ⎢ ⎢G ⎥ = ⎢-0.731 ⎢ B ⎥ ⎢-0.098 ⎣ ⎦ ⎣
-0.802 1.358 -0.187
-0.329 ⎤ ⎡ X ⎤ ⎥⎢ ⎥ 0..392 ⎥ ⎢ Y ⎥ 1.273 ⎥⎦ ⎢⎣ Z ⎥⎦ D 65
and the gamma conversion is given by: ⎡ R ⎤ ⎡ 3.2406 ⎢ ⎥ ⎢ ⎢G ⎥ = ⎢-0.9689 ⎢ B ⎥ ⎢ 0.0557 ⎣ ⎦ ⎣
-1.5372 1..8758 -0.2040
-0.4986 ⎤ ⎡ X ⎤ ⎥⎢ ⎥ 0.0415 ⎥ ⎢ Y ⎥ 1.0570 ⎥⎦ ⎢⎣ Z ⎥⎦
X ′ = 1.46 ⋅ X (
1.0 1.85
D 65
After this linear transformation the same known gamma correction is applied to each color channel, as in the following equation: ⎧⎪ 12.92 ⋅ X X′ = ⎨ (1.0 2.4) − 0.055 ⎩⎪1.055 ⋅ X
) − 0.46
X ≤ 0.0031308 else
A problem arises in color space conversions, due to quantization to be applied to the values. Since this results in the subdivision of a “light” range into a fixed number of steps (without considering the effect of gamma correction on the steps’ characteristics). Indeed, if we are interested in a limited area of the sRGB color space, we lose much color detail that is set aside to describe color space regions that our images will never use. This color loss is clearly observed in dermatoscopic images, which become less clear (lower contrast and dynamic range) when converted into an sRGB color space. For the above reasons, we decided to describe our images by a new color space, especially conceived for our instrument for daily use. The color settings usually provided by the instrument are satisfactory, so we just searched for a formal description of the color transformation from XYZ back to our instrument’s RGB. This was performed by converting a large number of colors randomly taken from real lesions to XYZ using the calibration procedure previously described and then finding the best transformation enabling us to go back to the original colors. This is not the same as simply inverting the first transformation, since real images are used instead of the ColorChecker. The color space obtained is extracted from an average characterization produced by many images, thus it is safely applicable to dermatoscopic images, since it is designed to achieve a better use of the color representation in the spectrum area occupied by this kind of images. Its known relation with XYZ enables a simple conversion into any other chosen color space, and the practice of viewing the images on an sRGB calibrated computer monitor (a common setting available in most modern monitors), allows a common evaluation of images, even if obtained from different sources. Our proposed conversion uses the following linear conversion:
The choice for the gamma values was made by averaging those of the three channels. This conversion produces not too saturated images and allows simple comparisons between different units.
76.3 AN EXAMPLE OF MULTIINSTRUMENT CALIBRATION For testing purposes, we decided to verify the effectiveness of the calibration procedure on two different FotoFinder units without carrying out a preliminary hardware calibration, i.e., attempting to subjectively adjust images. We thus decided to set the instruments at their default configuration (factory settings), choosing an Iris setting of -25 in order to avoid skin color saturation in light skinned patients. Moreover, with this setting, all colors from the GretagMacbeth ColorChecker (GMCC) can be included, ensuring a good fitting of the data with the least squares method. To verify the effects of the resulting calibration, we used a Home Made ColorChecker (HMCC), consisting of different squares of colored paper. An example acquisition of corresponding patches is provided in Figure 76.7. The calibration procedure was performed starting from the light compensation filter, going on to the gamma functions assessment and finally to the RGB to XYZ matrix computation. From the XYZ space we employed the known conversion to get CIE L*a*b* color values and from these we measured the Euclidean distances (E). Results are summarized in Table 76.2 and show a fairly good outcome, which is strongly influenced by the errors observed on the yellow and orange patches. These errors are caused by extremely low values on the blue channel that cannot be corrected by the estimated matrix. Looking directly at the RGB color space, we can compare the Euclidean difference before and after correction (Table 76.3), and this enables us to see how different values provided by different units of the same instrument can be. For a better understanding of reported values it is useful to remember that the maximum difference can be 255 3 ≅ 442 . The comparison shows that even on a surface that is not ideal (colored paper obviously has a different color response compared to the Color Checker and the skin), differences obtained with the HMCC were
660
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 76.7 (See color insert.) Comparison of the visual aspect of the same color patches acquired with two differen equipments.
6 times lower and those obtained with the GMCC 10 times lower. It is interesting to note that results are obviously better for the surface used for calibration, so for a real comparison, another target is needed. This is an important consideration, for a serious validation of results. Finally, we had to test the results on real world images, i.e. skin lesions. The problem here was that we couldn’t just compare the average values of the images, because this wouldn’t be sufficiently accurate (images are not uniform surfaces). So we decided to provide an initial measure distinguishing the skin from the lesion, then we tried to compare the distribution of colors in the RGB color space. To this aim the 16 million colors RGB cube was divided into uniformly distributed sub-areas, allowing each channel to have values from 0 to 7, i.e., a 3 bit per channel representation, that leads to 512 possible colors. A three-dimensional color histogram was produced as before using
h ( r, g, b ) = # {( x, y ) | R ( x, y ) = r ∧ G ( x, y ) = g ∧ B ( x, y ) = b} then a common histogram comparison technique, called histogram intersection [30], was used to provide a metric to assess color similarity. Even if this technique tends to underrate similarity [31], it is perfectly suitable to assess a measure of color similarity improvement, especially if we know the compared objects are the same. The similarity measure is simply given by: 7
(
)
HI h1, h2 =
∑ min { h (r, g, b ), h (r, g, b )} 1
2
r , g ,b = 0
The value is computed over the normalized versions of the two histograms (each value of the histogram is divided
Practical Color Calibration for Dermatoscopic Images
TABLE 76.2 Euclidean Distance in the CIEL*a*b* Color Space between Corresponding Patches Measured with the Two Different Instruments Patch Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Mean Maximum
GMCC
HMCC
0,73 2,21 1,27 3,29 1,45 0,60 21,71 1,35 1,41 6,61 1,52 13,73 6,10 2,35 10,55 10,54 0,73 1,39 4,31 1,16 1,24 2,35 1,11 1,85 4,15 21,71
11,59 25,45 2,75 37,42 2,07 2,20 3,70 3,31 2,49 1,47 2,64 2,08 3,97 1,88 2,69 12,73 3,65 3,86 10,17 30,44 31,89 14,56 13,55 6,61 9,71 37,42
by the image size), so that HI is limited between 0 (no correspondence) and 1 (perfect histogram match). Table 76.4 reports the resulting comparisons and in Figure 76.8 and Figure 76.9 it is possible to see the result on the images that gave the best (lesion 4) and worst (lesion 1) results.
76.4 CONCLUSIONS Future developments of dermoscopy will deal with the progress of image analysis systems, tied to automatic diagnosis, and with tele-dermatology, to obtain a remote diagnostic consultation. In both cases, color calibration is of fundamental importance for two reasons: to make the developed algorithms applicable on various instruments, i.e. diagnostic center independent, and on the other hand to allow expert evaluators to have reproducible diagnoses, not biased by color degradation [32,33]. We presented a complete workflow for dermatologic image calibration, taking into account some practical, but
661
TABLE 76.3 Euclidean Distances in the Instruments RGB Color Space between Corresponding Patches Measured with the Two Different Instruments, before and after Calibration GMCC
HMCC
Uncalibrated Calibrated Uncalibrated Calibrated 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Mean Maximum
57,29 71,32 77,12 60,72 75,55 66,82 69,97 85,41 76,24 61,15 77,93 76,48 56,09 80,38 58,06 59,39 81,17 66,33 62,41 72,95 68,75 79,14 59,99 46,93 68,65 85,41
2,92 5,56 6,10 3,65 8,06 4,23 5,97 6,34 6,39 9,15 5,40 10,28 6,52 7,59 12,71 4,07 4,43 2,81 0,02 1,25 7,18 13,31 4,21 5,34 5,98 13,31
76,64 57,17 56,60 71,00 62,58 56,10 76,23 70,71 82,38 78,25 61,27 51,08 65,94 83,82 72,00 40,47 73,52 68,11 74,57 60,20 70,12 63,87 47,17 71,75 66,31 83,82
29,17 9,68 3,43 15,70 10,24 2,24 12,89 13,08 14,21 6,40 6,78 5,68 0,25 13,01 10,14 3,33 5,94 6,23 19,25 12,51 15,50 6,37 11,06 16,61 10,40 29,17
very important issues such as camera temperature effects, illumination correction, easy camera gamma estimation and a specific color space generation. The system is simple, and after calibration, allows the user to continue using his own software and algorithms, but with a much higher informative content. Corrected images should be handled carefully to ensure conveyance of the final selected color space to whoever receives them, to enable their use in color calibrated contexts. At the time we didn’t explore the possibility of including the color space description into an ICC specification, but this could be the next step for a commercially suitable product. The analysis of smaller gamut color spaces (to better describe interest zones) is still a matter of debate in standardized color management documents. Encouraging the widespread use of color calibration in this field, will not only improve the quality of dermatoscopic digital libraries, but will also open the way to teleconsulting, remote analysis and result comparisons
662
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 76.4 Euclidean Distances and Histogram Intersections in the Instruments RGB Color Space between Corresponding Lesions Measured with the Two Different Instruments, before and after Calibration Lesion Uncalibrated 1 2 3 4 5 6 Mean Maximum
82,96 75,22 100,86 66,73 68,66 83,78 79,70 100,86
Skin
Calibrated 15,78 4,61 37,61 8,19 5,96 17,33 14,91 37,61
Uncalibrated 121,92 113,45 133,30 102,16 114,19 113,44 116,41 133,30
Histogram Calibrated 21,02 3,69 25,50 7,55 5,76 3,26 11,13 25,50
Uncalibrated
Calibrated
15,05% 13,94% 8,19% 23,14% 6,76% 0,11% 11,20% 23,14%
61,63% 82,78% 64,44% 90,56% 80,37% 77,30% 76,18% 90,56%
FIGURE 76.8 (See color insert.) Worst calibration result: first line shows the original images, second line shows the results of lighting correction, third line shows the calibrated images.
Practical Color Calibration for Dermatoscopic Images
663
FIGURE 76.9 (See color insert.) Best calibration result: first line shows the original images, second line shows the results of lighting correction, third line shows the calibrated images.
between different computer algorithms, in order to stimulate joint development or knowledge sharing.
REFERENCES 1. Kenet R.O., Kang S., Kenet B.J., Fitzpatrick T.B., Sober A.J., and Barnhill R.L., Clinical diagnosis of pigmented lesions using digital epiluminescence microscopy, Arch. Dermatol., 129, 157, 1993. 2. Bufounta M.L., Beauchet A., Aegerter P., and Saiag P., Is dermoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Arch. Dermatol., 137, 1343, 2001. 3. Pehamberger H., Steiner A., and Wolff K., In vivo epiluminescence microscopy of pigmented skin lesions. I. Pattern analysis of pigmented skin lesions, J. Am. Acad. Dermatol., 17, 571, 1987.
4. Nachbar F., Stolz W., Merkle T., Cognetta A.B., Vogt T., Landthaler M., Bilek P., Braun-Falco O., and Plewig G., The ABCD rule of dermatoscopy, J. Am. Acad. Dermatol., 30, 551, 1994. 5. Menzies S.W., Ingvar C., and McCarthy W.H., A sensitivity and specificity analysis of the surface microscopy features of invasive melanoma, Melanoma. Res., 6, 55, 1996. 6. Argenziano G., Fabbrocini G., Carli P., De Giorgi V., Sammarco E., and Delfino M., Epiluminescence microscopy for the diagnosis of doubtful melanocytic skin lesions. Comparison of the ABCD rule of dermoscopy and a new 7-point checklist based on pattern analysis, Arch. Dermatol., 134, 1563, 1998. 7. Mac Kie R.M., Fleming C., Mc Mahon A.D., and Jarret P., The use of the dermatoscope to identify early melanoma using the three-colour test, Brit. J. Dermatol., 146, 481, 2002.
664
Handbook of Non-Invasive Methods and the Skin, Second Edition
8. Cascinelli N., Ferrario M., Bufalino R., Zurrida S., Galimberti V., Mascheroni L., Bartoli C., and Clemente C., Results obtained by using a computerized image analysis system designed as an aid to diagnosis of cutaneous melanoma, Melanoma Res., 2, 163, 1992. 9. Schindewolf T., Stolz W., Albert R., Abmayr W., and Harms H., Classification of melanocytic lesions with color and texture analysis using digital image processing, Analyt. Quant. Cytol. Histol., 15, 1, 1993. 10. Green A.C., Martin N.G., Pfitzner J., O’Rourke M., and Knight N., Computer image analysis in the diagnosis of melanoma, J. Am. Acad. Dermatol., 31, 958, 1994. 11. Cucchiara R., Grana C., Seidenari S., and Pellacani G., Exploiting color and topological features for region segmentation with recursive fuzzy c-means, Machine Graphics and Vision, 11, 169, 2002. 12. Gutkowicz-Krusin D., Elbaum M., Szwaykowski P., and Kopf A.W., Can early malignant melanoma be differentiated from atypical melanocytic nevus by in vivo techniques? Part II. Automatic machine vision classification, Skin Res. Technol., 3, 15, 1997. 13. Seidenari S., Pellacani G., and Pepe P., Digital videomicroscopy improves diagnostic accuracy for melanoma, J. Am. Acad. Dermatol., 39, 175, 1998. 14. Pellacani G., Martini M., and Seidenari S., Digital videomicroscopy with image analysis and automatic classification as an aid for diagnosis of Spitz nevus, Skin Res. Technol., 5, 266, 1999. 15. Seidenari S., Pellacani G., and Giannetti A., Digital videomicroscopy and image analysis with automatic classification for detection of thin melanoma, Melanoma Res., 9, 163, 1999. 16. Seidenari S., Pellacani G., Grana C., Computer description of colours in dermoscopic melanocytic lesion images reproducing clinical assessment, Br. J. Dermatol., 149, 523, 2003. 17. Pellacani G., Grana C., and Seidenari S., Automated description of colours in polarized-light surface microscopy images of melanocytic lesions, Melanoma Res., 14, 125, 2004. 18. Seidenari S., Pellacani G., Righi E., and Di Nardo A., Is JPEG-compression of videomicroscopic images compatible with telediagnosis? Comparison between diagnostic performance and pattern recognition on uncompressed TIFF images and JPEG compressed ones, Telem. J. E-Health, 10, 294, 2004. 19. Pellacani G., and Seidenari S., Comparison between morphological parameters in pigmented skin lesion images acquired by means of epiluminescence surface microscopy and polarized light videomicroscopy, Clin. Dermatol., 20, 222, 2002.
20. Pellacani G., Grana C., Cucchiara R., and Seidenari S., Automated extraction and description of dark areas in surface microscopy melanocytic lesion images, Dermatology, 208, 21, 2004. 21. Vannoorenberghe P., Colot O., and De Brucq D., Dempster-Shafer’s Theory as an aid to Color Information Processing Application to Melanoma Detection in Dermatology, Proceedings of the Int. Conf. Image Analysis and Processing, 774, 1999. 22. Faziloglou Y., Stanley R.J., Moss R.H., Van Stoecker W., and McLean R.P., Colour histogram analysis for melanoma discrimination in clinical images, Skin Res. Technol., 9, 147, 2003. 23. Gerger A., Stolz W., Pompl R., and Smolle J., Automated epiluminecence microscopy-tissue counter analysis using CART and 1-NN in the diagnosis of Melanoma, Skin Res. Technol., 9, 101, 2003. 24. Korichi R., Provost R., Heusèle C., and Schnebert S., Quantitative assessment of properties of make-up products by video imaging: application to lipsticks, Skin Res. Technol., 6, 222, 2000. 25. Miyamoto K., Takiwaki H., Hillebrand G.G., and Arase S., Development of a digital imaging system for objective measurement of hyperpigmented spots on the face, Skin Res. Technol., 8, 227, 2002. 26. Setaro M., and Sparavigna A., Quantification of erythema using digital camera and computer-based colour image analysis: a multicentre study, Skin Res. Technol., 8, 84, 2002. 27. Herbin M., Venot A., Devaux J.Y., and Piette C., Color Quantitation Through Image Processing in Dermatology, IEEE T. Med. Imaging., 9, 262, 1990. 28. Haeghen Y.V., Naeyaert J.M.A.D., Lemahieu I., and Philips W., An Imaging System with Calibrated Color Image Acquisition for Use in Dermatology, IEEE T. Med. Imaging., 19, 722, 2000. 29. Berns R.S., Billmeyer and Saltzman’s Principles of ColorTechnology, 3rd ed. Wiley-Interscience, New York, 2000. 30. Swain M.J., and Ballard D.H., Color Indexing, Int. J. Comput. Vision., 7, 11, 1991. 31. Rubner Y., Tomasi C., and Guibas L.J., A Metric for Distributions with Applications to Image Databases, Proceedings of the 1998 IEEE Int. Conf. Comput. Vision, 59, 1998. 32. Seidenari S., Pellacani G., and Grana C., Computer description of colours in dermoscopic melanocytic lesion images reproducing clinical assessment, Br. J. Dermatol., 149, 523, 2003. 33. Pellacani G., Grana C., Cucchiara R., and Seidenari S., Automated extraction and description of dark areas in surface microscopy melanocytic lesion images, Dermatology, 208, 21, 2004.
of Erythema and 77 Measurement Melanin Indices Hirotsugu Takiwaki Department of Dermatology, The University of Tokushima School of Medicine, Tokushima, Japan
CONTENTS 77.1 Introduction............................................................................................................................................................665 77.2 Object and Methodological Principle....................................................................................................................665 77.2.1 Theoretical Aspect .....................................................................................................................................665 77.2.2 Instruments.................................................................................................................................................667 77.3 Sources of Error.....................................................................................................................................................667 77.4 Correlation with Other Methods ...........................................................................................................................668 77.5 Clinical and Experimental Applications................................................................................................................669 77.6 Recommendation ...................................................................................................................................................670 References .......................................................................................................................................................................670
77.1 INTRODUCTION
77.2.1 THEORETICAL ASPECT
Erythema, also referred to as hemoglobin (Hb), and melanin indices are the indicators that quantify the intensity of erythema and pigmentation, respectively, and are derived from reflectance data of the skin at specific wavelengths. Unlike color coordinates, such as CIE-L*a*b*,1 these indices are designed to show quantities that correlate linearly with the amounts of hemoglobin and melanin in the skin. Therefore, they can be handled as genuine physical quantities. Although these indices were first devised in the 1980s to extract information about the amounts of Hb and melanin from reflectance spectra of the skin,2 portable types of specialized instruments have been developed and are now commercially available and widely used in the fields of dermatology, cosmetic science, and pharmacology.
The erythema and melanin indices are both based on the reflectance of an object in a selected band of the spectrum. In order to understand the significance of the erythema and melanin indices, it is helpful to use a simplified, multilayered skin model. It is composed of three layers, with the uppermost layer containing only melanin, the middle layer only hemoglobin, and the bottom layer neither chromophore (Figure 77.1). These layers represent the epidermis, the plexus of blood vessels in the upper dermis, and the dermis below them, respectively. In this model, we assume that when each layer is placed separately on an ideal black background, there is no regular reflection from the surface, the diffuse reflectance of each of the upper two layers is nearly zero, and that of the bottom layer is high. According to Dawson et al.2 and Diffey et al.,3 the total reflectance, R, of this model at a given wavelength can be roughly expressed as
77.2 OBJECT AND METHODOLOGICAL PRINCIPLE The quantity of hemoglobin, which corresponds directly to the extent of erythema, is expressed as the erythema index or hemoglobin index, and that of melanin as the melanin index. The objective of the erythema index is to quantitatively evaluate anemic or hyperemic color changes of the skin, and that of the melanin index to assess hypoor hyperpigmented skin conditions.
R = I/I0 ≈ TM2 TH2 RD
(77.1)
where I0 is the intensity of incident light and I that of reflected light, TM is the transmittance of the uppermost layer and TH that of the middle layer, and RD is the diffuse reflectance of the bottom layer. By taking the logarithm to base 10 of the inverse reflectance (LIR), from Equation 77.1 we get 665
666
Handbook of Non-Invasive Methods and the Skin, Second Edition
80 Absorbance (a.u.)
Epidermis
Superficial plexus
Dermis
60 40 20 0 450
(a)
Transmittance
Melanin layer
TM
Hemoglobin layer
TH Diffuse reflectance RD
Chromophore free backing (b)
FIGURE 77.1 (a) Schematic structure of the skin. (b) Optical skin model of three-layered structure with an outer melanin layer, an inner hemoglobin layer, and a backing representing chromophore-free dermis.
log(1/R) ≈ 2log(1/TM) + 2log(1/TH) + log(1/RD) (77.2) Since it is assumed that the scattering within the upper two layers is extremely minor, the Beer–Lambert law of radiation absorption can be applied when the concentration of melanin and hemoglobin is low. Equation 77.2 can then converted to Log(1/R) ≈ dMEMCM + dHEHCH + log(1/RD) (77.3) where dM is the thickness of the melanin layer, EM the coefficient determined by the absorbance of melanin at a given wavelength, and CM the concentration of melanin, respectively, while dH, EH, and CH are their counterparts in the hemoglobin layer. Log (1/R) is referred to as the apparent absorbance, as this value is regarded as a double of the ordinary absorbance (= log (1/transmittance)). By substituting A for log (1/R), m for dMEM, h for dHEH, and d for log (1/RD), Equation 77.3 can be written as A ≈ mCM + hλCH + dλ
550 600 Wavelength (nm)
650
FIGURE 77.2 Spectra of apparent absorbance of oxyhemoglobin (●), deoxyhemoglobin (), and melanin (x). (Modified from Hagisawa, S. and Ferguson-Pell, M., Ikakikaigaku, 64, 299, 1994. With permission.)
Reflected light (I)
Incident light (Io)
500
(77.4)
where Δ equals the given wavelength. If we select two different wavelengths, 1and 2, and designate the respective apparent absorbance at these wavelengths as A1 and A2, we get A1 – A2 = ΔA ≈ (m1 – m2) CM + (h1 – h2) CH (77.5) since the human dermis has similar reflectance values in wavelengths of visible light. Figure 77.2 shows the absorbance spectra of reduced or oxygenated Hb and that of melanin. If we select a set of λ1 and λ2 for which m1 – m2 is nearly zero and h1 – h2 is substantially high, then A (h1 – h2) CH
(77.6)
A is therefore suitable for use as the erythema index, because this value is likely to be proportional to CH. Similarly, if we select a set of 1 and 2 for which h1 – h2 is nearly zero and m1 – m2 is substantially high, A is likely to be proportional to Cm, which therefore is suitable for use as the melanin index. In practice, a wavelength or narrowband wavelengths corresponding to green (around 550 nm) and red (around 650 nm) light are selected as λ1 and λ2, respectively.3–7 For the melanin index, a wavelength or narrowband wavelengths corresponding to red light (e.g., 620 and 720 nm)4,8 are selected as both λ1 and λ2 for some systems, while one instrument (Mexameter) uses red and near-infrared lights.7 Another instrument (DermaSpectrometer) uses only one narrowband wavelength centered at 670 nm instead of two separate wavelengths, as hλ is nearly zero in this wavelength.6 In this case, Equation 77.4 is written as A ≈ mλ CM + dλ
(77.7)
in which it is of note that dλ is not negligible, so that the index value is not zero, but appears to be high even if the skin contains no melanin. Most instruments and systems
Measurement of Erythema and Melanin Indices
use the definition of erythema and melanin indices outlined here. More complex definitions are also used, however, to produce a more accurate erythema (or melanin) index that is less influenced by factors other than Hb (or melanin) when a full-band reflectance spectrophotometer is used for derivation of the indices.2,9 The principle of index derivation, however, is basically similar to that underlying the absorbance of the multilayered model.
77.2.2 INSTRUMENTS
Erythema and melanin indices can be roughly assessed from digital images of the skin captured with a digital camera or videomicroscope.10 In this case, a white color chip should be included in the image. This chip is used for the white standard to calibrate brightness of illumination and to calculate the averaged reflectance of the region of interest selected in the image. By substituting apparent absorbance data in the red and green channels for those measured at λ1 and λ2, an erythema index and a DermaSpectrometer-type melanin index can be derived.
77.3 SOURCES OF ERROR Although these instruments are convenient for obtaining indices, there are some critical points that examiners should keep in mind. First, the erythema index may be affected by hyperpigmentation.4,10 Equation 77.5 indicates that the influence of melanin cannot be neglected if CM is high, which is the case with hyperpigmented skin. Figure 77.4 shows the relationship between the erythema index and the melanin index measured with DermaSpectrometer at the sites with various degrees of pigmentation induced by ultraviolet (UV) B irradiation in five Japanese subjects 14 days before the measurement. This result clearly indicates that the erythema index increases in an apparently linear fashion as the melanin index increases. Comparison of erythema indices between two sites showing quite different degrees of pigmentation should therefore be avoided. Second, the melanin index may be affected by the oxygen saturation level of hemoglobin. Since reduced hemoglobin absorbs more red light than oxyhemoglobin (Figure 77.2), the term hCH in Equation 77.4 cannot be disregarded when the concentration of reduced hemoglobin is relatively high, even if a wavelength corresponding to red light is selected. This implies that the melanin index apparently increases under static or cyanotic conditions.
Erythema index
When a full-band reflectance spectrometer is used, erythema and melanin indices can be calculated based on the various formulae that have been published.2–9 However, most of the reflectance spectrometers widely used in dermatology and cosmetology seem to be of the type that can measure reflectance at every 5 to 10 nm within the visible wavelength of light. If reflectance at the desired wavelength cannot be obtained, it is then necessary to calculate it by means of interpolation using data available for neighboring wavelengths. The most popular, commercially available instruments for measuring erythema and melanin indices seem to be the DermaSpectrometer (Cortex Technology, Hadsund, Denmark) and Mexameter MX16® and MX18® (Courage and Khazaka Electronic GmbH, Cologne, Germany) (Figure 77.3). The DermaSpectrometer is a handheld instrument equipped with two light-emitting diodes (LEDs) that emit green light centered at 568 nm and red light at 655 nm. The Mexameter is a portable instrument with three LEDs that emit green, red, and near-infrared light with respective centered wavelengths of 568, 660, and 880 nm. Both instruments detect reflected light at these narrowband wavelengths with photodiodes and use built-in microcomputers to calculate erythema and melanin indices with Equations 77.5 and 77.6.
667
FIGURE 77.3 Instruments for measuring erythema and melanin indices. From left to right, the Mexameter MX18, the reflectance spectrophotometer Minolta CM-2002, and the DermaSpectrometer. (From Kohno, H. and Takiwaki, H., Dermatol. Pract., 14, 78, 2002. With permission.)
20 18 16 14 12 10 8 6 4 2 0
25
30
35 40 Melanin index
45
50
FIGURE 77.4 Dependence of erythema index on melanin index. Both indices were measured with the DermaSpectrometer at test sites showing various degrees of pigmentation induced by UVB irradiation of different doses 14 d before the measurement.
668
Handbook of Non-Invasive Methods and the Skin, Second Edition
In fact, reduced hemoglobin easily increases during arm 20
500
Eds
Emx
400
15
300 10 200 5
100 E
0 0 60
10
20
30
40
50
E
0 0 500
Mds
50
10
20
30
40
50
Mmx
400
40 300 30 200 20 100
10 0
M 0
2
4
6
8
10
M 0
0
2
4
6
8
10
FIGURE 77.5 Correlations between the erythema index by Dawson et al.2 (E) measured with the reflectance spectrophotometer CM2002 and two types of erythema indices measured with the DermaSpectrometer (Eds) and Mexameter MX18 (Emx). Correlations between melanin indices (M, Mmx, and Mds) are also shown. Measurements were made at four different anatomical sites of 15 healthy Japanese subjects. (Redrawn from Kohno, H. and Takiwaki, H., Dermatol. Pract., 14, 78, 2002. With permission.)
lowering, which results in a paradoxical increase in the melanin index.11 These errors result inevitably from the definition of the erythema and melanin indices. In addition, errors due to incorrect techniques may occur when the measuring head is applied with too much pressure, or when the measurement area to be measured is too hairy. Measurement in direct sunlight should be avoided as well.
77.4 CORRELATION WITH OTHER METHODS Moderate to excellent linear correlations have been found between erythema indices of normal and erythematous skin measured with different instruments or systems.12 Our measurements at four anatomical locations of the normal skin of 15 Japanese volunteers showed that the correlation coefficients between Dawson’s erythema index2 and erythema indices measured with the DermaSpectrometer (DS) and Mexameter (MX) were 0.77 and 0.87, respectively (Figure 77.5).13 Clarys et al.12 reported a correlation
of 0.81 for the erythema indices determined with DS and MX. Unlike the erythema index, the definition of the melanin index differs considerably, depending on the equipment. Nevertheless, the correlation coefficients between Dawson’s melanin index2 and the melanin indices measured with DS and MX were 0.82 and 0.94, respectively, according to our evaluation (Figure 77.5).13 However, Clarys et al.12 reported the correlation for the melanin indices determined with DS and MX to be rather low (0.53). They also mentioned that the melanin index of the Mexameter MX16 was less sensitive; however, the manufacturer has developed an allegedly improved version, the MX18. As mentioned in Section 77.2, it should be noted that the melanin index of DS inevitably shows far higher values than that of other systems, even if no melanin is present in the measured skin, because the influence of the dermal factor on the melanin index has not been excluded. This raised-bottom value of the DermaSpectrometer melanin index corresponds to the value for vitiliginous skin. Our assessments of the color of normal skin and of psoriasis lesions measured with the DermaSpectrometer
Measurement of Erythema and Melanin Indices
669
ΔLDF (arbitray unit)
a∗ 30 25 20 15 10 : n = 230
5 0
10
20 Erythema index
: n = 50
–4
–2
Arm elevation
40
30
1 0.8 0.6 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 –1 –6
FIGURE 77.6 Correlation between erythema index and a* measured at 23 anatomical sites of 10 healthy Caucasian subjects (·) and that in 50 plaques of psoriasis of 10 patients (). Correlation coefficient r = 0.92 and r = 0.91, respectively. Regression lines are shown for each group. The DermaSpectrometer (erythema index) and Minolta Chromameter CR-200 (a*) were used. (Modified from Takiwaki, H. et al., Skin Pharmacol., 7, 217, 1994. With permission.)
80
0 2 ΔErythema index
4
6
8
Arm lowering
FIGURE 77.8 Relationship between the change in erythema index and that in laser Doppler blood flow following arm elevation () and lowering (●), where the change is expressed by the difference (Δ) between the value at heart and that in each position.
The correlation between the erythema index and laser Doppler flow (LDF) seems to depend upon what is the cause of erythema. The erythema index correlates positively with LDF in acute inflammation, such as seen in irritant or allergic patch test reaction15 and UV-induced erythema, whereas LDF decreases despite an increase in the erythema index during arm lowering (Figure 77.8).11 This is most likely caused by the fact that the erythema index reflects only the blood volume, while LDF reflects the movement of blood cells.16
L∗
70
77.5 CLINICAL AND EXPERIMENTAL APPLICATIONS
60
In comparison with tristimulus colorimetry, the instruments and methods discussed here have the following advantages:
50
FIGURE 77.7 Correlation between melanin index and L* measured at 23 sites of 10 Caucasians: r = –0.56, p < 0.001. (From Takiwaki, H. et al., Skin Pharmacol., 7, 217, 1994. With permission.)
1. The intensities of erythema and of pigmentation can be separately quantified. 2. The erythema index is likely to have a linear relationship with the content of red blood cells in the upper dermis, and the melanin index with that of melanin in the epidermis, unless the extent of erythema or pigmentation is intense.
and Minolta Chromameter CR-200®6,14 showed a strong linear correlation between the erythema index and a*, representing the red-green axis in the CIE-L*a*b* space (Figure 77.6). The melanin index also correlated negatively with L* values representing brightness (Figure 77.7). This correlation, however, does not mean that the melanin index is equivalent to L*, because L* is mainly determined by the reflectance of green light, and the melanin index by that of red light. The melanin index is more specific for the degree of pigmentation than L*, which is also affected by the degree of erythema.
These instruments are not suitable for measuring changes in color due to other chromophores, such as jaundice. Moreover, these indices may be inadequate for the quantification of the hemoglobin and melanin located deeply in the dermis in such lesions as cavernous hemangioma, subcutaneous hemorrhage, and nevus of Ota, because their spectral reflectance is affected by the scattering effect of the dermis, and differs from that of the usual erythema and pigmentation. Therefore, direct numerical comparison of index values should be avoided for skin diseases or conditions in which pigments are
40 10
20
30 Melanin index
40
50
670
Handbook of Non-Invasive Methods and the Skin, Second Edition
present at different depths (e.g., port-wine stain and cavernous hemangioma). Measurement of the erythema index on black skin was found to be unsuccessful.17 Various quantitative in vivo studies have been carried out by using these advantages of erythema and melanin indices. This is especially true for studies of various aspects of inflammatory skin response to ultraviolet light, such as the exact relationship between dose and response, time course of the response, regional or constitutional variations in sensitivity, dependency on wavelength, suppressive effects of anti-inflammatory agents, and the relationship between UV-induced erythema and postinflammatory pigmentation.18–24 The intensity and time course of skin response to irritants or allergens have also been assessed quantitatively.15,25,26 Furthermore, these instruments and methods can be readily applied to other assessments, such as blanching or the vasoconstrictive action of corticosteroids, evaluation of the severity of inflammatory skin diseases,14,27 and monitoring of the efficacy of topical treatments, chemical peeling, and laser surgery for various kinds of nevi and hemangiomas.28–31
77.6 RECOMMENDATION Considering how the erythema and melanin indices are defined, the following is recommended in order to make the best use of these instruments and methods: 1. Subjects should take a standardized position for each series of measurement, especially if it is performed in the extremities. 2. The measuring head should be placed very softly and perpendicularly onto the skin. Holding it too long at a test site should be avoided. Much or incorrect pressure makes the site anemic or congested. Measurements should not be performed under direct sunlight. 3. Avoid comparing erythema indices between two sites where the levels of pigmentation (melanin index) are considerably different. If obtainable, empirical corrections for the indices4,10 are desirable. 4. Skin color shows spontaneous diurnal variations32 and regional differences.6 When the intensity of skin test reaction is successively examined in a wide area, the data for the control should be obtained in the normal skin adjacent to the test site at each measurement.
REFERENCES 1. Robertson, A., The CIE 1976 color-difference formulae, Color Res. Appl., 2, 7, 1977.
2. Dawson, J.B., Barker, D.J., Ellis, D.J., Grassam, E., Cotterill, J.A., Fisher, G.W., and Feather, J.W., A theoretical and experimental study of light absorption and scattering by in vivo skin, Phys. Med. Biol., 25, 695, 1980. 3. Diffey, B.L., Oliver, R.J., and Farr, P.M., A portable instrument for quantifying erythema induced by ultraviolet radiation, Br. J. Dermatol., 111, 663, 1984. 4. Feather, J.W., Ellis, D.J., and Leslie, G., A portable reflectometer for rapid quantification of cutaneous haemoglobin and melanin, Phys. Med. Biol., 33, 711, 1988. 5. Pearse, A.D., Edwards, C., Hill, S., and Marks, R., Portable erythema meter and its application to use in human skin, Int. J. Cosmet. Sci., 12, 63, 1990. 6. Takiwaki, H., Overgaard, L., and Serup, J., Comparison of narrow-band reflectance spectrophotometric and tristimulus colorimetric measurement of skin color: 23 anatomical sites evaluated by the DermaSpectrometer and the Chromameter CR-200, Skin Pharmacol., 7, 217, 1994. 7. Edwards, C., The Mexameter MX 16™, in Bioengineering of the Skin: Methods and Instrumentation, Berardesca, E., Elsner, P., Wilhelm, K.P., and Maibach, H.I., Eds., CRC Press, Boca Raton, FL, 1995, p. 127. 8. Kollias, N. and Baqer, A., Spectroscopic characteristics of human melanin in vivo, J. Invest. Dermatol., 85, 38, 1985. 9. Feather, J.W., Hajizadeh-Saffer, M., Leslie, G., and Dawson, J.B., A portable scanning reflectance spectrophotometer using visible wavelengths for the rapid measurement of skin pigments, Phys. Med. Biol., 34, 807, 1989. 10. Takiwaki, H., Shirai, S., Kanno, Y., Watanabe, Y., and Arase, S., Quantification of erythema and pigmentation using a videomicroscope and a computer, Br. J. Dermatol., 131, 85, 1994. 11. Takiwaki, H. and Serup, J., Variation in color and blood flow of the forearm skin during orthostatic maneuver, Skin Pharmacol., 7, 226, 1994. 12. Clarys, P., Aleweaters, K., Lambrecht, R., and Barel, A.O., Skin color measurements: comparison between three instruments: the Chromameter, the DermaSpectrometer, and the Mexameter, Skin Res. Technol., 6, 230, 2000. 13. Kohno, H. and Takiwaki, H., Erythema and melanin index meter, Dermatol. Pract., 14, 78, 2002 (in Japanese). 14. Takiwaki, H. and Serup, J., Measurement of color parameters of psoriatic plaques by narrow-band reflectance spectrometry and tristimulus colorimetry, Skin Pharmacol., 7, 145, 1994. 15. Gawkrodger, D.J., McDonagh, A.J.G., and Wright, A.L., Quantification of allergic and irritant patch test reactions using laser-Doppler flowmetry and erythema index, Contact Derm., 24, 172, 1991. 16. Bircher, A., Boer, E.M.D., Agner, T., Wahlberg, J.E., and Serup, J., Guidelines for measurement of cutaneous blood flow by laser Doppler flowmetry. A report from the standardization group of European Society of Contact Dermatitis, Contact Derm., 30, 65, 1994.
Measurement of Erythema and Melanin Indices
17. Takiwaki, H., personal communication, 1991. 18. Farr, P.M. and Diffey, B.L., Quantitative studies on cutaneous erythema induced by ultraviolet radiation, Br. J. Dermatol., 111, 673, 1984. 19. Farr, P.M. and Diffey, B.L., The erythemal response of human skin to ultraviolet radiation, Br. J. Dermatol., 113, 65, 1985. 20. Farr, P.M., Besab, J.E., and Diffey, B.L., The time course of UVB and UVC erythema, J. Invest. Dermatol., 91, 454, 1988. 21. Farr, P.M. and Diffey, B.L., A quantitative study of the effect of topical indomethacin on cutaneous erythema induced by UVB and UVC radiation, Br. J. Dermatol., 115, 453, 1986. 22. Takiwaki, H., Shirai, S., Kohno, H., Soh, H., and Arase, S., The degrees of UVB-induced erythema and pigmentation correlate linearly and are reduced in a parallel manner by topical anti-inflammatory agents, J. Invest. Dermatol., 103, 642, 1994. 23. Park, S.B., Huh, C.H., Choe, Y.B., and Youn, J.I., Time course of ultraviolet-induced skin reactions evaluated by two different reflectance spectrophotometers: DermaSpectrometer and Minolta spectrophotometer, Photodermatol. Photoimmunol. Photomed., 18, 23, 2002. 24. Damian, D.L., Halliday, G.M., and Barnetson, R.S., Prediction of minimal erythema dose with a reflectance melanin meter, Br. J. Dermatol., 136, 714, 1997. 25. Held, E. and Agner, T., Comparison between 2 test models in evaluating the effect of a moisturizer on irritated human skin, Contact Derm., 40, 261, 1999.
671
26. Fluhr, J.W., Kuss, O., Diepgen, T., Lazzerini, S., Pelosi, A., Gloor, M., and Barardesca, E., Testing for irritation with a multifactorial approach: comparison of eight noninvasive measuring techniques on five different irritation types, Br. J. Dermatol., 145, 696, 2001. 27. Suh, D.H., Kwon, T.E., Kim, S.D., Park, S.B., Kwon, O.S., Eun, H.C., and Youn, J.I., Changes of skin blood flow and color on lesional and control sites during PUVA therapy for psoriasis, J. Am. Acad. Dermatol., 44, 987, 2001. 28. Gladstone, H.B., Nguyen, S.L., Williams, R., Ottomeyer, T., Wortzman, M., Jeffers, M., and Moy, R.L., Efficacy of hydroquinone cream (USP 4%) used alone or in combination with salicylic acid peels in improving photodamage on the neck and upper chest, Dermatol. Surg., 26, 333, 2000. 29. Hurley, M.E., Guevara, I.L., Gonzales, R.M., and Pandya, A.G., Efficacy of glycolic acid peels in the treatment of melasma, Arch. Dermatol., 138, 1578, 2002. 30. Miyamoto, H., Takiwaki, H., Yamano, M., Ahsan, K., and Nakanishi, H., Color analysis of nevus of Ota for evaluation of treatment with a Q-switched alexandrite laser, Skin Res. Technol., 3, 45, 1997. 31. Manuskiatti, W., Sivayathorn, A., Leelaudomlipi, P., and Fitzpatrik, R.E., Treatment of acquired bilateral nevus of Ota-like macules (Hori’s nevus) using a combination of scanned carbon dioxide laser followed by Q-switched ruby laser, J. Am. Acad. Dermatol., 48, 584, 2003. 32. Queille-Roussel, C., Poncet, M., and Schaffer, H., Quantification of skin-colour changes induced by topical corticosteroid preparation using Minolta Chroma Meter, Br. J. Dermatol., 124, 264, 1991.
78 Dynamic Capillaroscopy H.S. Yu,1 C.H. Lee,2 and C.H. Chang3 1
Department of Dermatology, College of Medicine, National Taiwan University Hospital, Taipei, Taiwan 2Department of Dermatology, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan 3Department of Dermatology, College of Medicine, Tzu-Chi Medical University, Hualien, Taiwan
CONTENTS 78.1 Introduction............................................................................................................................................................673 78.2 Methodological Principle ......................................................................................................................................673 78.2.1 Computerized Laser Capillary Microscopy ..............................................................................................673 78.2.2 Dynamic Capillaroscopy with Load: Pressure, Cold Provocation, and Iontophoresis ............................674 78.3 Application.............................................................................................................................................................674 78.3.1 Tetralogy of Fallot .....................................................................................................................................674 78.3.2 Raynaud’s Syndrome and Connective Tissue Diseases............................................................................675 78.3.3 Diabetes Mellitus .......................................................................................................................................675 78.3.4 Arteriosclerosis ..........................................................................................................................................675 References .......................................................................................................................................................................676
78.1 INTRODUCTION Cutaneous microcirculation can be divided into thermoregulatory shunt vessels and nutritive skin capillaries. Laser Doppler flowmetry, transcutaneous oxygen pressure measurement, and capillary microscopy are known noninvasive methods for assessing cutaneous microcirculation. Flux in nonnutritional shunt vessels dominates the signal recording of the laser Doppler flowmetry. Transcutaneous oxygen pressure primarily reflects the function of the nutritive skin capillaries. Presently, capillary microcopy is the best choice for studying the nutritional status of a certain skin area, and it is the only method that allows direct visualization of the nutritive dermal vessels in vivo.1,2 In 1964, Zimmer and Demis3 demonstrated a microscope–television system for studying dynamic blood flow in human skin capillaries. Bollinger and coworkers4 further refined this method and adapted a frame-to-frame analysis to measure both blood flow velocity and vessel diameters in nail-fold capillaries. The application of the cross-correlation technique5 for measuring the velocity of blood cells in the capillaries greatly improved this measurement. Thereafter, capillary microscopy has been coupled with a videophotometric system and used with software to analyze the capillary blood cell velocity (CBV).6 This technique makes it possible to noninvasively study human skin capillaries under physiological and pathophysiological conditions. Applications
include evaluation of the dynamic microcirculatory status in peripheral vascular disorders, including arterial occlusive disease, connective tissue disease, and diabetes mellitus. The purpose of this chapter is to introduce recent advances in dynamic capillaroscopy in the study of diseases, namely, tetralogy of Fallot, Raynaud’s syndrome, atherosclerosis, and diabetes mellitus (DM); a focus on clinical application will be presented.
78.2 METHODOLOGICAL PRINCIPLE 78.2.1 COMPUTERIZED LASER CAPILLARY MICROSCOPY During the past few years, a new type of computerized capillary microscope has been developed (Figure 78.1). A low-power near-infrared laser is focused to a 10-microndiameter beam that can be positioned onto a single capillary. The built-in high-resolution charge coupled device (CCD) camera allows continuous monitoring of the capillary position on either the computer screen or a separate video monitor. The laser beam is reflected by moving blood cells at the focal point. The frequency of the reflected beam is Doppler shifted. The detected shift is directly proportional to the velocity of the reflecting blood cells. Coupled with software for analysis, digital image 673
674
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 78.1 Computerized laser capillary microscope.
change and rest CBV (rCBV) in capillaries. However, in some pathological conditions, physical load is necessary to display the abnormality in capillaries. A 1-min arterial occlusion and cold provocation tests are useful methods for this purpose. An arterial occlusion in of digital arteries results in a reactive hyperemia, which indicates the myogenic response of the precapillary sphincter. The important parameters for postocculsive reactive hyperemia response include peak CBV (pCBV), time to pCBV (tpCBV), and the percent increase of CBV above rCBV during postocculsive reactive hyperemia (%PRH). The cold provocation test5–15 has been used for studying the disturbances of skin microvascular reactivity in different types of Raynaud’s syndrome. Using dynamic capillaroscopy and iontophoresis, Serne et al.7 demonstrated that systemic hyperinsulinemia in skin induces recruitment of capillaries, augments vasodilatation, and influences vasomotion in diabetic patients.
78.3 APPLICATION 78.3.1 TETRALOGY
FIGURE 78.2 Normal hairpin-like capillary loops in regular arrangement.
output, and an improved optical and electronic system, it provides an easy, stable, direct, and accurate measurement of cutaneous capillary dynamics. For CBV measurement, capillaries in the nail-fold area are suitable for this purpose because they lie parallel to the skin surface and can be visualized rather nicely in their full length, resembling a hairpin (Figure 78.2). A finger or toe is placed on the investigation plate, and a small bracket is allowed to lightly touch the distal end of the nail. A drop of immersion oil is applied to make the nail fold transparent and increase refraction. When capillary microscopy is performed using a video recording system, it permits realtime imaging of cutaneous capillaries and retrospective analysis of capillary dynamics. The whole process is performed automatically. Observation parameters consist of recording of capillary density; capillary morphology, including caliber, length, shape, and tortuosity; intercapillary distance; and CBV in afferent limb, apex, and efferent limb of the capillary loops.
OF
FALLOT
Tetralogy of Fallot (TF) is recognized as the most common congenital cyanotic cardiac malformation, consisting of a ventricular septal defect, pulmonary stenosis, aortic override, and right ventricular hypertrophy. Patients with TF usually suffer from generalized cyanosis, clubbing digits, and a high risk of thrombosis due to secondary polycythemia. Cutaneous microcirculation was studied, and we found that the nail-fold capillaries in TF patients became dilated, tortuous, and branched with an increase of total length and vascular area (Figure 78.3). The degree of dilation and vascularity was closely related to the hemoglobin (Hb) concentration. CBV declined with the increase of hematocrit (Hct), and a significant reduction was noted when Hb >19 g/dl. TF patients are good models for studying the effects of long-term hypoxia. Capillary dilation and vascularity increase are compensatory
78.2.2 DYNAMIC CAPILLAROSCOPY WITH LOAD: PRESSURE, COLD PROVOCATION, AND IONTOPHORESIS Computerized capillary microscopy is the most sensitive noninvasive method for evaluation of both morphological
FIGURE 78.3 Extremely dilated, torturous capillary loops observed in a patient of tetralogy of Fallot with hemoglobin of 22 g/dl.
Dynamic Capillaroscopy
675
reactions and are reversible. We can use capillary microscopy to evaluate the compensatory status of TF patients before operation and dynamic changes after operation.8
78.3.2 RAYNAUD’S SYNDROME TISSUE DISEASES
AND
CONNECTIVE
Raynaud’s syndrome is the paroxysmal constriction of small arteries of the extremities, usually precipitated by cold. When exposed to low temperatures, the digits become white (ischemic), then blue (cyanotic), and finally red (hyperemia).9 Over 90% of the cases are primary Raynaud’s syndrome. The principal clinical challenge is to distinguish idiopathic cases of primary Raynaud’s syndrome from secondary ones due to underlying disease of connective tissues, obstructive arterial disease, blood dyscrasia, drug toxicity, or artery injury. Maricq et al.10,11 first described the abnormality of capillary morphology in connective tissue disease, especially in scleroderma. It is a qualitative method to differentiate primary from secondary Raynaud’s syndrome. A quantitative morphological analysis of capillary microscopy was developed by Lefford and Edward.12 The pattern of capillary morphology consists of enlarged and deformed capillaries with dilation of both limbs of the loop, which is often engorged with blood (sausage loop). Marked disorganization of the loop is observed. Loss of capillaries produces many avascular areas and the disruption of the orderly appearance of the capillary bed. Capillary morphology of primary Raynaud’s disease is approximately normal, but the CBV decreased markedly after cold exposure.13 On the other hand, abnormal morphological changes were noted in secondary Raynaud’s syndrome, including an increase in intercapillary distance and ratio of torturous capillary loops, and a decrease in capillary loops, numbers, and total length.14,15 In addition, there is a significant dilation of afferent and efferent limbs and apex.16 The degree of abnormal morphological changes in systemic sclerosis (Figure 78.4) is more severe than lupus erythematosus. Enlarged, giant capillaries, a
FIGURE 78.4 In progressive systemic sclerosis (PSS), a continuous blood stream in capillary is replaced by a slow blood flow and a granular pattern (arrow) that reflects an aggregation of erythrocytes in capillaries.
reduced numbers of capillaries, severe avascularity, and hemorrhage were most commonly seen in systemic sclerosis than in other connective tissue diseases and normal controls.17 Patients with Raynaud’s phenomenon with avascularity or a mean of more than two megacapillaries per digit are likely to develop a scleroderma spectrum disorder.18 Dynamic capillaroscopy reveals its accuracy in differentiating systemic sclerosis from primary Raynaud’s syndrome and lupus erythematosus, as well as the clinical progressions of underlying diseases.
78.3.3 DIABETES MELLITUS Microcirculation is known to be disturbed in many organs of diabetic patients. A previous study revealed that prevalence of coiled and slightly enlarged microvessels is significantly increased in long-term diabetics at the nail folds of fingers19 and toes.20 In diabetes without retinopathy, no significant changes of nail-fold capillaries could be detected. In diabetes with background retinopathy, they showed dilatation and tortuosity, whereas in diabetes with proliferative retinopathy, they showed loss of loop and retardation of blood flow (Figure 78.5). It was also found that rCBV did not differ significantly in patients with diabetes compared to controls.21 However, dynamic capillaroscopy with a 1-min digital arterial occlusion test can detect early impairment of cutaneous microcirculation in DM.22 In nonretinopathy diabetes, functional impairments of capillary circulation include decreased rCBV, pCBV, and prolonged tpCBV. The degree of tortouosity of capillaries and impairment in pCBV and tpCBV of capillary circulation are significantly correlated with the gravity of retinopathy in diabetic patients.23 Dynamic capillaroscopy used in concert with ophthalmoscopy can facilitate a comprehensive examination of vasculopathy in DM.
78.3.4 ARTERIOSCLEROSIS Morphological changes of capillaries appear with the progression of peripheral vascular disease. Jorneskog et al.24
FIGURE 78.5 In diabetes with proliferative retinopathy, in addition to tortuosity, there is a loss of loop.
676
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 78.6 In arteriosclerosis, there are many dilated and tortuous loops of capillaries.
utilized dynamic capillaroscopy to measure capillary blood cell velocity during rest and following a 1-min arterial occlusion at the toe base. The skin microvascular reactivity was impaired in both diabetic and nondiabetic patients with peripheral vascular disorders.24 In patients with arteriosclerosis, the morphology analysis demonstrated increased numbers of tortuous loops of capillaries in affected toes (Figure 78.6). The rCBV of both toes and fingers was within normal range. However, there was a significant deterioration in the parameters of postocclusive reactive hyperemia response, i.e., a decrease in pCBV and an increase in tpCBV. These results suggest that dynamic capillaroscopy is a sensitive method for evaluating the cutaneous nutritive status in arteriosclerosis.25 Dynamic capillaroscopy provides a new approach to the early detection of circulatory disturbances resulting from different mechanisms such as collagen vascular diseases, diabetes mellitus, arteriosclerosis, and perhaps even some types of heart disease.
REFERENCES 1. Fagrell B. Vital capillary microscopy: a clinical method for studying changes of the nutritional skin capillaries in legs with arteriosclerosis obliterans. Scan J Clin Lab Invest, Suppl. 133, 1973. 2. Fagrell B. Advances in microcirculation network evaluation: an update. Int J Microcirc 15:34–40, 1995. 3. Zimmer JG, Demis DJ. The study of the physiology and pharmacology of the human cutaneous microcirculation by capillary microscopy and television cinematography. Angiology 15:232–235, 1964. 4. Bollinger A, Butti P, Barras JP, Trachsler H, Siegenthaler W. Red blood cell velocity in nailfold capillaries of man measured by a television microscopy technique. Microvasc Res 7:61–72, 1974. 5. Intaglietta M, Silverman NR, Tompkins WR. Capillary flow velocity measurements in vivo and in situ by television methods. Microvasc Res 10:165–179, 1975. 6. Fagrell B, Eriksson SE, Malmstrom S, Sjolund A. Computerized data analysis of capillary blood cell velocity. Int J Microcirc Clin Exp 7:276, 1988.
7. Serne EH, IJzerman RG, Gans RO, Nijveldt R, De Vries G, Evertz R, Donker AJ, Stehouwer CD. Direct evidence for insulin-induced capillary recruitment in skin of healthy subjects during physiological hyperinsulinemia. Diabetes 51:1515–1522, 2002. 8. Chang CH, Yu HS. Study of cutaneous microcirculation in tetralogy of Fallot. Microsvas Res 51:59–68, 1996. 9. Coffman JD. Rayanud’s phenomenon. Hypertension 17:593–602, 1991. 10. Maricq HR, LeRoy EC. Patterns of finger capillary abnormalities in connective tissue disease by wide-field microscopy. Arthritis Rheum 16:619–628, 1973. 11. Maricq HR, LeRoy EC, Dangelo WA, Medsger TA, Rodnan GP, Sharp GC, Wolfe JF. Diagnostic potential of in vivo capillary microscopy in scleroderma and related disorders. Arthritis Rheum 23:183–189, 1980. 12. Lefford F, Edward JCW. Nailfold capillary microscopy in connective tissue disease: a quantitative morphological analysis. Ann Rheum Dis 45:741–749, 1986. 13. Jacohs MJHM, Breslau PJ, Slaaf DW, Reneman RS, Lemmens JAJ. Nomenclature of Raynaud’s phenomenon: a capillary microscopic and hemorheologic study. Surgery 101:136–145, 1987. 14. Ohtsuka T, Ishikawa H. Graphic analysis of nailfold capillary in patients with collagen disease, especially in those with systemic scleorosis. Jpn J Clin Dermatol 45:637–644, 1991. 15. Caspary L, Schmees C, Schoetensack I, Hartung K, Stannat S, Deicher H, Creutig A, Alexander K. Alternations of the nailfold capillary morphology associated with Raynaud’s phenomenon in patients with systemic lupus erythematosus. Rheumatology 18:559–566, 1991. 16. Liu CG, Su W, Luo Y. Changes in cutaneous microcirculation, hemorheology and platelet aggregation function in dermatomyositis. J Dermatol Sci 2:346–352, 1991. 17. Kabasakal Y, Elvins DM, Ring EF, McHugh NJ. Quantitative nailfold capillaroscopy findings in a population with connective tissue disease and in normal healthy controls. Ann Rheum Dis 55:507–512, 1996. 18. Zufferey P, Depairon M. Prognostic significance of nailfold capillary microscopy in patients with Raynaud’s phenomenon and scleroderma-pattern abnormalities. A six-year follow-up study. Clin Rheumatol 11:536–541, 1992. 19. Rouen LR, Terry EN, Doft BH, Clauss RH, Redisch W. Classification and measurement of surface microvessels in man. Microvasc Res 4:285–292, 1972. 20. Chazan BI, Balodimos MC, Larine RI, Koncz L. Capillaries of the nailfold in diabetes mellitus. Microvasc Res 2:504–507, 1970. 21. Tooke JE, Lins PE, Ostergren J, Fagrell B. Skin microvascualr autoregulatoy response in type I diabetes: the influence of duration and control. Int J Microcirc Clin Exp 4:249–256, 1985. 22. Tooke JE, Ostergern J, Lins PE, Fagrell B. Skin microvascular blood flow control in long duration diabetes with and without complications. Diabetes Res 5:187–192, 1987.
Dynamic Capillaroscopy
23. Chang CH, Tsai RK, Wu WC, Kuo SL, Yu HS. Use of dynamic capillaroscopy for studying cutaneous microcirculation in patients with diabetes mellitus. Microvasc Res 53:121–127, 1997. 24. Jorneskog G, Brismar K, Fagrell B. Skin capillary circulation is more impaired in the toes of diabetic than non-diabetic patients with peripheral vascular disease. Diabet Med 12:36–41, 1995.
677
25. Yu HS, Chang CH, Chen GS, Yang SA, Yu CL. Study of dynamic microcirculatory problems in ‘blackfoot disease’: emphasizing its differences from arteriosclerosis. J Biomed Sci 2:183–188, 1995.
and 79 Capillaroscopy Videocapillaroscopy Assessment of Skin Microcirculation: Dermatological and Cosmetic Approaches Philippe Humbert,1 Jean-Marie Sainthillier,1 Sophie Mac-Mary,1 Adeline Petitjean,1 Pierre Creidi,1 and Tijani Gharbi2 1 2
Laboratoire de Biologie et Cutanée d’Ingénierie, Besançon France Laboratoire d’Optique P.M. Duffieux, University of Franche-Comté, Besançon France
CONTENTS 79.1 79.2 79.3 79.4 79.5 79.6 79.7 79.8 79.9 79.10
Anatomy of the Skin Microcirculation................................................................................................................679 Capillaroscopy......................................................................................................................................................679 Periungual Capillaroscopy ...................................................................................................................................680 Videocapillaroscopy .............................................................................................................................................680 Capillaries Morphology .......................................................................................................................................681 Hypertension Assessment.....................................................................................................................................682 Venous Insufficiency ............................................................................................................................................682 Age-Related Changes of the Cutaneous Microcirculation..................................................................................683 Pharmacological Inhibition of the Dermal Microcirculation ..............................................................................683 Quantitative Assessment.......................................................................................................................................684 79.10.1 Image Processing Techniques................................................................................................................684 79.10.2 Capillaries Detection..............................................................................................................................684 79.10.3 Geometrical Capillary Network Analysis..............................................................................................684 79.11 Cosmetical Example.............................................................................................................................................685 79.11.1 Couperose and Erythrosis Assessment ..................................................................................................685 79.12 Conclusion............................................................................................................................................................685 References .......................................................................................................................................................................686
79.1 ANATOMY OF THE SKIN MICROCIRCULATION The skin microcirculation is organized as two horizontal plexuses, one located 1 to 1.5 mm below the skin surface, and the other at the dermal-subcutaneous junction that comprises collecting veins. It consists in arterioles, which may divide into many capillary loops at the level of the papillary layer of the skin. Indeed, arterial capillaries rise to form the dermal papillary loops at this level.1,2 Then, capillaries converge into collecting systems of the venous plexus. The vascular network of the skin varies considerably from one area to another.3,4
79.2 CAPILLAROSCOPY Among the different techniques available5 to study skin microcirculation (Table 79.1), human skin capillaroscopy, a specialized form of intravital microscopy, is the only method that allows direct visualization of the capillary network in vivo. Its principle is easy. After skin transparency has been enhanced by a drop of oil, an optical magnifying system allows visualization of its vascular network directly through the skin. The optical devices used to examine the cutaneous capillaries in vivo are most of the time the light microscope and stereomicroscope,5 but also the videocapillaroscope. 679
680
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 79.1 Non-Invasive Bioengineering Techniques Used to Study Skin Microcirculation Skin temperature measurements Capillaroscopy Dynamic capillaroscopy Dynamic capillaroscopy with dye Laser Doppler flowmetry Isotope techniques (133xenon) Transcutaneous measurement of partial oxygen pressure Capillary pressure Photopulse plethysmography Infrared thermography Colorimetry
79.3 PERIUNGUAL CAPILLAROSCOPY Nail-fold capillaroscopy (Figure 79.1) is usually performed in order to search some capillary deformations, characterizing the presence of a pathological situation.6,7 In this area, capillaries lie in a horizontal plane, so that a large part of their loops can be observed. Their aspect is particular. They look like hairpins with a diameter of 8 to 15 μm. Each loop is parallel to the other one, and oriented to the extremity of the finger (Figure 79.2A). Some particular aspects of the capillary loops allow to detect some systemic conditions, such as progressive systemic sclerosis and lupus erythematosus (Figure 79.2B). Nail-fold capillaroscopy has a good sensitivity and a good specificity. Therefore, morphological abnormalities must systematically be looked for on all accessible fingers.
79.4 VIDEOCAPILLAROSCOPY
(a)
(b)
FIGURE 79.2 Nail-fold capillaroscopy images. (a) Normal aspect. (b) Dilated capillaries with heterogenous distribution.
Contact videomicroscopy systems appeared recently in the industry for nondestructive control processes. They require epi-illumination of the skin surface8 and image transmission to a video camera via the optics of a microscope. The sensor located at one end of a flexible cord easily allows investigation on the whole tegument’s surface. Optical fibers convey illuminating light, which is provided by the handheld probe. The images are then visualized on a screen and picture digitalization is thus performed. Image numerization can be done directly during the examination; therefore, the data quality can be controlled instantly. The video imaging systems (Scopeman®, FORT®, Microvision®, Microwatcher) consist of a video signal control unit and a mini-CCD camera. A manually adjusted
Videocapillaroscopy devices (Figure 79.3) tend to replace the more conventional capillaroscopy instruments.
FIGURE 79.1 Capillaroscopy equipment for the nail-fold analysis.
FIGURE 79.3 Contact videocapillaroscopy with monitoring on TV.
Capillaroscopy and Videocapillaroscopy Assessment of Skin Microcirculation
681
TABLE 79.2 Different Architectural Frameworks of Skin Capillary Network Parallel arrangement and regular meshes network
Parallel arrangement with irregular meshes network
Perpendicular arrangement and regular dot line
Perpendicular arrangement and irregular dot line
Special pattern with parallel arrangement
Forehead Cheekbone region Cheek Chin Internal surface of arms
Trunk (anterior and posterior aspects) Breast Arms (external surface) Legs (internal and external surfaces)
Fingertip Eminentia tenar Eminencia hipotenar Tip of toes
Palm of hands Back of the hand and the foot Nipple
Finger nail fold Labial mucosa
From Miniati, B. et al., Ital. J. Anat. Embryol., 106, 233–238, 2001.
focusing system coupled with the camera head allows the obtaining of a sharp image of the capillary network. Magnification ranges from ×100 to ×1000. A magnification higher than ×600 enables visualization of blood cells inside the capillary. Venous congestion increases the number of capillaries detected.
79.5 CAPILLARIES MORPHOLOGY Capillaroscopy does not visualize the capillary wall, but only the blood red cells that cast the vessel. Thus, only functional capillaries can be detected. In most skin body
areas, the vascular morphology showed differences (Table 79.2). On the forehead and crow’s-foot, line and network forms dominate; on the dorsum of the hand, the dot and comma ruled5; while on the inner forearm both types were basically equal. A normal architectural framework shows two main patterns, a parallel and a perpendicular arrangement of capillary loops with respect to the skin surface (Figure 79.4). Capillary loops with a parallel arrangement form a vascular network with meshes that may or may not be regular. In most skin body areas9,10 capillaries are perpendicular to the skin surface, so that only the summit can be
A
B
C
D
FIGURE 79.4 Vascular morphology examples in different skin body areas: (A) scalp, (B) forehead, (C) cheek, (D) forearm.
682
Handbook of Non-Invasive Methods and the Skin, Second Edition
79.6 HYPERTENSION ASSESSMENT TABLE 79.3 Pathological or Cosmetical Usefulness of Videocapillaroscopy Peripheral arterial obliterative disorders12 Venous insufficiency13 Diabetes mellitus14 Hypertension15 Psoriasis16
Aging17
Specific morphological changes Rarefaction and dilatation of the capillary loops Tortuous and dilated capillaries Capillary rarefaction Grossly dilated and tortuous capillaries; many more perfused capillaries Reduced dermal papillary loops
Effects of topical cosmetics or chemical agents
seen. It looks like a point or a comma. The caliber of capillaries varies from 15 to 20 μm in regions with parallel arrangement of capillary loops. The capillary density ranges from 14 to 30 capillary loops per square millimeter in skin regions where capillary loops are arranged perpendicularly to the skin surface.4 The availability of videomicroscopes together with advances in our knowledge about the importance of microcirculation in the pathogenesis of trophic complications of arterial and venous insufficiencies explains the present development of capillaroscopy on the skin. The determination of morphological or dynamic changes in the microcirculation belongs to the non-invasive techniques of the biometrological domain. In pathology, numerous conditions can be better examined with videocapillaroscopy (Table 79.3): peripheral arterial obliterative disorders, venous disorders such as venous insufficiency, diabetes mellitus, hypertension, couperosa (Figure 79.5A), and psoriasis (Figure 79.5B).
(a)
Capillary network analysis on the volar aspect of the forearm or on the fingers could be of interest in hypertension. Indeed, it was suggested that microvascular rarefaction represents an important mechanism in primary hypertension. Capillary rarefaction has been described in various tissues from patients with essential hypertension. The introduction of intravital videomicroscopy allowed the discovery of a 15 to 20% reduction in the capillary density of the nail-fold skin, and intravital fluorescein angiography found a 20% reduction in capillary density in the forearm skin of hypertensive subjects compared with normotensive subjects.11 Some other conditions, such as peripheral arterial obliterative disorders, venous disorders like venous insufficiency, diabetes mellitus, and psoriasis (Table 79.3), are candidates for microcirculatory evaluation.
79.7 VENOUS INSUFFICIENCY Videocapillaroscopy is mainly performed on the instep in venous insufficiency, and on the back of the foot (first intermetatarsian space and toes) in arterial insufficiency. Venous insufficiency is characterized by the decrease of the capillaries density, the widening of the dermal papilla, the size of which becomes heterogeneous, and the contours marked by the hemosiderine deposits of ochre dermatitis. The decrease of the capillaries density is partially compensated by their increased length, inducing an increased number of meanders, which can even take the shape of a glomerular cluster in the most severe forms (lipodermatosclerosis, white atrophy). As soon as the first trophic troubles occur, the venules are not visible any longer.18–20
(b)
FIGURE 79.5 Peripheral arterial obliterative disorders. (a) Couperosa (×50 = 27 mm2). (b) Psoriasis with dilated and tortuous capillaries (×200 = 1.73 mm2).
Capillaroscopy and Videocapillaroscopy Assessment of Skin Microcirculation
683
TABLE 79.4 Number of Capillary Loops in Five Age Groups on Different Examined Sites (n = 20) Sites
Group 1
Group 2
Group 3
Group 4
Group 5
Age range (years) (Mean age) Hand Arm Postauricular
20–29 (25.3) 48.5 23.3 12.6
30–39 (34.5) 41.4 23.1 16.5
40–49 (44.6) 49 26.8 13.3
50–59 (53.5) 38 26.8 15.5
60–69 (64.4) 28.5 20.1 2.6
Source: Zhu, W., Assessment of Cutaneous Microvasculature in Aging and Photoaging: A Videocapillaroscopic Study on Caucasian Women, personal data.
Fagrell has described three classes of increasing severity and validated their discriminating value for the local trophic prognosis:21,22 A, capillary dilatation; B, edema impairing capillary visualization; and C, absence of visible capillary (prenecrosis stage).
79.8 AGE-RELATED CHANGES OF THE CUTANEOUS MICROCIRCULATION A significant decrease of cutaneous capillary loop density is observed with age (Table 79.4). The microvasculature is regularly formed in young skin, with many orderly arranged capillary loops (dots) and some horizontal vessels (lines) (Figure 79.6A).23 It becomes thicker, twisted, and irregular in older skins (Figure 79.6B), with horizontal vessels that appear tortuous, elongated, disorganized, and dilated (W. Zhu, Assessment of Cutaneous Microvasculature in Aging and Photoaging: A Videocapillaroscopic Study on Caucasian Women, personal data). Thus, the parallel vasculature can be more easily observed with aging. This result accords with that of biopsy specimens
(a)
observed by light or electric microscope. As the epidermis becomes thinner, the transparency of the skin increases, facilitating the observation of the papillary vascular plexus. And as the vasculature expands and thickens, some of the microvasculature usually difficult to observe and the deeper vasculature can also be examined. Results also indicate that photodamaged vessels correspond to an increased formation of new vessels.24
79.9 PHARMACOLOGICAL INHIBITION OF THE DERMAL MICROCIRCULATION A pharmacological agent such as neosynephrin is able to induce dermal capillary density reduction after topical application.25,26 In this case, videocapillaroscopy is a reliable tool to quantify in vivo the capillary loop density in the dermis before and after the application of the solution. This pharmacological model could be useful to evaluate the effects of different cosmetic or drug agents able to inhibit the dermal adrenergic response in dermis.
(b)
FIGURE 79.6 Capillary patterns observed by videocapillaroscopy with young (a) and elderly subject (b) on the hand.
684
Handbook of Non-Invasive Methods and the Skin, Second Edition
79.10 QUANTITATIVE ASSESSMENT
A
Videocapillaroscopy is mainly used for the morphological information it provides, and it is different from indirect techniques (laser Doppler, transcutaneous PO2) also because it is free from interpretation artifacts, a considerable advantage in a field as complex as microcirculation. However, the absence of quantification and the subjective and operator-dependent character of capillaroscopy show its limits.
79.10.1 IMAGE PROCESSING TECHNIQUES Quantitative capillaroscopy is now possible due to the development of computerized systems, and series of advanced image processing methods have been developed.
79.10.2 CAPILLARIES DETECTION Finding capillary loops automatically in an image is a difficult yet important first step in order to achieve microcirculation analysis. An automatic counting of capillaries is now available27,28 with a detection rate of 82%. This detection system was based on a principal component analysis (PCA) and was associated with a retinally connected neural network. This filter system is capable of real-time processing, recognizes capillaries anywhere in an image, and operates successfully under a wide range of lighting and noisy conditions. It is assumed that the study of microcirculation must include all dynamic and cooperative processes between the capillaries. Indeed, the geometric complexity of the capillary structure can lead to heterogeneities in oxygen delivery. If a finite tissue element is not within the proximity of a capillary site, necrosis could develop in that tissue area.
79.10.3 GEOMETRICAL CAPILLARY NETWORK ANALYSIS For characterizing capillary ensembles, the statistical and geometrical properties of the network need to be explored.29 In fact, this network forms a map that is based on adjacent relationships among the capillaries. These relationships30,31 can be quantified by a distance parameter (distance between one capillary and all its neighbors) and a surface parameter (an influence area corresponding to a surface around the capillary). Zhong et al.29 described for the first time a method to construct the capillary network by using image processing with Delaunay triangulation and a Voronoï diagram. Delaunay triangulation32 was implemented to obtain the nearest neighbor for each capillary (Figure 79.7A). This representation permitted calculation of the minimal, maximal, or mean distance between capillaries. The Voronoi diagram33 was used to determine a surface parameter that
B
FIGURE 79.7 Delaunay triangulation (A) and Voronoï diagrams (B): 67 capilllaries, on the scalp (×200 = 1.73 mm2).
could be considered an oxygen diffusion area (Figure 79.7B). For each parameter, the distribution (thresholding or normalization) was analyzed in order to eliminate extreme values, which correspond to artifacts.34 We noticed that without these values, the proximity parameters distribution was normal. These algorithms have been implemented in a software platform (Capilab Toolbox®) providing functions and interactive tools for analyzing pictures of the skin. This graphical environment integrates complete statistical and geometrical computing, visualization, and image enhancement algorithms. The analysis could be improved by taking into account morphologic parameters (shape, roundness) of each capillary (Figure 79.8). Indeed, we do not know whether a detected capillary is physiologically active, or if an interaction exists between patterns with different shapes or sizes. The mathematical morphology could be a mean to focus the analysis on a specific group of capillaries, with the same properties, or to construct a network with shapebased relations.
Capillaroscopy and Videocapillaroscopy Assessment of Skin Microcirculation
A
B
C
D
685
FIGURE 79.8 Intermediate images showing the morphologic image processing technique.
79.11 COSMETICAL EXAMPLE 79.11.1 COUPEROSE
AND
ERYTHROSIS ASSESSMENT
Rosacea is a frequent disease that occurs mostly in women. Rosacea is heralded, around the age of 20 years, by intermittent facial erythema and by the gradual development of permanent erythema (erythrosis) with telangiectasia (couperose). Image processing techniques based on neural network algorithms can be used again to detect and quantify by a surface parameter the erythema in the pictures. A color neural network-based method was developed (based on an RGB and a lab code) that is capable of real-time processings, increasing the quality of videocapillaroscope images and minimizing the disturbance of artifacts.35 This system can selectively recognize regions where the couperose is heavy or light. The algorithms (implemented in Capilab Toolbox) can process any kind of color images of the skin (Figure 79.9).
79.12 CONCLUSION The determination of morphological or dynamic changes in the cutaneous microcirculation belongs to the non-invasive techniques of the biometrological domain. Every capillary modification due to topical cosmetic products or chemical agents can then be observed. In pathology,
numerous conditions can be better examined with this system. Capillaroscopy is now used routinely in different hospitals and companies. Although its employment requires some experience, this technique brings direct information on the capillary network morphology. It differs from the indirect methods used to explore microcirculation, such as laser Doppler or transcutaneous partial oxygen. The techniques to visualize skin capillaries enable getting an irreplaceable approach of the physiology and physiopathology of the skin capillary circulation. Compared to other heavy research methods, traditional capillaroscopy techniques, much simpler and cheaper, have shown their usefulness in the detection of connectivity microangiopathies and vascular acrosyndromes. Combined with the potential of numerical image analysis, they will probably extend their application fields to the assessment of the influence of arterial and venous diseases on the skin’s nutritional circulation. Capiflow®, Capiflow AB, Kista, Sweden FORT® Imaging Systems, Curno BG, Italy Scopeman ® Moritex, P.M. Industries Ltd., Cambridge, U.K. Microvision® MV 2100, Finlay Microvision Co. Ltd., Warwickshire, U.K.
686
Handbook of Non-Invasive Methods and the Skin, Second Edition
A
B
C
D
FIGURE 79.9 Examples of couperose image processing.
Microwatcher Model VS-10®, Mitsubishi Kasei Corp., Tokyo, Japan Capilab Toolbox®, LIBC, Besançon, France
REFERENCES 1. Rhodin JAG. Anatomy of the microcirculation. In Microcirculation: Current Physiologic, Medical and Surgical Concepts, Effros RM, Schmid-Schöbein H, Ditzel J, Eds. Academic Press, New York, 1981, pp. 11–7. 2. Braverman IM. The cutaneous microcirculation. J Invest Dermatol 5: 3–9, 2000. 3. Bongard O, Bounameaux H. Clinical investigation of skin microcirculation. Dermatology 186: 6–11, 1993. 4. Miniati B, Macchi C, Molino Lova R, et al. Descriptive and morphometric anatomy of the architectural framework of microcirculation: a videocapillaroscopic study on healthy adult subjects. Ital J Anat Embryol 106: 233–238, 2001. 5. Carpentier PH. Méthodes d’exploration vasculaire chez l’homme: microcirculation et veines. Therapie 54: 369–374, 1999. 6. Hu Q, Mahler F. New system for image analysis in nailfold capillaroscopy. Microcirculation 6: 227–235, 1999.
7. Allen PD, Taylor CJ, Herrick AL, Moore T. Image Analysis of Nailfold Capillary Patterns. Medical Image Understanding and Analysis 1998. http://www.isbe.man.ac.uk/~pa/NailFold.html. 8. Jairo J, Monari M. Human capillaroscopy by light emitting diode epi-illumination. Microvasc Res 59: 172–175, 2000. 9. Li L, Mac-Mary S, Sainthillier JM, et al. Cutaneous facial vascular network cartography. Ann Dermatol Venereol 129: 1172, 2002 (abstract). 10. Li L, Mac-Mary S, Sainthillier JM, et al. Vascular network cartography of the body skin. Ann Dermatol Venereol 129: 1173, 2002 (abstract). 11. Prasad A, Dunnill GS, Mortimer PS, MacGregor GA. Capillary rarefaction in the forearm skin in essential hypertension. J Hypertens 13: 265–268, 1995. 12. Hern S, Mortimer PS. Visualization of dermal blood vessels: capillaroscopy. Clin Exp Dermatol 24: 473–478, 1999. 13. Fagrell B. Advances in microcirculation network evaluation: an update. Int J Microcirc 15 (Suppl. 1): 34–40, 1995. 14. Chang CH, Tsai RK, Wu WC, et al. Use of dynamic capillaroscopy for studying cutaneous microcirculation in patients with diabetes mellitus. Microvasc Res 53: 121–127, 1997. 15. Serne EH, Gans RO, ter Maaten JC, et al. Impaired skin capillary recruitment in essential hypertension is caused by both functional and structural capillary rarefaction. Hypertension 38: 238–242, 2001.
Capillaroscopy and Videocapillaroscopy Assessment of Skin Microcirculation
16. Bull RH, Bates DO, Mortimer PS. Intravital video-capillaroscopy for the study of the microcirculation in psoriasis. Br J Dermatol 126: 436–445, 1992. 17. Kelly RI, Pearse R, Bull RH, et al. The effects of aging on the cutaneous microvasculature. J Am Acad Dermatol 33: 749–756, 1995. 18. Fagrell B. Vital microscopy and the pathophysiology of deep venous insufficiency. Int Angiol 14: 18–22, 1995. 19. Franzeck UK, Bollinger A, Huch R, Huch A. Transcutaneous oxygen tension and capillary morphologic characteristics and density in patients with chronic venous incompetence. Circulation 70: 806–811, 1984. 20. Stucker M, Schobe MC, Hoffmann K, Schultz-Ehrenburg U. Cutaneous microcirculation in skin lesions associated with chronic venous insufficiency. Dermatol Surg 21: 877–882, 1995. 21. Fagrell B, Hermansson IL, Karlander SG, Ostergren J. Vital capillary microscopy for assessment of skin viability and microangiopathy in patients with diabetes mellitus. Acta Med Scand, Suppl. 687: 25–28, 1984. 22. Bollinger A, Fagrell B. Clinical Capillaroscopy. Hogrefe et Huber, Toronto, 1990. 23. Li L, Mary S, Sainthillier JM, Degouy A, Gharbi T, De Lacharriere O, Humbert P. Changes of cutaneous microcirculation in the different anatomic sites with aging in women. China J Microcirc 2: 43–45, 2004. 24. Toyoda M, Nakamura M, Luo Y, et al. Ultrastructural characterization of microvasculature in photoaging. J Dermatol Sci 27: S32–S41, 2001. 25. Degouy A, Creidi P, Sainthillier JM, et al. In vivo transcutaneous capillaroscopy: assessment of dermal capillary density decrease after topical pharmacological agent applications. Ann Dermatol Venereol 129: 414, 2002 (abstract).
687
26. Sainthillier JM, Creidi P, Degouy A, et al. Topical application of a manganese gluconate preparation inhibits the effects of neosynephrin on the cutaneous microcirculation. Ann Dermatol Venereol 129: 459, 2002 (abstract). 27. Sainthillier JM, Bonnans V, Degouy A, et al. Application des réseaux neuronaux pour le traitement et l’analyse des images en bio-ingénierie cutanée. In Actualités en Ingénierie Cutanée, Vol. 3, Perrenoud D, Gabard B, Eds. Eska, Paris, 2003, pp. 117–124. 28. Sainthillier JM, Gharbi T, Muret P, Humbert P. Pattern recognition of skin capillary network by means of neural algorithms. Skin Res Technol, in press. 29. Zhong J, Asker CL, Salerud EG. Imaging, image processing and pattern analysis of skin capillary ensembles. Skin Res Technol 6: 45–57, 2000. 30. Robert JM, Toussaint GT. Computational geometry and facility location. In Proceedings of the International Conference on Operations Research and Management Science, Manila, The Philippines, 1990, pp. B1–B19. 31. De Berg M, Kreweld M, Overmars M, Schwarzkopf O. Computational Geometry: Algorithms and Applications. Springer-Verlag, Berlin, 2000. 32. Fortune S. Voronoi diagrams and Delaunay triangulations. In Computing in Euclidean Geometry, Lecture Notes Series on Computing, Vol. 4, Ding-Zhu D, Hwang F, Eds. World Scientific, Singapore, 1995, pp. 225–265. 33. Fortune S. Sweepline algorithms for Voronoi diagrams. Algorithma 2: 153–174, 1987. 34. Sainthillier JM, Degouy A, Gharbi T, et al. Geometrical capillary network analysis. Skin Res Technol 9: 312–320, 2003. 35. Degouy A, Sainthillier JM, Mac-Mary S, et al. Evaluation de l’activité clinique et vidéocapillaroscopique d’une formulation cosmétique sur la microcirculation cutanée des couperoses légères à modérées. Nouv Dermatol 22: 554–556, 2003.
Blood Flow, Vasomotion, and Vascular Functions
Doppler Measurement of Skin 80 Laser Blood Flux: Variation and Validation Andreas J. Bircher Department of Dermatology, University Hospital, Basel, Switzerland
CONTENTS 80.1 Introduction............................................................................................................................................................691 80.2 Anatomical and Physiological Factors..................................................................................................................692 80.2.1 Skin Microvasculature ...............................................................................................................................692 80.2.2 Cutaneous Blood Flow ..............................................................................................................................692 80.3 Methodological Principles: Technical Aspects .....................................................................................................692 80.4 Factors Influencing Cutaneous Blood Flow..........................................................................................................693 80.4.1 Subject-Related Variables ..........................................................................................................................693 80.4.2 Instrument-Related Variables.....................................................................................................................693 80.5 Correlation with Other Methods ...........................................................................................................................694 80.5.1 133Xenon Clearance....................................................................................................................................694 80.5.2 Venous Occlusion Plethysmography .........................................................................................................694 80.5.3 Photoplethysmography ..............................................................................................................................695 80.5.4 Thermographic Methods............................................................................................................................695 80.5.5 Other Lasers...............................................................................................................................................695 80.6 Conclusions............................................................................................................................................................695 References .......................................................................................................................................................................695
80.1 INTRODUCTION The cutaneous microcirculation plays an outstanding role in physiologic processes and is also involved in many pathologic conditions. Therefore, there is considerable interest to objectivate changes in the cutaneous blood flow (CBF) in physiologic conditions and pathologic disorders as well as to pharmacological stimuli. Blood flow has been measured on virtually all animal and human organ surfaces, such as the nasal and gastrointestinal mucosa, in the eye, and on bone and muscular tissue. The skin, however, remains the most readily accessible organ, and therefore a vast amount of literature on numerous aspects of CBF has been published in the last years. Several methods to identify and quantitate CBF fluctuations have been developed. Among these are more direct methods such as photoplethysmography, venous occlusion plethysmography, and 133xenon (133Xe) clearance, as well as indirect techniques such as the determination of skin temperature and transcutaneous pO2. A newer noninvasive technology is laser Doppler velocimetry or flowmetry (LDF), which is based on optical princi-
ples and which allows a more direct measurement of the cutaneous microcirculation. LDF has some advantages over the other methods in that it allows noninvasive, continuous recording of CBF in relatively superficial skin layers on virtually any reachable skin or mucous membrane surface. The application of the Doppler phenomenon and the use of the unique properties of laser light to detect the motion of macromolecules were suggested by Cummins et al.21 and first used in a biological system by Riva et al.22 Later the approach was improved and applied to human retinal vessels by Tanaka et al.23 Stern et al.9 was the first to determine blood flow in the intact cutaneous microvasculature of humans.1 Several instruments that were usable also in clinical settings were then developed. The first, the laser Doppler velocimeter, was designed and later improved by Holloway and Watkins.2 Subsequently, a modified version, the LDF, was developed by Nilsson and colleagues.3 Both instruments use basically the same principles based on light-bearing spectroscopy with some technical variations to improve the yield of the measurements. Detailed descriptions of the theoretical aspects of 691
692
Handbook of Non-Invasive Methods and the Skin, Second Edition
LDF and the available instruments that have later been designed have been published.1
80.2 ANATOMICAL AND PHYSIOLOGICAL FACTORS 80.2.1 SKIN MICROVASCULATURE The human epidermis contains no vessels and is therefore nourished by diffusion from the dermis. Its thickness, which is dependent on the anatomical site, varies from 50 to 200 mm. In the dermis, supplied by vessels from the subcutaneous tissue, there is a complex vascular network to perform the particular tasks, such as thermoregulation, nutrition, and metabolism. In the upper dermis a superficial plexus is present with a special capillary network extending into the dermal papillae, the so-called capillary loops. They have a mean length of approximately 0.2 to 0.4 mm and each supplies an average of 0.04 to 0.27 mm2 to the skin surface. A deeper plexus is situated at the dermal-subcutaneous border. The microcirculatory blood flow is regulated by smooth muscle cells that are mainly located in the walls of the small arteries and arterioles and, to a lesser extent, in venules and small veins. The true capillaries are free of smooth muscle cells; however, the pericyte cells located in the capillary vessel walls may induce to some extent capillary contractions. Arteriovenous anastomoses are present in some locations, particularly in the face and in the acral areas. They play an outstanding role in the thermoregulation of the organism.4
80.2.2 CUTANEOUS BLOOD FLOW Blood flow or flux implies the movement of blood, which is a heterogeneic mixture of liquid, soluble, and cellular elements, through the preformed channels of the arterial and venous system, which is linked by the above-mentioned capillary network or by arteriovenous anastomoses. Blood flow is far from being constant; on the contrary, it is dynamically regulated to fulfill the requirements of the organism. Dependent on the measuring technique, its resolution, and the investigated tissue volume, any method of measurement expresses blood flow only in relation to the type and localization of the examined vessels. In optical methods, further variables that influence the measurements are the structure of the skin surface, the skin thickness, and the pigmentation of the epidermis. Due to the small measuring area and the anatomical architecture of the dermal microvasculature, considerable variations in blood flow are present.5 Therefore, the regional variations of CBF have a considerable magnitude and have to be taken into account when using such measuring methods. Most of the current laser Doppler devices use helium–neon lasers with a wavelength l = 632.8 nm and small probes. It is estimated that the tissue volume in
which CBF is measured with such devices has an approximate surface area of 1 mm2, extending to an estimated depth of 1 mm, resulting in a theoretical total measured tissue volume of 1 mm3. The penetration depth of laser radiation, however, is rather variable and dependent on the above-mentioned factors. The incident laser light is absorbed, scattered, and only to a small extent reflected by the skin tissue structures. Stationary tissue, such as fibers, macromolecules, and vessel walls, scatter and reflect the incident radiation at the same frequency. Typically the largest fraction of the diffusely reflected laser light waves comes from stationary tissues. Tissue components such as red blood cells moving with a mean estimated speed of 1 mm/sec reflect the light with a shifted frequency (optical Doppler effect). The blood flow signal measured by laser Doppler instruments is an indicator of the cutaneous perfusion; however, due to the complex structure of the skin texture and somewhat random orientation of the cutaneous vessels, only semiquantitative, relative measurements of CBF can be made6 with these techniques. Commonly the CBF results are given in millivolts or in arbitrary units. Laser Doppler measurements of CBF from normal relaxed subjects have three major characteristics (Figure 80.1): (A) pulsatile flow synchronized with the cardiac cycle, (B) vasomotor waves of lower frequency (approximately 4 to 6 per minute), and (C) skin blood flow, i.e., the deviation from the instrument baseline.7
80.3 METHODOLOGICAL PRINCIPLES: TECHNICAL ASPECTS In laser Doppler measurements, a laser light source and the Doppler effect are used to generate an output proportional to the red blood cell movement through the skin vessels under investigation. Prerequisites for this technique are the characteristics of laser radiation, such as mV 300
A
B
C
250 200 150 100 50 0
FIGURE 80.1 Schematic response pattern in LDF measured blood flow: (A) pulsatile flow synchronized with the cardiac cycle, (B) vasomotor waves of lower frequency, and (C) relative blood flow, i.e., the deviation from the instrument baseline. (Redrawn from Karanfilian, R. et al., Am. Surg., 50, 641, 1984. With permission.)
Laser Doppler Measurement of Skin Blood Flux: Variation and Validation
693
80.4 FACTORS INFLUENCING CUTANEOUS BLOOD FLOW 80.4.1 SUBJECT-RELATED VARIABLES
A
B
C
FIGURE 80.2 Schematic presentation of laser-light interaction with tissue components. Transmitted light is reflected and shifted in frequency by (A) moving erythrocytes or (B) reflected without frequency shift from stationary tissue or (C) does not reach the receiving probe fiber.
monochromaticity and spatial and temporal coherence. In typical laser Doppler instruments, a 2- or 5-mW-powered helium–neon laser (l = 632.8 nm) is usually the light source. Other instruments make use of an infrared laser (l = 780 nm). The laser radiation is guided through an optical fiber to the probe head. The probes are usually held in contact with the skin by a double-sided adhesive ringshaped tape. The emitted radiation enters the skin and is reflected by stationary and moving tissue components (Figure 80.2). Stationary tissue scatters and reflects the incident radiation at the same frequency. Red blood cells moving with a certain speed reflect the light with a shifted frequency, the Doppler effect. Thus, the light waves returning to the instrument are composed of two components: the frequency-modulated spectrally broadened light, which is directly related to the number of erythrocytes times their velocity, and the nonshifted fraction, which has been reflected from nonmoving tissue. The reflected light is transmitted through one or more receiving optical fibers, and the nonshifted reference and the Doppler-shifted beam are then mixed on a photodetector and processed by optical heterodyning. The generated beat frequency is then converted into an electrical output. The signal is usually measured as a fluctuating voltage, expressed in millivolts. This flow signal measured by laser Doppler instruments is an indicator of cutaneous perfusion; however, due to the complex structure of the skin texture and the somewhat random orientation of the cutaneous microcirculation, only semiquantitative relative measurements of CBF are obtained.1
Recently, a position paper on guidelines for LDF has been published by the standardization group of the European Contact Dermatitis Society.8 In this paper variables influencing CBF have been extensively reviewed. A broad spectrum of aspects may influence the measurements of CBF. Such variables include individual, interindividual, environmental, and technical factors. The latter variables are discussed in Bircher et al.8 Individualrelated variables include age, gender, and ethnic background. Selectable individual-related variables are the anatomical location of the measurements, the skin temperature, the subject’s position, physical and mental activities, previous consumption of food, beverages, drugs, and nicotine, the menstrual cycle, possibly some laboratory values, and to some extent the considerable temporal variation of repeated measurements. Environmental variables include air convection, ambient temperature, and probably air humidity (Table 80.1).
80.4.2 INSTRUMENT-RELATED VARIABLES In the above-mentioned position paper guidelines for LDF have also been published.8 Therefore, these guidelines will only be briefly summarized here. Due to the wide range of instruments available that work with several techniques and principles, a universal standard procedure cannot be developed. However, some general recommendations for validation can be given. The instrument should be used in accordance with the recommendations of the manufacturer, particularly with regard to instrument setup and safety precautions. The operating procedure of every laboratory should be clearly stated, and a standardized procedure that includes validation under the laboratory’s own specific conditions is proposed. Some recommendations concerning the adjustments of the zero levels are briefly outlined. The instrument zero is the value obtained when the probe is held against a white surface, e.g., white porcelain. It may slightly deviate from the electrical zero of the instrument. The biological zero, which is higher than the instrumental zero, is obtained in an anatomical site, e.g., the forearm, under arterial occlusion. In the same procedure the dynamic range of the blood flow can be measured. This basic standard reactive hyperemia experiment should be performed in at least three healthy adults. These recommendations should facilitate the comparison of measurements obtained at different research facilities with different laser Doppler instruments. Naturally, the individual and environmental variables described should be controlled and mentioned. Subjects should be managed with regard to smoking, food and drug
694
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 80.1 Individual and Environmental Variables Affecting Cutaneous Blood Flow Variable Age Gender Menstrual cycle Ethnic background Subject position Anatomical location Skin temperature Laboratory hematologic values Temporal (same day) Temporal (day to day) Physical activity Mental activity Food and beverages Systemic drugs Topical drugs Nicotine Ambient temperature
Influence on Blood Flow Mostly minor Minor or none Minor or none None or minor Major Considerable variation Major None or minor Minor Considerable Considerable Considerable Considerable Considerable Major Major Major
Remarks Dependent on location — — Reflection of pigmented skin Orthostatic dependence Also age related Also environment related Only when major pathology is present — — Short influence Short influence e.g., caffeine, alcohol Vasoactive compounds Vasoactive compounds Vasoconstriction —
Modified from Bircher, A.J. et al., Contact Dermatitis, 30, 65, 1994. With permission.
intake, and physical and mental stress. When small probes are used, several blood flow determinations, which may be averaged, are recommended. The study environment should be controlled with respect to temperature, air movements, and other factors. Instrumental factors to be controlled include warming up of the instrument and particular precautions with the optical fiber and the application pressure of the probe.
80.5 CORRELATION WITH OTHER METHODS 80.5.1 133
133 ENON
X
CLEARANCE
Xe clearance was compared to LDF in ultraviolet-stimulated CBF in forearm skin.2,9 In both studies 133Xe was delivered by injection. To minimize the spatial and temporal variation in LDF measurements, the average of 5 and 10 measuring sites, respectively, was determined. In these experiments a linear relationship (r = 0.9) between the two methods was found. In a study using an atraumatic delivery technique for 133Xe in four subjects, fingertip and finger web blood flow was examined.10 With both methods, parallel changes of CBF in skin without arteriovenous anastomoses were observed; however, in areas with shunt vessels, no relation was present. Also, in healthy skin and in uninvolved skin of psoriatic patients11 a poor correlation between LDF and xenon washout was observed. A better correlation was obtained in psoriatic lesions, leading to the conclusion that although LDF allows a rough estimate of CBF at high perfusion rates, it is not reliable in areas
with low blood flow. A comparison of LDF and 133Xe washout in skin lesions of localized and generalized morphea showed a good correlation between the methods. The sclerotic plaques and the perilesional inflammatory lilac rings had a significantly greater CBF than normal skin. Older plaques had higher CBF than early lesions. However, in the most advanced burnt out plaques, CBF was not different from that in normal skin.12 Very similar results were obtained in another study of scleroderma.13 Again, in ultraviolet B-induced erythema of the forearm a significant correlation between 133Xe washout and LDF measurements of CBF was found. Surprisingly, LDF measurements showed a fourfold greater increase over 133Xe determinations. The importance of the use of the biological zero instead of the instrumental electrical zero was also emphasized.
80.5.2 VENOUS OCCLUSION PLETHYSMOGRAPHY Venous occlusion or strain gauge plethysmography is an established method to determine total blood flow, particularly in extremities. In thermally stimulated CBF by direct heating of the whole body surface14 a good correlation between CBF measured by LDF and total forearm blood flow determined by occlusion plethysmography (r = 0.94 to 0.98) was found. However, the correlation varied considerably between and within subjects. A nonlinear relationship between occlusion plethysmography and LDFCBF measurements, stimulated by exercise and heat stress, was reported, suggesting that the two methods measure different blood flow parameters.15 In a comparison of
Laser Doppler Measurement of Skin Blood Flux: Variation and Validation
LDF, occlusion plethysmography, and thermal clearance, considerable variability in the results was found, which was explained on the basis of the different skin volumes and types of blood flow measured by the methods.16
80.5.3 PHOTOPLETHYSMOGRAPHY The resting blood flow levels as a function of the anatomic position were determined in 52 body sites in 10 subjects17 by photoplethysmography and LDF. Photoplethysmographic and LDF measurements agreed well in the areas with low CBF, but differed considerably in highly perfused areas. This was interpreted to be the consequence of the different parameters measured by the two methods (blood volume in photoplethysmography, volume times velocity in LDF).
80.5.4 THERMOGRAPHIC METHODS Several thermal methods that have been compared to LDF were found to give acceptable correlations. Thermal measurements, however, were usually slower in the speed of the response and more dependent on ambient temperature.18 Also, upon an intravenous injection of naftidrofuryl, similar blood flow values, as measured by thermal conductivity and LDF, were observed. The skin temperature also decreased over time, however, with a temporally delayed response.19
80.5.5 OTHER LASERS In a study investigating histamine-induced CBF a helium–neon (l = 633 nm) and an infrared (l = 780 nm) laser were compared.20 Very similar results of CBF changes were found with the two lasers, although the infrared laser radiation has a greater penetration. This implies that probably only superficial vessels were mainly affected by the histamine effect.
80.6 CONCLUSIONS Laser Doppler flowmetry requires a sophisticated technology; however, the instrument’s handling is simple. It allows noninvasive measurements of skin blood flow on virtually any area of the body and provides a continuous recording of the actual, mostly superficial, blood flow. It is particularly suited to measure and follow stimulated blood flow. Major fields of interest and application have been the determination of the qualitative and quantitative effects of systemically and topically applied vasoactive drugs, the study of inflammatory mediators, the investigation of allergic and irritant skin reactions, particularly in combination with other noninvasive bioengineering methods, and the evaluation of vascular phenomena in skin and other diseases, and also effects of therapeutic interventions on skin disorders.
695
Because the instrument is easy to use, some precautions should be taken when performing measurements. One major limitation is the rather small measured tissue volume, which may be overcome by future technical improvements, e.g., of scanning probes, as described in the next chapter. A variety of factors, including the instrument, the subject, and the environment, influence the measurements. Because of the high variability of skin blood flow, such as inter- and intraindividual subject variability and environmental parameters, these factors have to be taken into consideration in the planning and realization of experiments.
REFERENCES 1. Shepherd, A. and Öberg, P., Laser-Doppler Blood Flowmetry, Kluwer Academic Publishers, Boston, 1990. 2. Holloway, G. and Watkins, D., Laser Doppler measurement of cutaneous blood flow, J. Invest. Dermatol., 69, 306, 1977. 3. Nilsson, G., Tenland, T., and Öberg, P., A new instrument for continuous measurement of tissue blood flow by light bearing spectroscopy, IEEE Trans. Biomed. Eng., 27, 12, 1980. 4. Tenland, T., On Laser Doppler Flowmetry. Thesis, Linköping University, Linköping, Sweden, 1982. 5. Braverman, I., Keh, A., and Goldminz, D., Correlation of laser Doppler wave patterns with underlying microvascular anatomy, J. Invest. Dermatol., 95, 283, 1990. 6. Holloway, G., Laser Doppler measurement of cutaneous blood flow, in Non-invasive Physiological Measurements, Rolfe, P., Ed., Academic Press, New York, 1983, p. 219. 7. Karanfilian, R., Lynch, T., Lee, B., Long, J., and Hobson, I.R., The assessment of skin blood flow in peripheral vascular disease by laser Doppler velocimetry, Am. Surg., 50, 641, 1984. 8. Bircher, A.J., de Boer, E., Agner, T., Wahlberg, J., and Serup, J., Guidelines for the measurement of cutaneous blood flow by laser Doppler flowmetry, Contact Derm., in press. 9. Stern, M., Lappe, D., Bowen, P., Chimosky, J., Holloway, G., Keiser, H., and Bowman, R., Continuous measurement of tissue blood flow by laser Doppler spectroscopy, Am. J. Physiol., 232, H441, 1977. 10. Engelhart, M. and Kristensen, J., Evaluation of cutaneous blood flow responses by 133 xenon washout and laser Doppler flowmeter, J. Invest. Dermatol., 80, 12, 1983. 11. Klemp, P. and Staberg, B., The effect of antipsoriatic treatment on cutaneous blood flow in psoriasis measured by 133xenon washout method and laser Doppler velocimetry, J. Invest. Dermatol., 85, 259, 1985. 12. Serup, J. and Kristensen, J., Blood flow of morphoea plaques as measured by laser Doppler flowmetry, Arch. Dermatol. Res., 276, 322, 1984.
696
Handbook of Non-Invasive Methods and the Skin, Second Edition
13. De Lacharrière, O. and Kalis, B., Measurement of cutaneous microcirculation in dermatology and dermatopharmacology, in Cutaneous Investigation in Health and Disease, Leveque, J., Ed., Marcel Dekker, New York, 1989, pp. 385–420. 14. Johnson, J., Taylor, W., Shepherd, A., and Park, M., Laser Doppler measurement of skin blood flow, comparison with plethysmography, J. Appl. Physiol., 56, 798, 1984. 15. Smolander, J. and Kolari, P., Laser Doppler and plethysmographic skin blood flow during exercise and acute heat stress in the sauna, Eur. J. Appl. Physiol., 54, 371, 1985. 16. Saumet, J., Dittmar, A., and Leftheriotis, G., Non-invasive measurement of skin blood flow: comparison between plethysmography, laser-Doppler flowmeter and heat thermal clearance method, Int. J. Microcirc. Clin. Exp., 5, 73, 1986.
17. Tur, E., Tur, M., Maibach, H.I., and Guy, R.H., Basal perfusion of the cutaneous microcirculation: measurements as a function of anatomic position, J. Invest. Dermatol., 81, 442, 1983. 18. Johnson, J., The cutaneous circulation, in Laser-Doppler Blood Flowmetry, Shepherd, A. and Öberg, P., Eds., Kluwer Academic Publishers, Boston, 1990, pp. 121–139. 19. Dittmar, A., Skin thermal conductivity, in Cutaneous Investigation in Health and Disease, Leveque, J., Ed., Marcel Dekker, New York, 1989, pp. 323–358. 20. Coulsen, M., Hayes, N., and Foreman, J., Comparison of infrared and helium-neon lasers in the measurement of blood flow in human skin by the laser Doppler technique, Skin Pharmacol., 5, 81, 1992. 21. Cummins et al. 22. Riva et al. 23. Tanaka et al.
of Periodic Fluctuations 81 Examination in Cutaneous Blood Flow Robert Gniadecki, Monika Gniadecka, and Jørgen Serup Department of Dermatology, Bispebjerg Hospital, Copenhagen, Denmark
CONTENTS 81.1 Physiology of Vasomotion.....................................................................................................................................697 81.2 Methods of Detection and Analysis of Vasomotion .............................................................................................698 81.2.1 Intravital Microscopy.................................................................................................................................698 81.2.2 Laser Doppler Technique ..........................................................................................................................699 81.2.3 Analysis of Waveform Patterns .................................................................................................................700 81.2.3.1 Fast Fourier Transform...............................................................................................................700 81.2.3.2 Prony Spectral Line Estimation .................................................................................................702 81.2.3.3 Autoregressive Modeling ...........................................................................................................702 81.3 Vasomotion in Pathologic Conditions ...................................................................................................................702 81.3.1 Increased Venous Pressure, Chronic Venous Insufficiency, and the Postthrombotic Syndrome .............702 81.3.2 Arterial Insufficiency .................................................................................................................................703 81.3.3 Sickle Cell Disease ....................................................................................................................................703 References .......................................................................................................................................................................704
81.1 PHYSIOLOGY OF VASOMOTION Small arteries and arterioles in the skin and the subcutaneous tissue exhibit vasomotion, i.e., rhythmic changes of vessel diameter due to a series of contractions and relaxations. The principal features of vasomotion have been described by Nicoll and Webb,1,2 who observed rhythmic changes of diameter of small vessels in bat wing in vivo. These authors reported that vasomotion induced variations of blood blow in the vessels (flow motion). Since that time vasomotion has been directly observed in various animal and human tissues: in the skin (hamster skinfold window preparation3 and cheek pouch4), nail-fold capillaries in man,5 skeletal muscle,6 testicle,7,8 conjunctiva,9 retina,10 brain,11 and heart.12 The origin of vasomotion has been studied mainly in the rabbit tenuissimus muscle model.6 It has been found that vasomotion is elicited by the rhythmic activity of smooth muscle pacemaker cells in which the spontaneous depolarization occurs.13–15 These cells are located in cushion-like thickening of vessel wall near the branching points and are supposed to provide the original trigger for vasomotion, which is eventually propagated downward to the larger arterioles (Figure 81.1).16–18 The frequency of vasomotion changes abruptly at bifurcation points and
gradually increases in the downstream direction, from 0.5 to 18 cycles per minute (cpm; 60 cpm = 1 Hz) in the larger arteries to 9 to 21 cpm in terminal arterioles. This downstream propagation of contractions and dilations causes superposition of waves, and the final vasomotion pattern in the distal elements of the vascular tree is the composite effect of signals that originate at various branching points in the microvasculature.18 The major role of arteriolar vasomotion and cyclic oscillations of blood flow in the microcirculation is probably the enhancement of blood passage in the capillaries. Poiseuille’s law, which describes flow of the fluid in the cylinder-like vessel, states that conductivity of the vessel is proportional to the fourth power of its radius. Therefore, the vessel of the oscillating diameter will be much more conductive than the vessel of the same mean but constant diameter.19,20 Additionally, vasomotion plays an important role in the control of vascular resistance. Secomb et al.19 and Secomb and Gross21 predicted that the increase of vascular resistance to four times the initial value would be impossible without the participation of vasomotion, because in this instance the vascular diameter must have been controlled within very tight and unrealistic limits, and the diameter of the vessel would be smaller than the critical minimum diameter for passage of red blood cells. 697
698
Handbook of Non-Invasive Methods and the Skin, Second Edition
Epidermis
1 min
Capillaries
Fluxmotion
Superficial arteriolar plexus
Vasomotion 5–20 cpm
Arterioles Arterioles
Vasomotion 2–5 cpm
Bifurcational thickening (smooth muscle cells) vasomotion pacemaker
FIGURE 81.1 Hypothetical origin of vasomotion in human cutaneous microcirculation. (Based on References 13–18, 75.) Vasomotion is triggered by pacemaker smooth muscle cells located at arterial bifurcations. The dominant vasomotor frequency (5–10 cpm) is generated at the origin of ascending arterioles, whereas vasomotion in the larger arteries has lower frequency. The complex pulsating flow in the superficial arteriolar plexus and the capillaries is a result of superposing of rhythms from several pacemakers.
Therefore, vasomotion enables red blood cell transfer in the conditions of increased vascular resistance. Other functions of vasomotion are listed in Table 81.1.
81.2 METHODS OF DETECTION AND ANALYSIS OF VASOMOTION The ideal method for the examination of cutaneous vasomotion would be the direct observation of the changes of the diameter of the arterioles in the skin. Since this cannot be easily accomplished in humans, methods have been developed that detect temporal variations of blood flow
TABLE 81.1 Role of Vasomotion in Microcirculation Function Enhancement of vessel conductivity Control of vascular resistance Improvement of oxygen delivery to the issues Maintenance of blood fluidity Removal of tissue edema Stimulation of lymph flow Detachment of cells adherent to endothelium
Ref. 19 21 19 91 92, 100 93, 101 90
(flow motion) or red blood cell flux* (flux motion) that appear secondarily to arteriolar vasomotion. Flow motion in skin capillaries can be analyzed with dynamic capillaroscopy, whereas the laser Doppler technique is used for the analysis of flux motion in the microvasculature.
81.2.1 INTRAVITAL MICROSCOPY Intravital dynamic capillaroscopy is based on obtaining a magnified image of skin capillaries and dynamic recording of red blood cell velocity in these vessels. The most often used system has been developed by Bollinger et al.5 and Butti et al.22 and further improved by Fagrell et al.23,24 In this system vessels are visualized with an in vivo microscope and the image is recorded with the aid of a closedcircuit television.25 Nail-fold capillaries (in the finger or toe) have been commonly investigated, but some studies were also done in a titanium chamber system.26 The alterations of the velocity of red cells are analyzed with video densitometric techniques with correlation of the photometric signals.27 Wayland and Johnson28 developed a dualslit method that is based upon the measurement of the delay between the two signals elicited by the same configuration of red blood cells as they pass two photosensors separated by a fixed distance. More recently, Intaglietta et * Flux is the product of the red blood cell concentration and speed.
Examination of Periodic Fluctuations in Cutaneous Blood Flow
al.29 described a video dual-window technique for measuring of blood velocity. The video signal passes through two independent square areas (windows) fitted in size to the capillary width. The velocity of blood is determined by measuring the intrawindow transit time. Simultaneously with blood velocity other parameters may be recorded, such as arterial pulsations in the finger, respirations, ECG, etc.23 However, the real-time online determination of blood cell velocity with these systems is complicated, especially when high velocities are to be measured. Slaaf et al.30 proposed an easy-to-operate system based on a three-stage prism grating technique. A microscopic image of a microvessel is projected on a grating of alternating transparent and opaque lines, and the light that passes through the grating is focused on a photosensor via the transfer lens. Moving erythrocytes modulate the intensity of light that is recorded by the photosensor. This allows the online determination of the velocity and the direction of flow in the capillary. Direct observation of blood flow velocity changes in human nail-fold capillaries revealed spontaneous fluctuations, in most cases at the frequency of 6 to 10 cpm, that were not related to normal respiration.5,23,31 These fluctuations of flow velocity are considered to be a result of arteriolar vasomotion31–33 and are correlated with changes of pressure of blood in a capillary, increased pressure being associated with the most rapid speed of flow.34 The principal drawback of examining vasomotion by direct observation of blood vessel is that capillary microscopy is restricted to nail-fold capillaries. Examination of vasomotion in other regions requires implantation of a titanium chamber and cannot be considered fully noninvasive.
81.2.2 LASER DOPPLER TECHNIQUE The laser Doppler method has recently become an attractive alternative to intravital dynamic capillaroscopy for studying cutaneous vasomotion. The laser Doppler technique enables simple, real-time monitoring of relative changes of red blood cell flux in the cutaneous microvascular bed in any region of the body. The pioneering laser Doppler studies of Holloway and Watkins,35 Tenland et al.,36 and Salerud et al.37 demonstrated spontaneous oscillations of blood flux (flux motion) in normal human skin in resting conditions. The average frequency of the oscillations recorded by these authors was 8.6 cpm, and this value was comparable with rhythmic variations of capillary blood pressure,34,38 capillary blood velocity,5,39,40 and vasomotion.3 Oscillations of flux could not be suppressed by a proximal nervous blockade, implying a local, myogenic mechanism. It therefore has been concluded that the flux motion recorded with the laser Doppler technique reflects changes of microcirculatory blood flow due to arteriolar vasomotion.37
699
It is not fully understood what actually generates the laser Doppler signal in the skin,32,41–44 and there is much debate to what extent flux motion recorded with laser Doppler technique reflects flow motion in the microvasculature due to arteriolar vasomotion. Tooke et al.43 compared the laser Doppler method with video dynamic capillaroscopy and found that with both techniques it was possible to record oscillations of 4 to 6 cpm. When the Pearson product momentum correlation coefficient, which predicts the relationship between two sets of paired data, was calculated, a linear relationship between the laser Doppler signal and red blood cell velocity was seen in 11 of 14 sets of records. In 8 of the 14 recordings the best correlation of flux motion and flow motion was found at no time delay. Spectral analysis showed that in 14 of 16 recordings the flow motion pattern had similar frequency spectra to the flux motion. The amplitudes of flow oscillations were significantly higher than the amplitudes of flux motion. Therefore, although laser Doppler measures blood flux not only in the nutritive capillaries, but also in deeper elements of the skin vascular tree,42–45 this is a useful tool for evaluation of cutaneous vasomotion. The issue of the correlation of the origin of laser Doppler signal with the microvascular segments of human skin has been further investigated by Braverman et al.46 They found that with a commonly used laser Doppler probe the maximum amplitude of rhythmic oscillations could be obtained when the probe was placed directly over an ascending elastic arteriole and its immediate branches. On the other hand, when the probe was moved to the site between ascending arterioles, the nonpulsatile laser Doppler signal of the flow flux value was recorded. The dependency of the laser Doppler flux pattern on the position of the probe on the skin may explain low reproducibility in the detection of flux motion among different authors. In some studies cyclic changes of cutaneous flux could not be detected in normal conditions in humans.47,48 Because of these difficulties, methods for amplification of vasomotion have been devised. Wilkin47 observed that cutaneous flux motion is provoked in the early phase of postocclusive reactive hyperemia. The hyperemia response has been obtained after occlusion of the brachial artery with a sphygmomanometer cuff for 6 min. After release of the occlusion, a prompt increase of flux is seen and blood flux oscillations can be recorded from the forearm skin during the return of blood flux to normal values (Figure 81.2). The average period of these oscillations is 9.6 ± 0.3 s (SE), in accordance with the reported rhythmic variations of blood flow observed microscopically in human nail-fold capillaries23 and flux motion in the forehead37 and leg49 obtained in the normal resting conditions. It is conceivable that amplified laser Doppler flux oscillations are due to the local synchronicity of oscillatory blood flow in a group of cutaneous capillaries in the period of hyperemia. The synchronicity is
700
Handbook of Non-Invasive Methods and the Skin, Second Edition
a
b
c
TABLE 81.2 Modulation of Oscillatory Blood Flow in Cutaneous Circulation Oscillatory Blood Flow Amplitude
Frequency
Dependent position (increased venous pressure) Post-thrombotic syndrome
↓ ↓
0 ↓
↓ ↓
Arterial insufficiency Sickle cell disease Diabetes mellitus General anesthesia Postocclusive reactive hyperemia Topical application of: Propionaldehyde Nitroglycerine Carbon dioxide Thermal challenge Hyperventillation Smoking Hypertension Aging
0 ↑ ↓ ↓ ↑
↓ ND ↑ ↑ 0 ↓ 0 ND
70a, 66, 67 68 69, 71 56, 73 48, 77 94b, 95c, 96 51, 97 47, 48, 51
↑ ↑ ↑ ↑ ↑ 0 ↑ ↓
ND ND ND 0 0 ↓ 0 0
50 51 53 52 54 26 98d 99
Flux
Condition
Time
FIGURE 81.2 Enhancement of cutaneous blood flow oscillations in the early phase of postocclusive hyperemia. (a) Baseline flux, (b) arterial occlusion, and (c) reactive hyperemia. Note prominent oscillations of blood flux on the descending arm of hyperemic response (c).
limited to a small area of the skin. Loss of synchronicity has been reported in the sites only 0.5 to 2 cm apart, implying a local origin of the oscillations.37,43,47 Besides postocclusive hyperemia, cutaneous flux oscillations can be induced by other stimuli that provoke a local increase in skin blood flow (Table 81.2). Wilkin50 applied topically a 5 M aqueous solution of propionaldehyde and reproducibly a tenfold increase of cutaneous blood flux. During the recovery to the resting flux level marked oscillations of flux could be detected. A similar hyperemic response and augmentation of flux motion in the recovery phase could be induced with a topical 2.2 × 10–2 M nitroglycerine.51 Kastrup et al.52 reported that during local skin heating to 42˚C flux motion is induced in the posthyperemic phase. The oscillations, with a mean frequency of 6.9 cpm (range, 5.2 to 10.4) were not suppressed during local or central nervous blockade, with lidocaine or trimethaphan campsylate, respectively. Some authors were also able to induce vasomotion by a local application of carbon dioxide53 and during hyperventilation.54
81.2.3 ANALYSIS
OF
WAVEFORM PATTERNS
Cyclic oscillations in the blood flow or flux tracings may be analyzed by manual determination of wave frequency, amplitude, and prevalence. It has been realized that the frequency and amplitude of flux motion waves are heterogeneous and may be further subdivided into discrete components. Kastrup et al.52 divided flux motion into two constituents: high-frequency regular oscillations (mean, 6.8 cpm) of the nonneurogenic (probably myogenic) origin and irregular low-frequency oscillations (mean, 1.5 cpm) that were caused by periodic changes in the autonomic tone. Similarly, Scheffler and Rieger,55 Seifert et al.,56 and Bongard and Fagrell57 found two categories of
Ref. 49 57
Note: ↓-decreased; ↑-increased; 0-no change; ND-not done. a
In legs with edema, vasomotion restored after edema removal. Based on the thermal clearance method (detection of arteriovenous anastomotic blood flow). c Based on venous occlusion plethysmography. d Studies in experimental animals. b
oscillations: large waves of large amplitude and low frequency (10 cpm) that are sometimes seen in PSD may reflect either vasomotor
40
20
0 6.24
6.32
6.40
6.48
6.56
7.4
7.12
7.20
7.28
7.36
7.44
7.52
8.0
8.8
(a)
2 × 102
1 2
3
1 × 102
0 × 102 0.000
0.312
0.625
0.938
1.2
(b)
FIGURE 81.3 Fast Fourier transform (FFT) analysis of cutaneous flux-motion in humans. (a) Laser Doppler recording taken from the lower extremity (horizontal position), x axis = time (min:sec), y axis = units of flux. (b) FFT analysis — PSD graph. x axis = frequency (Hz), y axis = power of harmonic components. Note two distinct vasomotion-specific peaks at 2.34 (1), and 4.68 (2) cpm, in addition to a smaller peak at 61 cpm (3), characteristic for pulse-related flux changes in the skin.
702
Handbook of Non-Invasive Methods and the Skin, Second Edition
activity or, alternatively, the ventilatory effect on venous return, respiratory sinus arrhythmia, or periodic blood pressure waves — all were shown to reside in the 9.6- to 15-cpm frequency band.61 81.2.3.2 Prony Spectral Line Estimation The PSLE technique has been proposed for analysis of vasomotion by Colantuoni et al.18 and Meyer and Intaglietta.62 The basic algorithm has been devised by Prony63 and later by Burkhardt.64 A finite block of time-domain data is modeled as the sum of finite numbers of nonharmonically related sinusoids in white (uncorrelated) noise, and the original waveform pattern is reconstructed by the iterative procedure. A first order of size approximations contains one sinusoid, and the correlation coefficient is computed at each increment of order size. If the correlation coefficient is larger than the previous order, then this solution is kept and the previous is discarded. Thus, a maximal correlation will be seen at a specific order.65 If the correlation is greater than the predetermined level (usually the correlation coefficient >0.8 to 0.95), a solution has been found. In certain circumstances PSLE is more powerful than FFT, because the resolution is not dependent upon the data length.58 The drawback of the PSLE method is computational complexity and the necessity to determine the order.58 Moreover, the PSLE graph may be difficult to interpret for an uninitiated individual due to a relatively large number of spectral lines. 81.2.3.3 Autoregressive Modeling Autoregressive PSD estimation employs the autocorrelation method to find the repetitive components in the input data.58 Autoregression gives good results for strong sinusoidal components; however, the power of amplitude of the components cannot be calculated.58 Other limitations of autoregression methods involve the degrading effect of observation noise, the presence of multiple spurious peaks, and the shifting of main frequencies toward higher values at low signal-to-noise ratio.58 The resolution of autoregressive modeling is often better than that of FFT. Moreover, this procedure is not dependent on the number of samples, and the results are not affected by windowing.
81.3 VASOMOTION IN PATHOLOGIC CONDITIONS Vasomotion is easily modulated in a variety of physiologic and pathologic conditions (Table 81.2). Changes of vasomotion in arterial insufficiency, venous hypertension, and sickle cell disease were investigated in some detail.
81.3.1 INCREASED VENOUS PRESSURE, CHRONIC VENOUS INSUFFICIENCY, AND THE POSTTHROMBOTIC SYNDROME The influence of the increased venous pressure in the lower extremity, caused either experimentally by leg lowering or by the pathologic process of venous thrombosis and valve injury (a postthrombotic syndrome), has been recently investigated by several groups with laser Doppler flowmetry.49,57,66–69 Increased leg venous pressure led to postural vasoconstriction and decrease of amplitude of vasomotion49,57 (Figure 81.4). It is not clear whether increased venous pressure causes changes of the frequency of vasomotion. Bongard and Fagrell57 found a decrease of vasomotion frequency from 3.5 to 2.1 cpm, whereas the Fourier spectral analysis performed by Gniadecki et al.49 did not reveal significant changes of the frequency of the dominant harmonic component of the laser Doppler flux signal. The postural attenuation of vasomotion could be suppressed by the application of leg compression,49 a finding providing further proof that increased venous pressure and venous distention are responsible for the suppression of vasomotion. The vasomotion patterns in patients with chronic venous insufficiency are not consistent. Gniadecka et al.70 reported that in these patients vasomotion could not be detected in the region of lipodermatosclerosis either in the horizontal or in the dependent position of the lower extremity. Removal of the leg edema, which coexists with chronic venous insufficiency, with long-term (4 weeks) compressive therapy restored normal cutaneous vasomotion. Therefore, it is likely that limb edema itself inhibits vasomotion in lipodermatosclerosis. Similar findings were reported by Pekanmäki et al.66 These authors were able to restore vasomotion (5 cpm; range, 2 to 8 cpm) in the patients with lipodermatosclerosis by a single treatment with intermittent pneumatic compression (Ventipress®, Lemi, Finland). However, in this study, the position of the patients during laser Doppler measurement has not been reported and leg edema has not been assessed. Therefore, it is difficult to conclude whether the restoration of vasomotion due to intermittent pneumatic compression of the leg could be attributed to the removal of leg edema or to other mechanisms. Decrease of the amplitude of blood flux oscillations in the postthrombotic syndrome has also been observed by Belcaro et al.68 In a marked contrast, Chittenden et al.69 and Cheatle et al.71 reported an enhancement of vasomotion in patients with chronic venous insufficiency in the areas of lipodermatosclerosis. These authors recorded statistically significant higher baseline laser Doppler flux, higher vasomotion frequency (3.2 vs. 1.92 cpm in the control), and higher vasomotion amplitude (8.5 vs. 1.2 mV, control) in the lipodermatosclerosis. Increased frequency and amplitude of vasomotion were unlikely to
Examination of Periodic Fluctuations in Cutaneous Blood Flow
induced by stiff sickle erythrocytes blocking blood flow in the microvessels (see Sections 81.3.2 and 81.3.3).
40 H
D
81.3.2 ARTERIAL INSUFFICIENCY
20 F1 F2 0 0.0
703
2.40
5.20
8.0
10.40
13.20
16.0
18.40 21
(a) 6 × 102
4 × 102
2 × 102
0 × 102 0.000
0.312
0.625
0.
(b) 3 × 102
2 × 102
1 × 102
0 × 102 0.000
0.312
0.625
0.
(c)
FIGURE 81.4 Effects of the increased venous pressure, caused by leg lowering, on cutaneous blood flux oscillations: (a) original laser Doppler recordings (H = horizontal position, D = dependent position), (b) power spectral densities (FFT) of blood flux in the horizontal position, and (c) in the dependent position. Note the marked decrease of power amplitudes in (c).
be attributed to high flux because increasing of flux with pilocarpine did not induce changes of vasomotion. These results are difficult to compare with the studies showing decreased vasomotion in chronic venous insufficiency, because the authors did not assess leg edema in their patients. In the absence of edema, vasomotion in lipodermatosclerosis and chronic venous insufficiency could be enhanced secondarily, due to capillary plugging by white blood cells,72 a mechanism similar to that seen in sickle cell disease, where cutaneous vasomotion is
The effects of leg ischemia due to peripheral atherosclerosis on vasomotion were studied by Yanar et al.73 and Seifert et al.56 Cutaneous vasomotion in the second and third toe were recorded with laser Doppler in four groups of patients with different stages of the disease (Table 81.3). The authors found increased prevalence of the high-frequency vasomotion laser Doppler waves proportional to the severity of the peripheral ischemia (Table 81.3). Increased vasomotion frequency was directly linked to peripheral hypoperfusion since the prevalence of highfrequency waves decreased after successful restoration of peripheral circulation by means of angioplasty or thrombolysis.74 The pathophysiologic basis of the changes of vasomotion in peripheral ischemia was studied by Bertuglia et al.75 in a hamster skinfold window model. These authors divided skin microvasculature according to Strahler classification,75,76 so that order 0 was capillaries and order 4 the largest skin arterioles. In the baseline situation the dominant vasomotion frequency was found in order 1 arterioles (4 to 18 cpm; mean, 9.1 ± 3.9 cpm) and decreased with the increasing arteriole order, a finding in accordance with earlier observations of Colantuoni et al. 18 However, during experimentally evoked tissue hypoxia caused by the inspiration of an 8 and 11% O2 mixture, the frequency of vasomotion increased in all arteriolar branches and the dominant frequency was generated in order 3 arterioles (25.5 ± 4.5 cpm at 8% O2 vs. 3.4 ± 1.8 cpm in the control). Increased frequency of vasomotion was accompanied by a decreased mean and effective vessel diameters and reduced capillary blood flow. It is probable that the laser Doppler findings of increased flux oscillation frequency during limb ischemia in humans could be explained by the increased activity of vasomotion pacemaker in larger cutaneous arteries. Such a phenomenon is likely to reduce resistance in the skin microcirculatory network and ensure adequate blood supply to the tissue.
81.3.3 SICKLE CELL DISEASE Cutaneous vasomotion in the patients with sickle cell disease has been investigated with laser Doppler fluxmetry by Rodgers et al.48 and Gniadecka et al.77 These authors found prominent oscillations of blood flux of the period 7 to 10 sec and peak-to-trough magnitudes about half the mean flow. The oscillations were apparently associated with the presence of the pathological hemoglobin S, since in two patients studied by Rodgers et al.48 blood transfusion resulted in the diminution or disappearance of the rhythmic variations in blood flux.
704
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 81.3 Fluxmotion in Patients with Peripheral Ischemia High Frequency Wavesa Degree of Limb Ischemia No (control) Light Moderate Severe a b
Ankle/Arm Pressure Ratioa 1.22 0.81 0.62 0.38
(0.2) (0.1) (0.2) (0.1)
Walking Distance
Frequency (Hz)
Amplitudeb
Prevelance (%)
— >200 m 89%), and the nutritional flow is a small component. New microprobes and new LDF flowmeters may reduce the measuring depth and selectively evaluate the most relevant superficial nutritional capillaries, but so far clinically relevant data are not available. Using 1-mm-thick plastic spacers with the same reflectance properties and light transmission characteristics of the skin, it is theoretically possible to evaluate the most superficial skin flux. However, the more distal LDF signal seems to contain a high proportion of noise components, and it appears to be weaker. Early results from our study in which skin spacers are used24 indicate that there is a significant important difference in skin flux values and in flux responses, i.e., venoarteriolar response, flow increase with local temperature increase, etc. This
FIGURE 82.3 The percentage of nutritional flow in the LDF tracing is possibly less than 10% as most of the signal component is due to the non-nutritional subpapillary flow.
difference is noted both with and without the spacers, but with a different, not correlatable extent in low-perfusion microangiopathy — peripheral vascular disease, hypertension associated with vasospasm, and Raynaud’s disease and phenomenon. However, in venous disease and diabetic microangiopathy (these conditions can be defined as highperfusion microangiopathy) the reflex responses and skin flux measurements are reduced with the spacers, but comparable and parallel to measurements obtained without them. This possibly indicates that the differentiation between the nutritional and the thermoregulatory component is more important in states of low-perfusion microangiopathy. In low-perfusion microangiopathy the increase in skin flux as measured by LDF, i.e., following administration of a vasoactive drug or superficial revascularization, may only reflect an increase in the thermoregulatory, nonnutritional flow. This often could be irrelevant in the evaluation of skin flow changes due to the treatment, and produce misleading conclusions in relation to the healing or development of skin necrosis.
82.6 VASOMOTION As LDF measures flux continuously, flux motion can also be easily evaluated. During LDF monitoring in normal conditions the skin vessels in the microcirculation fill with blood rhythmically as a consequence of pressure and flow changes due to cardiac action respiration and vasomotion. The LDF output therefore shows a continuous variation and the sample volume varies continuously. Flux motion patterns are markedly changed in patients with peripheral vascular disease, who frequently present a high-frequency flux motion component. The prevalence of high-frequency motion waves is significantly increased in low-perfusion states, and it is proportionally more evident with increased levels of ischemia. The relationship between the presence of high-frequency flux motion waves at the forefoot has been evaluated by Hoffman et al.11 before and after percutaneous transluminal angioplasty. In successfully treated patients, a significant decrease in high-frequency flux waves was observed after angioplasty. However, the
Laser Doppler Flowmetry: Principles of Technology and Clinical Applications
713
Scanner
Laser Processor Detector
Computer
Plotter
Averaging Profile Subtraction Single values
FIGURE 82.4 A diagram of the LDF scanning system.
prevalence of such waves after angioplasty indicated that high-frequency waves are related to severe chronic ischemia. Different patterns of vasomotor waves were detected by Hoffman et al.11 in different conditions of perfusion. Flux motion in normals was characterized by low-frequency and pulsatile flux waves. Occasionally, additional high-frequency wave components appeared in the recording. Flux motion patterns, in severe ischemia, showed almost no pulsatile flux waves, whereas highfrequency waves were frequently observed in more severe ischemia. In severe cases of ischemia, no flux motion was observed. By contrast, patients with intermittent claudication showed variable patterns (excluding the last pattern with no vasomotion). Therefore, using frequency analysis, it appears possible to qualitatively differentiate degrees of ischemia. However, while there is little doubt that the alteration of vasomotion in low- or high-perfusion states is clinically relevant to indicate microcirculatory disturbances, no definite clinically usable answer has been provided about the analysis of vasomotion. Sophisticated frequency analysis, not commercially available with the instrumentation but usable in postprocessing with the signal output, is needed to extract clinically meaningful data from the analysis and limits so far the application of the method. There is, however, little doubt that qualitative analysis of vasomotion may offer new and interesting concepts to evaluate microcirculatory disturbances.
82.7 NEW TECHNOLOGY LD imaging is a new method of evaluating skin perfusion (other tissue can be also studied). A diagram of the system is shown in Figure 82.4. The multiple skin sampling produces a color perfusion image of the tissue under evaluation and a perfusion profile (Figure 82.5). The color-coded perfusion scale clearly indicates absence of perfusion in the second finger from the bottom in the image due to occlusion with a rubber band (dark blue). The system has been developed very recently, and some clinical application of the system appears very promising.
82.8 CONCLUSIONS The use of LDF is progressively increasing in physiological, pharmacological, and practical clinical evaluation of vascular disease. The monitoring of the effects of treatment on the microcirculation appears to be one of the most promising fields of application of LDF. The technical development of LDF combined with extensive clinical research application and frequency analysis systems will make this method one of the future’s most interesting noninvasive fields of investigation in vascular disease. At this stage, there are some problems concerning standardization of the different systems and calibration problems. It would be useful to define a common standard so that measurements in different centers and different instruments could be comparable, if not universal.
714
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 82.5 An example of LDF scanning. The low flow in the second finger from below is due to occlusion with a rubber band and is indicated in the color-coded perfusion scan by the dark blue color.
REFERENCES 1. Almond NE, Jones DP, Bowcock SA, and Cooke ED. A laser Doppler blood flowmeter used to detect thermal entrainment in normal persons and patients with Raynaud’s phenomenon. In Practical Aspects of Skin Blood Flow Measurement, Spence VA, Sheldon CD, Eds. Biological Engineering Society, London, 1985, p. 31. 2. Almond NE. Laser Doppler flowmetry: instrumentation theory and practice. In: Belcaro G, Hoffman U, Nicolaides AN, Bollinger A, Eds. Med-Orion Publishing Co., in press. 3. Almond NE, Jones DP, and Cooke ED. Noninvasive measurement of the human peripheral circulation: relationship between laser Doppler flowmeter and photoplethysmograph signals from the finger. Angiology 39: 819, 1988. 4. Boggett D, Blond J, and Rolfe P. Laser Doppler measurement of blood flow in skin tissue. J Biomed Eng 7: 225, 1985. 5. Bonner RF and Nossal R. Model for laser Doppler measurements of blood flow in tissue. Appl Opt 20: 2097, 1981. 6. Bonner RF, Clem TR, Bowen PD, and Bowman RL. Laser Doppler continuous realtime monitor of pulsatile and mean blood flow in tissue microcirculation. In Scattering Techniques Applied to Supramolecular and Nonequilibrium Systems, Nato Advanced Study Institute, Series B, Physics, Chen S, Chu B, Nossal R, Eds. Plenum Press, New York, 1981, p. 685. 7. Borgos JA. TSI’s LDV blood flowmeter. In Laser Doppler Blood Flowmetry, Shepherd A, Oberg P, Eds. Kluwer Academic Publ., Boston, 1990, p. 73.
8. Engelhart M and Kristensen JK. Evaluation of cutaneous blood flow responses by 133Xenon washout and a laser Doppler flowmeter. J Invest Dermatol 80: 12, 1983. 9. Fagrell B. Problems using laser Doppler on the skin in clinical practice. In Belcaro G, Hoffman U, Nicolaides AN, Bollinger A, Eds. Med-Orion Publishing Co., in press. 10. Hellem S, Jacobsson LS, Nilsson GE, and Lewis DH. Measurement of microvascular blood flow in cancellous bone using laser Doppler flowmetry and 133Xe clearance. Int J Oral Surg 12: 165, 1983. 11. Hoffman U, Seifert H, Baider E, and Bollinger A. Skin blood flux in peripheral arterial occlusive disease. In Laser-Doppler Flowmetry Experimental and Clinical Applications, Belcaro G, Bollinger A, Franzeck U, Hoffman U, Nicolaides AN, Eds. Med-Orion Publishing Co., in press. 12. Holloway GA and Watkins DW. Laser Doppler measurement of cutaneous blood flow. J Invest Dermatol 69: 306, 1977. 13. Johnson JM, Taylor WF, Shepherd AP, and Park MK. Laser Doppler measurement of skin blood flow: comparison with plethysmography. J Appl Physiol Respirat Environ Exercise Physiol 56: 798, 1984. 14. Nilsson GE, Tenlan T, and Oberg PA. Evaluation of a laser Doppler flowmeter for measurement of tissue blood flow. IEEE Trans Biomed Eng 27: 597, 1980. 15. Obeid AN, Boggett DM, Barnett NJ, Dougherty G, and Rolfe P. Depth discrimination in laser Doppler skin blood flow measurement using different lasers. Med Biol Eng Comput 26: 415, 1988. 16. Obeid AN, Barnett NJ, Dougherty G, and Ward G. A critical review of laser Doppler flowmetry. J Med Eng Technol 14: 178, 1990.
Laser Doppler Flowmetry: Principles of Technology and Clinical Applications
17. Riva C, Ross B, and Benedek GB. Laser Doppler measurement of blood flow in capillary tubes and retinal arteries. Invest Ophthalmol 11: 936, 1972. 18. Stern MD. In vivo evaluation of microcirculation by coherent light scattering. Nature 254: 56, 1975. 19. Stern MD, Lappe DL, Bowen BD, Chimosky JE, Holloway GA, Keiser HR, and Bowman RL. Continuous measurement of tissue blood flow by laser Doppler spectroscopy. Am J Physiol 232: H441, 1977. 20. Tonneson KH and Pederson LJ. Laser Doppler flowmetry: problems with calibration. In Belcaro G, Hoffman U, Nicolaides AN, Bollinger A, Eds. Med-Orion Publishing Co., in press.
715
21. Wardell K, Jakobbson A, and Nilsson GE. A Laser Doppler Imager for Microcirculatory Studies. Paper presented at the 1st European Laser Doppler Users Meeting, Oxford, March 1991. 22. Weis GH, Nossal R, and Bonne RF. Statistics of penetration depth of photons re-emitted from irradiation tissue. J Mod Opt 36: 349, 1989. 23. Winsor T, Haumschild DJ, Winsor DW, Wang Y, and Luong TN. Clinical application of laser Doppler flowmetry for measurement of cutaneous circulation in health and disease. Angiology 10: 727, 1987. 24. Belcaro G and Nicolaides AN. Article in preparation.
83 Laser Doppler Imaging of Skin Karin Wårdell Department of Biomedical Engineering, Linköping University, Linköping, Sweden
CONTENTS 83.1 Introduction............................................................................................................................................................717 83.2 Object.....................................................................................................................................................................717 83.3 Methodological Principle ......................................................................................................................................718 83.3.1 Operating Principle and Instrumentation ..................................................................................................718 83.3.2 Data Analysis .............................................................................................................................................718 83.3.3 Design of a Measurement Procedure ........................................................................................................719 83.4 Sources of Error.....................................................................................................................................................720 83.4.1 Temporal Changes in Perfusion and Movement Artifacts........................................................................720 83.4.2 Distance, Angular, and Reflection Errors..................................................................................................720 83.5 Correlation with Other Methods ...........................................................................................................................721 Acknowledgment.............................................................................................................................................................721 References .......................................................................................................................................................................721
83.1 INTRODUCTION Skin blood flow is an important parameter to record and assess in a large number of clinical and experimental settings. Peripheral vascular disease often manifests itself as a disturbance in cutaneous microcirculation, while an early result of agents irritating the skin is an elevation in its perfusion. Measurement of skin blood flow, therefore, constitutes an important diagnostic procedure in the vascular laboratory as well as in the evaluation of consumer products and potential skin irritants. Methods for measuring skin blood flow should preferably be noninvasive, analytical, versatile, easy to use, and cost-effective. The noninvasiveness should, if possible, also be extended to imply that the measuring device does not have to be in physical contact with the tissue, because even the weakest external stimuli may disturb the flow conditions of the microvascular network under study. The methods should be analytical in the sense that the recorded blood flow signals may be stored for later analysis by means of dedicated software packages. Versatility implies that the same device should be applicable to studies of microvascular perfusion under clinical as well as laboratory conditions. Ease of use and cost-effectiveness are important features, especially in clinical situations where the measurements need to be done on a routine basis. Very few of the methods described in the literature over the years fulfill all these requirements. Laser Doppler
flowmetry, laser Doppler perfusion monitoring (LDPM),1 and laser Doppler perfusion imaging (LDPI)2 influence the microvascular network only to a minimal extent during measurements and are possible to apply to studies of many tissues of the body. Since these technologies are also easy to use and supported by analytical software packages, they have become increasingly important for studies of microcirculation both in the laboratory and in clinical settings.
83.2 OBJECT Skin blood flow generally possesses both substantial temporal and spatial variations. The temporal variations can be rhythmic in nature or show a more fluctuating and stochastic pattern.3 Depending on the architecture of the underlying microvascular network, skin blood flow shows a characteristic granular speckle pattern, and large variations in perfusion can be demonstrated even at adjacent skin sites.4 In addition, the blood flow in the skin is generally known to be compartmentalized. The superficial capillaries are perfused by slow-speed red cells that supply the tissue with oxygen and nutritive substances and remove waste metabolites. Deeper-lying arteriovenous anastomoses take an active part in body temperature regulation, while small arteries and veins constitute routes for the supply and drainage of blood. Taking all this into account, the ideal method for assessing skin perfusion should be able to capture both 717
718
Handbook of Non-Invasive Methods and the Skin, Second Edition
Scanner head Output signal
Laser beam Backscattered light
Signal processing
Computer
Light interaction with tissue Two-dimensional flow map
FIGURE 83.1 Operating principle of the LDPI system.
the temporal and spatial variations in skin blood flow as well as have the potential to separate the signals generated by the different compartments of the microvascular network. Conventional LDPM can readily track fast changes in perfusion, but the small measuring volume prevents assessment of the spatial variability in skin blood flow. In order to overcome this limitation, LDPI was developed in the late 1980s. With this method it is possible to create two-dimensional flow maps of a specific tissue and to visualize the spatial variation of its perfusion.
83.3 METHODOLOGICAL PRINCIPLE 83.3.1 OPERATING PRINCIPLE INSTRUMENTATION
AND
The laser Doppler perfusion imager is a data acquisition and analysis system that generates color-coded images of the tissue perfusion. Two different systems have been presented in the literature and are also available on the market today. One of these systems utilizes a continuously moving laser beam,5 and the other one a stepwise scanning beam.2 The latter will be referred to and described in more detail in the following sections. The optical scanner guides a low-power laser beam (1 mW, 670 nm) to the tissue surface. At each measurement point, the laser light interacts with the skin and its microvasculature. After interaction with the moving red blood cells, the light becomes spectrally broadened due to the Doppler effect. A fraction of this Doppler-broadened light is backscattered and detected by a photodetector positioned in the scanner head. The instantaneous light intensity is converted into an electrical signal and processed to form an output value proportional to the perfusion, defined as the product of the average blood cell speed
and concentration. A detailed description of the laser Doppler theory is found in Nilsson et al.6 For each measurement position, the perfusion value is stored in the computer memory for further signal processing, image generation, and data analysis. In addition to images, data can be captured in the duplex mode.7 Duplex LDPI makes possible the recording of time traces over a small measurement area, including up to 16 measurement sites. With this feature, a combination of spatial and temporal components of the microvascular blood flow can be studied. A schematic overview of the operating principle is presented in Figure 83.1.
83.3.2 DATA ANALYSIS During a measurement procedure, the total light intensity (TLI) and the Doppler components are sampled and stored separately. The TLI value is used by the system to automatically determine whether the light level is adequate enough for recording of a Doppler value and to display a photographic gray-scale image of the measurement area. In practice, this arrangement makes discrimination between the object and the background possible (assuming that a light-absorbing material is used as the background). During a measurement procedure each captured perfusion value is immediately color coded and an image is continuously generated on the monitor. To convert the set of perfusion values into more quantitative parameters, data analysis functions such as perfusion value profile generation, selection of regions of interest (ROIs) for statistical calculation, and image subtraction are incorporated as an integral part of the system. Figure 83.2 exemplifies this by a recording from a skin tumor.
Laser Doppler Imaging of Skin
719
2
FIGURE 83.2 LDPI recording from a skin tumor. Left perfusion image and right photographic image of the tumor. (a)
(b)
(c)
0.98–7.10 V
0–10 V
0–5 V
was applied. The same image is presented with three different fixed color scales, namely, the actual minimum and maximum values (0.89 to 7.10 V) of this image (A), the entire range of the system (0 to 10 V) (B), as well as the range 0 to 5 V (C). In D through F the corresponding interpolated images are displayed. Images recorded by the LDPI system may also be converted to standard formats such as TIFF and ASCII for further image analysis and statistical calculations in software packages of individual preference.
83.3.3 DESIGN (d)
(e)
(f )
FIGURE 83.3 The same image presented in three individual color ranges (0.89 to 7.10 V, 0 to 10 V, and 0 to 5 V) (A, B, and C, respectively). The corresponding interpolated images are presented below the corresponding color ranges (D to F).
Figure 83.3 demonstrates a recording performed on the dorsal side of the hand 15 min after a vasodilatating cream
OF A
MEASUREMENT PROCEDURE
Before a measurement is initiated it is advisable to establish a well-organized protocol. Such a protocol simplifies the analysis and evaluation, which often take place at several points in time following image capturing. The test subject should be sitting or laying down in a comfortable position during the entire measurement procedure. Since the ambient temperature is known to have a substantial influence on skin blood flow, it is recommended that this be recorded. When imaging of the skin perfusion of an
720
Handbook of Non-Invasive Methods and the Skin, Second Edition
extremity is to be performed, this extremity should preferably be attached to, e.g., a sandbag or an arm or leg rest with tape or Velcro® straps, in order to keep the tissue in the same position throughout the scanning procedure. Treat the skin area under investigation carefully, and avoid scratching or applying unintentional pressure or mechanical stimuli, since even the slightest stimulation may substantially influence the perfusion. In order to be able to differentiate the backscattered light as either tissue or background, the object should be placed on a light-absorbing material. To simplify the identification and orientation of the perfusion image, it is useful to mark at least two to three reference points on the skin. If the measurement situation allows, use either a black or green ink pen, dark paper, or tape that can be removed after the scanning procedure is completed. When the laser beam hits the dark background or markers, the light is absorbed instead of being backscattered to the detector. The markers will appear as nonperfusion pixels in the image. The spatial resolution of the captured image depends on the distance between the scanner head and the tissue surface as well as on the system parameter resolution, which can be set by the operator in the software. An image-capturing procedure starts by positioning the lower side of the scanner head in parallel with the tissue surface of interest. By keeping the scanner head parallel to the tissue surface, geometrical distortion of the image is reduced to a minimum. Some distortion, however, is inevitable if imaging is performed on a surface with sharp curvatures. When all parameters are set, the extension of the tissue area to be imaged may be marked by moving the beam along its boundaries. This procedure indicates the area to be imaged and allows the operator to adjust the scanner head or any of the parameters before a measurement is started. To reduce optical interference with the laser light to a minimum, direct ambient light should be avoided. Excessive ambient light levels may render the light-absorbing reference markers placed on the skin invisible in the image. Recordings of dark lesions or pigmented skin perfusion may require a reduced distance between the scanner head and tissue in order to increase the backscattered light level to above the threshold level for tissue–background discrimination. In summary, the most important steps in preparing a successful image of skin tissue are: 1. Use a well-designed protocol. 2. Position the test subject such that gross movement artifacts can be avoided. 3. Use light-absorbing material as the background. 4. Mark reference points on the skin surface. 5. Position the scanner head in parallel with the skin surface.
6. Mark a measurement area and adjust the scanner head if necessary. 7. Protect the scanner head from ambient light. Guidelines for visualization of cutaneous blood flow by laser Doppler imaging have been made available through the European Union 5th Framework Program, denoted Hirelado.8
83.4 SOURCES OF ERROR 83.4.1 TEMPORAL CHANGES IN PERFUSION MOVEMENT ARTIFACTS
AND
In LDPI it is assumed that the perfusion of the tissue to be imaged is stationary and does not show any temporal variations during the time required to capture an image. If this is not the case, temporal variations will manifest themselves as a false spatial heterogeneity in tissue perfusion. Rhythmical variations in tissue perfusion such as vasomotion or a temporary reduction in flow, due to, e.g., taking a deep breath, generally show up as isolated stripes of perfusion values different from the perfusion in adjacent spots. Isolated gross movements of the tissue will likewise appear as stripes of falsely elevated and sometimes saturated values. To avoid such artifacts, it is important that the test subject sit as still as possible during the scanning procedure, or in the case of animal experiments, to apply anesthesia to prevent muscle contraction and shivering. Rhythmical movements of the whole tissue caused by breathing may give rise to a periodic pattern in the image, which is partly caused by tissue movement artifacts and partly by respiratory-related changes in tissue perfusion. The respiratory tissue movement artifacts may be particularly substantial in small animals, and a careful selection of skin areas that are only minimally affected by these movements is recommended.
83.4.2 DISTANCE, ANGULAR, ERRORS
AND
REFLECTION
The system has been designed to be virtually independent of the distance between the skin surface and the scanner head. The dependence of the angle between the light beam incident on the detector surface and a line perpendicular to the detector surface for the calculated perfusion value is automatically compensated for by the system for each measurement site. Surfaces with a pronounced curvature, however, have a tendency to scatter away the light due to pure surface reflection from the object, at least at the image boundaries. A small fraction of the light beam is scattered directly on the surface of the skin due to a mismatch in the refractive indexes of air (n = 1) and tissue (n = 1.5). This surface reflection is, however, generally
Laser Doppler Imaging of Skin
limited to about 5% in normal skin. If the skin is covered by an optical semitransparent material, images of the underlying tissue can still be captured, but the sensitivity is generally reduced due to scattering and absorption in the material or reflections from the surface. Imaging of skin surfaces immersed in water, where the water has been thermostated to a well-defined temperature, can be performed. A geometrical distortion due to the differences in the refractive indexes of water and air is, however, inevitable. Care must also be taken to ensure that the water surface is kept still. If it is not, the light reflected from the water surface may include Doppler components that are difficult to separate from perfusion-generated Doppler shifts in the final perfusion image.
83.5 CORRELATION WITH OTHER METHODS LDPM and LDPI use a similar signal processing algorithm for the calculation of the perfusion value. LDPM, however, is intended for continuous recordings at a single site, whereas LDPI records the perfusion within a specific tissue area. Combining the two methods therefore facilitates studies of both the temporal and spatial variations in skin perfusion. Thus, it is not surprising that Sefalian et al.9 as well as Harrison and colleagues10 have been able to demonstrate a good agreement between the two methods when studying skin perfusion. This agreement was demonstrated despite the spatial variations that exist in normal skin perfusion between adjacent skin areas. However, if the LDPM probe is moved to an adjacent site, the slope of the LDPI/LDPM curve can be expected to change due to the heterogeneity in skin perfusion. This has been demonstrated by several studies.3,9 By the use of biopsy, and a special probe holder for the generation of LDPM topographic maps, Braverman et al.11 correlated different LDPM perfusion patterns with vessel type. Furthermore, this LDPM mapping technique has been compared to LDPI with the modification of using a longer sampling time at each measurement site than in the ordinary setup.12 Perfusion values recorded from ventral forearm skin areas with consistent high and low perfusion coincided when measurements were made by both systems. Harrison and coworkers10 also compared LDPI with thermographic mapping of the skin and found that the regional temperature profile resembled the perfusion values, but the extreme heterogeneity picked up as large pixel-to-pixel variations by LDPI could not be demonstrated by thermography. Thermography, however, records a parameter (skin temperature) that is only indirectly related to skin perfusion, while LDPI directly senses the speed and concentration of the blood cells in the microvascular network. Diverging results may therefore be expected, especially when measurements are made at the
721
fingertips or over ulcers with a high evaporative water loss that significantly reduces the tissue temperature. Since its introduction, LDPI has been used in a wide range of skin applications. Some examples are characterization of the cutaneous axon reflex,13 irritant and allergy patch testing,14,15 investigation of perfusion pattern in basal cell carcinoma treated by photodynamic therapy,16 burn treatment,17 evaluation of UVB reactions in phototesting,18 and study of perfusion in leg ulcers.19
ACKNOWLEDGMENT The author thanks Michail Ilias, Ph.D., at the Department of Biomedical Engineering for invaluable help by preparing the manuscript and illustrations.
REFERENCES 1. Nilsson, G.E., Tenland, T., and Öberg, P.Å., Evaluation of a laser Doppler flowmeter for measurement for tissue blood flow, IEEE Trans Biomed Eng, 27, 597, 1980. 2. Wårdell, K., Jakobsson, A., and Nilsson, G.E., Laser Doppler perfusion imaging by dynamic light scattering, IEEE Trans Biomed Eng, 40, 309, 1993. 3. Tenland, T., Salerud, G., and Nilsson, G.E., Spatial and temporal variations in human skin blood flow, Int J Microcirc Clin Exp, 2, 81, 1983. 4. Braverman, I.M. and Schechner, J.S., Contour mapping of the cutaneous microvasculature by computerized laser Doppler velocimetry, J Invest Dermatol, 97, 1013, 1991. 5. Essex, T.J. and Byrne, P.O., A laser Doppler scanner for imaging blood flow in skin, J Biomed Eng, 13, 189, 1991. 6. Nilsson, G.E., Salerud, E.G., Strömberg, T., and Wårdell, K., Laser Doppler monitoring and imaging techniques, in Biomedical Photonics Handbook, VoDinh, T., Ed., CRC Press , Boca Raton, FL, 2003, chap. 15, pp. 1–24. 7. Wårdell, K. and Nilsson, G.E., Duplex laser Doppler perfusion imaging, Microvasc Res, 52, 171, 1996. 8. Fullerton, A., Stücker, M., Wilhelm, K.-P., Wårdell, K., Anderson, C., Fischer, T., Nilsson, G.E., and Serup, J., Guidelines for visualisation of cutaneous blood flow by laser Doppler imaging, Contact Derm, 46, 129, 2002. 9. Seifalian, A.M., Stansby, G., Jackson, A., Howell, K., and Hamilton, G., Comparison of laser Doppler perfusion imaging, laser Doppler flowmetry, and the thermographic imaging for the assessment of blood flow in human skin, Eur J Vasc Surg, 1993. 10. Harrison, D.K., Abbot, N.C., Swanson Beck, J., and McCollum, P.T., A preliminary assessment of laser Doppler perfusion imaging in human skin using the tuberculin reaction as a model, Physiol Meas, 14, 241, 1993. 11. Braverman, M.I., Keh, A., and Goldmintz, D., Correlation of laser Doppler wave patterns with underlying microvascular anatomy, J Invest Dermatol, 3, 283, 1990.
722
Handbook of Non-Invasive Methods and the Skin, Second Edition
12. Wårdell, K., Braverman, I.M., Silverman, D.G., and Nilsson, G.E., Spatial heterogeity in normal skin perfusion recorded with laser Doppler imaging and flowmetry, Microvasc Res, 48, 26, 1994. 13. Wårdell, K., Naver, H.K., Nilsson, G.E., and Wallin, B.G., The cutaneous vascular axon reflex in humans characterized by laser Doppler perfusion imaging, J Physiol, 460, 185, 1993. 14. Fullerton, A., Benfeldt, E., Petersen, J.R., Jensen, S.B., and Serup, J., The calcipotriol dose-irritation relationship, 48 hour occlusive testing in healthy volunteers using Finn chambers, Br J Dermatol,138, 259, 1998. 15. Bjarnason, B. and Fischer, T., Objective assessment of nickel sulfate patch test reactions with laser Doppler perfusion imaging, Contact Derm, 39, 112, 1998.
16. Enejder, A.M., Klinteberg, C., Wang, I., AnderssonEngels, S., Bendsoe, N., Svanberg, S., and Svanberg, K., Blood perfusion studies on basal cell carcinomas in conjunction with photodynamic therapy and cryotherapy employing laser-Doppler perfusion imaging, Acta Derm Venereol, 80, 19, 2000. 17. Kloppenberg, F.W., Beerthuizen, G.I., and ten Duis, H.J., Perfusion of burn wounds assessed by laser doppler imaging is related to burn depth and healing time, Burns, 27, 359, 2001. 18. Ilias, M.A., Wårdell, K., Falk, M., and Anderson, C., Phototesting based on a divergent beam: a study on normal subjects, Photodermatol Photoimmunol Photomed, 17, 189, 2001. 19. Gschwandtner, M.E., Ambrozy, E., Maric, S., Willfort, A., Schneider, B., Böhler, K., Gaggl, U., and Ehringer, H., Microcirculation is similar in ischemic and venous ulcers, Microvasc Res, 62, 226, 2001.
Heat Wash-In and Heat Wash-Out 84 The Technique for Quantitative, NonInvasive Measurement of Cutaneous Blood Flow Rate Per Sejrsen and Mette Midttun Department of Medical Physiology, The Panum Institute, University of Copenhagen, Copenhagen, Denmark
CONTENTS 84.1 Introduction............................................................................................................................................................723 84.2 Object.....................................................................................................................................................................723 84.3 Methodological Principles .....................................................................................................................................724 84.3.1 Physical Principles.....................................................................................................................................724 84.3.2 The Measuring Probe ................................................................................................................................724 84.3.3 Registration and Data Management ..........................................................................................................724 84.3.4 The Wash-In, Wash-Out Model.................................................................................................................724 84.3.5 Calculation of Blood Flow Rates ..............................................................................................................725 84.3.6 Loss of Heat to the Surrounding Air and the Surrounding Tissue...........................................................725 84.4 Sources of Error.....................................................................................................................................................726 84.5 Correlation with Other Methods ...........................................................................................................................727 84.6 Experimental and Clinical Applications................................................................................................................728 84.6.1 Experimental Studies .................................................................................................................................728 84.6.2 Clinical Studies..........................................................................................................................................729 84.7 Recommendations..................................................................................................................................................730 Acknowledgment.............................................................................................................................................................730 References .......................................................................................................................................................................730
84.1 INTRODUCTION Measurement of blood flow rate in cutaneous tissue using heat as an indicator has the serious problem that heat diffuses about 100 times faster than gases. Therefore, heat has the possibility to escape by other routes than transport with the blood stream. The consequence is that the wash-in rates of heat after cooling will be dominated by the fast diffusion transport of heat into the measuring area from the surrounding air and tissue. After heating, the registered wash-out rate will be dominated by the fast diffusion transport of heat from the measuring area to the surrounding air and tissue. To minimize these problems, the measuring probe has been constructed with a cap that is thermostatically controlled to keep the same temperature as the measuring disc
in contact with the skin surface, in order to eliminate the temperature gradient between the measuring disc and the surrounding air and tissue. The cap is located to the outside of the probe directed against the surrounding air. It is in contact with the surface of the cutaneous tissue by a metal ring placed in the circumference of the central measuring disc at a distance of 2 mm.1,2
84.2 OBJECT The purpose of the present chapter is to describe the heat wash-in and the heat wash-out technique for noninvasive, quantitative measurement of blood flow rates in cutaneous tissue in areas without as well as with arteriovenous anastomoses. 723
724
Handbook of Non-Invasive Methods and the Skin, Second Edition
84.3 METHODOLOGICAL PRINCIPLES 84.3.1 PHYSICAL PRINCIPLES Heat has a diffusion coefficient of 10–3 cm2·s–1. This is, as mentioned above, about 100 times faster than for gases. The consequence is that heat has the possibility to escape from the measuring area by diffusion to the surrounding air and the surrounding tissue during the measuring period, with an elimination rate that is much faster than the elimination caused by the blood flow rate. Therefore, it is necessary to minimize the heat uptake from or the heat loss to the surrounding air and tissue, in order to obtain the heat wash-in rate or the heat wash-out rate separately due to the blood flow rate in the tissue area under study. The probe is constructed with a thermostatically controlled cap set to keep the same temperature as the central metal measuring disc by heating or cooling delivered by a Peltier element located in the cap. A thermistor is placed in the cap for registration of the temperature in the cap. The cap is in contact with the skin surface by a metal ring located in the circumference of the measuring disc at a distance of 2 mm from the central disc. By this construction, the thermostatically controlled cap will eliminate the temperature gradient between the measuring site and the surrounding air and the surrounding tissue area. During the primary heating or cooling a metal bridge is placed to give a thermic contact between the cap with the Peltier elements and the central measuring disc. After the primary period of cooling or heating to a steady-state temperature level the metal bridge is redrawn to the cap and the temperature of the cap is set to follow the temperature of the measuring disc.1,2
84.3.2 THE MEASURING PROBE The heat wash-in and heat wash-out probe are presented in Figure 84.1; 20a is the central measuring disc containing a thermistor 45 and 15 is the thermostated cap with thermistor 40. The thermal metal bridge is 35, with the part 35a, which can be brought in contact with the measuring disc by rotation of the finger screw, which is the top of the bridge. The values 25 and 30 are the Peltier elements for cooling or heating, 10 and 50 are cooling ribs for exchange of heat with the surrounding air, and 20b and 55 are made of a plastic material.1,2
84.3.3 REGISTRATION
AND
DATA MANAGEMENT
A 2 to 5°C change in temperature is introduced into the measuring disc and the tissue in contact with the probe by heating or cooling with the Peltier elements located in the cap of the probe via the thermal metal bridge in contact position.
35
10
50 25
30 40 55
15
35a 45 20a
20b
10 mm
FIGURE 84.1 Drawing of the heat wash-in, heat wash-out probe for measuring cutaneous blood flow rates.2 For details, see text.
When a steady-state temperature is obtained in the measuring disc, after a few minutes the thermic bridge is redrawn to the cap and the heating or cooling is interrupted. The cap is thereafter set to follow the temperature of the central measuring disc. The temperature change per time unit of the measuring disc is registered in 5-s periods. The temperature change is registered until the baseline temperature is obtained. The baseline temperature was measured in the steady-state temperature period before heating or cooling, and again after the end of heat washout or heat wash-in. The registered temperature values are corrected for the baseline temperature. A fine adjustment of the baseline temperature value can be made manually in 0.01°C intervals if necessary due to an eventual small change in baseline temperature. The temperature wash-in or wash-out curves corrected for the baseline temperature level follow after a few seconds a monoexponential wash-in or wash-out function. This monoexponential function is caused by the supply or elimination of heat by the blood flowing through the tissue. The initial phase of a few seconds duration is due to a temperature equilibration between the probe and the tissue. The slope of the following monoexponential function is equal to the blood flow rate in the cutaneous tissue. A heat wash-out curve measured on the forearm in a normal human subject is seen in Figure 84.2.2
84.3.4 THE WASH-IN, WASH-OUT MODEL The wash-out system is in analogy with an electric model, where Coulon = joule, voltage = temperature, capacitans = joule/degree Celsius, and ampere = joule/time = watt. Ohm = temperature/watt = 1/(flow rate) CD = heat in the measuring disc RI = temperature fall or increase per time unit due to the initial heat equilibration
The Heat Wash-In and Heat Wash-Out Technique
log T – Tb
10
1
0.1 0
2
4 6 Minutes
8
10
FIGURE 84.2 Heat wash-out from the forearm in a normal human subject after an initial heating of the measuring disc of the probe and the underlying tissue to 43°C.2 The results are plotted as the measured temperature minus the baseline temperature on a logarithmic ordinate scale against the time in minutes on a linear scale on the abscissa.
CC = heat in the skin tissue RF = temperature fall or increase per time unit due to blood perfusion in the cutaneous tissue See Figure 84.3. The temperature, U, of the measuring disc in contact with the skin surface is then U = U0(a1·e–t/T1 + b1· e–t/T2) Analogous electrical model Coulon = joule Voltage = temperature Capacitans = joule/degree Ampere = joule/time Ohm = temperature/watt = 1/flow CD = heat in the probe
RI = initial temp. fall
CC
CC = heat in the tissue
RF = temp. fall due to perfusion
FIGURE 84.3 A presentation of an electrical analog model. For details, see text.
725
where t is the time variable in the wash-in or wash-out process. After a curve resolution the resulting two components have the intercepts a1 and b1 and the rate constants 1/T1 and 1/T2 for the initial and following components, respectively. Just after cessation of the heating an equilibration of heat will take place between the measuring disc and the tissue area under study. This is a fast process conditioned on the very high diffusion coefficient of heat in the tissue and the metal disc. It is observed that this initial fast component has a duration of a few seconds, with a maximum of 10 to 20 s. It can be assumed that CD, the amount of heat in the measuring disc, is small in comparison to CC, the amount of heat in the cutaneous tissue area under study. Furthermore, it can be assumed that RI, the temperature fall per time unit in the cutaneous tissue area in the initial phase due to the diffusional equilibration of heat, is very fast and brief compared to RF, the following monoexponential temperature fall due to the heat wash-out by the blood perfusion. The rate constant of this function, 1/T2, is a direct expression of the elimination rate caused by the blood flow rate in the cutaneous tissue: T2 = RF·(CD + CC)
84.3.5 CALCULATION
OF
BLOOD FLOW RATES
Blood flow rates can be calculated from the rate constant of the monoexponential function (Figure 84.3) using the equation of Kety,5 f = ln2·(T1/2)–1·lambda·100 (ml·(100 g·min)–1), where ln2 is the natural logarithm to 2, lambda is the tissue-to-blood partition coefficient for heat in milliliters per gram of tissue, and T1/2 is the half-time of the monoexponential wash-in or wash-out function. The tissue-to-blood partition coefficient lambda for heat is 0.954 ml·g–1, indicating that the heat capacity for tissue and blood, respectively, is almost equal. For simplification, a lambda value of 1 ml·g–1 can be used.
84.3.6 LOSS AND
HEAT TO THE SURROUNDING AIR THE SURROUNDING TISSUE OF
Measurement of heat elimination during blood flow cessation was performed with a cuff on the upper arm. Furthermore, a lead shield was placed distally to the probe in order to produce a pressure of 40 mmHg to eliminate the blood flow in the venous rete on the forearm. When the pressure in the cuff was above the systolic blood pressure (200 mmHg), the diffusional, non-blood-flow-dependent combined loss of heat to the surrounding air and surrounding tissue was measured. The results given as ml·(100 g·min)–1 for measurements performed on the forearm were an average of 0.9 in seven measurements, range 0.5 to 1.5, and on the pulp
726
Handbook of Non-Invasive Methods and the Skin, Second Edition
of the thumb the results were an average of 2.9 in six measurements, range 2.7 to 3.1.2 Similar measurements with the probe placed on a block of styrene foam gave the average result of 5.2 in six measurements, range 5.1 to 5.3.2 As a consequence of the low values obtained in these measurements, a correction for this error will be unnecessary in most situations. However, as mentioned above, the heat loss in the area under study can be measured during blood flow cessation and a correction performed if necessary.
2. 3.
4.
84.4 SOURCES OF ERROR The basic assumptions of the heat wash-in, heat wash-out method are identical to those of the 133Xe wash-out method.3–5 The assumptions are the following:
5.
70
60 Blood flow rate, heat (ml⋅(100g⋅min)–1)
1. Blood flow rate shall be constant during the registration, and no trauma at the introduction of the indicator must occur; i.e., a steady-state blood flow rate shall be present during the measurement. 2. The indicator must be stable; i.e., it must not be metabolized or bound in the tissue. 3. The indicator must not leave or be supplied to the system by routes other than the flowing blood. 4. Equilibrium for the indicator between the tissue and the flowing blood shall be obtained during the passage of the blood through the tissue. 5. The system shall be linear; i.e., a doubling of the indicator input to the system shall result in a doubling of the response. 6. The system shall be homogeneous. 7. Registration of the slope of the wash-out curve shall be estimated correctly. 8. The partition coefficient between tissue and blood, lambda, shall be known for the indicator used.
20 min after heating with the probe to a given temperature level by the procedure used.1,14 The indicator heat is stable. The heat production in the tissue is relatively very small and constant. The supply of heat or elimination of heat by routes other than the blood is minimized by the use of a thermostatically controlled cap. This is seen from the very low wash-in or wash-out rate values measured during blood flow cessation.2 The fulfillment of the assumption concerning equilibrium for the indicator between tissue and blood during the passage of blood through the tissue is presumably obtained. This is seen from the monoexponential wash-in and wash-out of heat, and from the correlation of the results of the heat wash-out with the 133Xe wash-out method in simultaneously performed experiments on the human forearm. In this region, only nutritive capillary blood flow is present. A correlation diagram is presented in Figure 84.4.1,14 The system has been shown to be linear in experiments on the pulp of the thumb at various temperature levels from 38 to 45°C (Figure 84.6).1,14
50
40
30
20 Subject 1 10
The fulfillment of the above-listed assumptions for the heat wash-in, heat wash-out method is made probable by the following conditions:
0
Subject 2
0
10
20
30
40
50
60
70
Blood flow rate, 133Xe (ml⋅(100g⋅min)–1)
1. The subjects were kept in termoneutral surroundings, and the indicator was introduced to the tissue area under study with heating or cooling by the probe to a steady-state temperature level which was obtained for some minutes. By the 133Xe wash-out method,3,4 it was shown that the blood flow rate remained constant for about
FIGURE 84.4 A comparison of measurements with the 133Xe wash-out method and the heat wash-out method.1,14 The measurements were performed simultaneously and in the same area on the forearm in two normal human subjects. During the measurements blood flow cessation was performed in the venous rete on the forearm by compression with a lead shield. For details, see text.
The Heat Wash-In and Heat Wash-Out Technique
6. The wash-in and wash-out curve for heat shows monoexponential functions for the cutaneous tissue over more than two decades. 7. To make an accurate registration of the slope of the curve, the wash-in and wash-out functions have been followed over a range of about two decades. 8. The partition coefficient, lambda, between tissue and blood for heat was calculated as the ratio of the data for heat capacitance of the tissue and blood, respectively. As the basic assumptions of the heat wash-in and heat wash-out method seem to be fulfilled for all practical purposes, and furthermore, because the results were found closely correlated to results obtained by the 133Xe washout method in simultaneous experiments,3,4 the heat washin, heat wash-out method can be considered a quantitative measure of blood flow rate in the cutaneous tissue. By taking the average value of a heat wash-in and a heat wash-out measurement in the same tissue area and with the same temperature change from the baseline temperature, it is presumably possible to get a blood flow rate equal to that in undisturbed conditions. In areas with arteriovenous anastomoses, as in the thumb pulp, the heat wash-in, heat wash-out method will measure the sum of the nutritive blood flow rate in the capillaries and that in the arteriovenous anastomoses. The blood flow rate in the capillaries in these regions can be measured by the 133Xe wash-out method.3,4 By subtraction of the 133Xe results from the heat wash-in or heat wash-
727
out results, it is possible to get the blood flow rate separately in the arteriovenous anastomoses.
84.5 CORRELATION WITH OTHER METHODS A comparative study has been performed as a simultaneous study on the same cutaneous area on the forearm with the heat wash-out method and the 133Xe wash-out method in two normal human subjects (Figure 84.4).1,14 The equation for the regression line is y = 2.5 + 0.968·x, and the correlation coefficient is 0.986 in the temperature interval from 37 to 45°C. With a Bland Altman6 statistical treatment for assessing agreement between results obtained with the two methods (Figure 84.5), the mean difference is 0.86, with SEM = ±0.27. The SD = 1.22 and the 95% confidence interval is –1.58 to 3.30, all values given in ml•(100 g•min)–1. Despite this, the limits of agreement (–1.58 and 3.30) are small enough to be confident, so that the heat wash-out method can be used in place of the 133Xe washout method for clinical purposes with use of a lambda value of 1.00 (Figure 84.5). If the correct partition coefficient lambda for heat, 0.954 ml·g–1, is used in this statistical treatment, it gives the following values: mean difference = 0.01, SEM = ±0.32, SD = 1.50, 95% confidence interval = –0.31 to 0.33, and the limits of agreement = –2.99 to 3.01. These results shows only a little closer agreement between the results of the two methods than does the use of lambda = 1.0.
Difference in blood flow rate ml⋅(100g⋅min)–1
4 +2 SD 3 +1 SD
2
1
Mean
0 –1 SD –1 –2 SD –2
10
70 Average blood flow rate of the two methods ml⋅(100g⋅min)–1
FIGURE 84.5 A Bland Altman plot6 of the result in Figure 84.4. For details, see text.
728
Handbook of Non-Invasive Methods and the Skin, Second Edition
160 Blood flow rate (ml⋅(100g⋅min)–1)
1
140 120 100 80 60 40 20 0
0.1 0
2
4
6
8 10 Time (min)
12
14
36
16
FIGURE 84.6 Heat wash-out measurements from the pulp of the thumb in a normal human subject after heating the measuring disc of the probe and the underlying tissue to various temperatures.1,14 The curves presented are heat wash-out after initial heating to temperatures in the interval 38 to 45°C, with steps of 1°C. The results are plotted after subtraction of the baseline temperature on a logarithmic ordinate scale with a linear timescale on the abscissa. The crosses indicate the start and end of the calculated regression lines.
84.6 EXPERIMENTAL AND CLINICAL APPLICATIONS 84.6.1 EXPERIMENTAL STUDIES Blood flow rate in the pulp of the thumb measured by the heat wash-out method in a normal subject at various temperatures from 38 to 45°C is shown in Figure 84.6.1,14 The wash-out rates are almost the same in the eight measurements, as the subject is in warm conditions. The blood flow rate is in warm conditions dominated by the blood flow in the open arteriovenous anastomoses. A study with measurements on the pulp of the thumb in two normal human subjects has been performed to compare results obtained by the heat wash-out method, the 133Xe wash-out method on the skinfold between the thumb and the forefinger, and venous occlusion plethysmography on the distal part of the thumb. The results from one of the subjects are shown in Figure 84.7.1,14 The measurements were done at various temperatures from 36 to 45°C. The heat wash-out method shows much higher values than the two other methods. Figure 84.8 shows blood flow rate measured by the heat wash-out method and the 133Xe wash-out method in the thumb pulp in a normal human subject exercising on an ergometer bicycle.7,14 After 2 min of rest follows 7 min of exercise with a moderate load, and finally again resting conditions. Blood flow rate measured by the 133Xe washout method in the cutaneous capillaries does not change during the exercise period. The heat wash-out method
37
38
39 40 41 42 Temperature (°C)
43
44
45
FIGURE 84.7 Blood flow rates in the pulp of the thumb in a normal human subject measured in the temperature interval from 36 to 45°C by the 133Xe wash-out method (triangles), by venous occlusion plethysmography (squares), and by the heat wash-out method (dots).1,14
200
Blood flow rate (ml⋅(100g⋅min)–1)
ΔT = T – Tb (°C, log scale)
10
150
100
50 Rest
Exercise
Rest
0 0
5
10
15
Minutes
FIGURE 84.8 Blood flow rates in the pulp of the thumb measured in a normal human subject with the heat wash-out method. The blood flow rate in the skinfold between the thumb and the forefinger was measured by the 133Xe wash-out method. The measurements were done during rest for 2 min, exercise for 5.5 min, and finally rest for 3.5 min. The cutaneous capillary blood flow rate measured by the 133Xe wash-out method is shown by the lower line.8,14
shows the combined blood flow rate in the arteriovenous shunt vessels and the capillaries. It is demonstrated that the blood flow rate in the shunt vessels is reduced in the
The Heat Wash-In and Heat Wash-Out Technique
729
120
100 Series 1
100
Series 2
ml⋅(100g⋅min)–1
60
40
20
0 h.l.
80 60 40 20
45 cm elev. h.l. 70 cm elev. Position in or above heart level
h.l.
Series 1: Measurements on the pulp of the thumb at various positions. Series 2: Measurements on the pulp of the thumb with the other hand in ice-water.
FIGURE 84.9 Blood flow rate in the pulp of the thumb in a normal human subject during ortostatic maneuvers. Measurements were performed with the hand at the heart level, 45 cm above the heart level, at heart level, 70 cm above the heart level, and finally at the heart level. The lower line shows blood flow rates with the contralateral hand in ice water at heart level, 45 cm above heart level, and at heart level.2
initial 4.5 min of the exercise period. This is followed by an increase presumably triggered by the center of thermoregulation in hypothalamus in order to eliminate heat from the body. Figure 84.9 demonstrates the reduction in blood flow rate in the pulp of the thumb when the position of the hand is changed from heart level to positions 45 and 70 cm above heart level, respectively.2 Furthermore, the near blood flow cessation in the pulp of the thumb, as an effect of placing the contralateral hand in ice water, is demonstrated.2 Figure 84.10 shows the blood flow rate in the ear lobe and the nose in two normal subjects.8,14 By subtraction of the blood flow rate measured by the 133Xe wash-out method from that measured by the heat wash-out method, the blood flow rate in the arteriovenous shunt vessels has been obtained. Heat wash-out measurements on the pulp of the thumb and the first toe showed in 30 subjects (from 3 to 93 years old) a decrease with age.9,14 This is in accordance with the decrease in metabolic rate during the lifetime.
84.6.2 CLINICAL STUDIES The results of clinical examinations of blood flow rate in the pulp of the first toe during ortostatic maneuvers in claudicants, patients with critical ischemia, and normals are shown in Figure 84.11.10,14 Blood flow rate is low in patients with critical ischemia both at the heart level and in a position 50 cm below the heart level. In claudicants
0
Els
Elr
Nls
Nlr
Ells
Ellr
Nlls
Nllr
FIGURE 84.10 Average blood flow rates in the ear lobe, E, on the side of the nose, N, upon sitting, s, and in the recumbent position, r, in two normal subjects, n = 5. The results of measurements with the heat wash-out method are the total height of the columns.10,14 The hashed parts of the columns denote blood flow rate in the capillaries as measured by the 133Xe wash-out method. The white part of the columns denotes blood flow rate in the arteriovenous anastomoses obtained by subtraction of the 133Xe values from the heat wash-out values.
70
60
50 ml⋅(100g⋅min)–1
ml⋅(100g⋅min)–1
80
40
30
20
Norm. Claud. SL Claud. AS Crit. isch. SL Crit. isch. AS
10
0 +50 cm above
0 Heart level
–50 cm below
FIGURE 84.11 The median blood flow rate in the pulp of the first toe at ortostatic maneuvers in normal subjects (crosses); in claudicants, asymptomatic side (open squares); in claudicants, symptomatic side (filled squares); in patients with critical ischemia, asymptomatic side (open circles); and in patients with critical ischemia, symptomatic side (filled circles). The positions were 50 cm above heart level, at heart level, and 50 cm below heart level.11,14
730
Handbook of Non-Invasive Methods and the Skin, Second Edition
blood flow rate is low at the heart level, but it increases with a factor of 1.7, to a normal level, in the dependent position 50 cm below the heart level. Above the heart level the blood flow rate is reduced in both normals and claudicants. After surgical revascularization claudicants showed normal blood flow rate responses to ortostatic maneuvers.11,14 In patients with nonspecific aneurysmal disease of the infrarenal aorta the blood flow rate measured by the heat wash-out method on the pulp of the first toe showed values equal to those found in normal subjects. However, the blood flow rate in the subcutaneous adipose tissue on the forefoot, measured by the 133Xe wash-out method, was in these patients about three times higher than in normals at heart level and during ortostatic maneuvers.12 The results can be interpreted as a general degeneration of the elastic fibers in the arterial vessel walls, reducing the vascular resistence. Blood flow rate in the cutaneous tissue on the thorax was measured by the heat wash-in method in the initial phase just after placing the probe on the skin surface.13 By this procedure the blood flow to the area under study will increase the temperature until a steady-state level. By subtraction of the steady-state temperature from the registered temperatures a monoexponential wash-in function is obtained, and the rate constant of this function yields the blood flow rate in the cutaneous tissue. In this study on six subjects, it was shown that blood flow rate increased during thoracic epidural blockade from 13.6 (range, 10.6 to 14.6 ml•(100 g•min)–1) to 18.4 (range, 13.9 to 24.5 ml • (100 g • min) –1 ) (p < 0.05). During unblocked conditions, local heating to 40°C gave an increase in blood flow rate in five of the six patients, and the result was 27.9 on average (range, 20.8 to 34.6 ml•(100 g•min)–1) (p = not significant). Thus, heating to 40°C yields a blood flow rate that is not significantly different from that obtained during blockade. In a recent study blood flow rate in cutaneous tissue was measured by the heat wash-out method on the forehead in patients with partial obstruction in the common carotic artery. The measurements were done in an area over the medial part of the eyebrow. In this cutaneous area blood is supplied from the internal carotic artery. During entarterectomia it was shown that the blood flow to this region was increased by 18.6% after opening of the external carotic artery and by 81.4% after opening the internal carotic artery. Thus, it seems as if an indirect measure of blood flow rate to the brain can be obtained by using this cutaneous region supplied by the internal carotic artery.
84.7 RECOMMENDATIONS The heat wash-in, heat wash-out method, which is an atraumatic and quantitative method for measurement of cutaneous blood flow rate, seems to present good condi-
tions for use in experimental physiology and pharmacology and in clinical studies. Another possibility is to use the method for calibration of other methods giving qualitative measuring results. This is in accordance with the use of the 133Xe wash-out method for this purpose. The heat wash-in, heat wash-out method has the advantage, in comparison with the 133Xe wash-out method, to have no expense to buy the indicator 133Xe, with intervals of about 2 weeks. Furthermore, the time interval required for a measurement is much shorter, about 5 to 20 min with the heat wash-in, heat wash-out method, compared to about 25 to 90 min with the 133Xe wash-out method. The heat wash-in, heat wash-out method has a great potential for clinical use within the following fields: arterial and venous diseases in the legs before and after treatment, valuation of level for amputation, white fingers, burn and cold injuries, wound healing, skin diseases, and diabetes. Another field is validation of the sympathetic tonus in patients with heart diseases, in supervision of prematures, in patients during surgery, and in intensive care. Furthermore, the possibility for measuring blood flow rate in arteriovenous anastomoses is of special importance in the temperature regulation investigation.
ACKNOWLEDGMENT The authors acknowledge Jens D. Hove, M.D., Ph.D., MA. Phys., for his contribution by creating the analogous electrical model for the heat wash-in, heat wash-out technique.
REFERENCES 1. Midttun, M., Sejrsen, P., and Colding-Jørgensen, M., Heat-washout: a new method for measuring cutaneous blood flow rate in areas with and without arteriovenous anastomoses, Clin Physiol, 16, 259, 1996. 2 Sejrsen, P. and Midttun, M., A Method and an Apparatus for Measuring Flow Rates, International Publication WO 01/43629 A1, published under the Patent Cooperation Treaty (PCT), and U.S. Patent Application 20030139676 A1, http://appft1.uspto.gov/netacgi/nphParser. 3. Sejrsen, P., Measurement of cutaneous blood flow by freely diffusible radioactive isotopes. Methodological studies of the washout of krypton-85 and xenon-133 from the cutaneous tissue in man, Dan Med Bull, Suppl. 18, 9, 1971. 4. Sejrsen, P., The 133xenon wash-out technique for quantitative measurement of cutaneous and subcutaneous blood flow rates, chap. 17.5, ibid. 5. Kety, S.S., Measurement of regional circulation by the local clearance of radioactive sodium, Am Heart J, 38, 321, 1949.
The Heat Wash-In and Heat Wash-Out Technique
6. Bland, J.M. and Altman, D.G., Statistical methods for assessing agreement between two methods of clinical measurement, Lancet, 1, 307, 1986. 7. Midttun, M. and Sejrsen, P., Cutaneous blood flow rate in areas with and without arteriovenous anastomoses during exercise, Scand J Med Sci Sports, 8, 84. 8. Midttun, M. and Sejrsen, P., Blood flow rate in arteriovenous anastomoses and capillaries in thumb, first toe, ear lobe, and nose, Clin Physiol, 16, 275, 1996. 9. Midttun, M., Blood flow rate in arteriovenous anastomoses: from the cradle to the grave, Clin Physiol, 20, 5, 360, 2000. 10. Midttun, M., Sejrsen, P., and Paaske, W.P., Blood flow rate during orthostatic pressure changes in the pulp skin of the first toe, Eur J Vasc Endovasc Surg, 13, 278, 1997.
731
11. Midttun, M., Sejrsen, P., and Paaske, W.P., Peripheral blood flow rates and microvascular responses to orthostatic pressure changes in claudicants before and after revascularisation, Eur J Vasc Endovasc Surg, 17, 225, 1999. 12. Midttun, M., Sejrsen, P., and Paaske, W.P., Is non-specific aneurysmal disease of the infrarenal aorta also a peripheral microvascular disease? Eur J Vasc Endovasc Surg, 19, 625, 2000. 13. Nygård, E., Sejrsen, P., and Kofoed, K.F., Thoracic sympatholysis with epidural blockade assessed by quantitative measurement of cutaneous blood flow, Acta Anaest Scand, 46, 1037, 2002. 14. Midttun, M., Heat-washout: a new method for measuring cutaneous blood flow in areas with and without arteriovenous anastomoses. Physiological and patophysiological examinations, Dan Med Bull, Suppl., 2004.
85
The 133Xenon Wash-Out Technique for Quantitative Measurement of Cutaneous and Subcutaneous Blood Flow Rates Per Sejrsen Department of Medical Physiology, The Panum Institute, University of Copenhagen, Copenhagen, Denmark
CONTENTS 85.1 Introduction............................................................................................................................................................733 85.2 Object.....................................................................................................................................................................733 85.3 Methodological Principles .....................................................................................................................................733 85.3.1 Physical Principles.....................................................................................................................................733 85.3.2 Atraumatic Local Labeling........................................................................................................................733 85.3.3 Registration and Data Management ..........................................................................................................734 85.3.4 The Washout Model...................................................................................................................................735 85.3.5 Calculation of Blood Flow Rates ..............................................................................................................738 85.3.6 Loss of 133Xe from the Skin Surface.........................................................................................................738 85.4 Sources of Error.....................................................................................................................................................739 85.5 Correlation with Other Method.............................................................................................................................739 85.6 Recommendations..................................................................................................................................................739 References .......................................................................................................................................................................740
85.1 INTRODUCTION Measurement of blood flow rates in cutaneous and subcutaneous tissues is of interest in human physiology, pathophysiology, and in control of the therapeutic effect. It is especially of interest in the understanding of the distribution of cardiac output to the skin during test, orthostatic maneuvers, and dynamic exercise, and in thermoregulation. Most of the methods developed for this purpose have been qualitative in nature. The introduction of the 133Xe washout method after epicutaneous labeling has made it possible to measure the cutaneous and subcutaneous blood flow rates quantitatively during atraumatic conditions.1
85.2 OBJECT The purpose of the present chapter is to describe the measurement of cutaneous and subcutaneous blood flow
rates by the washout of 133Xe after atraumatic local epicutaneous labeling using external residue detection.
85.3 METHODOLOGICAL PRINCIPLES 85.3.1 PHYSICAL PRINCIPLES 133
Xe is a radioactive inert gas isotope with a physical half-life of 5.3 days. The radiations emitted by disintegration of 133Xe are x-ray and emission. By a NaI (T1) scintillation detector coupled to a γ spectrometer, it is possible to register the 133Xe γ emission of 81 keV, with an incidence of 35.5%, and the x-ray of about 40 keV, with an incidence of 64.5%, by setting the window to include these two energy peaks. The distance between the 133Xe deposit and the detector shall be kept constant throughout the total period of registration to measure the relative washout rate. The collimation shall be so wide that registration is obtained from the total 133Xe depot area — also when this expands by diffusion to the surrounding tissue area. A 733
734
Handbook of Non-Invasive Methods and the Skin, Second Edition
Mylar membrane
Seen from the side
Adhesive area
Cutis
Mylar membrane Seen from above
Adhesive area
5 cm
FIGURE 85.1 Technique for epicutaneous application of Physiol., 24, 570, 1968. With permission.)
133
Xe gas or
suitable distance between deposit and detector is 15 to 20 cm. This distance will minimize the effect on the counting efficiency of the expansion of the 133Xe deposit by diffusion, which will increase the distance between the detector and the labeled area. Another detector suitable for registration of 133Xe activity is a cadmium telluride (chloride) detector (Cd Te (Cl)).2 As this detector type can be fixed to the region by adhesive plaster, keeping the counting geometry almost constant, it is a portable solution of the registration. It is important to note that the labeled area, by the short distance used with this detector type, shall be either so small that 133Xe cannot leave the counting area by diffusion or so large that a constant concentration is present in an area somewhat larger than the area of registration. The short counting distance used by this detector type has been from 1 to 20 mm, which makes it very important to correct for the elimination rate due to expansion by diffusion, if present. This can be done by subtracting the elimination rate measured during blood flow cessation from that obtained with intact blood flow. As 133Xe is a gas, it is freely diffusible in the tissues, and equilibrium between tissue and blood is obtained for 133Xe during the passage of blood through the tissue. This has been shown in experiments on semi-isolated, autoperfused gastrocnemius muscles in cats, where the 133Xe washout method was compared to the directly measured outflow rate of blood.3 On this basis, 133Xe is a suitable indicator for measurement of blood flow rates, as the 133Xe washout rate is proportional to blood flow rate.
85.3.2 ATRAUMATIC LOCAL LABELING Atraumatic local labeling of the tissue with 133Xe can be done by application of 133Xe gas or a 133Xe in saline
133
Xe dissolved in isotonic saline. (From Sejrsen, P., J. Appl.
solution on the skin surface for a few minutes, e.g., 3 min.4 In practice this is performed by the following technique. A deposit of 133Xe is placed on the skin surface in a chamber formed by the skin surface and a circular gastight Mylar® membrane 3 to 8 cm in diameter and 20 mm thick. The membrane is attached to the skin surface by a ringshaped, 0.7- to 1.5-cm-wide, adhesive membrane with adhesive material on both sides (Figure 85.1).1 The dimension of the central chamber will then be from 1.6 to 5 cm in diameter. A thin injection needle is placed between the Mylar and the adhesive membrane, leading from the outside into the chamber. From a syringe it is then possible to introduce a deposit of 133Xe gas or 133Xe in isotonic saline solution into the chamber. After labeling by diffusion from the deposit on the skin surface into the skin for about 3 min, the deposit is redrawn to the syringe. The membrane with adhesive ring, needle, and syringe is removed, the region is dabbed with a piece of soft tissue paper, and the surplus of 133Xe is blown away.
85.3.3 REGISTRATION
AND
DATA MANAGEMENT
The registration of the 133Xe activity is then performed as a residue detection by external counting in time intervals of, e.g., 20 sec (from 1 sec to 1 min), dependent on the purpose of the measurement and the level of activity. The data obtained are then plotted in a semilogarithmic diagram after subtraction of the background activity. The x axis is time in a linear scale, and the y axis is the activity in a logarithmic scale. The washout of 133Xe after atraumatic local labeling of a skin area shows a biexponential course (Figure 85.2). This is due to diffusion of 133Xe from the cutaneous venous blood out through the walls of the venous vessels during the passage of this bloodstream through veins located in
The 133Xenon Wash-Out Technique for Quantitative Measurement
Application
Counts/minute
105
104
0
30
60 Minutes
90
120
FIGURE 85.2 Washout curve after epicutaneous labeling with 133Xe for 3 min on the lateral side of crus. Solid circles show the registered activity with time. The open circles show the result of graphic curve resolution. The mathematical expression of the two exponentials obtained by the graphic curved resolution are presented in Equation 85.8. The washout curve separate for the cutaneous tissue is constructed by drawing a line parallel to the straight line through the open circles from the top of the registered curve. By subtracting the values given by this line from those of the registered curve the separate curve for the subcutaneous tissue is constructed, here denoted by crosses. The mathematical expressions of these separate washout curves for the two tissues are given in Equation 85.7. (From Sejrsen, P., Circ. Res., 25, 215, 1969. With permission.)
the subcutaneous tissue. 133Xe has about 10 times higher solubility in subcutaneous adipose tissue than in blood. This has the effect that the subcutaneous tissue acts as a zinc for 133Xe, resulting in an accumulation of 133Xe in the subcutaneous tissue with time. This is illustrated in Figure 85.3a and b, showing the distribution of 133Xe in the tissue after atraumatic local labeling by autoradiographic technique. After 2 min of in vivo washout the 133Xe is located almost exclusively in the cutaneous tissue, and after 70 min almost exclusively in the subcutaneous tissue. The result of the very limited transport by diffusion alone without blood flow after 70 min is shown in Figure 85.3c.5 A diffusion of 133Xe directly from cutaneous to subcutaneous tissue over the contact area between these two tissues thus seems of lesser importance than the above-mentioned transport by convection with venous blood flow combined with diffusion out through the venous vessel walls and into the surrounding subcutaneous tissue. A transport in the opposite direction from subcutaneous tissue to cutaneous tissue later in the washout process seems to be negligible due to the following reasons. The very high solubility of 133Xe in subcutaneous tissue compared to that in blood and cutaneous tissue will counteract an exchange by diffusion from the subcutaneous tissue. The
735
higher linear velocity of blood in the arterial vessels, and the lower contact area between blood and tissue in these vessels compared to that of the venous vessels, will also minimize the exchange between the subcutaneous and cutaneous tissues. The biexponential washout of 133Xe is on the abovementioned conditions a combined washout curve including an initial, fast washout component from the cutaneous tissue and an accumulation in the subcutaneous tissue, followed by a washout from this tissue. The accumulation in the subcutaneous tissue is determined by the convective transport with the cutaneous venous blood. Thus, the registered curve contains only two washout rate constants — the cutaneous and the subcutaneous — and by a graphic curve resolution these two rates can be obtained (Figure 85.2).5 In a special region of the skinfold between the extended thumb and the forefinger it was possible to measure the washout of 133Xe separately from cutaneous tissue. This was done after atraumatic local labeling with 133Xe gas of the region and a shielding of the rest of the hand by a 3-mm-thick lead shield. By registration of the activity from the distal, unshielded 3 to 4 mm of the skinfold, being solely cutaneous tissue, a monoexponential washout curve was obtained over 3.5 decades (Figure 85.4). A similar result has been obtained for a skinfold raised on the back of the hand. A monoexponential washout of 133Xe from subcutaneous tissue has also been observed. After an atraumatic local labeling of subcutaneous fatty tissue in an autoperfused inguinal fat pad preparation in cats, a monoexponential washout was followed over 2.5 h (Figure 85.5).5
85.3.4 THE WASHOUT MODEL On the basis of these observations of monoexponential washout of 133Xe from cutaneous and subcutaneous tissues the following combined in-series and in-parallel washout model is described.5 The model assumes that under steadystate conditions a constant fraction, E, of the 133Xe in the cutaneous venous blood is extracted, as it passes through the subcutaneous tissue, due to the 10-fold higher solubility of 133Xe in this tissue than in blood. The model consists of two homogeneous compartments symbolized by C and S for the cutaneous and subcutaneous compartments, respectively. The input to the system is in the form of an impulse into C. With the initial amount of 1 U, the output from C is divided into two fractions: (1) the extracted fraction, E, and (2) the transmitted fraction, 1 – E. The extracted fraction, E, of the output from C is reaching the subcutaneous compartment, and thus C is in-series with S for this fraction. The complementary fraction, 1 – E, is transmitted solely through the vascular volume by the flowing blood. The transport
736
Handbook of Non-Invasive Methods and the Skin, Second Edition
C S
(a)
C S
(b)
C S
(c)
FIGURE 85.3 Radioautograms of cutaneous (C) and subcutaneous (S) tissues. The tissue boundaries are illustrated in the schematic drawings to the right, (a and b) were taken after epicutaneous labeling with 133Xe gas for 3 min. (a) was taken after 2 min of in vivo washout, (b) after 70 min. (c) is the distribution of 133Xe in the tissue after intracutaneous injection of 0.1 ml of 133Xe in isotonic saline taken 70 min later demonstrating the minimum of exchange by diffusion between the cutaneous and the subcutaneous tissue without blood flow. The exposure time of the film emulsion was 20 h in (a), and 120 h in (b) and (c). (From Sejrsen, P., Circ. Res., 25, 215, 1969. With permission.)
of this fraction is thus an output from C, which is inparallel with the output from S (Figure 85.6). The total amount 1 is initially located in compartment C, and the cumulative output from C at time t is called H1. The amount retained in compartment C at time t, Rc, can then be written as Rc = 1 – H1. It is observed that compartment C is a well-mixed compartment with an exponential washout. The expression Rc = 1 – H1 = e–kc·t
(85.1)
can then be written for compartment C, where kc is the elimination rate constant for C.
At time zero there is no indicator in compartment S. The rate of input of tracer to S is a constant fraction, E, of the output rate from C. This can be obtained by differentiation of H1 with change of the sign. The input rate to S, Is, is then Is = E · kc · e–kc·t
(85.2)
It is assumed that the output from S follows a monoexponential function, which can be described by ks·e–ks·t for 1 U of indicator. Corresponding to the input rate to S, given by Equation 85.2, the output, Os, from S is the input rate, Is, convoluted by the impulse response for S:
The 133Xenon Wash-Out Technique for Quantitative Measurement
737
t
105
Os =
Application
∫E⋅k 0
⋅ e – kc⋅τ ⋅ k s ⋅ e – ks⋅( t – τ ) dτ
E ⋅ kc ⋅ ks ⋅ e – ks⋅t – e – kc⋅t kc – ks
(
Os = 104
c
(85.3)
)
(85.4)
A combined expression for the amount of indicator contained in C plus S, Rc plus Rs is called Q(t):
Counts/minute
Q(t) = Rc + Rs
(85.5)
By inserting Equations 85.1, 85.2, and 85.4 in Equation 85.5, the following expressions are obtained:
103
t
t
∫
∫
Q(t ) = R c + I s dt – O s dt 0
(85.6)
0
⎤ ⎡ E – kc Q(t ) = e – kc⋅t + ⎢ ⋅ e – ks⋅t – e – kc⋅t ⎥ k k – s ⎦ ⎣ c
102
(
)
(85.7)
⎛ E ⋅ k c ⎞ – kc⋅t E ⋅ kc ⋅e + ⋅ e – ks⋅t (85.8) Q(t ) = ⎜1 – k c – k s ⎟⎠ kc – ks ⎝ 101 0
30
60 Minutes
FIGURE 85.4 Washout curve separate from cutaneous tissue after epicutaneous labeling with 133Xe gas in 3 min. (From Sejrsen, P., Circ. Res., 25, 215, 1969. With permission.)
Counts/minute
104
Thus, it is possible by graphic curve resolution to determine the rate constants for the cutaneous and subcutaneous components. Furthermore, it is possible to calculate the E fraction from the intercepts of the two curves and their rate constants. The rate constants are the blood flow rate-to-partition coefficient ratios for the two tissues. On average, E was observed to be 0.50 in 10 washout experiments on the lateral side of the lower leg.
Application
103
102 0
30
60
90
120
150
Minutes
FIGURE 85.5 Washout curve separate from cutaneous tissue after local labeling with Res., 25, 215, 1969. With permission.)
133
Xe gas in 1 min. (From Sejrsen, P., Circ.
738
Handbook of Non-Invasive Methods and the Skin, Second Edition
S E C 1-E
FIGURE 85.6 Schematic diagram of the washout model. (C) is the cutaneous and (S) the subcutaneous component. (E) is the extracted fraction, which is washed out via the subcutaneous tissue. (1 – E) is the complementary fraction, which is washed out from the cutaneous tissue and transmitted by convection through the vascular volume. (From Sejrsen, P., Circ. Res., 25, 215, 1969. With permission.)
85.3.5 CALCULATION
OF
BLOOD FLOW RATES
Blood flow rates can be calculated from the rate constants using the equation introduced by Kety:6,7 f = ln2 · T1/2–1 · λ · 100 (ml · (100 g · min)–1)(85.9) where ln2 is the natural logarithm to 2 and l is the tissueto-blood partition coefficient for 133Xe in milliliters per gram (0.7 for cutaneous tissue and 10 for subcutaneous tissue).1 t1/2 is the half-time of the two monoexponential components as obtained by graphic curve resolution for the two tissues. The factor 100 is introduced to give the results per 100 g of tissue, which is the conventional term.
85.3.6 LOSS
OF 133XE FROM THE
SKIN SURFACE
The question concerning loss of 133Xe by diffusion out through the intact skin surface during the washout period has been elucidated by experiments with blood flow cessation. This was done by a cuff placed on the upper arm
60000
Counts/minute
Application
and inflated to a pressure of 230 mmHg, a pressure chosen well above the systolic pressure. Under this condition the very slow elimination rate is solely due to a loss of 133Xe out through the epidermal membrane (Figure 85.7).4 This has been demonstrated by placing a gastight Mylar membrane, 20 μm thick, over the deposit area with a drop of water interposed. After this gastight sealing of the surface the curve has an almost horizontal course with a decline equal to the physical decay. Without this gastight sealing of the surface the observed very slow elimination rate is about 1 to 2% of that observed during intact blood flow at normal, thermoneutral conditions (Figure 85.7). However, under sweating conditions, it can account for as much as 20 to 25% of the measured elimination rate of 133Xe from a cutaneous deposit, as measured just after the end of the labeling period. In such situations a correction can be performed to give the cutaneous blood flow rate. This can be done by subtraction of the elimination rate measured during blood flow cessation from that obtained for cutaneous tissue during intact blood flow after curve resolution. Thus, the elimination rates are in the order of 10–4 –1 min due to the physical decay of 133Xe, 10–3 min–1 due to diffusional loss of 133Xe out through the intact epidermal membrane, 10–2 min–1 due to profuse sweating, and from 0.07 to 0.7 min–1 due to cutaneous blood flow, corresponding to blood flow rates of about 6 to 50 ml · (100 g · min)–1 at normal conditions and at local heating to 45°C, respectively.5,8 On this basis the errors due to physical decay and loss of 133Xe out through the intact epidermal membrane are considered negligible when calculating blood flow rate from 133Xe washout curves. Blood flow rates in subcutaneous tissue are at normal conditions and during heating of the skin surface to 45°C measured to about 3 and 50 ml · (100 g · min)–1, respectively.1,8
Tourniquet
k = 0.0007 min−1 −1
k = 0.0430 min
30000
Environmental temperature 20°C
15000 0
10
20
30 Minutes
k = 0.1160 min−1
40
50
FIGURE 85.7 Washout of 133Xe after epicutaneous labeling with and without bloodflow cessation. (From Sejrsen, P., Circ. Res., 25, 215, 1969. With permission.)
The 133Xenon Wash-Out Technique for Quantitative Measurement
85.4 SOURCES OF ERROR The loss of 133Xe out through the skin surface is, as mentioned, negligible with an intact epidermal membrane, but correction can be necessary during sweat secretion. The gastight epidermal membrane can be removed by 40 to 50 times of stripping with adhesive plaster. This procedure removes the dry part of the epidermis, stratum corneum, which normally has a water content of about 4%. By this procedure the basal, living cell layers in stratum germinativum are left in situ. After 40 or 50 times of stripping with adhesive plaster the loss of 133Xe from the skin surface will increase about 30 to 40 times.9 Epidermal desquamation or denudation due to pathological processes can therefore give rise to a severe loss of 133Xe out through the skin surface. However, by placing a 20-mm-thick Mylar membrane on the skin with a drop of water interposed, it is possible to reestablish a gastight sealing of the surface. Steady state of blood flow is an assumption for the method. This is the reason for a demand of constant thermal conditions during a measurement. Alsom traumatic influence has to be avoided. The reason for not using a labeling technique with intracutaneous injection of 133Xe dissolved in isotonic saline is just the trauma effected by the injection, leading to hyperemia in the following 10 to 30 min. An uptake of 133Xe in rubber and plastic materials has been observed. When a cadmium telluride detector mounted with a rubber cap is placed in contact with the labeled skin area, the 133Xe uptake in the rubber invalidates the method. The use of standard values for the tissue-toblood partition coefficient, λ, for cutaneous and subcutaneous tissues are presumably acceptable in many regions. However, in regions with a thin subcutaneous layer, the lipid contents can be reduced, and a lower value has to be used for this tissue. In such a region it is possible to make an estimate of the relative contents of lipid, water, and protein in a tissue biopsy, and from these values and the corresponding values for the blood, to calculate a λ value for the tissue in question.
85.5 CORRELATION WITH OTHER METHODS Other methods have been employed in attempts to measure cutaneous blood flow rates. 85Krypton washout has been used with registration of the β activity by a Geiger–Müller tube after an intra-arterial injection of 85Kr dissolved in isotonic saline.10 This method is invalidated by diffusion processes in the tissue due to the short halfvalue thickness in the tissue of the emitted β radiation, only 0.25 mm.11 Other radioactive isotopes have been used, such as 24Na,6,7,12 131iodine,13 125I-antipyrine, and 131I-antipyrine,14
739
given as local injections with the isotopes dissolved in isotonic saline. The problems have been the trauma of injection, and that equilibrium between tissue and the flowing blood cannot be obtained for most of these indicators, as they are not freely diffusible in the tissues. Helium uptake through the skin has been used on the extremities. This method underestimates the cutaneous blood flow rate due to the existence of the epidermal diffusion barrier to gases.15,16 In accordance with this valuation, values obtained by this method have been low, about 3 to 4 ml · (100 g · min)–1). Heat conduction has been employed as a qualitative measure of cutaneous blood flow rate.17,18 The loss of heat to the surroundings invalidates this method as a quantitative method. From the amount of heat dissipated from the skin, a blood flow rate in the order of magnitude 2 to 10 ml(100 gmin)–1 seems likely.19,20 Venous occlusion plethysmography can only give a rough evaluation of the cutaneous blood flow rate, as it is based on measurements of blood flow rates before and after iontophoresis of epinephrine into the skin combined with complicated subtraction procedures. These procedures are necessary to correct for blood flow in subcutaneous and muscle tissue.21,22 By using values of the ratio between the weight of the cutaneous and subcutaneous tissues and the blood flow rate in subcutaneous tissue, a rough estimate gives a cutaneous blood flow rate in the order of 6 to 9 ml · (100 g · min)–1. Measurements with the laser Doppler technique are determined by both the velocity and the contents of the red blood cells in the tissue, causing this method to be of a semiquantitative nature. The change in blood cell contents in the vessels during orthostatic maneuvers and heat stress limits use of this method.23
85.6 RECOMMENDATIONS It is recommended to make control experiments with blood flow cessation to exclude loss of 133Xe from the skin surface or change in counting geometry during the registration. By this type of measurement it is possible to estimate a rate constant for loss of 133Xe to the surrounding air, and thereby to make a correction for this non-bloodflow-dependent elimination. As the 133Xe washout method with graphic curve resolution is based on the assumption of a steady-state blood flow rate, it is important to maintain a constant temperature in the body and in the surroundings during the registration. Also, the body and the region under study shall be kept in a constant position during the measurement to obtain steady-state conditions. Changes in blood flow rate in cutaneous tissue during the measurement can be registered quantitatively by the 133Xe washout method in the skinfold on the hand. A similar possibility is present for the subcutaneous tissue in the later part of the washout curve, which is separate
740
Handbook of Non-Invasive Methods and the Skin, Second Edition
from this tissue. It is important to use a sufficiently high 133Xe activity and to follow the washout for a sufficiently long time (1.5 to 2 h) to get an acceptably low standard deviation of the subcutaneous washout rate. This is a necessary basis for a reasonable graphic curve resolution and determination of the cutaneous washout rate.
REFERENCES 1. Sejrsen, P., Measurement of cutaneous blood flow by freely diffusible radioactive indicators, Dan Med Bull, Suppl. 18, 1, 1971. 2. Bojsen, J., Staberg, B., and Kølendorf, K., Subcutaneous measurements of 133Xe disappearance with portable CdTe(Cl) detectors: elimination of interference from combined convection and diffusion, Clin Physiol, 4, 309, 1984. 3. Sejrsen, P. and Tønnesen, K.H., Inert gas diffusion method for measurement of blood flow using saturation techniques: comparison with directly measured blood flow in isolated gastrocnemius muscle of the cat, Circ Res, 22, 679, 1968. 4. Sejrsen, P., Atraumatic local labelling of skin by inert gas: epicutaneous application of xenon-133, J Appl Physiol, 24, 570, 1968. 5. Sejrsen, P., Blood flow in cutaneous tissue in man studied by washout of xenon-133, Circ Res, 25, 215, 1969. 6. Kety, S.S., Quantitative measurement of regional circulation by the clearance of radioactive sodium, Am J Med Sci, 215, 352, 1948. 7. Kety, S.S., Measurement of regional circulation by the clearance of radioactive sodium, Am Heart J, 38, 321, 1949. 8. Jaszczak, P. and Sejrsen, P., Determination of skin blood flow by 133Xe washout and by heat flux from a heated tc-Po2 electrode, Acta Anaest Scand, 28, 482, 1984. 9. Sejrsen, P., Epidermal diffusion barrier to xenon-133 in man and studies of clearance of xenon-133 by sweat, J Appl Physiol, 24, 211, 1968. 10. Jacobsson, S., Studies of the blood circulation in lymphoedematous limbs, Scand J Plast Reconstruct Surg, Suppl. 3, 1, 1967.
11. Sejrsen, P., Diffusion processes invalidating the intraarterial krypton-85 beta particle clearance method for measurement of skin blood flow in man, Circ Res, 21, 281, 1967. 12. Braithwaite, F., Farmer, F.T., and Herbert, F.I., Observations on the vascular channels of tubed pedicles using radioactive sodium, III, Br J Plast Surg, 4, 38, 1951. 13. Alpert, J.S. and Coffman, J.D., Effect of intravenous epinephrine on skeletal muscle, skin, and subcutaneous blood flow, Am J Physiol, 216, 156, 1969. 14. Kövamees, A., Skin blood flow in obliterative arterial disease of the leg, Acta Chir Scand, Suppl. 397, 1, 1968. 15. Behnke, A.R. and Willmon, T.L., Cutaneous diffusion of helium in relation to peripheral blood flow and the absorption of atmospheric nitrogen through the skin, Am J Physiol, 131, 627, 1940/41. 16. Klocke, R.A., Gurtner, G.H., and Farhi, L.E., Gas transfer across the skin in man, J Appl Physiol, 18, 311, 1963. 17. Hensel, H., Messkopf zur Durchblutungsregistrierung an Oberflauachen, Arch Physiol, 268, 604, 1959. 18. Golenhofen, K., Die Hautdurchblutung des Menschen; Möglichkeiten zur Objektivierung von Hautreaktionen, Fette, Seifen Anstrichmittel, 3, 177, 1968. 19. Hardy, J.D. and Soderstrom, G.F., Heat loss from the nude body and peripheral blood flow at temperatures of 22 to 35C, J Nutr, 16, 493, 1938. 20. Stewart, H.J. and Evans, W.F., The peripheral blood flow under basal conditions in normal male subjects in the third decade, Am Heart J, 26, 67, 1943. 21. Cooper, K.E., Edholm, O.G., and Mottram, R.E., Blood flow in the skin and muscle of the human forearm, J Physiol (London), 128, 258, 1955. 22. Kontos, H.A., Richardson, D.W., and Patterson, J.L., Blood flow and metabolism of forearm muscle in man at rest and during sustained contraction, Am J Physiol, 211, 869, 1966. 23. Klemp, P. and Staberg, B., The effect of antipsoriatic treatment on cutaneous blood flow in psoriasis measured by 133Xe washout method and laser Doppler velocimetry, J Invest Dermatol, 85, 259, 1986.
86 Evaluation of Lymph Flow P.S. Mortimer St. George’s and Royal Marsden Hospitals, London, United Kingdom
CONTENTS 86.1 Introduction............................................................................................................................................................741 86.2 Background ............................................................................................................................................................742 86.2.1 Lymphangiography ....................................................................................................................................742 86.2.1.1 Fluorescence Microlymphangiography12 ...................................................................................742 86.2.1.2 Indirect Lymphography13............................................................................................................742 86.2.2 Lymphoscintigraphy ..................................................................................................................................742 86.2.2.1 Historical Perspective.................................................................................................................742 86.2.2.2 Lymph Transport Kinetics..........................................................................................................742 86.2.3 Skin Lymph Flow ......................................................................................................................................743 86.3 Object.....................................................................................................................................................................743 86.3.1 Radiolabeled Tracers .................................................................................................................................743 86.3.2 Procedure ...................................................................................................................................................743 86.3.3 Analysis of Data ........................................................................................................................................743 86.3.4 Reliability and Reproducibility .................................................................................................................744 86.3.4.1 Animal Studies ...........................................................................................................................744 86.3.4.2 Human Studies ...........................................................................................................................744 86.3.4.3 Studies in Pathological Skin ......................................................................................................745 86.4 Sources of Error.....................................................................................................................................................747 86.4.1 Tracer Migration ........................................................................................................................................747 86.4.2 Blood Clearance ........................................................................................................................................747 86.4.3 Injection Trauma........................................................................................................................................747 86.4.4 Injection Depth ..........................................................................................................................................748 86.4.5 Volume of Distribution ..............................................................................................................................748 86.4.6 Extrinsic Forces .........................................................................................................................................748 86.5 Correlation with Other Methods ...........................................................................................................................749 86.6 Recommendations..................................................................................................................................................749 References .......................................................................................................................................................................749
86.1 INTRODUCTION Few techniques exist for the functional assessment of skin lymphatics; yet, it is in this area of microcirculation research that most questions relating to the role of lymphatics in disease remain unanswered. The lymphatic vessels provide an important “limb” to the microcirculation of the skin and, together with the blood vessels, cater for the constant recirculation of protein- and lymph-borne cells, e.g., Langerhans cells1 and T lymphocytes. It is the essential function of the lymphatic system to return to the vascular compartment extravascu-
lar protein molecules, colloids, and particulate matter too large to reenter the blood capillaries directly.2 The rate at which labeled protein molecules or colloids are removed from the interstitial tissues has been regarded as an index of lymphatic function.3–5 Measurement of skin and subcutaneous lymph flow has employed the same principle of isotope clearance as measurement of skin and subcutaneous blood flow.6,7 However, the interpretation of the clearance of tracers from the skin in disease states appears difficult and unreliable.8,9 741
742
Handbook of Non-Invasive Methods and the Skin, Second Edition
86.2 BACKGROUND 86.2.1 LYMPHANGIOGRAPHY In vivo visualization of lymphatic vessels (lymphangiography) using x-ray contrast medium10 remains the gold standard for lymphatic vessel abnormalities. The technique, however, is invasive, difficult to perform, and provides only anatomical detail with no functional information. Only subcutaneous lymphatics as large, or larger than, collectors can be opacified, except in pathological circumstances when dermal backflow occurs and small skin lymphatics become visible. Intravital dyes, e.g., patent blue, used to delineate subcutaneous lymphatics prior to direct cannulation for x-ray lymphography, can be used to visualize dermal lymphatics but not capillaries.11 The results, however, are transitory. Two new methods of lymphangiography have been developed in recent years. 86.2.1.1 Fluorescence Microlymphangiography12 This technique enables the superficial lymph capillary network of the skin to be seen under the vital microscope by means of fluorescing macromolecules (FITC-Dextran, Sigma) injected subepidermally and cleared exclusively by lymphatics. Information regarding the morphology of lymphatic capillaries and precollectors (initial lymphatics) and the extent of tracer propagation within the dermal lymphatic network can be recorded on video for analysis. 86.2.1.2 Indirect Lymphography13 Indirect lymphography employs water-soluble nonionic xray contrast media that can be administered via an interstitial injection without recourse to direct access to lymphatics. Iotralan® or Iotasol® (Schering AG, Berlin) is infused by a motor pump into the skin; 2 to 3 ml injected intradermally leads to considerable local skin distention and is not without discomfort. Dermal and subcutaneous collecting lymphatics can be visualized by x-ray using the mammography film method. In the presence of incompetent valves and dermal backflow, initial lymphatics can also be seen. All lymphangiographic methods are limited in their ability to evaluate lymph flow, as the techniques are essentially for demonstrating the anatomy of lymph vessels.
86.2.2 LYMPHOSCINTIGRAPHY The development of lymphoscintigraphy was aimed mainly at imaging the lymphatic system and in particular the lymph nodes.14,15 Lymphoscintigraphy has proved more useful in the determination of lymph flow. Because of the close interrelationship between lymph formation
and flow, lymphoscintigraphy theoretically provides a much more comprehensive and functional examination of lymph drainage than does x-ray lymphography. 86.2.2.1 Historical Perspective The earliest studies involved measurement of lymph flow by external counting following the subcutaneous injection of 131I-labeled plasma protein into the hind limb of healthy dogs.16 The author, using an external scintillation counting technique, concluded incorrectly that the major route of the removal from the tissue spaces of crystalloid and protein molecules is via the blood capillary bed. Taylor et al.,3 using 131I-human serum albumin (HSA), concluded that the behavior of radioactive proteins injected subcutaneously was consistent with removal by the lymphatic route and that the rate of removal was slower in patients with lymphedema than in normal subjects. Hollander et al.4 studied patients with and without edema by 131I-HSA clearance from the subcutaneous tissue and concluded that the rate of removal was significantly reduced in edema caused by lymphatic obstruction, but significantly increased in edema caused by venous obstruction, congestive cardiac failure, or hypoproteinemia. Similar results were obtained by Sage et al.17 using radioactive gold (198Au), but the absorbed dose of radioactivity was unacceptably high. Emmett et al.18 stated that 131I-protein clearance studies are of little value for the initial assessment of individual patients with swollen limbs but may well prove to be the most sensitive method for evaluating the response to treatment. Although further studies on lymphatic tracer clearance were performed, their usefulness for clinical measurement was of some doubt. 86.2.2.2 Lymph Transport Kinetics Lymphoscintigraphy has proved to be a sensitive and specific method for the study of lymph transport kinetics19,20 and useful for repeat examinations.21 Computerization has allowed data on depot clearance, colloid transit, and nodal uptake to be correlated. Investigation times have been reduced owing to extrapolation of data collected over 1 to 2 hours. 86.2.2.2.1 Lymph Node Uptake Lymph node uptake has proved to be a more reliable measurement than tracer clearance.22 Most studies have employed a subcutaneous injection. Intradermal administration of tracer seems to encourage increased migration of tracer, possibly due to higher interstitial pressures in the dermis than the subcutis. 86.2.2.2.2 Transit Times The speed of passage of tracer along main lymphatic vessels immediately following the administration of tracer can be measured. Such transit times indicate patency of
Evaluation of Lymph Flow
lymph drainage routes as well as being an indirect measurement of lymph flow. 86.2.2.2.3 Fractional Removal Rate (Depot Clearance) The amount of an injected tracer deposited in a tissue principally decreases along an exponential curve, the slope of which is expressed as a clearance constant.23 The tracer must be freely diffusible through the tissue, and to equate with lymphatic clearance, the tracer must be removed solely through the lymphatic route.
86.2.3 SKIN LYMPH FLOW Measurement of lymphatic function specifically in the skin by the disappearance rate of 131I albumin from the dermis was first reported in 19708; 50 patients were investigated and radioactivity declined exponentially, giving a linear plot over 50 hours. Radioactivity fell more rapidly during the first 4 hours, giving an initial curve on a semilogarithmic plot. Results were reproducible and clearance rates varied according to the injection site. The only further study of skin lymph flow examined albumin clearance from psoriatic skin.9 Again, small quantities (0.1 ml) of 131I-HSA containing 10 mCi of radioactivity were administered intradermally, and the radioactivity of each depot was measured by sodium iodide scintillation detectors. Half clearance times were calculated by regression analysis by the least squares method. Clearance was shown to be monoexponential and increased in involved psoriatic skin, indicating increased lymph drainage.
86.3 OBJECT If interstitial protein clearance is the essential function of the lymphatic system, can skin lymph flow be measured reliably using the principle of isotope clearance and give meaningful results? A measure of lymphatic function is the efficiency of protein removal from the tissues. Lymph flow is considered here as equivalent to the movement of protein and accompanying fluid for the purpose of lymph drainage. (Flow refers to bulk transport per unit volume of tissue per unit time. Strictly, it is not possible to measure absolute lymph flow in vivo.) Skin lymph flow relies on several interdependent steps, which include lymph formation, its entry into lymph capillaries followed by transit through noncontractile lymphatics, and then propulsion by subcutaneous contractile lymphatics. Movement of solid matter need not necessarily relate to that of fluid, and lymph flow is driven by many extrinsic forces. Colloidal (or protein) clearance from the dermis does involve every step of skin lymph flow and is theoretically the ideal test. Perhaps a more appropriate
743
term would be skin lymph drainage instead of lymph flow. For this reason, lymph flow was expressed as a half clearance time (t1/2) and calculated from the slope of the exponential clearance curve.
86.3.1 RADIOLABELED TRACERS Macromolecules and colloids of a certain size are transported exclusively in the lymph. The ideal tracer is one that migrates freely from the injection site through the tissues and away in the lymphatics. It was discovered that the optimal colloid was one with a particle size of a few nanometers with a small dispersion around that value.24 Too large a particle resulted in poor absorption from the injection depot, and too small a particle risked blood clearance. 99mTc-labeled agents offer significant advantages in terms of radiation dose and energy characteristics. Much lower doses of radioactivity are possible with external scintillation counting, but image formation using scintiscanner or gamma camera demands higher radioactivity. In a study25 comparing 131I-HSA, 198Au, and 99mTccolloid (TCK 17, Cis), t1/2 values for 198Au were extremely long and variable. The clearance of 99mTc-colloid was slightly faster, but not significantly so, than 131I-has, and more consistent (Figure 86.1). To reduce leakage, the needle was inserted obliquely through the skin. Tracer was injected slowly with minimal force. The needle entry site was wiped once with a cotton wool swab after withdrawal of the needle.
86.3.2 PROCEDURE Injections of tracer were made into the dermis at a superficial level (subepidermal). An injection volume of 0.03 ml was used using a 30-gauge needle. Albumin may be more physiological, but the main function of the lymphatics is the removal of macromolecules, including exogenous colloids, from the interstitium. A well-collimated sodium iodide detector was positioned over the injection site with the collimator surface 10 mm from the skin. Additional detectors can be used to monitor other injection sites when paired studies are to be performed or to monitor uptake in the regional lymph nodes. Each detector was connected to a dual-channel interface analyzer operating in the multiscaler mode to give a digital output of radioactive counts with time. Counts were integrated over periods of 10 to 50 seconds and data recorded on a computer for a total time of 30 min. Observations up to periods of 90 min were not found to give any greater consistency.
86.3.3 ANALYSIS
OF
DATA
Changes in radioactive counts were plotted as the percentage of the maximum count rate against time. The resulting clearance curves were analyzed using a single exponential equation.25 The data points were fitted to this equation
744
Handbook of Non-Invasive Methods and the Skin, Second Edition
5
Half clearance times - T½ (hr)
4
3
2
1
0 Pig 1
Pig 2
Pig 1
99MTc-colloid
Pig 2 1311-HSA
99MTC-Colloid
1311-HSA
Cocktail
FIGURE 86.1 Variations in t1/2 values for the clearance of 99mTc-colloid with 131I-HSA (●) when injected as single tracers or together as a cocktail into the flank skin of the pig. The individual results of separated measurements on two pigs are given along with the mean ± SE for each set of results.
using a nonlinear least squares method. The rate of removal was corrected for the decay of the isotope (99mTc = 6 hours) and lymph flow expressed as a half clearance time (t1/2) in hours.
86.3.4 RELIABILITY
AND
REPRODUCIBILITY
86.3.4.1 Animal Studies The ICT proved to be a reliable and reproducible method when performed as a group test under controlled laboratory conditions in anesthetized large white pigs.25 Data were found to fit best to a monoexponential equation with good correlation coefficients (>0.86), indicating that the clearance was a mono- rather than biexponential function. This is in keeping with the previously published human work.8,9 Long investigation times are not only impractical, but fluctuations in lymph flow may be expected, particularly from movement in an unanesthetized animal. Differences were observed in lymph flow between the two pigs studied (Figure 86.1) and between skin sites within each animal. Differences in lymph vessel density and distribution and in tissue compliance were the likely explanations. Age was not a major factor in these studies, but has been recently shown to result in a decline in limb lymph drainage in humans.26
86.3.4.2 Human Studies As in the controlled studies performed in pig skin,25 similar studies performed in human skin revealed essentially monoexponential clearance.27 There was, however, very little consistency of t1/2 in repeat basal studies (basal = without interference, e.g., lymph flow enhancement), and serious doubts must be raised regarding the value of single lymph flow determinations at rest (basal). Fractional removal rates were not significantly different in pretibial skin compared with the skin of the thigh or foot. The wide range of t1/2 values witnessed in normal human skin differed from the reproducible results seen under controlled conditions in pig skin. This was considered to indicate real differences in lymph flow rather than technical error, particularly as results varied in repeat studies in the same subject on consecutive days, whereas right and left legs showed similarity when examined together. Lymph flow at rest is slow and subject to instant fluctuations, and its measurement is clearly prone to error unless carried out under strictly controlled conditions. All components of lymph movement depend upon changes in local tissue and hydrostatic pressure. These changes are produced by external compression,28 muscular activity,29,30 skin surface massage,31,32 passive movements,33 and local arterial pulsation.34 Movement of macromolecules from
Evaluation of Lymph Flow
745
1.0 0.9 0.8
Half clearance - T½ (hr)
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Control
Massage
Pig 1
Control
Massage
Pig 2
FIGURE 86.2 The effect of local massage on the half clearance time for 99mTc-colloid from skin on the flank of pigs. The individual results are given for separate measurements on two pigs (●, unmassaged sites; , massaged sites) along with the mean values ± SE.
interstitium toward, into, and through peripheral lymphatics would appear to be a predominantly passive process dependent on many extrinsic forces rather than an active process generated by the lymphatic itself. A study of lymph flow enhancement comparing vibration with local massage demonstrated greater colloid clearance from massage. The response to vibration was disappointing. A possible explanation for this failure would be that the rapid movement of vibration was too fast to allow adequate lymph vessel filling. Local massage significantly enhanced colloid clearance in both normal pig25 (Figure 86.2) and human skin (Figure 86.3). Massage performed some distance away in the leg from the injection depot according to the principle of manual lymphatic drainage35 invoked a significant increase in clearance superior to that generated by pneumatic compression therapy (Talley Medical Equipment Ltd.) (Figure 86.4). The proposed theory is that such massage has a milking or siphoning effect on distal lymph. Stimulation of the intrinsic contractility of the main lymphatic collecting vessels in the limb would pump lymph proximally, so generating a pressure gradient that draws lymph from peripheral lymphatics, including in the skin. The lack of efficacy from pneumatic compression therapy was a surprise as the 10-chamber inflatable garment produced a pressure wave moving repeatedly up the limb that
was far stronger than the massage. These machines are widely used for the treatment of lymphedema, but may do little to mobilize protein via lymphatics.36 Comparison of colloid clearance from normal skin on the dorsum of the foot with the same site in a lymphedematous leg showed no difference under basal conditions with the subject supine. Clearance of 99mTc-colloid from the dermis, as a measure of skin lymph flow, could only differentiate normal subjects from patients with lymphedema by the response to massage. Single lymph flow determinations using 99mTc-colloid clearance from the dermis over a short investigation time are therefore only meaningful when attempts are made to enhance lymph flow and so test lymph transport capacity. Massage by stimulating lymph flow exposes the deficiency in lymph transport, which examination at rest may miss. Only then can lymphatic insufficiency be distinguished from normal function. 86.3.4.3 Studies in Pathological Skin Because edema and connective tissue changes are wellknown sequelae of lymphatic damage, and because such changes occur following radiation to the skin, lymph flow studies using the ICT (99mTc-colloid) were undertaken in pig skin following single doses of x-rays.37 Paired estimates of lymphatic clearance were performed in irradiated
746
Handbook of Non-Invasive Methods and the Skin, Second Edition
10
8
T½ (hrs)
6
4
2
0 Basal
Vibration
Basal
Stimuloval
Basal
No vibration/ stimuloval
FIGURE 86.3 The clearance of 99mTc-colloid from normal skin of the lower leg expressed as t1/2, in normal subjects at rest (basal), and following a period of lymph flow enhancement (vibration or surface massage).
35
30
T½ (hrs)
25
20
15
10
5
0 Basal
PT101
Basal
Manual massage
FIGURE 86.4 The clearance of 99mTc-colloid, expressed as t1/2, from the skin of the dorsum of the foot during rest, basal, (0 to 30 min post-injection) and following a period (30 to 60 min) of lymph flow enhancement with either pneumatic compression (PT101) or manual massage.
and unirradiated sites on the flank skin of anesthetized large white pigs at 3, 6, 9, 12, 26, 39, 52, 64, and 78 weeks after a single dose of 18 Gy of x-rays. The results demonstrated good consistency of results relative to site and time examined. The results demonstrated two waves of
impaired lymphatic clearance with time, which correlates well with the gross morphological changes observed (Figure 86.5). It was concluded that impaired lymphatic drainage probably contributes to the gross and histological changes observed in the skin following x-irradiation. The
Evaluation of Lymph Flow
747
6
Half clearance time - T½ (hr)
5
4
3
2
1
0
0
12
24
39 54 Time after irradiation (weeks)
66
78
FIGURE 86.5 Time-related changes in the clearance of 99mTc-colloid from the dermis of the flank of the pig following irradiation with a dose of 18Gy of x-rays. Results are expressed as the mean half clearance items (± SE) for irradiated (●) and nonirradiated () fields. The shaded area represents the mean t1/2 value (± SE) in normal skin. Error bars indicated.
study demonstrated the value of the ICT for lymph flow measurements even in pathological skin, providing experimental conditions are well controlled.
86.4 SOURCES OF ERROR The advantage of the ICT is that it utilizes the principal function of the lymphatic, namely, the removal of protein or colloid from the interstitium, and examines it dynamically. As such, it explores the capacity of the lymphatic system to absorb material from the interstitial space and transport it to the regional node. This corresponds to the route taken by protein leaked from blood capillaries and subsequently drained from the interstitium by the lymphatics. Of the whole circulating plasma protein pool, 50 to 100% leaves the circulation daily. It is the lymphatic system that maintains this “extravascular circulation of plasma proteins.”38
86.4.1 TRACER MIGRATION Flow can be calculated from clearance provided that (1) the tracer leaves by only one route and (2) it reaches an instant diffusion equilibrium between lymph and tissue. The rate-limiting step for measuring lymph flow is almost certainly the poor migration of tracers from the injected site. Injected proteins behave differently from native
plasma proteins. When an isotope in its colloidal form is injected into the tissues, approximately 90% is precipitated on the local tissue proteins. Only a small proportion is attached to the mobile proteins and to phagocytes, and so taken up by the lymphatics.39 This obviously limits the measurable quantity of 99mTc-colloid available for clearance. In the studies described, an average of 15% of colloid injected was absorbed by lymphatics in the first 30 min.
86.4.2 BLOOD CLEARANCE Isotope clearance, as a measure of lymph flow, has been criticized because of the risk of significant amounts of radioactivity escaping by the bloodstream, thus invalidating clearance values interpreted as lymph flow. This problem was investigated by comparing blood clearance of 99m Tc-colloid with its total (lymphatic) clearance. 25 Results revealed that the percentage of tracer cleared by the bloodstream was never more than 1.5% of total clearance. Lymphoscintigraphic studies confirmed that blood clearance was negligible.
86.4.3 INJECTION TRAUMA Injection into the skin is obviously traumatic and nonphysiological. The insertion of a needle will undoubtedly on
748
Handbook of Non-Invasive Methods and the Skin, Second Edition
10 9
Half clearance times - T½ (hr)
8 7 6 5 4 3 2 1 0 S/E
S/C
S/E
D/D
Depth of injection
FIGURE 86.6 Variations in the t1/2 values for the clearance of 99mTc-colloid from the normal flank of the pig following injections at different depths. Injections were subepidermal (S/E, ) subcutaneous (S/C, ●), and deep-dermal (D/D, ). The individual resuls of separate measurements are given along with the mean ± SE for each set of results..
occasions disrupt a lymph capillary mesh 600 to 1000 mm wide. Disruption of blood capillaries is more likely, but bleeding at the entry point can be minimized if care is taken. Obvious bleeding should lead to termination of the measurements. Laceration of the lymph capillary network effectively results in direct injection into lymphatics with increased uptake and filling of the vessels.11 Nevertheless, this clearly does not prevent continuation of satisfactory clearance, and with good injection techniques the disturbance is both minimal and consistent.40
86.4.4 INJECTION DEPTH Studies of injection depth in pig skin demonstrated the importance of an accurate injection for dependable results.25 Subepidermal localization of the tracer produced faster and more consistent clearance (Figure 86.6). Similar results were observed in blood flow studies using the same technique,7 where clearance rates correlated with local vascular density. The much denser network of lymphatic capillaries existing in the subpapillary region of pig and human skin41 provides a greater surface area for absorption and would satisfactorily explain the faster subepidermal clearance of colloid than deep dermal and subcutaneous sites.
86.4.5 VOLUME
OF
DISTRIBUTION
Consideration must also be given to injection pressure and volume when interpreting results. A change in injection volume did not significantly influence t1/2.25 Clearance rates were slower with a larger volume. However, the clearance rate (KT) is only proportional to lymph flow (FL) when the volume of distribution (Vi) remains the same. Therefore, no change in t1/2 despite an increase in volume suggests an increased lymph flow. This is possibly the most serious source of error when interpreting clearance rates as lymph flow. Clearly, therefore, it is not possible to compare directly clearance from normal skin with lymphedematous skin because Vi remains unknown.
86.4.6 EXTRINSIC FORCES The differences in the consistency of lymph flow measurements performed in pig skin compared with the wide variation seen in human skin were considered a reflection of the influences of extrinsic forces. By controlling through the laboratory conditions for ambient temperature active and passive movements and pulse rate, these problems were largely overcome. This is not easily possible in human studies unless the extrinsic forces are specifically used to enhance lymph flow. Only by increasing lymph flow in response to standardized stimuli, e.g.,
Evaluation of Lymph Flow
massage, could impaired lymph drainage from the skin be detected.25
86.5 CORRELATION WITH OTHER METHODS The scintillation detector system with simultaneous measurement of two comparative sites, or depot clearance and nodal uptake, provides a portable and low-radiation method for the functional assessment of peripheral lymphatics in humans. The relative, simple, inexpensive equipment permits the technique to be used at the bedside. This has benefits in centers without gamma camera or whole-body scintillation scanners. An increased number of external detectors connected in series at intervals along the lymph drainage route from the injection site could be used to improve the sensitivity of the technique, particularly in relation to speed of lymph movement. Small detectors strapped to the limb in a method similar to that of pressure transducers would be a possibility. Isotope clearance is the only method currently available that provides objective and dynamic information on skin lymph flow. Most clinical methods that examine the lymphatic system — histology and electron microscopy, lymphangiography, and lymphoscintigraphy — either focus on large lymphatic vessels outside the skin or provide static, anatomical, or structural detail. Lymphoscintigraphic studies supported the results of the external detector studies and demonstrated that clearance of approximately 99% of the tracer was lymphatic. Blood clearance and tracer diffusion were negligible in normal skin. Results using external scintillation detectors were in broad agreement with the published findings using a gamma camera, although most examinations employed a subcutaneous injection of tracer. Gamma camera studies demand a 20-fold greater dose of radioactivity, but nevertheless still fall within category I for radiation risk. External detector studies permit multiple repeat lymph flow measurements with safety. Gamma camera measurements of isotope clearance from the injection site have been considered unreliable, but Pecking et al.21,42 in extensive studies has shown significant differences between lymphedema and normal limbs, according to clearance, with good reproducibility.
86.6 RECOMMENDATIONS Physiological and clinical measurement of the microcirculation is important for understanding the dynamic changes that occur with pathology. So often functional questions are answered incorrectly by extrapolation of data from static studies. The ICT has the major advantage that it utilizes the principal function of the lymphatic,
749
namely, the removal of protein or colloid from the interstitium, and examines it dynamically. Skin lymph flow can only be reliably measured when conditions are controlled. For lymph flow at rest, this means controlling for all extrinsic influences as well as intersite and intersubject variations. Extrinsic factors such as massage strongly influence lymph flow. Greater sensitivity in detecting lymphatic insufficiency may be achieved if a standardized stimulus to lymph flow is administered. The response in clearance to the lymph flow enhancement will be the best indicator of lymph drainage abnormalities.
REFERENCES 1. Silberberg-Sinakin, I., Thorbecke, G.J., Baer, R.L., Rosenthal, S.A., and Berezowsky, V., Antigen bearing Langerhans cells in skin, dermal lymphatics and in lymph nodes, Cell. Immunol., 25, 137, 1976. 2. Drinker, C.K. and Field, M.E., The protein content of mammalian lymph and the relation of lymph to tissue fluid, Am. J. Physiol., 97, 32, 1931. 3. Taylor, G.W., Kinmonth, J.B., Rollinson, E., and Rotblat, J., Lymphatic circulation studied with radioactive plasma protein, Br. Med. J., 1, 133, 1957. 4. Hollander, W., Reilly, P., and Burrows, B.A., Lymphatic flow in human subjects as indicated by the disappearance of 131I-labelled albumin from the subcutaneous tissue, J. Clin. Invest., 40, 222, 1961. 5. Fernandez, M.J., Davies, W.T., Owen, G.M., and Tyler, A., Lymphatic flow in humans as indicated by the clearance of 125I-labelled albumin from the subcutaneous tissue of the leg, J. Surg. Res., 35, 101, 1983. 6. Engelhart, M. and Kristensen, J.K., Evaluation of cutaneous blood flow responses by 133-xenon washout and a laser Doppler flowmeter, J. Invest. Dermatol., 80, 12, 1983. 7. Young, C.M. and Hopewell, J.W., The evaluation of an isotope clearance technique in the dermis of pig skin: a correlation of functional and morphology parameters, Microvasc. Res., 20, 182, 1980. 8. Ellis, J.P., Marks, R., and Perry, B.J., Lymphatic function: the disappearance rate of 131I albumin from the dermis, Br. J. Dermatol., 82, 593, 1970. 9. Staberg, B., Klemp, P., Aasted, M., Worm, A.M., and Lund, P., Lymphatic albumin clearance from psoriatic skin, J. Am. Acad. Dermatol., 9, 857, 1983. 10. Kinmonth, J.B., Lymphangiography in man, Clin. Sci., 11, 13, 1952. 11. Hudack, S.S. and McMaster, P.D., Lymphatic participation in human cutaneous phenomena, J. Exp. Med., 57, 751, 1933. 12. Bollinger, A., Jager, K., Sgier, F., and Seglias, J., Fluorescence microlymphography, Circulation, 64, 1195, 1981. 13. Partsch, H., Wenzel-Hora, B., and Urbank, A., Differential diagnosis of lymphoedema after indirect lymphography with Iotasul, Lymphology, 16, 12, 1983.
750
Handbook of Non-Invasive Methods and the Skin, Second Edition
14. Battezzati, M. and Donini, I., The use of radioisotopes in the study of the physiopathology of the lymphatic system, J. Cardiovasc. Surg., 5, 691, 1964. 15. Anghileri, L.J., Lymph nodes distribution of several radio colloids: migration ability through the tissues, J. Nucl. Biol. Med., 11, 180, 1967. 16. Jepson, R.P., Simeone, F.A., and Dobyns, B.M., Removal from skin of plasma protein labelled with radioactive iodine, Am. J. Physiol., 175, 443, 1953. 17. Sage, H.H., Sinha, B.K., Kizilay, D., and Toulon, R., Radioactive colloidal gold measurements of lymph flow and functional patterns of lymphatics and lymph nodes in the extremities, J. Nucl. Med., 5, 626, 1964. 18. Emmett, A.J., Barron, J.N., and Veall, N., The use of 131I-albumin tissue clearance measurements and other physiological tests for the clinical assessment of patients with lymphoedema, Br. J. Plast. Surg., 20, 1, 1967. 19. Kleinhaus, E., Baumeister, R., Hahn, D., Siuda, S., Bull, U., and Moser, R., Evaluation of transport kinetics in lymphoscintigraphy. Follow-up study in patients with transplanted lymphatic vessels, Eur. J. Nucl. Med., 10, 349, 1985. 20. Stewart, G., Gaunt, J., Croft, D.N., and Browse, N.L., Isotope lymphography: a new method of investigating the role of lymphatics, Br. J. Surg., 72, 906, 1985. 21. Pecking, A., Cluzan, R., Desprez-Curely, J.P., and Guerin, P., Functional study of the limb lymphatic system, Phlebology, 1, 129, 1986. 22. Mostbeck, A., Kahn, P., and Partsch, H., Quantitative lymphography in lymphoedema, in The Initial Lymphatics, Bollinger, A., Partsch, H., and Wolf, J.H., Eds., Thieme Verlag, Stuttgart, 1985, p. 123. 23. Kety, S.S., Measurement of regional circulation by the clearance of radioactive sodium, Am. Heart J., 38, 321, 1949. 24. Strand, S.E. and Persson, B.R., Quantitative lymphoscintigraphy. I. Basic concepts for optimal uptake of radiocolloids in the parastemal lymph nodes of rabbits, J. Nucl. Med., 20, 1038, 1979. 25. Mortimer, P.S., Simmonds, R., Rezvani, M., Robbins, M., Hopewell, J.W., and Ryan, T.J., The measurement of skin lymph flow by isotope clearance: reliability, reproducibility, injection dynamics, and the effect of massage, J. Invest. Dermatol., 95, 677, 1990. 26. Bull, R.H., Gane, J., Evans, J., Joseph, A., and Mortimer, P.S., Abnormal lymph drainage in patients with chronic venous leg ulceration, J. Am. Acad. Dermatol., in press. 27. Mortimer, P.S., Measurement of Skin Lymph Flow by an Isotope Clearance Technique, M.D. thesis, University of London, London, 1990.
28. Miller, G.E. and Seale, J.L., The mechanics of terminal lymph flow, J. Biomech. Eng., 107, 376, 1985. 29. Yoffey, J.M. and Courtice, F.G., Lymphatics, Lymph and Lymphomyeloid Complex, Academic Press, New York, 1970. 30. Barnes, J.M. and Trueta, J., Absorption of bacteria, toxins and snake venoms from the tissues, Lancet, I, 623, 1941. 31. Calnan, J.S., Pflug, J.J., Reis, N.D., and Taylor, L.M., Lymphatic pressures and the flow of lymph, Br. J. Plast. Surg., 23, 305, 1970. 32. Olszewski, W.L., Peripheral Lymph: Formation and Immune Function, CRC Press, Boca Raton, FL, 1985. 33. Jacobsson, S., Lymph flow from the lower leg in man, Acta Chir. Scand., 133, 79, 1967. 34. Parsons, R.J. and McMaster, P.D., The effect of the pulse upon the formation and flow of lymph, J. Exp. Med., 68, 353, 1938. 35. Stijns, H.J. and Leduc, A., The contribution of physical therapy in the treatment of lymphoedema, in Lymphoedema, Clodius, L., Ed., Thieme, Stuttgart, 1977, p. 27. 36. Partsch, H., Mostbeck, A., and Leitner, G., Experimentelle untersuchungen zur wirkung einer druckwellenmassage (lymphapress) bein lymphodem, Lymphologie, V, 35, 1981. 37. Mortimer, P.S., Simmonds, R.H., Rezvani, M., Robbins, M.E., Ryan, T.J., and Hopewell, J.W., Time related changes in lymphatic clearance in pig skin after a single dose of 18Gy of X rays, Br. J. Radiol., 64, 1140, 1991. 38. Mayerson, H.S., The physiologic importance of lymph, in Handbook of Physiology, Vol. 2, Hamilton, W.F. and Dows, P.G., Eds., Waverly, Baltimore, 1963, p. 1035. 39. Haagensen, C.D., Methods of study of the lymphatic system, in The Lymphatics in Cancer, Haagensen, C.D., Ed., W.B. Saunders, Philadelphia, 1972, p. 14. 40. Courtice, F.C., Lymph and plasma proteins: barriers to their movement throughout the extracellular fluid, Lymphology, 4, 9, 1971. 41. Mortimer, P.S., Jones, R.L., and Ryan, T.J., Human skin lymphatics: regional variation and relationship to elastin, in Immunology and Haematology Research: Progress in Lymphology, Vol. 2, Heim, L., Ed., Immunology Research Foundation, Inc., Nearburgh, 1984, p. 59. 42. Pecking, A., Cluzan, R., Desprez-Curely, J.P., and Guerin, P., Indirect lymphoscintigraphy in patients with limb oedema, Phlebology, 1, 211, 1986.
Temperature and Thermoregulation
and Handheld Devices for 87 Sensors Surface Measurement of Skin Temperature Roderick A. Thomas Snell International, Tyn-Y-Coed, Pontardulais, Swansea, United Kingdom
CONTENTS 87.1 87.2 87.3 87.4 87.5
Introduction............................................................................................................................................................753 Sensors and Handheld Devices for Surface Measurement of Skin Temperature ................................................754 Temperature Measurement ....................................................................................................................................754 Measurement Location ..........................................................................................................................................754 Contact Temperature Measurement.......................................................................................................................754 87.5.1 Thermocouples...........................................................................................................................................754 87.5.2 Thermometers ............................................................................................................................................754 87.6 Noncontact Temperature Measurement.................................................................................................................756 87.6.1 Infrared Theory..........................................................................................................................................756 87.6.2 Infrared Thermometers ..............................................................................................................................756 87.6.3 Fixed Monitoring Systems ........................................................................................................................757 87.6.4 Infrared Radiometers .................................................................................................................................757 87.6.5 Mechanical Scanning and Focal Plane Arrays..........................................................................................758 87.6.6 Detector Arrays..........................................................................................................................................760 87.7 Detectors ................................................................................................................................................................760 87.7.1 Photon Detectors........................................................................................................................................760 87.7.2 Short-Wave and Long-Wave Photon Detectors.........................................................................................761 87.7.3 Photovoltaic Detectors...............................................................................................................................761 87.7.4 Photoconductive Detectors ........................................................................................................................762 87.7.5 Quantum Well Infrared Photodetectors.....................................................................................................762 87.7.6 Thermal Detectors .....................................................................................................................................763 87.8 Emerging Technology............................................................................................................................................764 87.8.1 Computer Systems .....................................................................................................................................764 87.8.2 Expert Systems ..........................................................................................................................................764 87.9 The Future..............................................................................................................................................................764 Useful Terms ...................................................................................................................................................................765 Acknowledgments ...........................................................................................................................................................767 References .......................................................................................................................................................................767
87.1 INTRODUCTION With the current technological advancements associated with infrared (IR) thermography resulting in the development of a number of different thermal imaging devices
(radiometers), the operation of which is dictated by the type of detector used, this section sets out to introduce the main infrared systems and the operation of detectors therein.
753
754
Handbook of Non-Invasive Methods and the Skin, Second Edition
87.2 SENSORS AND HANDHELD DEVICES FOR SURFACE MEASUREMENT OF SKIN TEMPERATURE The establishment, development, and consequential success of a medical infrared thermographic (MIT) intervention is primarily based on the understanding of the following: 1. Problem/condition to be monitored 2. Setup and correct operation of infrared system 3. Appropriate conditions during the monitoring process 4. Evaluation of activity and development of standards and protocol In 2, setup and correct operation of infrared system, an understanding of the various methods of skin temperature is useful. Also adopting the correct training is critical to the efficacy of monitoring.
87.3 TEMPERATURE MEASUREMENT Traditional temperature measurement of the internal body and the external skin is an important medical practice. A thermal balance at a temperature of 37°C (±0.75°C) must be maintained for optimum cellular function. Physiological thermoregulation is a complex central and peripheral interaction that attempts to balance body heat against loss. For example, heat production is a result of motor activity, shivering, and metabolic thermoregulation, while heat loss/transfer includes conduction, convection, radiation, and evaporation: • • • •
Conduction through objects by direct contact Convection through air or liquid as contact medium Radiation through space without contact Evaporation through liquid, then to air
Body heat is dynamic, moving across tissue boundaries. Heat transfer occurs when there is a difference in heat content of adjacent areas. The sum of differences between one area and another is known as a gradient. Therefore, what is actually measured, as skin temperature, is energy in motion in search of equilibrium from warmer to cooler. Temperature measurement of skin is captured using a number of different devices, as described in the next section. However, skin temperature is often captured at a moment in time and, for example, with a thermometer positioned in such a way that it can capture the transfer of heat from one space to another — at the same time minimizing the effects of the surrounding environment. An infrared ear thermometer, for example, measures radiated heat in the form of infrared energy to the space around it.
Cold junction T1 + ΔV
Hot junction T2
T + ΔT
ΔV = αΔT
Where α = Seebeck coefficient
FIGURE 87.1 Seebeck effect.
87.4 MEASUREMENT LOCATION The location of the temperature measurement device is dependent upon the condition being monitored. There are 13 different locations on the human body where clinical temperature is measured. Some are internal and some external: oral cavity, rectum, axilla, ear, tympanic membrane, nasopharynx, inguinal area, forehead, esophagus, pulmonary artery, bladder, vagina, and the great toe. Table 87.1 illustrates advantages and disadvantages of various devices and locations.
87.5 CONTACT TEMPERATURE MEASUREMENT 87.5.1 THERMOCOUPLES Thermocouples function due to the Seebeck effect, which is a combination of the Peltier effect, by which a small voltage exists at the junction of two unlike metals, and a second effect credited to Lord Kelvin, which produces a small voltage along a conductor in a temperature gradient. The Seebeck effect is an emf ΔV generated due to a temperature difference ΔT (Figure 87.1). Both effects are proportional to the temperatures involved. The total voltage produced in a circuit, including a number of thermocouples, is zero, as there is no temperature difference around the loop. Thus, two thermocouples are normally employed. One is maintained at a reference temperature (e.g., the freezing point of water), and the other acts as the thermometer. The sensitivities of common thermocouples range from 6.5 to 80 μV/dC, with accuracies from 0.25 to 1%. Several thermocouples can be arranged in series to form a thermopile to increase the sensitivity. The advantages of thermocouples are their relatively fast response (down to 1 ms), small size (down to 12 μm diameter), ease of fabrication, and long-term stability. Their disadvantages are small output voltage, low sensitivity, and need for reference temperature. Small thermocouples can be inserted into catheters and hypodermic needles. Some clinical electronic thermometers employ thermocouples.
87.5.2 THERMOMETERS There are a number of devices available; some are more common than others:
Sensors and Handheld Devices for Surface Measurement of Skin Temperature
755
TABLE 87.1 Advantages and Disadvantages of Traditional Temperature Measurement Devices Site
Devices Used
Advantages
Oral cavity
Glass mercury Electronic predictive Phase change (dot matrix)
Easy access Familiar Minimally invasive
Rectum
Glass mercury Electronic predictive Phase change (dot matrix) Deep rectal probe Glass mercury Electronic predictive Phase change (dot matrix) Electronic sensor
Preferred by MDs
Axilla
Disadvantages Affected by eating, drinking, etc. Temperature varies within oral cavity Hard to keep thermometer in place Site records highest temp in body; lags behind other core sites when temp is changing rapidly
Ear
Infrared ear thermometer
Tympanic membrane Nasopharynx
Contact electronic sensor
Easy access Familiar Minimally invasive Preferred by American Academy of Pediatrics for use in infants Easy access Familiar Minimally invasive Two sites available Reflective of brain temperature Reflective of brain temperature
Electronic sensor
Reflective of brain temperature
Affected by breathing Invasive and uncomfortable
Glass mercury Electronic predictive Phase change (dot matrix) Electronic sensor Phase change (dot matrix) Electronic sensor
Easy access in infants and small children
Skin temperature Requires leg to be drawn up against abdomen Normal range not well documented Skin temperature Affected by environment
Esophagus
Electronic sensor
Pulmonary artery Bladder
Electronic sensor
Reflects temperature of body core Reflects temperature of body core Reflects temperature of body core
Groin
Forehead
Electronic sensor
Reflects skin temperature Not always a good indicator of core temperature Must be held in place Takes long time to reach equilibrium
Easy access Noninvasive
Vagina
Glass mercury Electronic
Preferred for basal temperature
Great toe
Electronic sensor
Easy access Noninvasive Can be informative if used with core temperature
Uses Most common site in adults and children over 5 Often requested by MDs as the most accurate site for core temperature Most common site in children under 5 Sometimes used during surgery
Requires thorough training and attention to technique
Commonly used in hospitals and clinics
Invasive and uncomfortable
Used during anesthesia Used during anesthesia, but not very common Used in infants and neonates
Temperature varies according to depth of probe placement Affected by temperature of infused fluids Affected by amount of throughput Lags behind other core body temperature sites Invasive
Peripheral skin temperature very remote from body core
Used during surgery for rough monitoring Used during anesthesia Used in surgery and critical care Used in surgery and emergency or critical care Usually used by women tracking their fertility Used during surgery to reflect peripheral circulation and temperature
From www.graduateresearch.com/thermometry/sites.htm.
•
•
Liquid-in-glass thermometer, which relies upon the expansion of a liquid or solid as the temperature rises. Mercury in glass is best known, and there are a number of variations, notably the maximum reading clinical thermometer. Digital thermometers.
• • • •
Electronic sensor/thermometers. Chemical thermometers. Dot matrix or phase change thermometers. Radiation thermometers, infrared ear thermometers.
756
Handbook of Non-Invasive Methods and the Skin, Second Edition
87.6 NONCONTACT TEMPERATURE MEASUREMENT There are essentially three types of infrared thermographic temperature measurement equipment:4 spot thermometers (sometimes known as single point devices), fixed monitoring (sometimes known as line scanners), and portable infrared thermal imaging cameras (including radiometers). The progressing development of high-quality imaging optics and novel IR detectors has substantially improved the thermal and spatial resolution of thermal imagers, resulting in excellent-quality images with quantitative information. The basis for accurate quantitative measurements is reliability, repeatability, and comparability of data. In the case of temperature measurements, as a consequence of Europeanization and globalization, the trend toward generally acknowledged quality assurance in accordance with ISO/IEC 17025 and the worldwide equivalence of measurements require traceability to the SI unit of temperature, the Kelvin, according to the International Temperature Scale of 1990 (ITS-90).
where E = energy (J) h = Planck’s constant (6.625 × 10–34 Js) ƒ = frequency (Hz) c = speed of light (m/sec) λ = wavelength (m) Planck derived (empirically) a formula to relate the radiated power spectral density from a blackbody radiating into cold space at any temperature. Bλ(T) is the energy in Joules emitted per second per unit wavelength from 1 m2 of a perfect blackbody at a temperature T (Kelvin):
( )
Bλ T =
2 πhc 2 / λ 5 e hc / λkT − 1
(87.2)
where h = Planck’s constant (6.625 × 10–34 Js) k = Boltzmann’s constant (1.3804 × 10–23 J/K) λ = wavelength (m) c = speed of light (3 × 108 m/s) T = temperature (K)
87.6.1 INFRARED THEORY In most objects, at a temperature above absolute zero (0 Kelvin or –273.16°C), every atom and every molecule vibrate. According to the laws of electrodynamics, a moving electric charge is associated with a variable electric field that in turn produces an alternating magnetic field. This vibration produces an electromagnetic wave that radiates from the body at the speed of light. A blackbody is defined as an object that absorbs all radiation that impinges on it at any wavelength. Kirchoff’s law states that any body that is capable of absorbing all radiation is equally capable of the emission of radiation. An example of a blackbody radiator is a lightproof box with a small hole with the following characteristics: •
•
Any radiation that enters the hole is scattered and absorbed by continued reflection from within the box, resulting in almost no energy escaping. The box when heated becomes a cavity radiator that radiates energy, the characteristics of which are determined only by the temperature.
Planck proposed that the frequency of the radiation emitted, and hence the energy, should be quantized to specific values determined by the frequency:
E = hf =
hc [J ] λ
(87.1)
Even though the radiation energies are quantized, there are so many that they continue into the microwave band. The most probable frequency is determined by equating to zero the first derivative (with respect to λ) of Planck’s equation:
λm =
2898 [μm ] T
(87.3)
This is known as Wein’s law. It states that the higher the temperature, the shorter the radiated wavelength. There is a known relationship between the surface temperature of an object and its radiant power. This principle makes it possible to measure the temperature of a body without physical contact with it. Medical thermography is a technique whereby the temperature distribution of the surface of the body is mapped within a few tenths of a Kelvin. The human skin approximates to within 1% of a blackbody radiator, and so a radiation thermometer can accurately detect the temperature of the skin.
87.6.2 INFRARED THERMOMETERS These devices are designed to yield the average amount of infrared energy (often an average temperature) over a small area, often referred to as a single point. The size being sensed depends on the design of the instrument’s optics and the distance from the surface to be measured (Figure 87.2).
Sensors and Handheld Devices for Surface Measurement of Skin Temperature
0.6 cm @ 0.61 m
2.22 cm @2m
87.6.4 INFRARED RADIOMETERS
5.55 cm @5m
S
0.24 in @ 2 ft
0.87 in @ 6.6 ft
757
2.18 in @ 16.4 ft
D
FIGURE 87.2 Distance-to-spot size (D:S) ratio.
The optical system of an infrared thermometer collects the infrared energy from a circular measurement spot and focuses it on the detector. Optical resolution is defined by the ratio of the distance from the instrument to the object compared to the size of the spot being measured (D:S ratio). The larger the ratio’s number, the better the instrument’s resolution, and the smaller the spot size that can be measured. The laser sighting included in some instruments only helps to aim at the measured spot. Examples of some infrared thermometers are illustrated in Figure 87.3.
87.6.3 FIXED MONITORING SYSTEMS Fixed monitoring systems consist of scanning using a thermal image camera capable of producing surface temperature profiles of objects, producing a picture made up of hundreds of points across the detector’s surface. Line scanners provide a single-dimensional view or line of comparative radiation. There are also examples of thermal imaging in surgery.1,2 Recently, increased awareness in this field due to development associated with the charge coupled device (CCD) camera, has greatly reduced in detector physical size and now extends vision accurately through the smallest of openings, with the added capabilities of recording and trending. These charge coupled devices also form the basis of electronic optical systems for noncontact visual measurement and inspection in medical applications.
FIGURE 87.3 Examples of modern infrared thermometers.
Infrared thermal imagers are instruments for detecting, measuring, and recording the thermal emissions from a surface without contacting it, at a safe distance and at varying speeds. The term radiometer is generally, though not always, applied to devices that measure infrared radiation. Figure 87.4 illustrates a number of these devices, many of which resemble conventional video camcorders. There are currently a myriad of infrared radiometers available operating within different wavelengths of the infrared spectrum, dependent primarily on the type of detector used (Figure 87.5). These devices can for the benefit of this section be classified as follows: • • •
Long-wave infrared (LWIR; 7.5 to 14 μm) Mid-wave infrared (MWIR; 3 to 5 μm) Short-wave infrared (SWIR; 0.9 to 1.7 μm)
The optimum wavelength of an infrared radiometer is therefore determined by the wavelength distribution of the emitted radiation and type of detector. Another consideration is the transparency of the atmosphere to the transmission of infrared radiation between radiometer and object. At particular wavelengths there is a lack of radiative transparency within segments of the infrared spectrum.3 There are high levels of infrared transparency, at 3 to 5 μm (with poor transmission at 4.2 μm, due to carbon dioxide absorption) and 7.5 to 14 μm. This seems to coincide as the two common wavelengths of a number of different radiometers, although not exclusively. For thermal measurements over short ranges, such as those found in the examining room, laboratory, or operating theater, it is possible to work outside these wavelengths. As well as two common wavelengths there are two types of infrared radiometer: mechanical scanning and focal plane arrays (FPAs). Mechanical scanners are slowly being superseded by focal plane arrays for a number of reasons, partly highlighted in Table 87.2. Plassman and Jones5 have identified that there are currently a number of possible deficiences surrounding some thermal imaging cameras, which need to be
758
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 87.4 Examples of modern infrared radiometers. (From Snell Infrared.)
Gamma rays 0.1 A
X-rays 1A
Ultraviolet
1U A 100 A 0.1μ
Radio EHF SHF UHF VHF HF
Infrared 1μ
10 μ 100 μ 0.1 cm
1 cm 10 cm 1 m
MF
LF
VLF
10 m 100 m 1 km 10 km 100 km Wavelength
Infrared 0.4
0.6
0.8
1
1.5
Measurement 2
3
Region 4
6
8
10
15
20
30 Wavelength μm
FIGURE 87.5 Infrared measurement region. (From Raytek.)
monitored especially when used in medical applications. Summarized these are:
• •
• •
•
Specifications — offset ±2°C even after manufacturer calibration. Stability • Short term: drift after internal calibration possible. • Medium term: even after switch-on time specified by manufacturer. • Long term: over several hours/days. Range — fluctuation of around ±1°C possible within the “human” range.
Uniformity — optical limitations typically ±0.5°C. Scene — flooding effect ±0.2°C.
87.6.5 MECHANICAL SCANNING ARRAYS
AND
FOCAL PLANE
The design of thermal imaging cameras has moved away from mechanical scanning technology to focal plane arrays. All radiometers have a detector array or array detectors, and optics to form an image on the detector (Figure 87.6). The term staring array is also associated with FPAs, but refers specifically to the use of array detectors, each of which looks at one point of the total image (Figure 87.7).
Sensors and Handheld Devices for Surface Measurement of Skin Temperature
759
TABLE 87.2 Radiometer Types: Advantages and Disadvantages
Advantages
Disadvantages
Mechanical Scanner (Cooled Detector)
Focal Plane Array (Cooled Detector)
Microbolometer (Uncooled Detector)
Uses a few relatively cheap detectors Cooler tends to be more reliable Excellent measurement accuracy Detector needs cooling Requires electric power for operating mechanical scanner Not very portable
Less weight due to a reduction in mechanical parts (no mechanical scanner) — more portable, lighter, reliable with improved battery life Excellent resolution (dependent on number of pixels) Generally has fast frame rates, 30–60 Hz Complex expensive detector arrays with multiplexing electronics Detector needs cooling Start-up times; cooler has to work hard to maintain lower temperatures because of the larger mass of detector material
Mechanical simplicity Improved reliability of camera — no cooling required
Internal temperature stability Compensation required for internal thermal noise Fill factors can be as low as 40%
From Thomas, R.A., Handbook on Thermography, Coxmoor Publishers, Oxford, 1999.
Lens
Filter
Detector
Electronics
Outputs
Object
FIGURE 87.6 Simplified layout of a radiometer.
Detector
Lens
Operator controls Video display
Video processing
FIGURE 87.7 Focal plane array. (From FLIR.)
The advantages and disadvantages between these systems are reviewed below: •
•
Scanning systems generally provide single- or two-dimensional images and often are fixed in particular locations. Focal plane arrays (staring systems) are predominantly used in industry and medicine as portable IR camcorders with cooled or
uncooled detectors, most providing radiometric capabilities. Recent technology has resulted in staring systems being used in fixed industrial locations providing three-dimensional image technology. As mentioned previously, there exist a plethora of thermographic cameras, some offering similar performance characteristics. More recently the uncooled
Handbook of Non-Invasive Methods and the Skin, Second Edition
microbolometer detector has become a popular option in industry. To appreciate the implications of these devices, Table 87.2 attempts to illustrate some of the general advantages and disadvantages of a mechanical scanner and focal plane array with and without cooling.
87.6.6 DETECTOR ARRAYS Detector arrays generally have the advantage of no mechanical moving parts. The spatial resolution, and ultimately picture quality, is determined by the number of pixels within the detector array. For example, there are currently two formats commonly employed: 256 × 256, providing 65,536 detectors, and 320 × 240 (320 columns × 240 rows = 76,800 detectors). More recently, a large 512 × 512 pixel platinum silicide (PtSi) focal plane array has been fabricated with CCD/CMOS technology with high performance and good pixel-to-pixel uniformity. Typical pitches between pixels are in the range 20 to 50 μm. Lack of uniformity of the detector elements across the array can affect performance. Individual pixel response characteristics differ considerably across the array. Therefore, pixel correction is required prior to final camera image. This amounts to calibrating each individual pixel, by exposing the array to calibrated surfaces of known temperature.
87.7 DETECTORS In all thermographic systems the type of detector used to convert the incident radiation into a meaningful signal ultimately shapes the functionality of the infrared camera. The actual detector used will depend on the wavelengths to be detected. Thermal sensitivity/resolution or noiseequivalent temperature difference (NETD) is considered to be one of the most useful measurement parameters in medical thermography, particularly with reference to measuring small temperature differences on the skin surface. NETD is defined as the temperature difference that will produce a signal-to-noise ratio of unity. It is a measure of the performance of detector and processing electronics. Grenn measured the NETD for first- and second-generation infrared cameras (mechanical scanners and focal plane arrays) from the analog video outputs with wide field-of-view optics (Figure 87.8). From Figure 87.8 it can be seen that the new detectors reveal a tenfold increase in sensitivity. There is no single figure of merit that will measure the quality of an infrared image, but NETD is widely adopted. The infrared detector can be divided into two groups known as thermal detectors and photon (quantum) detectors. From Figure 87.9 it can be seen that each detector has a different response profile compared to the operating wavelength.
0.35 0.30 Min: max NETD (K)
760
0.25 0.20 0.15 0.10 0.05 0.00 Serial scanner PtSI Uncooled QWIP
InSb
HdCdTe
Camera technology
FIGURE 87.8 Various minimum and maximum noise-equivalent temperature differences. (From Grenn, M.W., Recent advances in portable infrared imaging systems, Proc. 18th Intl. Conf. IEEE Eng. Med. Biol. Soc., Amsterdam, Paper 1091, 1996.)
Different detectors each have a maximum response, but at different wavelengths; the question often asked is: Should I use a short-wave or long-wave camera? Figure 87.9 illustrates that each detector can achieve an optimum performance level (albeit at different wavelengths), and provided the selected camera meets the general requirements concerning picture quality, thermal sensitivity, and accuracy, then cost is probably the determining factor. As mentioned previously, atmospheric absorption of infrared energy generally limits the useful band of infrared detectors; this is why the predominantly used thermal imaging systems operate at 3 to 5 μm and 7.5 to 14 μm, respectively (Figure 87.9).
87.7.1 PHOTON DETECTORS Photon detectors (sometimes known as quantum detectors) convert radiation directly to an electrical signal. In Figure 87.10 the detector and cooler are combined. For example, the absorption of long-wave radiation results directly in some specific quantum event, such as photoelectric emission of electrons from a surface, or electronic interband transitions in semiconductor materials. The output of photon detectors is governed by the rate of absorption of photons and not directly on the photon energy. It is usual to cool the detector down to cryogenic temperatures (77 K, liquid nitrogen; 4 K, liquid helium) to reduce any excessive dark current, resulting in improvedperformance larger detectors, but with smaller response times. Photon detectors exhibit the following characteristics: • • •
Need cooling Very sensitive Very stable
Sensors and Handheld Devices for Surface Measurement of Skin Temperature
761
System response curves
1 0.9
PtSi
0.8
Response
0.7 0.6
Microbolometer
0.5 0.4
GaAs Qwip
0.3 0.2 0.1 0.0
3
4
5
6
7
8
9 10 11 12 Wavelength (μm)
13
14
15
16
17
18
FIGURE 87.9 Typical infrared detector response curves. (From FLIR.)
FIGURE 87.10 Example photon detector with sterling cooler (SC1000). IR photon radiation
Electrical current flow
Electrons
Upper band
Energy gap
Electron movement – lower band Power supply
FIGURE 87.11 Photon detector operation.
When infrared photon radiation falls on the detector, electrons can move freely in the conductive band and contribute to the electrical signal (Figure 87.11). At room temperature all the electrons have thermal movement jumping up and down. The higher the temperature, the more jumping around, resulting in a large number of electrons jumping from the lower to higher bands, contributing to a large signal current called noise current. If the detector is cooled down, the movement of the electrons gets much lower, with just a few of sufficient
energy to jump the gap. The result is low-noise current — significantly reduced unwanted noise current. To derive a suitable signal from the detector and to help the electrons jump the gap, infrared photon radiation is sent onto the detector cell. The photons penetrate the detector material and hit the electrons, which are sent over the gap to the upper region, where they contribute to the electrical signal. This signal will be proportional to the number of photons with enough energy hitting the detector. The energy of the photons needs to be high enough to get the electrons over the gap.
87.7.2 SHORT-WAVE DETECTORS
AND
LONG-WAVE PHOTON
Short-wave photon detectors generally have higher energies and can easily kick electrons over the gap. Therefore, the gap can be made wider; this means that the detector does not need to be cooled down as much as in the long wave, allowing the photons to jump around quite freely because they want to jump the gap anyway (Figure 87.12). A common detector material used in short wave is platinum silicide (PtSi), and quantum wells in the long wave. There are two types of photon detectors characterized by the response caused with interaction to photons of radiation: photovoltaic (generate a potential difference across a junction) and photoconductive (generate free carriers in a semiconductor that in turn increases the conductivity).
87.7.3 PHOTOVOLTAIC DETECTORS As infrared radiation passes near a junction, it is absorbed and gives an electron enough energy to reach the conductive band. This generates an electron in the conduction band and leaves behind a hole that can also
762
Handbook of Non-Invasive Methods and the Skin, Second Edition
Photons require greater energy than the energy in the gap
Photons require greater energy than the energy in the gap
Free electrons
Free electrons Energy gap
Energy gap Trapped electrons Longwave
Trapped electrons Shortwave
FIGURE 87.12 Energy gap in photon detectors. Responsivity
8.75 μm Peak With grating Without grating
6
7
9 10 8 Wavelength (μm)
11
FIGURE 87.13 Photomicrograph of a QWIP detector array with grating (SC3000) and its spectral response curve. (From FLIR.)
contribute to conduction. This process is sometimes referred to as kicking the electron into the conduction band or creating an electron–hole pair. The presence of an electron–hole pair changes the current–voltage relationship of the diode. That change can be monitored to provide infrared detection. Photovoltaic devices need an internal potential barrier with built-in electric field in order to separate photogenerated electron–hole pairs. Such potential barriers can be created by the use of Scottky barriers. Whereas the current–voltage characteristics of photoconductive devices are symmetric with respect to polarity of the applied voltage, photovoltaic devices exhibit rectifying behavior. Typically, photovoltaic detector materials are: • • •
Platinum silicide (PtSi) Mercury cadmium telluride (MCT or HgCdTe) Indium antimonide (InSb)
87.7.4 PHOTOCONDUCTIVE DETECTORS The operation of photoconductive detectors is based on the photogeneration of charge carriers (electron–hole pairs). These charge carriers increase the conductivity of the device material. Typically photoconductive detectors materials are:
• • • • •
Mercury cadmium telluride (MCT) Indium antimonide (InSb) Lead sulphide (PbS) Lead selenide (PbSe) Quantum well infrared photodetector (QWIP)
In most cases photon detectors are cooled to cryogenic temperatures, with the exception of thermoelectric cooling, where 200 K seems to be sufficient, as in 3 to 5 μm MCT detectors.
87.7.5 QUANTUM WELL INFRARED PHOTODETECTORS One of the more recent detectors to appear particularly useful is the quantum well infrared photodetector (QWIP). These devices consist of quantum wells in semiconductor material, where resultant electronic levels can be tailored to absorb radiation in the 3- to 20-μm-wavelength region. Special grating structures are necessary in order to achieve a high quantum efficiency of the detector (Figure 87.13). Some of the characteristics associated with QWIPs are: • • •
Thermal sensitivity, 20 mK at 30°C Spatial resolution, 1.1 mrad Real-time 14-bit digital output
Sensors and Handheld Devices for Surface Measurement of Skin Temperature
763
TABLE 87.3 Characteristics of Three Different Detector Materials QWIP (AlGaAs/GaAs/InGaAs) Wavelength NETD Uniformity Optics Cooling Operability
• •
MWIR/LWIR 50%) Moderate increase (>20%) No change/effect Moderate increase (>20%) Minor increase (10%) Moderate decrease (80%) Moderate decrease (>30%) Minor decrease (0.5 cm; (Q)uality (scored 1–3), (D)ensity (0–3) and (P)roportion of zone covered with hair (0–1); total score = Q × D × P Quantitative score based on distribution of hair on each site Quantitative score based upon density and extent of involvement Quantitative assessment of length and density of terminal hairs > 0.5 cm
It could be concluded that the difference is due to variation in the degree of hair growth in different populations of hirsute women, but it is more probably due to a lack of observer conformity. It should be noted that in the myriad of publications that have employed subjective measurement of body hair growth, there are few attempts to evaluate the precision of the method used. In our hands, the Ferriman and Gallwey scoring system has a repeatability coefficient of 3.2 (Ferriman and Gallwey units) at a mean score of 26.12
101.3 OBJECTIVE MEASUREMENTS Quantitative evaluation of scalp hair requires techniques that are sensitive enough to assess fundamental variables such as hair density, fiber diameter, proportion of anagen
hair (i.e., actively growing), and linear growth rate. Such information provides essential details for determining normal morphology as well as understanding changes arising from disease.
101.3.1 PRESAMPLING CONSIDERATIONS Presampling factors are probably the most difficult problems to standardize. These include shampooing, combing, and other cosmetic procedures, all of which must be standardized to provide unbiased measurements. Unfortunately, ideal protocols for hair measurements are impractical since they would include 3-month shampoo-free periods for telogen counts. Therefore, subjects should continue with their normal daily routines, which are used for background data. The shampooing interval prior to sampling should be kept standard.
Ferriman & Gallwey score
40
30
20
10
0
FIGURE 101.1 Comparison of Ferriman and Gallwey scores of hirsute women from 11 studies (mean ± SD). It is assumed that the severity of hirsuties of women attending different centers is similar and therefore this graph illustrates the variation in hirsuties grading perceived by different investigators. (Redrawn from Diseases of the Hair and Scalp, Rook and Dawber, Eds., Blackwell Scientific, Oxford, 1991.)
Measurement of Hair Growth
101.3.2 PRESAMPLING RECOMMENDATIONS Hair should be shampooed daily or on alternate days during the month prior to sampling, and not less than on alternate days throughout the study period. On the day of sampling specimens should be obtained, or visual imaging performed, within 3 hours of shampooing. Such actions produce minimal influences upon the derived anagen estimate, while providing excellent conditions for photographic or image analysis reproductions. It should also be pointed out that an inadequate shampooing action can have a detrimental influence upon the derived anagen estimate. Inadequate and inefficient shampooing can be found in subjects experiencing excessive hair loss, which results in poor cleansing and only partial removal of the hairs due to be shed. As excessive amounts of hair are seen at each wash, the shampooing frequency is decreased in a belief that this will reduce future loses. When this situation is found, sufficient time must elapse to allow the reestablishment of shedding levels representative of the problem under investigation.
101.3.3 SAMPLING CRITERIA The selection of area to be sampled is also of critical importance. Sites must be chosen carefully to represent the changes occurring. Problems arising from patterned presentations within the sampling distribution require careful attention to avoid biased estimates. No real problems should be encountered where the distribution of hair density is uniform, assuming sufficient hairs are obtained. Approximately 100 hairs are required to estimate variables by proportion; consequently, a sufficiently hairy area will need to be sampled. Two sample sites providing the required total number of hairs give better estimates (smaller sampling variation) than a single site.13 However, the surrounding area should be indistinguishable (with respect to hair density) from the sample sites (Figure 101.1). This arrangement allows resampling within ±5 mm of the original sites without the need to find the exact initial spot. Disturbances such as male pattern baldness often present problems because of patterned distributions in affected areas. Hair density within the vertex or frontal hair line can sometimes vary tremendously, and exact resampling is required. This can be achieved by placing a small tattoo between the two original sample sites or within the center of a single site. Where tattooing is not possible, location coordinates are essential for accurate relocation at some future date. Resampling frequencies depend upon the variables being estimated and the method of evaluation being employed. Where plucking techniques are used, hair density cannot be reestimated for 6 months; with noninvasive methods (hair weights or phototrichograms), resampling can be performed within 1 month. However, estimates for the per-
871
centage of hair in the anagen growth phase need to be made 12 months apart due to seasonal variation14 and are a requirement independent of the method of evaluation.
101.3.4 THE UNIT AREA TRICHOGRAM The unit area trichogram evaluates scalp hair variables with excellent reproducibility. It is a semi-invasive (plucking) technique in which all the hairs within a defined area (usually >30 mm2) are counted and measured.15,16 The area to be sampled should first be degreased with an acetone/isopropanol (60:40 v/v) mixture to remove surface lipids, which can cause blurring of the delineating sample line. The area to be sampled should be identified prior to epilation; a roller ball pen produces the sharpest line compared to circles drawn with fiber, felt, or ballpoint pens. The sample area can be quantitatively measured from an enlarged black-and-white photograph containing a scale bar or a computer image. The hairs should be rapidly epilated in a single smooth action in the direction of growth in order to minimize trauma to the roots.
101.3.5 EXAMINING HAIR ROOT STATUS All epilated hairs are placed directly onto double-sided tape (15 mm wide) attached to a 25-mm microscope slide. The collected hairs are subsequently realigned so that their root ends protrude from the tape edge toward the center, thereby allowing easy visualization within the microscope (Figure 101.2). Each hair is classified microscopically (magnification, ×40) according to its growth phase (anagen, catagen, or telogen), diameter, and length. The typical microscopical (dry-mounted) appearance of the various hair growth stages, anagen, catagen, and telogen, are shown in Figure 101.3. Catagen hairs are classified with the telogen population for data analysis since catagen is effectively the first stage of telogen. The catagen bulb structure is fully keratinized and encased in a shriveledup translucent outer root sheath (Figure 101.3), and for these reasons we classify catagen fibers with the telogen population.
101.3.6 PROBLEMS WITH ANAGEN, CATAGEN, TELOGEN CLASSIFICATIONS
AND
A possible area of confusion and contention is the reporting of dysplastic or dystrophic hairs, which are usually observed in plucked samples. Apart from a few rare congenital diseases, dysplastic or dystrophic hair shafts are never seen on the scalps of men, women, or children, nor in scalp biopsy. It is our belief that these terms are in reality descriptions for traumatic features produced by the procedure of hair epilation. Classification of such roots is preferable to their exclusion from data analysis, and a few simple guidelines should ensure that difficult root presentations are assigned to the correct population.
872
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 101.2 Plucked hairs mounted on double-sided tape and ordered by length. The plucked hairs are initially collected onto another strip of adhesive tape and then during transfer into the illustrated format have their length measured.
The principal feature to keep in mind is that actively growing hairs (anagen) have soft prekeratinized tissues up to approximately 2.0 mm above the apex of the dermal matrix.13 This soft tissue bends easily and distorts upon epilation. Catagen and telogen hairs differ since they are fully keratinized, and are therefore rigid and do not easily bend. Moreover, the diameters of catagen and telogen fibers taper toward the bulb, and this zone is devoid of, or has reduced, pigmentation. The following therefore serves as a guide to assigning difficult fibers to the anagen or telogen populations: If the root end exhibits bending with or without pigmentation, we assign it to the anagen population. If not, we assign it to the telogen population. Broken hairs arising from the epilation procedure also cause difficulties with interpretation, but should be assigned to the telogen population if the proximal end exhibits tapering or loss of pigment; otherwise, the fiber should be classified as anagen. Fortunately, the occurrence of such hairs is 10% broken-off hairs cannot be evaluated correctly and should be repeated.20 Trichogram measurements should always be performed and evaluated by the same experienced investigator to maintain optimal and comparable examination conditions. A short training period with an experienced colleague is recommended.
102.5 COMPARISON WITH OTHER TECHNIQUES The trichogram is a semi-invasive technique with plucking of the entire hair. This makes it unsuitable for the monthly follow-up of patients, for studying the hair growth rate and the duration of the growing stage of individual hairs, and for studying seasonal variations. Noninvasive microscopic techniques such as optical microscopy with image analysis, 31 phototrichogram, 32,33 or the unit area trichogram34 should be chosen for this purpose. The trichogram technique is accurate and reliable for the diagnosis of hair diseases and is highly suitable because of its handiness.35 Although trichogram measurements are only
TABLE 102.2 Normal Distribution Pattern of Hair Roots in the Trichogram of the Scalp20 Hair Root Anagen Dysplastic anagen Catagen Telogen Dystrophic Broken-off hairs
% 60–80 5–20a 1–3 12–15 50%.
confined to two small scalp regions, earlier studies provided evidence that, except in alopecia areata, the trichogram technique of one site is representative of the neighboring areas.30
880
Handbook of Non-Invasive Methods and the Skin, Second Edition
Trichogram evaluation of hair root pattern is particularly useful at the beginning of abnormal hair loss, when hair density still seems normal at clinical examination. In this case, trichogram results may permit an early diagnosis, early determination of prognosis, and an early start of treatment to prevent further continuation of effluvium. In addition, normal trichogram results may allow a favorable prognosis and help to avoid various expensive and useless treatments.
REFERENCES 1. Kligman, A.M., The human hair cycle, J. Invest. Dermatol., 33, 307, 1959. 2. Uno, H., Biology of hair growth, Sem. Reprod. Endocrinol., 4, 131, 1986. 3. Ebling, F.J.G., The hair, in Textbook of Dermatology, Rook, A., Wilkinson, D.S., Ebling, F.J.G., Champion, R.H., and Burton, J.L., Eds., Blackwell Scientific, Oxford, 1986, p. 25. 4. Ebling, F.J.G., The biology of hair, Dermatol. Clin., 5, 467, 1987. 5. Rook, A. and Dawber, R.P.R., The comparative physiology, embryology and physiology of human hair, in Diseases of the Hair and Scalp, Rook, A. and Dawber, R.P.R., Eds., Blackwell Scientific, Oxford, 1982, p. 1. 6. Sato, Y., The hair cycle and its control mechanism, in Biology and Disease of the Hair, Koboti, T. and Montagna, W., Eds., University Park Press, Baltimore, 1986, p. 3. 7. Braun-Falco, O. and Kint, A., Dynamik des normalen und pathologischen Haarwachstums, Arch. Klin. Exp. Dermatol., 221, 75, 1966. 8. Orentreich, N., Scalp hair replacement in man, in Advances in Biology of Skin, Vol. 9, Montagna, W. and Dobson, R.L., Eds., Pergamon Press, Oxford, 1967, p. 99. 9. Barman, J.M., Astore, J., and Pecoraro, V., The trichogram of people over 50 years but apparently not bald, in Advances in Biology of Skin Hair Growth, Vol. 9, Montagna, W. and Dobson, R.L., Eds., Pergamon Press, Oxford, 1964, p. 211. 10. Trotter, M., The life cycles of hair in selected regions of the body, Am. J. Phys. Anthropol., 7, 427, 1924. 11. Blume, U., Ferracin, J., Verschoore, M., Czernielewski, J.M., and Schaefer, H., Physiology of the vellus hair follicle: hair growth and sebum excretion, Br. J. Dermatol., 124, 21, 1991. 12. Randall, V.A. and Ebling, F.J.G., Seasonal changes in human hair growth, Br. J. Dermatol., 124, 146, 1991. 13. Van Scott, E.J., Reinertson, R.P.A., and Steinmüller, R., The growing hair roots of the human scalp and morphologic changes therein following amethopterin therapy, J. Invest. Dermatol., 29, 197, 1957.
14. Pecoraro, V., Astore, J., Barman, J., and Araujo, C.S., The normal trichogram in the child before the age of puberty, J. Invest. Dermatol., 42, 427, 1964. 15. Pecoraro, V., Astore, J., and Barman, J., The normal trichogram of the pregnant woman, in Advances in Biology of Skin, Vol. 9, Montagna, W. and Dobson, J.M., Eds., Pergamon Press, New York, 1967, p. 203. 16. Barman, J.M., Astore, J., and Pecoraro, V., The normal trichogram of the adult, J. Invest. Dermatol., 44, 233, 1965. 17. Grosshans, E., Che pfer, M.P., and Maleville, J., Le trichogramme. A propos d’une méthode d’étude des cheveux, J. Med. Strasbourg, 378, 1972. 18. Metz, H.G. and Landes, E., Der Haarausfall und seine Untersuchung, Fortschr. Med., 88, 1327, 1970. 19. Zaun, H. and Ludwig, E., Zur Definition ungewöhnlicher Haarwurzeln im Trichogramm, Hautarzt, 27, 606, 1976. 20. Orfanos, C.E., Androgenetic alopecia, in Hair and Hair Diseases, Orfanos, C.E. and Happle, R., Eds., SpringerVerlag, Berlin, 1991, p. 485. 21. Jost, B., Meiers, H.G., Schmidt-Elmendorff, H., and Pfaffenrath, V., Trichometrische Quantifizierung und Verlaufsbeurteilung des Hirsutismus, Dtsch. Med. Wschr., 99, 2395, 1974. 22. James, K.C. and Rushton, D.H., Evaluation techniques for male pattern baldness, J. Am. Acad. Dermatol., 14, 849, 1983. 23. Orfanos, C.E. and Hertel, H., Haarwachstumsstörungen bei Hyperprolaktinämie, Z. Hautkr., 63, 23, 1988. 24. Sterry, W., Konrads, A., and Nase, J., Alopecie bei Schilddrüsenerkrankungen: Charakteristische Trichogramme, Hautarzt, 31, 308, 1980. 25. Orfanos, C.E., Meiers, H.G., Friedrich, H.C., Ludwig, E., Mahrle, G., and Zaun, H., Haarausfall, Trichogramm und hormonelle Haartherapeutika, Dtsch. Arztbl., 3603, 1974. 26. Blume, U., Föhles, J., Gollnick, H., and Orfanos, C.E., Genotrichoses: clinical manifestations and diagnostic techniques, in Proceedings of the 18th World Congress of Dermatology, New York, in press. 27. Barth, J.H., Measurement of hair growth, Clin. Exp. Dermatol., 11, 127, 1986. 28. Braun-Falco, O. and Fischer, C., Über den Einflub des Haarewaschens auf das Haarwurzelmuster, Arch. Klin. Exp. Dermatol., 227, 419, 1966. 29. Bassukas, I.D. and Hornstein, O.P., Effects of plucking on the anatomy of the anagen hair bulb. A light microscopic study, Arch. Dermatol. Res., 281, 188, 1989. 30. Braun-Falco, O. and Heilgemeir, G.P., The trichogram. Structural and functional basis, performance and interpretation, Sem. Dermatol., 1, 40, 1985. 31. Hayashi, S., Miyamoto, I., and Takeda, K., Measurement of human hair growth by optical microscopy and image analysis, Br. J. Dermatol., 25, 123, 1991. 32. Fiquet, C. and Courtois, M., Une technique originale d’appréciation de la croissance et de la chute des cheveux, Cutis, 3, 975, 1979.
Microscopy of the Hair: The Trichogram
33. Van Neste, D., Dumortier, M., and De Coster, W., Phototrichogram analysis: technical aspects and problems in relation to automated quantitative evaluation of hair growth by computer-assisted image analysis, in Trends in Human Hair Growth and Alopecia Research, Van Neste, D., Lachapelle, J.M., and Antoine, J.L., Eds., Kluwer Academic, Dordrecht, The Netherlands, 1989, p. 147. 34. Rushton, H., James K.C., and Mortimer, C.H., The unit area trichogram in the assessment of androgen-dependent alopecia, Br. J. Dermatol., 109, 429, 1983.
881
35. Meiers, H.G., Trichogramm (=Haarwurzelstatus, =Haarbild). Methode und Aussagefähigkeit, Akt. Dermatol., 1, 31, 1975. 36. Astore, J., Pecoraro, V., and Pecoraro, E.G., The normal trichogram in pubic hair, Br. J. Dermatol., 101, 441, 1979. 37. Saitoh, M., Uzuka, M., and Sakamoto, M., Human hair cycle, J. Invest. Dermatol., 54, 65, 1970. 38. Seago, S.V. and Ebling, F.J.G., The hair cycle on the human thigh and upper arm, Br. J. Dermatol., 113, 9, 1985.
and Computerized 103 Photographic Techniques for Quantification of Hair Growth D. Van Neste Skinterface, Tournai, Belgium
CONTENTS 103.1 Introduction ..........................................................................................................................................................883 103.2 Basics about Hair Structure and Function...........................................................................................................883 103.3 Hair Photography .................................................................................................................................................884 103.3.1 Search for Golden Standards.................................................................................................................884 103.3.2 Improving Hair Photography for Computerized Measurements ..........................................................887 103.3.3 Global Vision and Imaging Methods ....................................................................................................888 103.3.3.1 Categorical Classification Systems ......................................................................................888 103.3.3.2 Calibrated Scoring Systems .................................................................................................889 103.3.3.3 Global Photographs ..............................................................................................................889 103.3.4 Analytical Methods................................................................................................................................890 103.3.4.1 Phototrichogram: From Conventional PTGs to Contrast-Enhanced PTGs .........................890 103.3.4.2 Variants of PTG Methods.....................................................................................................891 103.3.4.3 Future Trends in Computerized Methods ............................................................................891 103.3.4.4 Detection of Nonvisible Hair: Less Contrasted and Thinning Hair....................................892 103.4 Conclusion............................................................................................................................................................892 References .......................................................................................................................................................................893
103.1 INTRODUCTION The present review will focus on basic principles involved in human hair evaluation using photographic and computerized methods. Needless to say, visible hair, hair growth, and regrowth are different in nature. Indeed, the visibility of hair will depend not only on the resolution power of the optics and camera, but also on the natural contrast between the object, i.e., hair (with every variation of the fiber components and optical interfaces), and background (including skin as a heterogeneous and variable background and extraneous material). Hair growth is a dynamic process that results in visible hair production at the skin surface reflecting cell proliferation, differentiation, and migration of trichocytes, i.e., epithelial cells that are under the influence of skin-deep dermal papilla. Hair regrowth is more related to the phenomenon of hair cycling (see later). All these biological events happen to occur in the deeper layers of the skin, i.e., in the hair follicle. As far
as hair is concerned, these processes result in the appearance and maintenance of one or more hair fibers at the surface of the skin. Finally, this is what matters as well to the patient as to scientists or observers interested in capturing the hair and analyzing its vital characteristics. Whether or not this biological event makes the person happy is a complex matter that will not be considered here.
103.2 BASICS ABOUT HAIR STRUCTURE AND FUNCTION Here we shall briefly review some basics about hair follicle structure and function necessary for understanding the fine-tuning of technical aspects of hair photography. Those who are acquainted with growth of hair and hair follicle cycling may easily proceed to the next section. Scalp hair is an appropriate example to introduce the concept of global and analytical methods while “zooming” into the field of hair (Figure 103.1). Scalp hair appears as 883
884
Handbook of Non-Invasive Methods and the Skin, Second Edition
1
2
1S
1R
2S
2R
SA n/cm2 RA
200 150 100 50 0
0
50
100
150
global maintenance. The key to the appropriate assessment of hair maintenance lies first in a thorough understanding of the hair cycle (for a review, see Reference 1). This is shown schematically in Figure 103.2 for one single hair follicle. On the scalp there are about 100,000 follicles. Importantly, not all follicles are active at the same time: some produce a hair, while others are resting. After this, hair is shed (exogen hair follicle) and eventually a new hair shows up. After a short time necessary to reach the scalp surface, hair becomes visible. Each follicle appears to function independently from its neighbors; i.e., the process of scalp hair growth is not synchronized. However, follicles in a certain field may express a common phenotype after exposure of some compounds, i.e., sex hormones. Some scalp areas, in genetically predisposed individuals, will show a phenotype of defective hair replacement, i.e., patterned balding (Figure 103.3). In other areas (axilla, face, chest, genital skin, etc.) other patterns will be expressed: the very thin hair follicles will now grow longer and coarser hair. As such, the clinical appearance of hairiness depends as much on regional modulation of the hair follicle activity, and of its cycle in particular, as on the number of hair follicles.
MA n/cm2
FIGURE 103.1 Zoom concept from global view to hair folliculogram. Global view of the top of the head in two male subjects with male pattern baldness.1,2 The top of the head is fragmented into smaller areas (squares) for scalp coverage scoring (for more details on SCS, see text and other figures). Clipping hair allows a more detailed record using close-up (or macro) photographs. Depending on the technique, the area of the target site ranges from 5 to 1 cm2 (white arrows; clipping and close-up photographs not shown). A small tattoo allows relocation of the target area (middle of white ring). Such tattoos can also be removed by a small 4-mm punch biopsy (not shown). We show images taken with a videomicroscope of the surface of the biopsy. This shows a very small number of hairs (1S less than 2S). The same scalp specimens are marked (red dots) and processed for the isolation of the dermal part of the hair fiber (1R and 2R). In the lower left panel we show the results of a validation test of a macrophotographic equipment and analysis system. After taking a macrophotograph for anagen hair counts (MA, n/cm2 46), tattoos were punched out and specimens were tested for anagen hair counts with a stereomicroscope on the surface or on the root view (respectively SA and RA, n/cm2). In most cases (3/4), counts of SA and RA were higher than those obtained with a close-up photograph (MA, n/cm2), showing that the macrophotographic system was less than optimal to detect all growing hair in vivo on the human scalp.4 Finally, the light microscope (lower right panel) gives a very detailed view of the hair as it emerges at the scalp surface (thick arrow) from the deeper parts of the hair infundibulum (three small arrows).
a stable mass of hair. It actually represents the cumulative end result of discrete changes of individual hair follicle dynamics: hair shedding and replacement resulting in
103.3 HAIR PHOTOGRAPHY 103.3.1 SEARCH
FOR
GOLDEN STANDARDS
There is not a single technical modality that will encompass with a sufficient degree of precision the many dimensions of hair.2,3 At the one end of the spectrum, as shown in Figure 103.1 and Figure 103.3, global viewing is very often used for demonstration purposes, i.e., to document health authorities or individual patients. Global viewing, even when highly standardized procedures are being used, does not resolve the question of hair cycling changes. At the other end of the spectrum (Figure 103.1), the microscope has a superb resolution power when a single hair fiber or follicle is concerned. A comparison between stereomicroscopic viewing of scalp samples and photographs of the same sites taken with commercially available macrophotographic equipment showed that the number of growing hair counted on the surface view or from the root view were always higher than the counts generated from printouts.4 However, one may argue that those small numbers generated from scalp biopsies (4 mm diameter) may not be clinically representative. A larger field containing over 100 hair fibers would certainly be more representative,5 but many clinical research programs and investigators do forget that one may not expand the size of the target area without affecting the resolution of the image. A proper balance between sample size and the optical resolution power must be established before launching a hair photographic method.
Photographic and Computerized Techniques for Quantification of Hair Growth
885
A 1 2 3 4 5 6 a
b
c
d
e
f
g
h
i
a
b
c
d
e
f
g
h
i
B 1 2 3 4 5 6
FIGURE 103.2 Hair cycle. The hair growth phase (anagen) during which hair is visible at the skin surface and growing is shown in A, while the apparent resting phase of the hair cycle (telogen) is shown in B. The follicle is represented schematically and the essential components are numbered in the legend: A: pigmented (1) and less pigmented (2) hair shaft produced during the growth phase. Inner root sheath (IRS) (3), club hair (4), stelae (5), dermal papilla (6). B: Inner root sheath (1), tip of new anagen hair (2), release of old club hair (4), new anagen grows out of IRS (5) and is more pigmented (6). A: From growth to rest. The same hair follicle is represented at various times (a to i) at the very end of its growth phase. At the skin surface, there is normal pigmented hair production (time between a–b and b–c is 24 hours). The increased length of visible hair represents the daily hair production from which the linear growth rate can be calculated (μm/24 hours). Growth takes place at the bottom of the hair follicle. The space immediately below the bottom of the hair fiber is where cell proliferation takes place, i.e., hair matrix. This is very close to the dermal papilla. Then the pigmentation of the newly synthesized hair shaft is decreased (c). This early event announces the hair follicle regression and is followed by terminal differentiation of cells in the hair matrix (d) that will turn into the club hair formation (d to g). The shrinking dermal papilla (d) begins an ascending movement together with the hair shaft (time between d and h, 21 days). This characterizes the catagen phase (d to h) with an apparent elongation of the hair fiber. This is not growth, but reflects the outward migration of the hair shaft. What is left over after disappearance of the regressing follicle is usually referred to as streamers or stelae (f to i). The true resting stage begins when catagen is completed, i.e., when the dermal papilla abuts to the bottom of the club hair. As for now, no hair elongation is observed at the surface (g to i). B: From rest to growth. During this stage, one notices absence of hair growth at the skin surface (a to e), but significant changes occur in the deeper parts of the hair follicle. The dermal papilla expands and attracts epithelial cells from the bulge (stem cell zone) in a downward movement (a and b). To create space, previously deposited materials have to be digested (a to c). The epithelial cells then start differentiation in an orderly fashion, starting with the inner root sheath (IRS forms the initial cone in b and then a cuff in c) containing the tip of the newly formed hair fiber. This tip is made of the cuticle and a very thin hair cortex, usually non- or less pigmented (c to e). The hair fiber is released from the IRS cuff when it reaches the isthmus (f to g). The old resting hair remains in the hair follicle for approximately 1 to 3 months (a to e), then a detachment process transforms the old follicle into exogen and, as a consequence, releases an exogen hair (f). Such exogen hair may stick in the follicle for a short time before being shed (f and g). The shiny root end of the shed hair is the club, visible with the naked eye. Before, during, or after hair shedding there may be replacement by a new, gradually thicker and more pigmented hair shaft (e to g). Indeed, under physiological conditions, the follicle may proceed immediately or only after some lag time with new hair production (from b to g; up to 90 days). In conditions like androgenetic alopecia, the shed hair may stick in the follicle for a longer time. Stasis of exogen hair may turn into trichostasis. This reflects an abnormal accumulation of nonadherent or loosely attached elements. Also, there may be a much longer interval before regrowth is recorded at the scalp surface (example, steps b to g may take 6 months). At the earliest visible stages, i.e., when the matrix of the new anagen follicle is again deeply set into the dermis, one notices at the scalp surface a thin, usually nonpigmented hair tip that is seen first (h), soon followed by a thicker, more pigmented and faster-growing hair fiber (i). This sequence, of course, depends on the many regulatory factors controlling the hair follicle activity during its replacement.
886
Handbook of Non-Invasive Methods and the Skin, Second Edition
FIGURE 103.3 Patterned hair loss. The present classification shows patterns that affect the scalp of genetically predisposed male subjects. Such patterns develop after puberty, when deficient hair replacement ultimately results in bald appearance. They are subdivided in six categories of increasing severity (I to VI), from mild to severe balding. The anterior pattern (A) indicates a backward progression of hair follicle miniaturization that starts from the frontal hairline. The vertex type (V) indicates isolated regression occurring on the vertex, but this is usually combined with some involvement of the frontal temporal areas. As in Figure 103.1, fragmentation (square areas) help in evaluating the severity — in terms of extension or progression — of the balding process.
One way to evaluate 100% of hair growth potential is to take an exhaustive sample. This means taking a skin biopsy; i.e., there is no way for a single hair follicle to
escape the analysis. After fixation and sectioning the scalp sample from top to bottom, one may examine and analyze all sections with a light microscope. A trained observer
Photographic and Computerized Techniques for Quantification of Hair Growth
FIGURE 103.4 Phototrichogram in a young male with patterned alopecia. The contrast enhanced phototrichogram (CE-PTG) technique (upper panels) matches perfectly histology (lower panels) for hair growth measurement. Evaluation of hair growth with the phototrichogram: At time 0 (day 0), the hairs are clipped close to the scalp surface and photographed. After a given time (48 hours in the author’s experience, day 2), the same scalp site is photographed again. Substantial elongation at day 2 reflects hair growth and indicates anagen in thick (1, 2, 5) and thin (7) hair. Moderate elongation reflects catagen in a thinning hair (3). No elongation reflects resting in thick (4) and thinning (6) hair. After shedding the thick (4) or the typical tiny (9) resting hair, an empty follicle in the exogen stage will remain visible (8 arrows at day 0 and day 2 and upper arrow in histology 8, lower panel). The formation of new follicles may initiate anagen, for a thick (lower arrow) and thin hair (upper right arrowhead) can be seen only with histology (8). Tracking of such empty follicles with CEPTGs will show newly growing hair in some days or weeks (see Figure 103.2B: steps f, g, h, i).
887
will easily stage the hair follicles into anagen, telogen, or catagen. After tracking hair follicles through several hundreds of sections, a complete picture of the hair follicles will be available. Because of its high magnification and the exhaustive sampling, quantitative microscopic analysis of scalp biopsy specimens may serve the purpose of validating other measurement methods such as counting hairs and evaluating growth on scalp hair photographs6 (as shown already in Figure 103.1 and further detailed in Figure 103.4). Taking repeat biopsies is definitely not practical for hair growth monitoring purposes. Hair density or number of hair per unit area is usually reported as number per square centimeter. It reflects the number of functionally active follicular units whether growing (anagen) or not (telogen). Under physiological conditions, i.e., the long duration of anagen phase and the comparatively short duration of telogen, we know that most scalp follicles are engaged in anagen. This will produce long and clearly visible hair at the scalp surface, but this may change dramatically in some hair disorders, leading to hair loss and balding. The percentage of anagen follicles properly reflects the time during which hair follicles are engaged into the growth phase. The anagen/telogen ratio is also often found in the literature. Many clinicians are not aware that this is true only when an exhaustive count of hair or hair follicles is made. This means including those follicles that are growing a fiber that is not yet visible at the scalp surface (anagen III to V; see Figure 103.2B, steps c to f), those that are subject to miniaturization, or those that produce fibers that may not be seen with conventional photography. Therefore, classification of follicles into terminal, intermediate, and miniaturized has been proposed. There is no consensus, however, on the definitions of these categories. Many based their assumptions on the thickness of hair fibers, but we argue that not enough attention has been devoted in the past to variables such as growth rate, natural pigmentation,7 and eventually combined effects such as relative resistance to clipping of thinner fibers influencing the remaining length after clipping, clustering of thin and thick fibers, etc.7–9 From the above sections it becomes clear that the results of an assay largely depend on the technology that was used for measuring hair parameters. Before discussing further the aims of computerizing, we wish to illustrate some tools that are commercially available or systems that can be delivered in the context of the installation of a hair clinic with licensing of technology or on the occasion of a clinical trial project (Figure 103.5).
103.3.2 IMPROVING HAIR PHOTOGRAPHY COMPUTERIZED MEASUREMENTS
FOR
Computers may help in performing routine tasks, but they need to be fed very high quality documents in order to
888
Handbook of Non-Invasive Methods and the Skin, Second Edition
Ga
Gb Ma
Mb
Mc
FIGURE 103.5 Equipment for global and close-up hair photography. Equipment for global scalp viewing (Ga from Skinterface or Gb from Canfield) or for macrophotography (Ma from Canfield or Mb (historical) and Mc from Skinterface). Ga also shows a model used for calibration purposes and the comb that is always necessary to organize the mass of scalp hair properly. Such materials are commercially available at Canfield or can be licensed by Skinterface on the occasion of a clinical trial project. Ga and Mc (developed in 2002) show the recent trend toward miniaturization (compare with Gb and Ma, Mb).
generate acceptable numbers from figures: there is no brain other than the human observer to interpret the clinical value of an image. In the description of noninvasive photographic methods that can be used in vivo in human subjects, we propose to distinguish between global and analytical methods. Global methods apprehend at once various factors responsible for the hair area under examination but cannot resolve the details, as opposed to the analytical methods that shall be described later on. The latter have a unique advantage over the former, because they provide a series of individual measures that reflect the structure/function of the hair follicle as an organ. Such data are subject to critical analysis. By combination of all the analytical data, one can generate a global value, but the reverse is not true. Also, another disadvantage of published global methods is that they are usually not calibrated. As such, it remains difficult to derive clinical relevance of statistically significant findings.
103.3.3 GLOBAL VISION
AND IMAGING
METHODS
103.3.3.1 Categorical Classification Systems Distinctive patterns of defective scalp skin coverage or alopecia have been identified by clinicians as patterns. You could practice in any public space (mall, theater, etc.) and see for yourself what the limits of this method are (Figure 103.1 and Figure 103.3). In the author’s hands, there is a very large variation between sessions when the same views of the top of the head of male subjects with patterned hair loss were presented several times in a randomized sequence (unpublished observations). Also, intervals between successive severity scales
10
20
30
11
21
31
FIGURE 103.6 How do we standardize global scalp hair imaging? In this setting, great care has been taken in the preparation of the background and positioning of the patient (1). Imaging (2) involves adjusting the subject with the headgear to the positioning device. Once the light box is switched on and the camera properly inserted into the light box, one captures an image appearing on the computer screen. Examples show the views of the top of the head (10 to 30), on which a grid can be placed as an overlay (11 to 31).
may not be equal. The consensus among experienced clinicians is that the schemes are of little help in measuring the dynamics of hair changes — whether growth or loss — over time. Based on these patterns, fragmentation of the scalp was proposed on the basis of anatomical criteria.10 However, no evidence was presented to support this new belief. The use of a noncalibrated density scale casts some doubts as to the practical application in real time or “as a bedside” measurement tool.10 In order to circumvent several of the weaker aspects of these methods (lack of standardization, lack of calibration, transposing shaved hair density skinhead type of density patterns onto real heads of patients with a great diversity of hairstyles, etc.), we proposed, as shown in Figure 103.1, Figure 103.3, and Figure 103.6, to fragment the top of the head into a number of fixed and predefined subunits. This helps to focus the attention of the observer in clearly outlined spaces where patterning results in various degrees of defective scalp coverage. We
Photographic and Computerized Techniques for Quantification of Hair Growth
will describe results obtained using this novel approach in greater detail in the next section. 103.3.3.2 Calibrated Scoring Systems A system has been developed where examination of the scalp surface proceeds through fixed external standards. The top of the head is separated into small units (equal size in the projection plane) with no anatomical correspondence of the bones of the skull (e.g., frontal, parietal, etc.) or the immense variation of patterning in human subjects. The relative difficulty to detect scalp skin between the hair is translated into scalp coverage scores (scalp coverage scoring (SCS) score 0 means there is no difficulty and score 5 means that it is barely possible to detect scalp skin through the hair). This is rated — under strict distance and angle control — against objective rulers of density (method and equipment for measuring coverage density of a surface, Skinterface, application patent PCT/EP 01/06970, 2001). The reproducibility (intra- and interobserver) is very high (correlation factors >0.9).11 Several studies using scores generated in the clinic (real-time measurement) and on global photographs have now been performed either as a single center or in the context of multicenter placebo-controlled trials (H.A.I.R. Technology® protocols 02P08, 03P17, and 03P22) using known drugs or new compounds (phases II and III). When all identified sources of variation are kept under control, the variation of scores is less than 5%.11 During calibration studies, the SCS was also correlated with clinically relevant hair parameters such as proportion of anagen hair or density of thinning hair.12 As such, SCS allows quantification of clinically relevant changes (Figure 103.3). The main advantage of the SCS method is its easiness to use in the hair clinic, where it helps in identifying subjects who respond to specific treatment regimens in vivo. In a 1-year period, a clinician is able to disclose hair loss in untreated (randomized placebo-controlled study in vivo) patients.13–15 In the longer run, the clinician is more accurate in detecting clinically relevant changes due to lack of active treatment (e.g., placebo effect), as compared to maintenance or improved scalp coverage under active treatment. Besides the real-time quantitative observations at the hair clinic or at bedside, SCS can also be applied to standardized global scalp photographs, as will be discussed in the next section. 103.3.3.3 Global Photographs Global photography has been a significant step forward in scalp hair documentation by creating a permanent record.
889
Global photography apprehends all factors involved in hairiness at once and can be used for drug efficacy evaluation providing that adequate scalp preparation and hairstyle are maintained throughout the study. This is the most patient-friendly photographic method. This method is used in the clinic under standardized conditions of exposure.16 Processing and rating have to be performed under controlled (e.g., blinded as to treatment or time) conditions. Trained experts could generate reproducible data. It appears that paired comparison of global photographs is more realistic in its appreciation of hair growth after drug treatment than subjective evaluations of investigators and patients.17 As sets of photographs can also be scored individually with the SCS system — one photograph at a time and in circumstances one usually encounters in day-to-day practice (see earlier) — the SCS could provide real-time scalp coverage values. Such a quality control before inclusion of subjects into clinical trials may be of great value. Indeed, quantifying disease severity and precise evaluation of distribution patterns are immediately available, as opposed to global photographs that need expert evaluation after completion of the study. Indeed, in practice, the quality control happens only after processing in a special laboratory. The quality of photographs is evaluated in terms of optics (focus, contrast, etc.), but not in terms of hair patterning or alopecia. The techniques of global picturing and density documentation appear to be important issues for future development, and work is in progress in our laboratory. Some computer approaches have already been published,18,19 but can hardly be used on real documents. In general, paired rating of global photographs and SCS are two techniques that are significantly correlated inasmuch as all factors of variation (hair length, room and scalp light, standardized combing, etc.) are well controlled (for some examples of technological problems, see Figure 103.7). SCS has the advantage of helping the clinician in his real-time evaluation of efficacy in vivo before analyzing comparatively scalp views on global photographs. Incidentally, the accuracy of the image setting used for the SCS three-dimensional control method (as demonstrated in Figure 103.6) allowed us to detect subtle changes in distance or angle of vision between pairs of before and after photographs. Such deviations remain usually unnoticed even to a panel of internationally recognized independent experts (as a practical exercise, trace a grid on a transparency and apply the grid on the photographs appearing in panels 1 and 2 of Figure 103.7; personal unpublished data).
890
Handbook of Non-Invasive Methods and the Skin, Second Edition
1
2
3
4
FIGURE 103.7 Coping with technical and styling variability of global photographs. These two sets of before and after pictures illustrate difficulties encountered in evaluation of time or treatment-related changes of scalp coverage. The commercially available Canfield system (1 and 2) and clinical trial set developed at Skinterface (3 and 4) illustrate some differences between the systems (backscattering of light, background, inclusion of color codes during imaging, etc.). The vertex view (1) turned into a top-of-the-head view (2) at the follow-up visit. The technician did not properly position the head of the patient on the two sessions (estimated angle error, ± 45˚). The two top-of-the-head views (3 and 4) illustrate the difficulty of managing properly the hairstyle. Although the technical quality is almost perfect (see overlay of targets, distance, angle, color codes, etc.), any possible change of hair growth is masked by the change in hairstyle.
103.3.4 ANALYTICAL METHODS 103.3.4.1 Phototrichogram: From Conventional PTGs to Contrast-Enhanced PTGs 103.3.4.1.1 What Did We Learn with the Conventional PTG? The basic principle of the phototrichogram (PTG) consists of taking a close-up photograph of a certain area of the scalp. The hair is cut very close in preparation for the first photograph, followed by repeat photographic documentation after a certain period. This period of time should be long enough to allow the growth of a hair segment (time window that is usually between 24 and 72 hours), but not too long, in order to prevent outgrowth or too much overlapping of growing hair. The growth is then evaluated by comparing the two pictures. Hairs that have grown are in
the anagen phase, and those that have not are in the telogen phase (Figure 103.4). Analytical methods that document major aspects of the hair cycling process have been developed over the years20–26 and are subject to continuous reevaluation and improvement.27–29 Some PTG data have been computed for mathematical modeling so as to develop virtual patterns of defective hair replacement, mimicking those observed in the hair clinic.30 Nevertheless, neither all hair fibers nor all productive hair follicles are taken into account during the conventional PTG procedures. In some assays, we noticed that the length of hair significantly affected its visibility. As growing hair will become more visible on the second photograph, a bias was detected (unpublished data). In 1989, we devised scalp immersion photography,31 or proxigraphy, 22 well before epiluminescence microscopy became the accepted term. After comparison with another, more invasive method,28 some weaknesses have been considered with great care and considered contrast enhanced (CE) as a further improvement.6,27 This brought the phototrichogram technique (CE-PTG) to a resolution almost equal to that of transverse microscopy of scalp biopsies, which are usually considered the golden standard.32–37 So far, the CE-PTG remains the only method that has documented all transitions of thick and thinning follicles, from anagen through catagen into telogen phases, on a follicular basis.6 In early stages of androgenetic alopecia (AGA) in man, this sensitive method12 was able to detect a subclinical phase of AGA with obvious shortening of the anagen phase in the absence of hair miniaturization. This preclinical stage evolves into patterning, i.e., the fullblown phenotype associated with a further shortening of the growth phase along with reduction in hair diameter.22–29,38–42 The follicular regression process finally results in production of clinically nonvisible hair.43 The CE-PTG method, as a refined, noninvasive, and validated technique, could be used for calibration purposes for any new method that would be developed for use in the skin and hair clinic. Hence, we identified a new global measurement method that integrates cumulative hair growth and reflects clinically relevant scalp skin coverage, as described earlier (patent application PCT/EP 01/06970, June 2001). 103.3.4.1.2 What Can We Learn to Measure with the CE-PTG? The assessment is made on one or a number of predefined scalp sites considered representative of the condition. The data that can be generated from a PTG are total number of hair present in a certain area, i.e., hair density (n/cm2), the percentage of hair in the growth phase (anagen %), the linear hair growth rate (LHGR; mm/day), and the hair thickness.
Photographic and Computerized Techniques for Quantification of Hair Growth
Thickness can be measured on hair clippings, on scalp biopsies, and on scalp hair photographs. It may seem trivial to state that the hair diameter evaluation is more precise with the microscope. These measures reflect diameter whenever the fiber section is circular, and in all races inasmuch as the hair remains thinner than 40 μm.44 In Caucasians, thicker hair loses this property and the section becomes elliptical. In Africans, the hair flattens even more, and we are not aware of any published data on the value of hair thickness evaluation on scalp photographs of those subjects. The same holds true for other body sites where curly and flattened hair are the rule rather than the exception. Nevertheless, we found a significant correlation between paired microscopic measurements and thickness evaluation on Caucasian scalp hair photographs in the range of 100-μm-thick hair fibers (less than 1% variation in a test using multiple technicians; personal unpublished data). By this approach measurements relate to sample population on a hair-to-hair basis and temper the apparent statistical superiority of microscopic measures. Indeed, we noticed that the sample collection and processing of clipped hair for microscopic display (and probably any other type of measurements) may be difficult to standardize and become sources of error.45 The method was applied with success to various genetic conditions, including male pattern baldness in vivo.3 This technique was able to document the earliest changes of hair growth in AGA in man. From shortening of anagen duration of thick hair, regression appears to evolve in sequence into a phase of further shortening associated with slower growth rates and more severe thinning, turning finally into a stage of reversible miniaturization without production of any visible hair. At this stage, one speculates that drug response is still possible before the follicle drops into total irreversible atrophy, though without scarring. 103.3.4.2 Variants of PTG Methods Subtle modifications in the preparation of the target site can help identification of the hair in the growing from resting phases, especially when less than optimal magnification lenses are used (e.g., less than 3×24). Indeed, after clipping the hair short (first step of the PTG as control for density or hair counts), a close shave will further reduce the visible length of the hair fiber. Then, usually 3 rather than 2 days later, the second photograph is taken. A new hair count of the long hair fibers reflects anagen hair follicles. This procedure has been used to monitor changes occurring after finasteride in man with AGA, demonstrating a significant induction of growth compared with placebo.46 When the photographic camera is replaced by a video or charge coupled device (CCD) camera equipped with
891
specific lenses, other variants of PTG recording are obtained. In fact, reports in Orientals and Caucasians were published. In the latter subjects, the contrast between hair and scalp seems favorable for the application of this method, and the low figures of hair density could possibly be racial in origin.21 As shown already,27,45 the use of CE is advisable — especially for Caucasian hair — even with the use of computer applications for textile fiber analysis47 or hair recognition software already developed for other applications.48 Such a system has been proposed49 and the promoters of the method recognize that the explored area is small (less than 0.25 cm2). This is yet another source of variability because the number of hairs so analyzed is definitely less than 100 and drops below statistically acceptable population sample size.5 Furthermore, all of these automated systems generate numbers without securing that all hairs in the field were detected. Various commercial brands have been proposed, such as Trichoscan49 or Capillicare (data in file), and customers reported lack of satisfaction with such methods because the dream of an easy and fast method vanishes as it remains time-consuming. The method analyzes only a limited number of clinical situations, and even when all clinical conditions are suitable for the computer, it does not generate all data that may be of interest to the observer in terms of diagnosis, prognosis, or any other clinically relevant information. More importantly, independent clinical observers50 also state that “the potential of computerassisted technology in this field is yet to be maximized and the currently available tools are less than ideal.” 103.3.4.3 Future Trends in Computerized Methods More than a decade ago, we listed the different problems arising when automated computer-assisted image analysis (ACAIA) comes into the scope of hair growth measurement.51 Some problems have been solved. Accurate analysis still requires expert human intervention during the sampling and processing of images. Today we are easily coping with the three dimensions involved in hair; we mastered the requirements in terms of enlargement factor or pixel size, immersion, and polarized light. We also wiped out the problem of backscattering of light and other problems linked to presence of sebum droplets, sweat, and scales. More recent developments helped us to clear the scene from any other loosely attached elements, such as exogen hair (application patent process, method and apparatus for removing nonadherent elements from the skin of living beings and measuring the hair loss of living beings, Skinterface, patent PCT/EP 02/06434, June 12, 2002), and add information on trichostasis, a largely unsuspected phenomenon that significantly influences hair counts. Increasing contrast between scalp and hair allows us to express data in a single case observation or to perform accurate
892
Handbook of Non-Invasive Methods and the Skin, Second Edition
time-course studies of hair growth changes using highly standardized operating and monitoring procedures. As an example of the technological resolving power, we recently discussed hair pigmentation during aging and how it might affect the outcome of a clinical trial.7,52 Initially, nonvisible hair under conventional PTG may become visible for drug effects such as increased pigmentation, initiation of shortlived but thick anagen hair, etc. The spectrum of follicular bioresponses is actually underestimated and requires sophisticated investigations.
1
3
5
2 4 a
b
c
d
e
6 78
103.3.4.4 Detection of Nonvisible Hair: Less Contrasted and Thinning Hair Clinicians occasionally express their fears about new information generated by these up-to-date technologies, such as taking into account that thin hair is becoming visible.10 This clearly illustrates that the message that contrast enhancement also makes thick hair more visible did not come through. There is only one way to know whether miniaturized hair has any potential meaning or is just useless as a potential site for drug action: use a method that is able to measure their presence and monitor their potential modification over time. Anything else should be considered as speculative issues contributing to maintaining hair science and technology in the marshlands of “tricho-quackery.” As a practical example of a critical comparative study, we refer to the illustrations and comments in Figure 103.8. Therefore, we conclude that fully automated analysis systems,31 notwithstanding the apparent easiness of more recent systems,49 remain generally unsatisfactory when a detailed description of hair variables, such as hair counts, diameter — especially for thinning hair — and growth rate, is required. Biologically significant data in terms of follicle distribution, productivity, cycling, regression, or progression are not yet published — but will be in the near future — with our updated technologies. We do hope that the work done during the last decade will definitely bring such systems to the status of medically acceptable as diagnostic, prognostic, and therapeutic monitoring tools.
103.4 CONCLUSION A bold statement would be to say that the assessment of hair loss requires some experience and a lot of technological effort in order to grasp all the parameters involved in hair measurement. There was a time when some colleagues argued that measurement methods were not necessary, as the patient could tell when hair was growing or not. It is obvious that this is untrue as soon as one enters the continuum of the hair replacement process, especially when some hair is still present. Such a statement also looks outdated when one acknowledges the
FIGURE 103.8 Quality process evaluation comparing manual vs. automated hair analysis software. Here we show a typical example of quality process evaluation — as it is usually practiced in our laboratory — comparing results of our routine hair processing method vs. those generated by the automated hair analysis software. Image processing results: An image (a, source image) was downloaded in July 2003 from a website (trichoscan.com) displaying a full automatic system for hair analysis. Accordingly, after the automated hair selection (b, yellow outline of selected elements by software) and analysis (c, binary image of selected elements), the same source document (a) was also submitted to a technician for routine processing in our laboratory. The technician had no information about the origin of the document or results of automated analysis. This modality is part of our routine processing quality control (d). Our technician identified 43 hair segments in the circular outline of that image. Fourteen hair fibers had a thickness estimated to be less than 40 μm and 29 were considered thicker hair (40 μm). Quality analysis of process: At the final revision by the author, discrepancies were noted between the two processing modalities. They are outlined visually (e) against a pink background. One very thin hair partly hidden within a trio (white rectangle with arrowhead) was missed by both measurement systems. A display of hair results from automated analysis of image (a) is shown in (f): 29 hairs were counted in the 0.227 cm2 area (parts of circular outline), generating a density of 127.8 hair/cm2. The automated system also missed seven other hairs that were identified by our technician (empty numbered spaces in (e)). Automated counts also included three hairs crossing the edges (white rectangles, top (one hair) and bottom (two hairs) of (e)). Hairs that stuck together were not individualized by the automated system. Results: While the automated analysis counted 29 hairs, we found 42 in the same area. The technician’s error (1/43) is much lower than the automation error (12, or 13/43), i.e., 2.3% instead of 30.2% of the total sample. Conclusion: Hair evaluation as performed with automated systems does not match the quality standards prevailing in our laboratory. Many problems identified in 198931 still cause problems for automatic evaluation in 2004.
Photographic and Computerized Techniques for Quantification of Hair Growth
importance of the placebo effect. Indeed, subjective evaluation may reach 60% or more satisfaction, while significantly decreased hair counts clearly document the natural worsening of the condition.17 Our experience points to the fact that a combination of a highly sensitive and precise analytical approach with a global calibrated method seems advisable in the context of kinetic monitoring of hair growth and hair loss in the hair clinic in general, and this is warmly recommended in the context of efficacy analysis of new (and recognized) compounds in future clinical trials.
REFERENCES 1. Stenn KS, Paus R. Controls of hair follicle cycling. Physiol Rev 81:449–494, 2001. 2. Sinclair R, Jolley D, Mallari R, Magee J, Tosti A, Piracinni BM, Vincenzi C, Happle R, Ferrando J, Grimalt R, Leroy T, Van Neste D, Zlotogorski A, Christiano AM, Whiting D. Morphological approach to hair disorders. J Invest Dermatol Symp Proc 8:56–64, 2003. 3. Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A. Hair Science and Technology. Skinterface, Tournai, 2003. 4. Van Neste D. Folliculogram demonstrates more anagen hair roots in male androgenetic alopecia after one year treatment with finasteride 1mg/d. In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 311–316. 5. Van Neste DJJ. Hair growth evaluation in clinical dermatology. Dermatology 187:233–234, 1993. 6. Van Neste DJJ. Contrast enhanced phototrichogram (CE-PTG): an improved non-invasive technique for measurement of scalp hair dynamics in androgenetic alopecia — validation study with histology after transverse sectioning of scalp biopsies. Eur J Dermatol 4:326–331, 2001. 7. Van Neste D. Thickness, medullation and growth rate of female scalp hair are subject to significant variation according to pigmentation and scalp location during ageing. Eur J Dermatol 1:1–2, 2004. 8. Van Neste D, Hughes TC, Herd H, Gibson WT. Thickness, medullation and growth rate of human scalp hair are subject to significant variation according to pigmentation and scalp location. Australas J Dermatol 38 (Suppl. 2):19, 1997 (abstract). 9. Van Neste D, Demortier Y. Detailed monitoring of hair cycle transitions in vivo using contrast enhanced phototrichogram (CE-PTG). In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 211–222. 10. Olsen EA. Current and novel methods for assessing efficacy of hair growth promoters in pattern hair loss. J Am Acad Dermatol 48:253–262, 2003.
893
11. Van Neste D, Leroy T, Sandraps E. Validation and clinical relevance of a novel scalp coverage scoring method. Skin Res Technol 9:64–72, 2003. 12. Sandraps E, Leroy T, Demortier Y, Van Neste D. Validation and clinical relevance of a novel scalp coverage scoring (SCS) method. In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 255–270. 13. Van Neste D. My management plan of the male patient with androgenetic alopecia. In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 301–310. 14. Van Neste D. Scalp coverage scoring (SCS) documents natural worsening within less than 12 months in a majority of men with androgenetic alopecia (AGA). In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 271–276. 15. Van Neste D, Leroy T, de Ramecourt A. Hair removal and hair follicle targeting. In Hair Science and Technology, Van Neste D, Blume-Peytavi U, Grimalt R, Messenger A, Eds. Skinterface, Tournai, 2003, pp. 401–412. 16. Canfield D. Photographic documentation of hair growth in androgenetic alopecia. Dermatol Clin 14:713–721, 1996. 17. Kaufman KD. Long-term (5-year) multinational experience with finasteride 1 mg in the treatment of men with androgenetic alopecia. Eur J Dermatol 12:38–49, 2002. 18. Gibbons RD. Computer-aided quantification of scalp hair. Dermatologic Clinics 4, 627–640, 1986. 19. Gibbons RD, Fiedler-Weiss VC, West DP, Lapin G. Quantification of scalp hair: a computer-aided methodology. J Invest Dermatol 86:78–82, 1986. 20. Guarrera M, Ciula MP. A quantitative evaluation of hair loss: the phototrichogram. J Appl Cosmetol 4:61–66, 1986. 21. Hayashi S, Miyamoto I, Takeda K. Measurement of human hair growth by optical microscopy and image analysis. Br J Dermatol 125:123–129, 1991. 22. Van Neste DJJ, Dumortier M, De Brouwer B, De Coster W. Scalp immersion proxigraphy (SIP): an improved imaging technique for phototrichogram analysis. J Eur Acad Dermatol Venerol 1:187–191, 1992. 23. Courtois M, Loussouarn G, Hourseau C, Grollier JF. Hair cycle and alopecia. Skin Pharmacol 7:84–89, 1994. 24. Van Neste D, De Brouwer B, De Coster W. The phototrichogram: analysis of some technical factors of variation. Skin Pharmacol 7:67–72, 1994. 25. Guarrera M, Rebora A. Anagen hairs may fail to replace telogen hairs in early androgenic female alopecia. Dermatology 192:28–31, 1996. 26. Guarrera M, Semino MT, Rebora A. Quantitating hair loss in women: a critical approach. Dermatology 194:12–16, 1997. 27. Blume U, Ferracin I, Verschoore M, Czernielewski JM, Schaefer H. Physiology of the vellus hair follicle: hair growth and sebum excretion. Br J Dermatol 124:21–28, 1991.
894
Handbook of Non-Invasive Methods and the Skin, Second Edition
28. Rushton DH, De Brouwer B, De Coster W, Van Neste D. Comparative evaluation of scalp hair by phototrichogram and unit area trichogram analysis within the same subjects. Acta Derm Venereol (Stockh) 73:150–153, 1993. 29. Van Neste DJJ, Rushton DH. Hair problems in women. Clin Dermatol 15:113–125, 1997. 30. Halloy J, Bernard BA, Loussouarn G, Goldbeter A. Modeling the dynamics of human hair cycles by a follicular automaton. Proc Natl Acad Sci USA 97:8328–8333, 2001. 31. Van Neste D, Dumortier M, De Coster W. Phototrichogram analysis: technical aspects and problems in relation to automated quantitative evaluation of hair growth by computer-assisted image analysis. In Trends in Human Hair Growth and Alopecia Research, Van Neste D, Lachapelle JM, Antoine JL, Eds. Kluwer Academic Publishers, Lancaster, U.K., 1989, pp. 155–165. 32. Headington JT. Histological findings in androgenic alopecia treated with topical minoxidil. Br J Dermatol 107(Suppl. 22):20–21, 1982. 33. Headington JT, Novak E. Clinical and histologic studies of male pattern baldness treated with topical minoxidil. Curr Ther Res 36:1098–1106, 1984. 34. Whiting DA. The value of horizontal sections of scalp biopsies. J Cutan Aging Cosmet Dermatol 1:165–173, 1990. 35. Headington JT. Telogen effluvium. New concepts and review. Arch Dermatol 129:356–363, 1993. 36. Whiting DA. Diagnostic and predictive value of horizontal sections of scalp biopsy specimens in male pattern androgenetic alopecia. J Am Acad Dermatol 28:755–763, 1993. 37. Whiting DA. Scalp biopsy as a diagnostic and prognostic tool in androgenetic alopecia. Dermatol Ther 8:24–33, 1998. 38. Tsuji Y, Ishino A, Hanzawa N, Uzaka M, Okazaki K, Adachi K, Imamura S. Quantitative evaluations of male pattern baldness. J Dermatol Sci 7:136–141, 1994. 39. Courtois M, Loussouarn G, Hourseau C, Grollier JF. Ageing and hair cycles. Br J Dermatol 132:86–93, 1995. 40. Courtois M, Loussouarn G, Hourseau S, Grollier JF. Periodicity in the growth and shedding of hair. Br J Dermatol 134:47–54, 1996.
41. Ishino A, Uzuka M, Tsuji Y, Nakanishi J, Hanzawa N, Imamura S. Progressive decrease in hair diameter in Japanese with male pattern baldness. J Dermatol 24:758–764, 1997. 42. Rushton DH. Androgenetic alopecia in men: the scale of the problem and prospects for treatment. Int J Clin Pract 53:50–53, 1999. 43. Dawber R, Van Neste D. Hair and scalp disorders . In Common Presenting Signs, Differential Diagnosis and Treatment. Martin Dunitz, London, 1995. (Note: New edition planned in 2004.) 44. Rushton DH. Chemical and Morphological Properties of Scalp Hair in Normal and Abnormal States. University of Wales, Cardiff, 1988. 45. Leroy T, Van Neste D. Contrast enhanced phototrichogram pinpoints scalp hair changes in androgen sensitive areas of male androgenetic alopecia. Skin Res Technol 8:106–111, 2002. 46. Van Neste D, Fuh V, Sanchez-Pedreno P, Lopez-Bran E, Wolff H, Whiting D, Roberts J, Kopera D, Stene JJ, Calvieri S, Tosti A, Prens E, Guarrera M, Kanojia P, He W, Kaufman K. Finasteride increases anagen hair in men with androgenetic alopecia. Br J Dermatol 143:804–810, 2000. 47. Van Neste D, De Coster W. Phototrichogram: technical problems in relation with automated quantitative evaluation of hair growth by computer assisted image analysis. Nouv Dermatol 7(Suppl. 1):56, 1988. 48. Lee T, Vincent NG, Gallagher R, Coldman A, McLean D. Dullrazor®: a software approach to hair removal from images. Comput Biol Med 27:533–543, 1997. 49. Hoffmann R. TrichoScan: combining epiluminescence microscopy with digital image analysis for the measurement of hair growth in vivo. Eur J Dermatol 11:362–368, 2001. 50. Chamberlain AJ, Dawber RP. Methods of evaluating hair growth. Australas J Dermatol 44:10–18, 2003. 51. Van Neste D. Dynamic exploration of hair growth: critical review of methods available and their usefulness in the clinical trial protocol. In Trends in Human Hair Growth and Alopecia Research, Van Neste D, Lachapelle JM, Antoine JL, Eds. Kluwer Academic Publishers, Lancaster, U.K., 1989, pp. 143–154. 52. Van Neste D, Tobin DJ. Hair cycle and hair pigmentation: dynamic interactions and changes associated with aging. Micron 1:1–2, 2004.
of the Mechanical 104 Measurement Strength of Hair R. Randall Wickett College of Pharmacy, University of Cincinnati, Cincinnati, Ohio
CONTENTS 104.1 Introduction ..........................................................................................................................................................895 104.1.1 Stress–Strain Curves ..............................................................................................................................895 104.1.2 The Two-Phase Model of the Hookean Region....................................................................................896 104.1.3 α-β Transformation in the Yield Region...............................................................................................896 104.1.4 The Series Zone Model and the Postyield Region ...............................................................................896 104.1.5 Variations among Fiber Types ...............................................................................................................897 104.2 Object ...................................................................................................................................................................897 104.2.1 Tensile Measurements of Hair Damage ................................................................................................897 104.2.2 Bending and Torsional Measurements ..................................................................................................897 104.2.3 Chemical Relaxation Methods...............................................................................................................897 104.3 Methodological Principle .....................................................................................................................................898 104.3.1 Overview ................................................................................................................................................898 104.3.2 Modulus Calculation..............................................................................................................................898 104.3.3 Tensile Testers........................................................................................................................................898 104.4 Sources of Error ...................................................................................................................................................898 104.4.1 Variability in Hair ..................................................................................................................................898 104.4.2 Relative Humidity ..................................................................................................................................899 104.4.3 Other Sources of Error ..........................................................................................................................899 104.5 Recommendations ................................................................................................................................................899 104.5.1 Reducing Variability ..............................................................................................................................899 104.5.2 Determination of Hair Diameters ..........................................................................................................899 104.5.3 Gripping the Hair...................................................................................................................................899 104.5.4 Breaking Strength from an Inexpensive Device ...................................................................................900 References .......................................................................................................................................................................900
104.1 INTRODUCTION 104.1.1 STRESS–STRAIN CURVES The mechanical behavior of hair and other keratin fibers is most conveniently, and thus most frequently, measured in extension. Figure 104.1 illustrates typical stress–strain curves for adjacent sections of the same hair obtained either in air at 40% relative humidity (RH) or immersed in water. These curves can be characterized by three different regions. In the first region the curve is approximately linear and a slope can be determined. This is called the Hookean region, and it extends to about 102% of the equilibrium length of the fiber (2% strain). Between 2 and
4% strain the curve “turns over” into the yield region. In the yield region, very little increase in force is required to increase extension. In the postyield region, which begins at about 25% strain in the dry fiber and 28% strain in the wet fiber, the force again increases markedly with strain. For this particular hair, under the conditions tested, the slope in the postyield region was about one fifth of that in the Hookean region of the dry fiber. There is little difference in postyield slopes between the wet and dry sections of the fiber. Published reports on the mechanical properties of keratin fibers date back to the work of Speakman1–3 in the 1920s. Since that time, extensive research on hair and wool has led to an interpretation of each region of the 895
896
Handbook of Non-Invasive Methods and the Skin, Second Edition
is close to that of ice, as expected for a hydrogen-bonded network.17 The matrix contributes viscous forces that decay with time, causing stress relaxation. The viscosity of the matrix decreases greatly as the water content of the fiber increases.4,11 The two-phase model of keratin fibers accounts well for the effects of water on the mechanical properties,2,11,18 the effect strain rate on Young’s modulus,4,19 the stress relaxation behavior in the Hookean region,13,20 and the behavior of wet, dry, and permanently set fibers in torsion.2,9,21
50 Hookean slopes
45 40
40% RH
Grams force
35
Water
30 25 Post yield region
20 15
Yield region
10 5 0
0
10
20
30 % Strain
40
50
60
FIGURE 104.1 Stress/strain curves for different sections of the same hair at 40% RH and in water.
stress–strain curves of keratin fibers in terms of changes occurring in the molecular structure. Feughelman4 has reviewed this work very thoroughly and his review is highly recommended to anyone with a serious interest in the mechanical properties of hair. The discussion below is confined to a brief overview.
104.1.2 THE TWO-PHASE MODEL OF THE HOOKEAN REGION In the region of the force extension curve below 1.5% extension keratin fibers are generally considered to behave as Hookean springs. Bendit5,6 has pointed out that the curve is not truly Hookean in this region. However, the curve is approximately linear and Young’s modulus of elasticity can be calculated.4,7 Since this region has been referred to as Hookean for at least 50 years, it is likely that the terminology will persist for the foreseeable future.4 The mechanical properties of hair or wool in the Hookean region are well explained by the two-phase model of Feughelman.4,8–14 This model considers the mechanical properties of the fiber to be determined by a water-impenetrable phase, C, the microfibrils, and a waterpermeable phase, M, the matrix. The microfibrils are primarily composed of a-helical proteins aligned parallel to the fiber axis,15 and the matrix is composed of water and high-sulfur proteins that may be globular.16 The composite may be modeled mechanically as a fixed Hookean spring in parallel with a spring and viscous dashpot in series.4,10 The spring contributes about 1.4 × 109 Pa to the Young’s modulus and is contained in the water-impenetrable microfibrils. The main resistance to extension of the microfibrils probably comes from the hydrogen bond network in the a-helical proteins. When a keratin fiber is immersed in liquid nitrogen to prevent segment mobility, the modulus
104.1.3 α-β TRANSFORMATION REGION
IN THE
YIELD
Above extensions of about 2 to 3% the stress–strain curve turns over into the yield region. The stress does not increase markedly until about 25% extension. The mechanical properties of a fiber extended into this region can be recovered by relaxing the fiber in water overnight if the fiber is not held too long in extension3 and the extension is carefully confined to the yield region. As we shall see, this fact is of great practical importance in designing protocols to measure hair strength. High-angle x-ray diffraction results have demonstrated that there is a progressive loss of α-helical content and a concomitant increase in β-sheet as a fiber is extended through the yield region.24 By the end of the yield region about 30% of the original α-helix has been unfolded. It has also been shown24,25 that the mechanical behavior of keratin in the yield region can be completely accounted for by application of a Burte–Halsey26 model. The fiber is considered to contain a continuum of units that can exist in a short state, A (α-helix), or an extended state, B (βsheet), with an energy barrier between the states. The yield region corresponds to a phase transition between state A and state B at constant stress. This first-order phase transition, producing a large length change at constant stress and temperature, is analogous to the transformation of water to steam, producing a large-volume change at constant temperature and pressure.
104.1.4 THE SERIES ZONE MODEL POSTYIELD REGION
AND THE
Speakman3 found the postyield slope to be independent of the water content of the fiber. The increase in stiffness in the postyield region was shown to result from a covalently bonded network involving cystine. The postyield slope has been shown to be dependant on the disulfide content of the fibers.27,28 Extension to about 50% strain leads to loss of all a-helical structure in the fiber24 as judged by x-ray diffraction. The behavior of keratin fibers in the yield and postyield regions has been interpreted in terms of a series
Measurement of the Mechanical Strength of Hair
zone model.29,30 The model postulates two alternating zones, X and Y, along the microfibrils. The X zones contain the 30% of the α-helices that unfold reversibly in the yield region. The Y zones contain regions of α-helix that cannot be unfolded without the breakdown of disulfide bonds. Thus, unfolding of the Y zones in the postyield region is irreversible.
897
treatments42 on hair. These studies have all shown tensile measurements to be very useful for the study of damaging treatments that break disulfide bonds in the cortex. However, Robbins and Crawford43 have shown that treatments that cause severe damage to hair cuticle may have little or no effect on tensile properties.
104.2.2 BENDING 104.1.5 VARIATIONS
AMONG
FIBER TYPES
Much of the work on the mechanical properties of keratin fibers has been carried out with wool. In the discussion above, the term keratin fiber has been used to refer to either hair or wool. The mechanical behavior of the two fibers is very similar. A comparative study by Menkart et al.31 found the elastic modulus and stress at 20% extension of hair to be somewhat higher than that of wool, while Chaiken and Chemberlain32 found the dynamic elastic modulus of hair to be nearly equivalent to that of wool. Tolgyesi et al.33 found beard hair to behave similarly to head hair in extension, but to have a slightly lower elastic modulus and stress at 20% extension. In general, head hair, beard hair, and wool may be considered similar enough in behavior that all of the conclusions about structure and mechanical properties discussed above can be considered to be equally valid for each structure.
104.2 OBJECT 104.2.1 TENSILE MEASUREMENTS
OF
HAIR DAMAGE
Measurements of hair tensile properties are most frequently made to assess the effects of chemical treatments on hair strength.7 The mechanical properties of hair are greatly affected when the number of disulfide cross-links is reduced. This is especially true of wet hair in the yield and postyield regions of the stress–strain curve.4,7,27,28,34–37 A typical protocol is to strain an untreated hair into the yield region and measure either the force or the work of extension. The work of extension is the area under the force vs. extension curve. Beyak et al.38 assessed the effects of bleaching and permanent waving on the relative stress to extend a hair to 15% strain before and after treatment, the 15% index (I15), in water. They found that a 30-min bleach treatment reduced I15 by an average of 10% and a 5-min permanent wave treatment reduced I15 by about 13%. Wolfram et al.39 reported that a 30-min bleaching treatment reduced the yield stress by about 12%, and Robbins7 presented data showing that a commercial permanent wave caused an 18% reduction in the work to extend a hair to 20% in the wet state compared to an 11% decrease in the work to extend hair to 20% in the dry state. Tensile measurements have also been used to characterize the effect of ultraviolet radiation,40 surfactant binding,41 and chlorine
AND
TORSIONAL MEASUREMENTS
A further objective of mechanical measurements on hair is to understand the processes involved in setting and permanent waving. Measurements of extensional properties are not necessarily the best way to achieve this goal. Bogaty44 pointed out that the behavior of hair under torsional and bending strains is very important to the permanent waving process because forming a curl from straight hair involves a combination of twisting and bending deformations. He found that permanent waving decreased the torsional rigidity of hair in the wet state, but actually increased it slightly at 65% RH. Wolfram and Albrecht45 made torsional measurements on hair and concluded that the cuticle is very stiff in the dry state and may make a significant contribution to the torsional rigidity, especially for fine hairs. However, in the wet state the cuticle was found to be so plasticized as to make no contribution to mechanical behavior. Scott and Robbins46,47 described a practical, balanced fiber method for measuring the bending stiffness of hair. A long hair is draped over a small wire with small weights attached to each end. The bending stiffness can be calculated from the distance between the two ends. It is also possible to measure bending strength by a three-point beam deflection method, and this method has been applied to measuring the stiffness of beard hairs.48 The balanced fiber method has the disadvantage of requiring a relatively long fiber, but in the author’s experience it is far easier to use than three-point bending methods.
104.2.3 CHEMICAL RELAXATION METHODS The dramatic effect that breaking disulfide bonds has on the tensile properties of hair can be used to study the kinetics of the reduction reaction. If a hair is stress relaxed to a constant level of force at a constant extension in water or buffer, and then the solution is switched to reducing agent, the force decays with time, due to the breaking of disulfide bonds,49–52 and kinetic parameters can be determined. An alternate method is to repeatedly stretch the hair in the linear region, measuring the reduction in elastic modulus as the reaction proceeds. 53 Wortman and coworkers54,55 have used a combination of these methods to attempt to predict permanent set based on relaxation parameters. I have recently reviewed these methods along with other work on the effects of permanent waving on the physical properties of hair.56
898
Handbook of Non-Invasive Methods and the Skin, Second Edition
104.3 METHODOLOGICAL PRINCIPLE 104.3.1 OVERVIEW This discussion of methodology will focus on use of extensional measurements to evaluate hair strength in tension, including determination of elastic modulus, yield stress, and breaking strength. These are the most widely used means to evaluate treatment effects on hair. Robbins7 has provided a thorough discussion of other methods for the evaluation of hair physical properties.
104.3.2 MODULUS CALCULATION Young’s modulus of elasticity, E (Y in some older papers), is calculated from the slope of the Hookean region of the stress–strain curve (Figure 104.1). E is defined as follows: E = ΔF*L/ΔL*A
(104.1)
where DF is the change in force induced by a change in length, DL, L is the equilibrium length of the fiber, and A is the cross-sectional area. For example, assume that a cylindrical hair of 80 mm (8 × 10–5 M) diameter, 5 × 10–9 M2 cross-sectional area, requires 10.2 grams-force (0.1 Newtons) to extend 1%: E = 100*0.1N/5 × 10–9 M2 = 2 × 109 N/M2 (104.2) A Newton/M2 is 1 Pa, so the modulus is 2 × 109 Pa. Older papers report E or Y in dynes/cm2. A pascal is 10 dynes/cm2, so the modulus of this fiber is 2 × 1010 dynes/cm2. When reporting E, the rate of extension must be specified because the slope of the Hookean region is known to vary with extension rate7,19 due to stress relaxation in the matrix.4 Young’s modulus is typically about 1.5 to 2.0 × 109 Pa for wet hair and 3.5 to 4.5 × 109 Pa for dry hair, depending on strain rate and hair source used.
104.3.3 TENSILE TESTERS Stress–strain measurements on hair are usually made with commercially available tensile testers. Most measurements reported in the literature over the years have been made using one of the several varieties manufactured by the Instron® Corporation, usually a table model such as the 4201. The Instron® uses a large, very robust screw drive system to move either a crosshead containing a load cell or a large mechanical stage at a constant rate of extension. Modern Instrons® are computer controlled, and a wide variety of stress–strain protocols can be programmed by using “canned” software provided with the instrument. By using a programming language such as PASCAL and drivers provided with the instrument, soft-
ware can be created to collect data using virtually any protocol that the programmer imagines. Advantages of the Instron® are its high precision and flexibility. Forces from less than 0.1 g to several kilograms can be measured, depending on the load cell selected, and extensions from less than a millimeter to more than a meter can be accurately performed. Disadvantages are its cost, $50,000 and up, depending on options, and its rather large size. Even a table model Instron® weighs a few hundred pounds and takes up a considerable amount of lab space. It is definitely not a portable instrument. A recently available option for measuring the mechanical properties of hair is the Dia-stron Rheometer (DR). The DR is portable. The measuring jig weighs just 3 kg, and the control unit is about the size of a typical personal computer box. The cost of the DR is about one fourth of that of an Instron®. The curves in Figure 104.1 were obtained using the DR. While completely adequate for obtaining stress–strain curves from hair, the DR has some limitations. The light weight of the DR measuring jig that makes it portable also makes it susceptible to vibrational noise, such as might arise from general activity in the lab. Stress–strain and stress relaxation protocols are entered into one of two methods available to the user at any one time. Each protocol allows one extension, and one compression phase with data collection after each, and each extension–compression cycle can be repeated several times. This gives considerable flexibility to the operator, but not the extreme flexibility provided by the Instron®. I should also point out that the software we obtained with the DR does not calculate Young’s modulus correctly. Whenever using a software package provided with an instrument, it is wise to check all calculations by hand.
104.4 SOURCES OF ERROR 104.4.1 VARIABILITY
IN
HAIR
A major source of error in measurements on hair is the inherent variability in the thickness and shape of hair. This can be compensated for to some extent by determining the dimensions of each hair and normalizing the result to cross-sectional area. Measurements of breaking strength are particularly prone to variability, and even different sections of the same hair will often break at different strengths. Another source of error when evaluating treatment effects is the inherent difference in treatment response between individuals. For example, reaction rates to reducing agent may vary by a factor of 10 between individuals who do not have a history of chemical treatment, and may vary more if chemically treated hair is used.50,52 Some possible solutions to the variability problem are discussed below.
Measurement of the Mechanical Strength of Hair
899
104.4.2 RELATIVE HUMIDITY When making dry measurements, exact control of relative humidity is very important. This is well illustrated by the fact that simple room relative humidity gauges often use a horsehair to move the dial as the room RH changes. Thus, variations in RH can be a significant source of error when making dry measurements of elastic modulus or yield stress. As can be seen in Figure 104.1, postyield slope and breaking strength are much less sensitive to RH.
104.4.3 OTHER SOURCES
OF
ERROR
Other sources of error are slipping of the hair gripping system, damage to the hair by the gripping system, and slack in the hair resulting in an inaccurate value of equilibrium length. Slack can be easily accounted for by modern computerized methods, which can recalculate the equilibrium length when significant force is first sensed. Gripping problems are discussed in detail below.
TABLE 104.1 Dia-stron® Hair Breaking Compared to Lever Breaking Hair # 3 5 1 10 9 2 8 7 6 4 Average
Dia-stron® 183.0 164.0 126.0 106.0 105.0 102.0 101.5 93.5 79.5 70.0 113.05
Lever
Difference Diast – Lever
193.0 159.0 150.0 111.0 97.0 89.0 95.0 95.0 65.0 82.0 113.6
+10 –5 +24 +5 –8 –13 –6.5 +1.5 –14.5 +12 0.55
104.5 RECOMMENDATIONS
variations in the hair, as described below. The only other alternative is to run literally hundreds of samples to get statistical significance.
104.5.1 REDUCING VARIABILITY
104.5.2 DETERMINATION
The best way to account for variability between hairs when evaluating treatment effects is to use each hair as its own control. As discussed above, the mechanical properties of a keratin fiber can be recovered if the hair is not strained into the postyield region.1,4,7,22 A typical protocol is that used by Beyak et al.38 Hairs were strained in water to 15% extension on an Instron® tensile tester and then soaked in water for 16 hours, treated, and rerun. The change in grams-force required to reach 15% extension was evaluated. Force values at 15% extension from 25 hairs ranged from 11.4 to 35.5 g in the first run. A second run without a treatment between showed an average change in force for each hair of only 0.33% of the original force, with a standard deviation of 2.68%. The changes observed ranged from +5.9% to –5.7%. Bleach treatments caused an average reduction in force at 15% strain of 10 to 20%, depending on treatment time. This study and many others like it7,39–42 clearly show the value of using each hair as its own control. When doing breaking or chemical relaxation studies, it is not possible to use each hair segment as its own control. The next best approach is to cut the hairs into different sections and compare results between sections, as was done to obtain the data in Table 104.1, discussed below. Breaking values on different sections are not as reproducible as reruns of stress–strain curves to 20% extension on the same section of hair, but still provide a large improvement over simply comparing different hairs. When it is not possible to either use each hair as its own control or compare different sections of the same hair, one must attempt to at least normalize for the dimensional
Determining the cross-sectional dimensions of a hair is a difficult problem. Not only is hair a fine fiber, but it is also not necessarily uniform in cross section. While Caucasian hair is generally considered elliptical in shape, significant variations from ellipticity can occur. Robbins7 has reviewed methods for diameter determination and recommends the linear density method as the method of choice. I concur with his recommendation. To use the linear density method, a hair is cut to a given length and weighed on a microbalance. The fiber density is assumed to be 1.32 g/cm3 (Reference 7) and the cross-sectional area, A, in square centimeters, is then given by
OF
HAIR DIAMETERS
A = W/(1.32*L)
(104.3)
where W is the weight of the hair in grams and L is the length of the segment measured in centimeters. If necessary, the effective diameter of the hair can then be determined from simple geometry if the hair is assumed to be circular in cross section.
104.5.3 GRIPPING
THE
HAIR
An annoying difficulty when making tensile measurements on hair is the problem of obtaining a good grip on the hair without causing damage at the gripping point. This is especially vexing when one wants to stress the hair all the way to breaking. The plastic fiber grips provided by Instron® are totally inadequate, as most hairs will slip from the grips well before they break.
900
Handbook of Non-Invasive Methods and the Skin, Second Edition
I have previously crimped each end of the hair into short sections of fine-diameter aluminum tubing, which can then be gripped by any good spring-loaded clamp, but the sections must be cut and filed carefully to avoid having any sharp edges that could cut the hair. What seems to be a reasonable solution to the gripping problem has recently been provided by Dia-stron®. It sells 14-mm sections of 2-mm-diameter metal tubing with plastic tubing inside. These sections can be crimped onto the hair to provide good handles for gripping. Dia-stron® also sells a crimping tool for this purpose that is convenient for crimping onto each end of a standard 3-cm hair section. We have found that an adequate crimp can also be obtained by careful use of an electrician’s crimping tool, available in nearly any hardware store. Even using the Dia-stron® crimping system one must be careful to get a good crimp if making a hair-breaking measurement. We usually move the sample block slightly and recrimp at least twice to ensure a good crimp. Otherwise, the hair may slip before breaking.
104.5.4 BREAKING STRENGTH FROM AN INEXPENSIVE DEVICE If you want to screen a treatment for its effect on hair strength and do not have access to a tensile tester, I recommend trying a simple and inexpensive device we have developed to measure the dry-breaking strength of hair.57 The device is based on a lever principle and is shown schematically in Figure 104.2. A meter stick pivots on a bolt through a hole in its center. The hair is fastened to one end and a weight on a sliding hanger is moved slowly and carefully along the other side, away from the center hole. The breaking strength is Wt* (L2/L1) – Wc, where Wt is the total weight of the sliding hanger and attached weight, L2 is the distance from the center hole to the slide at break, L1 is the distance from the center hole to the hair, and Wc is the weight of the clip used to hold the hair. I have made sliding weight hangers from either a L1 L2 Sliding mount Pivot rod Meter stick Upper grip Hair Lower grip Weight
FIGURE 104.2 Schematic diagram of a simple device to measure hair breaking strength.
mirror mount bracket or a closet door mount. No doubt other objects could be used for this purpose. Our current hanger weighs 30 g, and we use an additional 150 g of balance weights attached using wire. If the hairs are mounted into the Dia-stron® crimps discussed above, the grips can be made from alligator clips. Ten hairs were cut into two sections each, and the breaking strength of one section was determined on the DR and the other on the lever device. The results are shown in Table 104.1. The agreement between the average values for breaking strength obtained from each device is remarkable. Of course, the lever device does not allow one to produce a stress–strain curve, but it seems to work well as a potential screening tool considering that it took about a half hour and $12 worth of materials purchased from a local hardware store to construct.
REFERENCES 1. Speakman, J.B., The gel structure of the wool fibre, J. Text. Inst., 17, T457, 1926. 2. Speakman, J.B., The rigidity of wool and its changes with adsorption of water-vapor, Trans. Faraday Soc., 25, 92, 1929. 3. Speakman, J.B., The intracellular structure of the wool fibre, J. Text. Inst., 18, T431, 1927. 4. Feughelman, M., The physical properties of alpha keratin fibers, J. Soc. Cosmet. Chem., 33, 385, 1982. 5. Bendit, E.G., Properties of the matrix in keratins. II. The “Hookean” region in the stress-strain curves of keratins, Text. Res. J., 48, 717, 1978. 6. Bendit, E.G., There is no Hookean region in the stressstrain curve of keratin or other viscoelastic fibers, J. Macromol. Sci. Phys., B17(1), 129, 1980. 7. Robbins, C.R., Chemical and Physical Behavior of Human Hair, 2nd ed., Springer-Verlag, New York, chap. 8. 8. Feughelman, M., A two phase structure for keratin fibers, Text. Res. J., 29, 223, 1959. 9. Mitchell, T.W. and Feughelman, M., The torsional properties of single wool fibers. I. Torque-twist relationships and torsional relaxation in wet and dry fibers, Text. Res. J., 30, 662, 1960. 10. Feughelman, M., The relation between structure and the mechanical properties of keratin fibers, Appl. Polym. Symp., 18, 757, 1971. 11. Feughelman, M. and Robinson, M.S., Some mechanical properties of wool fibers in the “Hookean” region from zero to 100% relative humidity, Text. Res. J., 41, 469, 1971. 12. Feughelman, M., Keratin, in Encyclopedia of Polymer Science and Engineering, Vol. 8, 2nd ed., John Wiley & Sons, New York, 1987. 13. Feughelman, M. and Robinson, M.S., Stress relaxation of wool fibers in water at extensions in the Hookean region over the temperature range 0˚–90˚C, Text. Res. J., 39, 196, 1969.
Measurement of the Mechanical Strength of Hair
14. Feughelman, M., A note on the water impenetrable component of a-keratin fibers, Text. Res. J., 59, 739, 1989. 15. Fraser, R.D.B., MacRea, T.P., and Suzuki, E., Structure of the a-keratin microfibril, J. Mol. Biol., 108, 435, 1976. 16. Fraser, R.D.B., Macrae, T.P., and Rogers, G.E., Molecular organization in alpha keratin, Nature, 193, 1052, 1962. 17. Feughelman, M. and Robinson, M.S., The tensile behavior of wool fibers in liquid nitrogen, Text. Res. J., 37, 705, 1967. 18. Breuer, M.M., The binding of small molecules to hair. I. The hydration of hair and the effect of water on the mechanical properties of hair, J. Soc. Cosmet. Chem., 23, 447, 1972. 19. Sikorski, J. and Woods, H.J., the effect of rate of extension on Young’s modulus of keratin fibers, Leeds Phil. Soc., 5, 313, 1950. 20. Wortmann, F.J. and De Jong, S., Analysis of the humidity-time superposition for wool fibers, Text. Res. J., 55, 750, 1985. 21. Feughelman, M., Microfibril/matrix relationships in the mechanical properties of keratin fibers. I. The torsional properties of “melted” and permanently set keratin fibers, Text. Res. J., 48, 518, 1978. 22. Feughelman, M., A note on the recoverability of mechanical properties of wool, J. Text. Inst., 59, T548, 1968. 23. Bendit, The α-β transformation in keratin, Nature, 179, 535, 1957. 24. Feughelman, M., Creep of wool fibres in water, J. Text. Inst., 45, T630, 1954. 25. Feughelman, M. and Rigby, B.J., A two energy state model for the stress relaxation and creep of wool fibres in water, in Proceedings of the International Wool Textile Conference, Australia, 1955, p. D-62. 26. Burte, H. and Halsey, G., A new theory of non-linear viscoelasticity, Text. Res. J., 17, 465, 1947. 27. Feughelman, M., The mechanical properties of permanently set and cystine reduced wool fibers at various relative humidities and the structure of wool, Text. Res. J., 33, 1013, 1963. 28. Cannell, D.W. and Carothers, L.E., Permanent waving: utilization of the post-yield slope as a formulation parameter, J. Soc. Cosmet. Chem., 29, 685, 1978. 29. Feughelman, M. and Haly, A.R., Structural features of keratin suggested by its mechanical properties, Biochim. Biophys. Acta, 32, 596, 1959. 30. Feughelman, M., The post-yield region and the structure of keratin, Text. Res. J., 34, 539, 1964. 31. Menkart, J., Wolfram, L.J., and Mao, I., Caucasian hair, Negro hair and wool: similarities and differences, J. Soc. Cosmet. Chem., 17, 769, 1966. 32. Chaiken, M. and Chemberlain, W.H., The propagation of longitudinal stress pulses in textile fibers, J. Text. Inst., 46, T44, 1955. 33. Tolgyesi, E., Coble, D.W., Fang, F.S., and Kairinen, E.O., A comparative study of beard and scalp hair, J. Soc. Cosmet. Chem., 34, 361, 1983.
901
34. Weigman, H.D., Rebenfield, L., and Danziger, C., The Role of Sulfhydryl Groups in the Mechanism of Permanent Setting of Wool, III Cirtel, Paris, 1965, Sec. 2, p. 319. 35. Hermans, K.W., Hair keratin, reaction, penetration and swelling in mercaptan solutions, Trans. Faraday Soc., 59, 1633, 1963. 36. Weigman, H.D. and Danziger, C.J., Effects of crosslinks on the mechanical properties of keratin fibers, Appl. Polym. Symp., 18, 795, 1971. 37. Robinson, M.S. and Rigby, B.J., Thiol differences along keratin fibers: stress/strain and stress-relaxation behavior as a function of temperature and extension, Text. Res. J., 55, 597, 1985. 38. Beyak, R., Meyer, C.F., and Kass, G.S., Elasticity and tensile properties of human hair. I. Single fiber test method, J. Soc. Cosmet. Chem., 20, 615, 1969. 39. Wolfram, L.J., Hall, K., and Hui, I., The mechanism of hair bleaching, J. Soc. Cosmet. Chem., 21, 875, 1970. 40. Beyak, R., Kass, G.S., and Meyer, C.F., Elasticity and tensile properties of human hair. II. Light radiation effects, J. Soc. Cosmet. Chem., 22, 667, 1971. 41. Breuer, M.M., The interaction between surfactants and keratinous tissue, J. Soc. Cosmet. Chem., 30, 41, 1979. 42. Fair, N.B. and Gupta, B.S., The chlorine-hair interaction. II. Effect of chlorination at varied pH levels on hair properties, J. Soc. Cosmet. Chem., 38, 371, 1987. 43. Robbins, C.R. and Crawford, R.J., Cuticle damage and tensile properties of human hair, J. Soc. Cosmet. Chem., 42, 59, 1991. 44. Bogaty, H., Torsional properties of hair in relation to permanent waving and setting, J. Soc. Cosmet. Chem., 18, 575, 1967. 45. Wolfram, L.J. and Albrecht, L., Torsional behavior of human hair, J. Soc. Cosmet. Chem., 36, 87, 1985. 46. Scott, G.V. and Robbins, C.R., A convenient method for measuring fiber stiffness, Text. Res. J., 39, 975, 1969. 47. Scott, G.V. and Robbins, C.R., Stiffness of human hair fibers, J. Soc. Cosmet. Chem., 29, 469, 1978. 48. Savenije, E.P.W. and De Vos, R., Mechanical properties of human beard hair, Bioeng. Skin, 2, 215, 1986. 49. Reese, C. and Eyring, H., Mechanical properties and the structure of hair, Text. Res. J., 20, 743, 1950. 50. Wickett, R.R., Kinetic studies of hair reduction using a single fiber technique, J. Soc. Cosmet. Chem., 34, 301, 1983. 51. Wickett, R.R. and Barman, B.G., Factors affecting the kinetics of disulfide bond reduction in hair, J. Soc. Cosmet. Chem., 36, 75, 1985. 52. Wickett, R.R. and Mermelstein, R., Single fiber stress decay studies of hair reduction and depilation, J. Soc. Cosmet. Chem., 37, 461, 1986. 53. Szadurski, J.S. and Erlman, G., The hair loop test: a new method of evaluating perm lotions, Cosmet. Toilet., 41(12), 41, 1984. 54. Wortman, F.J. and Souren, I., Extensional properties of hair and permanent waving, J. Soc. Cosmet. Chem., 38, 125, 1987.
902
Handbook of Non-Invasive Methods and the Skin, Second Edition
55. Wortman, F.J. and Kure, N., Bending relaxation properties of human hair and permanent waving performance, J. Soc. Cosmet. Chem., 41, 123, 1990. 56. Wickett, R.R., Disulfide bond reduction in permanent waving, Cosmet. Toilet., 106(7), 37, 1991.
57. Wickett, R.R., An inexpensive device to measure hair breaking strength, manuscript in preparation.
the Strength of Human 105 Evaluating Hair Sidney B. Hornby Neutrogena Corporation, Los Angeles, California
CONTENTS 105.1 Introduction ..........................................................................................................................................................903 105.2 Methodology.........................................................................................................................................................903 105.2.1 Preparation of the Sample .....................................................................................................................903 105.2.2 Importance of Hair Fiber Dimensions ..................................................................................................904 105.3 Instruments ...........................................................................................................................................................904 105.3.1 Cyclic Tester ..........................................................................................................................................904 105.3.2 Impact Loading ......................................................................................................................................905 105.3.3 Flexabrasion Method .............................................................................................................................905 105.4 Analysis of Hair Failure Data..............................................................................................................................905 105.4.1 Characteristic Life or Alpha ..................................................................................................................905 105.4.2 Shape Parameter or Beta .......................................................................................................................905 105.4.3 Calculating Characteristic Life Values from Experimental Data..........................................................905 105.4.4 Survival Probability Curves...................................................................................................................906 105.5 Concluding Remarks ............................................................................................................................................907 References .......................................................................................................................................................................907
105.1 INTRODUCTION Strong, healthy-looking hair is almost as important as the condition of the skin to many people. However, most people engage in styling routines that lead to progressive damage to the hair fibers, which frequently manifests itself in undesirable hair splitting and breakage. In Chapter 104 the measurement of the mechanical break strength of hair by tensile testing was described. Reduction in the break strength can be observed in hair that has been severely damaged, especially by bleaching and permanent waving.1 However, damage in the surface regions of the hair that does not involve significant changes to the structure of the hair cortex cannot easily be distinguished by tensile testing. Nor can the benefit of conditioners and certain treatments always be demonstrated.2 Accelerated wear tests such as cyclic or fatigue testing are routinely employed to analyze the propensity of everything from aerospace composites to mechanical motors to fail in ordinary use. Items are repeatedly stressed until they break or fail. The failure data are then analyzed to determine whether the item can withstand the rigors of regular usage over time. In this chapter, the applications
of three accelerated wear methodologies to human hair fibers are described.
105.2 METHODOLOGY 105.2.1 PREPARATION
OF THE
SAMPLE
Human hair can be obtained from volunteers or purchased from a supplier. The effects of applied conditioning treatments can be best distinguished if the substrate hair is predamaged. A commercial bleach treatment is a good method of predamaging hair, since this is a very common procedure. Single fibers should be randomly selected from the hair bundle. Then a suitable length of each fiber (depending on the instrument used) should be affixed at each end. An easy mounting system consists of polyvinyl chloride (PVC)-lined brass ferrules that are crimped over each end of the single hair filaments. The PVC lining provides a cushion so that the hair fiber is not cut or crushed. Typically, 20 single hair fibers are tested in each sample, and the experiment should always include a control sample. Since the viscoelastic properties of hair are greatly 903
904
Handbook of Non-Invasive Methods and the Skin, Second Edition
Failure cycles
2000 1500 1000 500 0 0.01
0.015 0.02 0.025 Stress applied (grams/square micron)
FIGURE 105.1 The relationship between stress applied to hair fibers under a constant cyclic load of 60 g and the corresponding cycles to break.
affected by moisture in the environment,3 the relative humidity in the test area should be carefully controlled.
105.2.2 IMPORTANCE
OF
HAIR FIBER DIMENSIONS
It is known that the thickness of individual hair fibers can vary widely even if they are obtained from the head of one volunteer.3 Therefore, since the load applied to each fiber is constant, the applied stress on each fiber varies with different fiber dimensions. The relationship between the failure cycles of hair fibers and the stress applied during cyclic testing has been explored for chemically unaltered (virgin) hair, bleached hair, and bleached hair treated with two different conditioning systems.2 It was found that, except at the very low stress values (high-fiber cross-sectional area) and very high stress values (low-fiber cross-sectional area), there was no clear systematic relationship between the applied stress and cycles required to break each hair (Figure 105.1). One explanation for this result is that at the high stress values, preexisting fiber damage in the form of cracks or flaws propagates easily and rapidly through the hair fiber, resulting in breakage at a relatively low number of stress cycles. Therefore, for small-diameter hairs, the number of cycles to break each hair fiber is more strongly dependent upon the number of significant faults in the hair than the number of applied cycles. Conversely, the applied stress in fibers with larger than usual diameters may not be high enough to cause preexisting flaws to rapidly progress right through the fiber. As a result, much higher numbers of cycles will be required before the hair fails. Therefore, at low applied stress levels, the numbers of cycles at break becomes more strongly dependent on the duration of cyclic fatiguing endured by the fiber, and less dependent on the presence of surface damage. Therefore, the cross-sectional areas of each hair fiber in the sample should be determined microscopically or by laser micrometer before testing. Fibers with cross-
FIGURE 105.2 Cyclic tester for evaluating hair strength by resistance to break.
sectional areas between 2500 and 5000 μ2 can be included in the sample, and fibers outside of this range should be discarded.
105.3 INSTRUMENTS 105.3.1 CYCLIC TESTER The cyclic tester (Dia-Stron, Ltd., Andover, U.K.) is an apparatus that is designed to stretch single fibers to a selected load at a given speed and then relax them, with these cycles repeating until the hair breaks.2,4 Each hair fiber is placed in turn into the machined sample holder of the cyclic tester either automatically or by hand. The sample holder is designed so that one end of the hair fiber is attached to a sensitive load cell and the other to a movable sophisticated drive system that employs a closed-loop controlled linear actuator (Figure 105.2). This drive system can be programmed to achieve a wide range of movement distances, elongation speeds, and accelerations, and is capable of positional accuracy of better than 10 μ. Each stress cycle applied to the hair filament consists of stretching at the selected speed (10 to 20 mm/s) until the load cell registers the predetermined load. The load should be chosen to avoid stretching hair fibers well into their yield region. Typically, the load can be set to between 40 and 60 g at an elongation speed of 20 mm/s.4 Once the appropriate load is detected, the direction of the moving arm reverses, removing the load from the fiber. These oscillations are repeated until the fiber breaks. The computer then records the number of elongation cycles elapsed at break. The data obtained are in the form of number of cycles at break of each hair fiber, which can be sorted in order of breaking. In addition, the computer can record the actual load elongation curves at selected intervals. Such data allow us to observe the changes in the viscoelastic behavior of the hair fiber as it is stressed.
Evaluating the Strength of Human Hair
105.3.2 IMPACT LOADING This method (TRI Fatigue Tester, TRI/Princeton, Princeton, NJ)5,6 can stress up to 40 hair fibers simultaneously, rather than sequentially. The fibers are suspended from individual holders in a canopy with a weight on each one. At the beginning of the experiment each weight rests in a pocket equipped with a switch on a movable platform underneath the fibers. Then the platform oscillates up and down with amplitude of movement sufficient to relax and then load the hairs in a cyclic fashion. As each hair fiber breaks, the corresponding switch is triggered and the computer records the cycles elapsed. At the end of the experiment the data obtained are in an array of number of cycles at break of each hair fiber, but no load elongation data are recorded.
905
on a normally distributed data cannot be used to compare different samples of hair. Instead, the experimental data are usually exponentially distributed and are best analyzed using the twoparameter form of the Weibull distribution,10–15 shown in Equation 105.1:
F ( x) = 1 − e
⎛ x ⎞ −⎜ ⎝ alpha ⎟⎠
beta
, for x > 0
(105.1)
F(x) is the probability that a given hair fiber will break at x cycles, alpha is the Weibull characteristic life, and beta is the Weibull shape parameter. To compare the relative breakage resistance of the hair fiber samples, we first need to understand alpha and beta.
105.3.3 FLEXABRASION METHOD
105.4.1 CHARACTERISTIC LIFE
This technique is different from the cyclic or impact fatigue testing described above. In addition to the stress imparted to each hair fiber by loading it, this apparatus simulates the abrasive effect of a comb or brush rubbing against individual hair fibers during grooming.7,8 The Flexabrasion apparatus (Croda, Inc., Edison, NJ) can test 20 fibers at once. Typically, 14-mm lengths are cut from adjacent sections of a single hair fiber to minimize the effect of fiber-to-fiber variations in cross-sectional area. Each specimen is glued onto mounts. Each fiber is then repeatedly run over a drawn tungsten wire under a weight at a 90˚ angle, as diagrammed in Figure 105.3.9 Typically a 2- to 4-mm section of the fiber is abraded at a frequency between 0.25 and 7 Hz in a temperature- and humiditycontrolled cabinet. At the end of the experiment the data obtained are in the form of an array number of cycles at break of each hair fiber.
The characteristic life of failure data is analogous to the mean of normally distributed data and is rigorously defined as the number of stress cycles elapsed when 63.2% of the hair fibers in the sample have broken. Therefore, a larger characteristic life of a sample suggests that it has a greater resistance to breaking during everyday brushing, combing, and styling than a sample of hair with a lower characteristic life value.
105.4 ANALYSIS OF HAIR FAILURE DATA The data obtained from each of the techniques described above are in the form of an array of the cycles required to break each hair fiber, from first fiber to break to the last. Failure data are usually not normally distributed, so calculating the average failure cycles of the sample based Reciprocating motion
105.4.2 SHAPE PARAMETER
OR
OR
ALPHA
BETA
The shape parameter of the Weibull distribution is related to the failure rate of the hair fibers in the sample. Generally, shape factors greater than 1 indicate that the failure rate is increasing with the time that the fibers are subjected to fatigue testing. This behavior is typical of items that wear out during use, such as tires or clothing.11 Shape parameters less than 1 suggest that items in the sample are breaking early during the stress cycling. Very often, as the cyclic testing continues, the shape parameter may become equal to 1, indicating the onset of constant failure rate. This behavior would be typical in populations where flawed specimens are eliminated early in the testing. An example would be light bulb filaments, which experience a burn-in period where a number will burn out quickly during normal use, after which, the surviving filaments will burn for extended periods. A shape parameter of less than 1 is usually observed in hair samples, indicating that previously flawed hairs break quickly, leaving hair fibers that can withstand stressing for extended times.
Wire W
FIGURE 105.3 Schematic drawing of a hair fiber specimen being subjected to flex fatigue on the Flexabrasion apparatus.
105.4.3 CALCULATING CHARACTERISTIC LIFE VALUES FROM EXPERIMENTAL DATA Obviously, the form of Equation 105.1 does not conveniently lend itself to the calculation of characteristic life
906
Handbook of Non-Invasive Methods and the Skin, Second Edition
from the array of failure data generated during one of the three techniques. Therefore, Equation 105.1 should be transformed into a linear relationship of the form Y = mx + b by taking the natural log of each side twice, yielding Equation 105.2:
TABLE 105.1 Characteristic Life Values of Treated Hair
Treatment
⎡ ⎛ 1 ln ⎢ ln ⎜ ⎜ ⎢ ⎝ 1− F x ⎣
()
⎞⎤ ⎟ ⎥ = beta ln x − beta ln alpha ⎟⎠ ⎥ ⎦ (105.2)
( )
( (
))
It is recommended to estimate the probability of failure, F(x), by using the method of median ranks.11–15 First the data are ordered from first failure to last. Then F(x) is derived using Equation 105.3: F ( x) ≈
failure order number − 0.3 N + 0.4
(105.3)
N is the total number of broken specimens. Once F(x) is calculated for each fiber in the ordered array, ⎡ ⎛ ⎞⎤ 1 ln ⎢ ln ⎜ ⎟ ⎥ can be calculated and plotted vs. the ⎢ ⎜⎝ 1 − F x ⎟⎠ ⎥ ⎣ ⎦ natural log of the failure cycles, ln(x), obtained from the experiment. Figure 105.4 shows a typical plot of the linearly transformed Weibull distribution data obtained from a sample of bleached hair. The points lie close to the bestfit regression line, indicating that the two-parameter form of the Weibull distribution is a suitable model for the experimental data.
()
Unaltered (virgin) Bleached Bleached and leave-on conditioner Bleached and rinse-off conditioner
Characteristic Life, alpha (cycles)
Shape Parameter, beta
34,187 1,213 10,979
0.51 0.31 0.41
5,021
0.33
A poor fit to a linear regression line suggests that an alternative form of the Weibull equation that utilizes more than two parameters is required. This is rarely observed for hair fibers. The slope of the regression line is the Weibull shape parameter, beta, and its Y-axis intercept is –beta (1n (alpha)) from inspection of Equation 105.2. Once these two values are identified, it is easy to calculate alpha, the characteristic life. Examples of characteristic life values and shape parameters of unaltered hair and that of hair after bleaching are given in Table 105.1.2 The characteristic life of bleached hair is substantially less than that of virgin hair. Bleaching weakens the fiber structure,1 and it becomes less resistant to breakage during styling routines. The increase in characteristic life values after applying conditioner to the bleached hair suggests that the treatment helps it resist breaking. The comparison of alpha values also can be used to rate the effects of a leave-on or a rinse-off conditioner.
105.4.4 SURVIVAL PROBABILITY CURVES 1.5 1 0.5
1
ln ln
1 − F(x)
0 0
5
10
15
−0.5 −1 −1.5 −2 −2.5 −3 ln(x)
FIGURE 105.4 A typical plot of the linearly transformed Weibull distribution data obtained from a sample of bleached hair. The solid line is the best-fit regression line of the calculated data points.
A powerful feature of the Weibull analysis of failure data is the ability to predict the probability of hair fibers to break under stress cycling, which simulates everyday grooming and styling stresses. The procedure is to rewrite Equation 105.1 using the characteristic life value and shape parameter derived from the experimental data. Then the equation is used to calculate F(x), the probability of survival as a function of the number of applied stress cycles (ln[x]). The output can be plotted as demonstrated in Figure 105.5 for the data of Table 105.1. From the curves, it is apparent that the bleached hair fibers that were tested have about a 49% chance of surviving 400 cycles of stress under the conditions of the experiment. Unaltered hair has approximately a 90% probability of surviving the same number of cycles. The application of the leave-on conditioner on the bleached hair raises its survival probability to 77%. Therefore, plots of survival curves are a powerful tool for assessing the beneficial effects of treatments on damaged hair.
Evaluating the Strength of Human Hair
907
1 0.9
Survival probability
0.8 0.7
Unaltered hair
0.6
Bleached hair
0.5
Bleached + leave in conditioner
0.4
Bleached + rinse off conditioner
0.3 0.2 0.1 0 0
400
800 1200 Cycles to failure
1600
2000
FIGURE 105.5 Survival probability curves calculated using the alpha and beta values shown in Table 105.1.
105.5 CONCLUDING REMARKS Direct measurement of the mechanical strength of hair fibers is usually insufficient to elucidate the benefit of conditioning treatments on damaged hair. Hair strength is not just the amount of direct force to break the fiber; it is also reflected by the ability of the hair fiber to resist breakage under relatively low stress levels encountered during combing or brushing. Therefore, cyclic fatigue testing is an invaluable tool in assessing hair damage and evaluating the beneficial effect of formulations.
REFERENCES 1. Tate, M.L., Kamath, Y.K., Ruetsch, S.B., and Weigmann, H.-D., Quantification and prevention of hair damage, J. Soc. Cosmet. Chem., 44, 155, 1993. 2. Hornby, S.B., Cyclic testing: demonstrating conditioner benefits on damaged hair, Cosmet. Toilet., 116, 35, 2001. 3. Robbins, C.R., Chemical and Physical Behavior of Human Hair, 4th ed., Springer-Verlag, New York, 2002, p. 395. 4. Hornby, S.B., Wiunsey, N.J.P., and Bucknell, S., New technique to capture viscoelastic changes in hair induced by mechanical stress, IFSCC Magazine, 5, 93, 2002.
5. Kamath, Y.K., Hornby, S.B., and Weigmann, H.-D., Mechanical and fractographic behavior of Negroid hair, J. Soc. Cosmet. Chem., 35, 21, 1984 6. Kamath, Y.K., Hornby, S.B., and Weigmann, H.-D., Effect of chemical and humectant treatments on the mechanical and fractographic behavior of Negroid hair, J. Soc. Cosmet. Chem., 36, 39, 1985. 7. Swift, J.A., Chahal, S.P., and Challoner, N., Flexabrasion: a method for evaluating hair strength, Cosmet. Toilet., 116, 53, 2001. 8. Leroy, F. et al., Flexabrasion: A New Test for Predicting Human Hair Resistance, poster at the 1st Tricontinental Meeting of Hair Research Societies, Brussels, Belgium, October 8–10, 1995. 9. Personal communication. 10. Weibull, W., A statistical distribution function of wide applicability, J. Appl. Mech., 9, 292, 1951. 11. Kececioglu, D., Reliability and Life Testing Handbook, Vol. 1, Prentice Hall, Englewood Cliffs, NJ, 1993. 12. Collins, J., Failure of Materials in Mechanical Design, 2nd ed., John Wiley & Sons, New York, 1993. 13. Griffith, A.A., The phenomena of rupture and flow in solids, Phil. Trans. R. Soc. London, 163–197, 1921. 14. Epstein, B., Statistical aspects of fracture, J. Appl. Physics, 19, 140, 1948. 15. Evans, G. and Jones, R.L., Evaluation of a fundamental approach for the statistical analysis of fracture, J. Am. Ceram. Soc., 61, 156, 1978.
Nail Structure and Growth
for Nail Assessment: 106 Methods An Overview David de Berker Bristol Dermatology Centre, Bristol Royal Infirmary, Bristol, United Kingdom
CONTENTS 106.1 Methods for Nail Assessment: An Overview ...................................................................................................911 106.2 Photography.......................................................................................................................................................911 106.3 Photodermatoscopy and Dermatoscopy............................................................................................................912 106.4 Capillaroscopy ...................................................................................................................................................913 106.5 Profilometry .......................................................................................................................................................913 106.6 Surface Replicas ................................................................................................................................................913 106.7 Magnetic Resonance Imaging ...........................................................................................................................913 106.8 Microscopy: Light, Scanning, and Transmission Microscopy .........................................................................914 106.9 Other Techniques...............................................................................................................................................914 106.10 Scoring Systems ................................................................................................................................................914 106.11 Strength..............................................................................................................................................................915 106.12 Measuring Nail Strength ...................................................................................................................................915 106.13 Permeability.......................................................................................................................................................916 References .......................................................................................................................................................................916
106.1 METHODS FOR NAIL ASSESSMENT: AN OVERVIEW Much science applied to nails is a microcosm of dermatology in general. Nails are a specialized appendage that call for adaptation and experimentation of standard techniques. Only in some areas do they command techniques special to themselves, such as in the measure of longitudinal growth or their strength. In this chapter we cover how the nail can be assessed with the use of techniques described in detail in other parts of the book. This will lead to many brief sections, best read in conjunction with the core chapters. Many of these techniques are very simple, but small tips can make a big difference to the success of your methods. We also cover the borderline topic of constituent analysis. Nails and hair afford themselves to providing “tissue” without invading the body. In legal terms, it is not clear whether obtaining a nail clipping or a cut hair sample is an invasive process. However, in scientific terms and for medicine in general, most people would be happy to provide such samples.
106.2 PHOTOGRAPHY There are three points that have great importance when undertaking photography of the nail. First is focus. Although this might sound banal, it is important to realize that the nail is a limited field of focus usually within a larger picture. It is also curved in two axes. The depth of focus of digital cameras is generally poor and there is little tolerance. The autofocus should be set on central spot focus or manual to minimize the chance of having a beautifully focused background and a blurred digit. The second point is illumination (Figure 106.1). Most digital cameras have their own flash. This introduces problems with macrophotography, as you will want to be as close as possible. The angle between the origin of the flash and the point of focus means that the light is slanted and introduces a shadow down one side of the digit. This is avoided if you use ambient light, but that leads to slow exposures and blur. The problem can be overcome to some extent by taking high-resolution images at some distance and then blowing up the area of interest as part of the digital editing (Figure 106.2). The ideal situation is to work with the small number of cameras with high 911
912
Handbook of Non-Invasive Methods and the Skin, Second Edition
A
B
C
D
FIGURE 106.1 Toe photographed with digital camera (A) in bright ambient light in the longitudinal axis of the toe. (B) In the same light as (A), but with flash. (C) In poor ambient light and in the same light with flash. The skin tones are superior in (A) and (D). The proximity of the camera in (B) and (D) means that there is low exposure adjusted for flash, but the light does not reach the subject. The zoom can accommodate this to some degree, but the figure illustrates a point.
F E
A
G
FIGURE 106.2 Where there is difficulty obtaining a good close-up image due to depth of focus, it is possible to take a more distant image (E), crop it (F), and blow it up 4× (G), which compares favorably with the original detailed shot (A) from Figure 106.1.
tolerance of low light or a separate light source. Separate light sources can be proper studio lights or a handheld flash. The first is preferable, as it will allow you to gauge the distortion introduced by reflection from the nail surface. However, while it may be okay for standardized laboratory photographs, it is not so good for recording clinical scenarios. It is also not so flexible for controlling the angle of any shadow you may wish to introduce.
FIGURE 106.3 A patient with evolving pitting has mascara rubbed into the nail plate prior to photography to assist in demonstration of the pattern of pits.
Remember that if you do not want shadow, you can use a ring flash or orient the flash longitudinally up the digit. Thirdly is the artifact of reflection from the nail surface. This makes it difficult to visualize detail of the surface or color characteristics of the subungual tissues. This problem is resolved with the use of ambient or studio lighting. All flash techniques are vulnerable to production of surface glare from the nail plate. Subtle surface irregularities are often difficult to discern with photography, even with the above techniques. These features can be enhanced by using mascara rubbed into the nail surface and then partially removed with a soft cloth. This leaves mascara within the surface impressions and outlines them for photography (Figure 106.3). Video is less commonly used as a nail research tool. Goyal and Griffiths1 employed images grabbed from video to make serial measurements for nail dimensions for morphometric analysis.
106.3 PHOTODERMATOSCOPY AND DERMATOSCOPY The dermatoscope is a useful tool for close inspection and can be used with or without intervening liquid medium. Nikon makes attachments that provide compatibility with standard dermatoscopes and allow digital photography. When not using a medium, the dermatoscope can be used to assess fine surface detail with the benefits of the 10× magnification, most effectively delivered using the lightemitting diode (LED) dermatoscope head. The tool allows analysis of nail surface characteristics in vivo and fine changes in periungual epidermal morphology, as can be seen with defining the margins for excision of Bowen’s disease of the nail unit (Figure 106.4). The most convenient medium for nail assessment is aqueous clear gel as used for skin ultrasound or
Methods for Nail Assessment: An Overview
FIGURE 106.4 The patient is undergoing Mohs’ micrographic surgery. Intraoperative examination with the dermatoscope assists in defining the clinical margin that was previously obscured by the nail plate. The subtle border of change of skin markings can be seen.
lubrication. This comes in a tube and can be easily wiped off. The alternative of mineral oil is not so good, as it creates only a thin film and it is difficult for the flat lens of the dermatoscope to form a good seal. With this technique, many nail bed and matrix features can be more closely assessed. This is of particular value in pigmentary changes. The dermatoscope allows a high level of clinical confidence in the differentiation of blood, melanin of matrix origin, and pigment of fungal origin. In the nail fold, blood vessels are easily qualitatively assessed with this technique.2
106.4 CAPILLAROSCOPY Modified forms of the dermatoscope can be used for quantitative capillaroscopy in normal and disease (Figure 106.5). Some of these have a video element,3 which in turn requires computer analysis. Diseases include connective tissue disorders4 and psoriasis.5 Where basic morphology is being measured, descriptive terms are used according to taxonomies created within each study,6 as well as some more accepted terms.7
106.5 PROFILOMETRY Profilometry is the technique of measuring the profile of a surface. It can be used on the nail surface to assess pitting, grooves, and trachyonychia using the measures of roughness, mean depth of roughness, and number of peaks or crests.8 Changes in these characteristics can be equated with disease activity and used as measures of response to therapy, such as in psoriatic trachyonychia during lowdose cyclosporin9 and the rate of nail growth during itraconazole treatment.10
913
FIGURE 106.5 Nail-fold capillaries viewed with dermatoscope, ultrasound gel, and using a Nikon attachment.
106.6 SURFACE REPLICAS It is possible to make surface replicas of the nail with a range of materials. The two that are easiest to obtain are cyanoacrylate11 and silicone molding material. The first provides a replica that is translucent and allows examination by normal light microscopy as well as scanning electron microscopy. However, it is a difficult balance between obtaining a dry replica on the surface of a glass slide apposed to the nail and finding that the slide is irrevocably stuck to the nail. Given that the technique requires a glass slide to allow subsequent light examination, it can be a problem if significant force is needed to remove the slide from the nail surface. The rate of drying is a function of the temperature, force applied, and characteristics of the glue. It takes some practice and experimentation to master the technique — something that I have not managed. Silicone is more flexible and forgiving, but also introduces small artifacts not seen with cyanoacrylate. It is a good alternative to photography where a three-dimensional record needs to be kept of a morphological characteristic of the nail or periungual tissues (Figure 106.6). Where I have used this, I have glued the mold to a glass slide and used it to attach identification. The replica method has been evaluated alongside scanning electron microscopy, indicating a high degree of agreement between techniques.12
106.7 MAGNETIC RESONANCE IMAGING Magnetic resonance imaging (MRI) produces very clear images of the nail, phalanx, and periungual tissues. It has been used for anatomical delineation of structures in normal digits.13 Where there is unexplained pain or dystrophy of an isolated digit, MRI may reveal a soft tissue tumor where traditional radiography is normal.14 It is most useful when the tumor contrasts with surrounding tissues with respect to density, fluid, or fat content. The most marked example of this is with myxoid pseudocysts.15 It is of
914
Handbook of Non-Invasive Methods and the Skin, Second Edition
A
B
C
FIGURE 106.6 (A) Silicone rubber can be used as a molding material to make an imprint of the nail surface. (B) Detail of the imprint reveals the pattern of nail pitting. (C) The mold is glued to a slide, which in turn is labeled on the reverse.
particular value when there is pain after an initial operation for an entity such as a glomus tumor, characterized by pain. In this setting it is not certain whether the pain is due to recurrence and warrants further surgery or whether it is related to the surgery itself, which makes further surgery a bad idea.16
106.8 MICROSCOPY: LIGHT, SCANNING, AND TRANSMISSION MICROSCOPY Nail plate is amenable to examination by all forms of light and electron microscopy. Light microscopy is sometimes of higher quality if thinner sections (6 μ) are cut after embedding in epoxy, but the improvement over standard techniques is not great.17 Standard techniques entail thicker sections, cut from wax-embedded nail. However, one thing that does make a difference is how long the nail is left in formalin. There is a case for avoiding this stage altogether, as formalin hardens an already hard structure and makes it even more difficult to cut high-quality sections. It can be helpful to practice whatever staining procedure you anticipate on both fixed and unfixed specimens. Cutting sections typically causes problems with the microtome, leaves a jagged profile on sections and makes the blade blunt. This makes nail histology unpopular with laboratory technicians. A further problem is getting the sections to stick to the glass slides. Where there is no soft tissue as part of the specimen, there is little if any natural adhesiveness to keep the section on the slide. Some pretreated slides will minimize this problem, but it is still advisable to produce multiple sections in order that a percentage of failures can be tolerated.
Once ready for staining, sections can be stained to demonstrate different things. A periodic acid Schiff is useful for delineating the outline of the nail plate cells throughout the nail and also highlights fungus. Histological evaluation of nail clippings is one of the main supplements to routine mycology that can clarify uncertain diagnoses.18 Scanning electron microscopy can be used on nail fragments — on the dorsal or ventral aspect or on sections. The main use of this technique has been to explore the natural tendency of the lamellar fabric of the nail to separate at the free edge in onychoschizia.19 At the magnification used, there is not really any advantage over using light microscopy, which is much less expensive. However, scanning electron microscopy is superior at examining the nail plate surface and defining the features of individual corneocytes. Transmission electron microscopy has been used in the past for basic ultrastructural analysis of nail plate and its soft tissue attachments.20 It can be used on nail fragments and demonstrates the details of intercellular and intracellular structures within the nail plate.
106.9 OTHER TECHNIQUES Laser Doppler can be used to assess the blood flow in the nail unit and has been applied alongside capillaroscopy in diabetics21 and in Raynaud’s disease22 and normals.23 This combined approach provides an opportunity to try to correlate anatomical and functional aspects of blood flow. Dynamic aspects of flow are accentuated by cooling and rewarming to broaden the range for evaluation. Optical coherence tomography produces a series of cross-sectional images down to a depth of 1 mm, separated by 15 mm. It has some potential for examining periungual tissues, but has been little explored as a technique.24 Confocal microscopy is in a similar category, where light penetrates the unsectioned tissue to give three-dimensional information and has been used to examine the normal nail plate25 and to evaluate onychomycosis.26 Older techniques for the assessment of shape, and clubbing in particular, include brass templates,27 shadow graphs,28 plaster casts and planimetry,29 and plethysmography.30
106.10 SCORING SYSTEMS A range of scoring systems have been developed to measure disease or suffering in nail disease. Objective measures of disease mainly pertain to psoriasis and onychomycosis. Both of these are the subject of therapeutic trials, and it is important to have some objective measure of the disease activity. de Jong et al.31 has used a nail area severity (NAS) score, consisting of the separate parameters nail
Methods for Nail Assessment: An Overview
A
B
C
D
E
FIGURE 106.7 Five nails demonstrating the manner of dividing the nail into eight zones, representing 12.5% surface area per zone. This can lead to problems with interpretation — for instance, where the partial nail loss of (A) leads to uncertainty concerning where to place the transverse division. In (E), where there is superficial involvement, it is apparent why this scoring system might not work — although it is a score of disease as detected at the surface. In (D), part of every upper section is clear and part abnormal. If the nail was clipped vigorously, the fractions would alter.
TABLE 106.1 Scoring System for Evaluation of a Pigmented Streak in a Nail and Diagnosis of Subungual Melanoma A B C D E F
Age of patient (fifth to seventh decade) Breadth of streak > 3 mm Color change from brown to black High-risk digit, such as thumb or big toe Extension of the pigment onto periungual tissues (Hutchinson’s sign) Family history of melanoma or dysplastic nevi
From Levit, E.K. et al., J. Am. Acad. Dermatol., 42, 269–274, 2000.
pitting area, number of nail pits, subungual keratosis, onycholysis, oil spots, and a score for overall improvement. Baran32 has used a similar system, rating features on a scale of 1 to 3. A further scoring system for the same disease evaluates some of the subjective aspects of the disease, including pain and disability.33 In onychomycosis, the main scoring variable is percentage of involved nail surface. A common standard for judging this is to bisect the nail transversely, then longitudinally, giving quadrants. Each lateral segment is then bisected again with a longitudinal line, resulting in
915
division of the nail into eight segments, representing 12.5% of surface area each (Figure 106.7). There can be problems of interpretation with this method, as illustrated in the figure. When the nail is naturally small or partially disintegrated, this will mean that there is less diseased tissue than in someone who may have bigger nails but a smaller percentage score. Attempts have also been made to develop an onychomycosis disease-specific questionnaire that assesses functional impact, social stigma, and psychological distress associated with the disease.34 As with most quality of life indicators, the group is preselected by having sought treatment. This will tend to increase measures of subjective suffering, especially if working within a health care system where it is in the interests of the clinician that the patient pursues treatment. Where there are pigmented streaks of the nail, Levit et al.35 has proposed an ABC scoring system for clinical detection of streaks likely to be subungual melanoma.
106.11 STRENGTH Nail strength is partly due to its composition and partly its anatomical relationships with the underlying bone and surrounding soft tissues. The main protein constituent is keratin, the most abundant of intracellular proteins in epithelial cells. The nail plate is a modified epithelium and has a different complement of keratins from those in skin. In addition to those often termed soft or epithelial keratins, there are hard or hair/nail keratins. These contain a higher fraction of sulfur-rich amino acids than their soft counterparts. These enhance disulfide bridging between the threedimensional convolutions of the keratin molecule. This increases the rigidity and strength of the protein and creates a molecular environment that excludes water. This in turn makes the nail strong and resistant to chemical breakdown. One measure of this strength is the in vitro preparation needed for nail specimens when undergoing amino acid analysis.
106.12 MEASURING NAIL STRENGTH Several techniques have been developed to study the physical properties of nails.36–38 Techniques have been described to test the tearing, flexural, and tensile strength of nails.37 Finlay et al.38 devised a nail flexometer able to repeatedly flex longitudinal nail sections through 90˚, recording the number it took to fracture the nail. In this way, the strength could be quantified. Soaking the nails in water increased their flexibility.38 Wessel et al.’s work39 suggests that this may be due to the loosening of the alpha helix conformation seen with Raman spectroscopy in vivo after soaking in water for 10 min, where water occupies the interstices.
916
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 106.2 Different Methods of Nail Constituent Analysis Method
Element
Reference
I. Structural and Mineral Constituents Raman spectroscopy Immunohistochemistry Near infrared spectrometer Polymerase chain reaction Electron microscopy X-ray diffraction Colorimetry
Water, proteins, and lipid Keratin Water Deoxyribonucleic acid Cystine Mg, Cl, Na, Ca, S, Cu Fe
46 47 48 49, 50 51 52, 53 54
II. Exogenous Materials Atomic absorption spectometry Mass fragmentography Gas chromatography–mass spectroscopy High-performance liquid chromatography Flow injection hydride generation atomic Inductively coupled plasma–mass spectrometry Absorption spectrometry
Cd, Pb, Zn, Ca, Cr, Fe, Cu, Mn, Ni, Co, Na, K Methamphetamine Amphetamine, cocaine, cannabinoids Nicotine, terbinafine, testosterone, pregnenolone Arsenic Arsenic
53, 55, 56 57 58–61 62–64 65 66
DNA Furosine (glycosylated keratin) Lipid: triglyceride Ni Steroid sulfatase Zinc, selenium
67 68 51 69 70 71–73
III. Biological Markers Polymerase chain reaction High-performance liquid chromatography Microscopy Adsorption differential pulse voltametry Enzymic assay Neutron activation analysis
Zaun40 has used a method of assessment of brittleness that relies on the swelling properties of nail, employing the technique before and after therapy for brittle nails. Splitting can be partially overcome by applications of emollient after soaking the nails in water.
106.13 PERMEABILITY Nail permeability is relevant to topical drugs on the dorsal surface and systemic drugs from the ventral surface. Transonychial water loss can be measured in vivo41,42 and in vitro,43 where the nail appears to be 1000 times more permeable to water than is skin. This means that watersoluble drugs are likely to penetrate the nail as long as the molecular size is not large.44 However, most drug delivery processes across nail need more than fragments for their assessment and are not covered here. Nevertheless, it is possible to assess movement of systemic drugs into nail by obtaining material taken either from bore holes in the nail plate or from the free edge. This is particularly relevant for systemic antifungals taken for onychomycosis.45
REFERENCES 1. Goyal S, Griffiths WAD. An improved method of studying fingernail morphometry: application to the early detection of fingernail clubbing. J Am Acad Dermatol 39:640–642, 1998. 2. Bergman R, Sharony L, Schapira D, Nahir MA, BalbirGurman A. The handheld dermatoscope as a nail-fold capillaroscopic instrument. Arch Dermatol 139:1027–1030, 2003. 3. Cutolo M, Pizzorni C, Tuccio M, Burroni A, Craviotto C, Basso M, Seriolo B, Sulli A. Nailfold videocapillaroscopic patterns and serum autoantibodies in systemic sclerosis. Rheumatology (Oxford) 43:719–726, 2004. 4. Nagy Z, Czirjak L. Nailfold digital capillaroscopy in 447 patients with connective tissue disease and Raynaud’s disease. J Eur Acad Dermatol Venereol 18:62–68, 2004. 5. Bhushan M, Moore T, Herrick AL, Griffiths CE. Nailfold video capillaroscopy in psoriasis. Br J Dermatol 142:1171–1176, 2000. 6. Jones BF, Oral M, Morris CW, Ring EF. A proposed taxonomy for nailfold capillaries based on their morphology. IEEE Trans Med Imaging 20:333–341, 2001.
Methods for Nail Assessment: An Overview
7. Hu Q, Mahler F. New system for image analysis in nailfold capillaroscopy. Microcirculation 6:227–235, 1999. 8. Nikkels-Tassoudji N, Piérard-Franchimont C, de Doncker P, Piérard GE. Optical profilometry of nail dystrophies. Dermatology 190:301–304, 1995. 9. Piérard GE, Piérard-Franchimont C. Dynamics of psoriatic trachyonychia during low dose cyclosporin A treatment: a pilot study on onychochronobiology using optical profilometry. Dermatology 192:116–119, 1996. 10. de Doncker P, Piérard GE. Acquired nail beading in patients receiving itraconazoleaan indicator of faster nail growth? A study using optical profilometry. Clin Exp Dermatol 19:404 –406, 1994. 11. Marks R. Histochemical applications of skin surface biopsy. Br J Dermatol 86:20–26, 1972. 12. Hashimoto K. New methods for surface ultrastructure: comparative studies of scanning electron microscopy, transmission electron microscopy and replica method. Int J Dermatol 13:357–381, 1974. 13. Drapé JL, Wolfram-Gabel R, Idy-Peretti I, et al. The lunula: a magnetic resonance imaging approach to the subnail matrix area. J Invest Dermatol 106:1081–1085, 1996. 14. Drape JL. Imaging of the tumors of the perionychium. Hand Clin 18:655–670, 2002. 15. Drapé JL, Idy-Peretti I, Goettmann S, et al. MR imaging of digital mucoid cysts. Radiology 200:531–536, 1996. 16. Theumann NH, Goettmann S, Le Viet D, Resnick D, Chung CB, Bittoun J, Chevrot A, Drape JL. Recurrent glomus tumors of fingertips: MR imaging evaluation. Radiology 223:143–151, 2002. 17. de Berker D, Mawhinney B, Sviland L. Quantification of regional matrix nail production. Br J Dermatol 134:1083–1086, 1996. 18. Reisberger EM, Abels C, Landthaler M, Szeimies RM. Histopathological diagnosis of onychomycosis by periodic acid-Schiff-stained nail clippings. Br J Dermatol 148:749–754, 2003. 19. Wallis MS, Bowen WR, Guin JD. Pathogenesis of onychoschizia (lamellar dystrophy). J Am Acad Dermatol 24:44 –48, 1991. 20. Parent D, Achten G, Stouffs-Vamhoof F. Ultrastructure of the normal human nail. Am J Dermatopathol 7:529–535, 1985. 21. Meyer MF, Pfohl M, Schatz H. Assessment of diabetic alterations of microcirculation by means of capillaroscopy and laser-Doppler anemometry. Med Klin (Munich) 96:71–77, 2001. 22. Creutzig A, Hiller S, Appiah R, Thum J, Caspary L. Nailfold capillaroscopy and laser Doppler fluxmetry for evaluation of Raynaud’s phenomenon: how valid is the local cooling test? Vasa 26:205–209, 1997. 23. Lutolf O, Chen D, Zehnder T, Mahler F. Influence of local finger cooling on laser Doppler flux and nailfold capillary blood flow velocity in normal subjects and in patients with Raynaud’s phenomenon. Microvasc Res 46:374–382, 1993.
917
24. Welzel J, Lankenau E, Birngruber R, Engelhardt R. Optical coherence tomography of the human skin. J Am Acad Dermatol 37:958–963, 1997. 25. Kaufman SC, Beuerman RW, Greer DL. Confocal microscopy: a new tool for the study of the nail unit. J Am Acad Dermatol 32:668–670, 1995. 26. Hongcharu W, Dwyer P, Gonzalez S, Anderson RR. Confirmation of onychomycosis by in vivo confocal microscopy. J Am Acad Dermatol 42:214–216, 2000. 27. Stavem P. Instrument for estimation of clubbing. Lancet 2:7–8, 1959. 28. Bentley D, Moore A, Schwachman H. Finger clubbing: a quantitative survey by analysis of the shadowgraph. Lancet 2:164–167, 1976. 29. Regan GM, Tagg B, Thomson ML. Subjective measurement and objective measurement of finger clubbing. Lancet 1:530–532, 1967. 30. Cudowicz P, Wraith DG. An evaluation of the clinical significance of clubbing in common lung disorders. Br J Tuberculous Dis Chest 51:14–31, 1957. 31. de Jong EM, Menke HE, van Praag MC, van De Kerkhof PC. Dystrophic psoriatic fingernails treated with 1% 5fluorouracil in a nail penetration-enhancing vehicle: a double-blind study. Dermatology 199:313–318, 1999. 32. Baran RL. A nail psoriasis severity index. Br J Dermatol 150:568–569, 2004. 33. de Jong EM, Seegers BA, Gulinck MK, Boezeman JB, van de Kerkhof PC. Psoriasis of the nails associated with disability in a large number of patients: results of a recent interview with 1,728 patients. Dermatology 193:300–303, 1996. 34. Turner RR, Testa MA. Measuring the impact of onychomycosis on patient quality of life. Qual Life Res 9:39–53, 2000. 35. Levit EK, Kagen MH, Scher RK, Grossman M, Altman E. The ABC rule for clinical detection of subungual melanoma. J Am Acad Dermatol 42:269–274, 2000. 36. Baden HP. The physical properties of nail. J Invest Dermatol 55:115, 1970. 37. Maloney MJ, Paquette EG. The physical properties of fingernails. I. Apparatus for physical measurements. J Soc Comp Chem 28:415, 1977. 38. Finlay AF, Frost P, Keith AC, Snipes W. An assessment of factors influencing flexibility of human fingernails. Br J Dermatol 103:357–365, 1980. 39. Wessel S, Gniadecka M, Jemec GB, Wulf HC. Hydration of human nails investigated by NIR-FT-Raman spectroscopy. Biochim Biophys Acta 17:210–216, 1999. 40. Zaun H. Brittle nails. Objective assessment and therapy follow-up. Hautarzt 48:455–461, 1997. 41. Jemec GBE, Agner T, Serup J. Transonychial water loss: relation to sex, age and nail plate thickness. Br J Dermatol 121:443–446, 1989. 42. Spruit D. Measurement of water vapor loss through human nail in vivo. J Invest Dermatol 56:359–361, 1971. 43. Walters KA, Flynn GL, Marvel JR. Physicochemical characterization of the human nail. 1. Pressure sealed apparatus for measuring nail plate permeability. J Invest Dermatol 76:76–79, 1981.
918
Handbook of Non-Invasive Methods and the Skin, Second Edition
44. Mertin D, Lippold BC. In vitro permeability of the human nail and of a keratin membrane from bovine hooves: influence of the partition coefficient octanol/water and the water solubility of drugs on their permeability and maximum flux. J Pharm Pharmacol 49:30–34, 1997. 45. Munro CS, Shuster S. The route of rapid access of drugs to the distal nail plate. Acta Dermato-Venereol 72:387–388, 1992. 46. Gniadecka M, Nielsen O, Christensen D, et al. Structure of water, proteins and lipids in intact human skin, hair and nail. J Invest Dermatol 110:393–-398, 1998. 47. Heid HW, Moll I, Franke WW. Patterns of expression of trichocytic and epithelial cytokeratins in mammalian tissues. Differentiation 37:215–230, 1988. 48. Egawa M, Fukuhara T, Takahashi M, Ozaki Y. Determining water content in human nails with a portable near-infrared spectrometer. Appl Spectrosc 57:473–478, 2003. 49. Kaneshige T, Takagi K, Nakamura S, et al. Genetic analysis using fingernail DNA. Nucleic Acid Res 20:5489–5490, 1992. 50. Tahir M, Watson N. Typing of DNA HLA-DQa alleles extracted from human nail material using polymerase chain reaction. J Forens Sci 40:634–636, 1995. 51. Salamon T, Lazovic-Tepavac O, Nikulin A, et al. Sudan IV positive material of the nail plate related to plasma triglycerides. Dermatologica 176:52–54, 1988. 52. Sirota L, Straussberg R, Fishman P, Dulitzky F, Djaldetti M. X-ray microanalysis of the fingernails in term and preterm infants. Pediatr Dermatol 5:184–186, 1988. 53. Forslind B. Biophysical studies of the normal nail. Acta Dermato-Venereol 50:161–168, 1970. 54. Jacobs A, Jenkins DJ. The iron content of finger nails. Br J Dermatol 72:145–148, 1960. 55. Wilhelm M, Hafner D, Lombeck I, Ohnesorge FK. Monitoring of cadmium, copper, lead and zinc status in young children using toenails: comparison with scalp hair. Sci Total Environ 103:199–207, 1991. 56. Nowak B. Occurrence of heavy metals, sodium, calcium and potassium in human hair, teeth and nails. Biol Trace Elem Res 52:11–22, 1996. 57. Suzuki S, Inoue T, Hori H, Inayama S. Analysis of methamphetamine in hair, nail, sweat and saliva by mass fragmentography. J Anal Toxicol 13:176–178, 1989. 58. Suzuki O, Hattori H, Asano M. Nails as useful materials for detection of methamphetamine or amphetamine abuse. Forens Sci Int 24:9–16, 1984. 59. Cirimele V, Kintz P, Mangin P. Detection of amphetamines in fingernails: an alternative to hair analysis. Arch Toxicol 70:68–69, 1995.
60. Miller M, Martz R, Donnelly B. Drugs in keratin samples from hair, fingernails and toenails. In Second International Meeting on Clinical and Forensic Aspect of Hair Analysis, Genoa, Italy, June 6–8, 1994, p. 39 (abstract). 61. Raharjo TJ, Verpoorte R. Methods for the analysis of cannabinoids in biological materials: a review. Phytochem Anal. 15:79–94, 2004. 62. Al-Delaimy WK, Mahoney GN, Speizer FE, Willett WC. Toenail nicotine levels as a biomarker of tobacco smoke exposure. Cancer Epidemiol Biomarkers Prev 11:1400–1404, 2002. 66. Dykes PJ, Thomas R, Finlay AY. Determination of terbinafine in nail samples during treatment for onychomycoses. Br J Dermatol 123:481–486, 1990. 67. Choi MH, Yoo YS, Chung BC. Measurement of testosterone and pregnenolone in nails using gaschromatography-mass spectrometry. J Chromatogr B Biomed Sci Appl 25:495–501, 2001. 68. Das D, Chatterjee A, Badal K, et al. Arsenic in ground water in six districts of West Bengal, India: the biggest arsenic calamity in the world. Analyst 120:917–924, 1995. 69. Chen KL, Amarasiriwardena CJ, Christiani DC. Determination of total arsenic concentrations in nails by inductively coupled plasma mass spectrometry. Biol Trace Elem Res 67:109–125, 1999. 70. Oz C, Zamir A. An evaluation of the relevance of routine DNA typing of fingernail clippings for forensic casework. J Forens Sci 45:158–160, 2000. 71. Sueki H, Nozaki S, Fujisawa R, et al. Glycosylated proteins of skin, nail and hair: application as an index for long-term control of diabetes mellitus. J Dermatol 16:103–110, 1989. 72. Gamelgaard B, Anderson JR. Determination of nickel in human nails by adsorption differential-pulse voltametry. Analyst 110:1197–1199, 1985. 73. Matsumoto T, Sakura N, Ueda K. Steroid sulphatase activity in nails: screening for X-linked ichthyosis. Pediatr Dermatol 7:266–269, 1990. 74. Rogers M, Thomas DB, Davis S, et al. A case control study of oral cancer and pre-diagnostic concentrations of selenium and zinc in nail tissue. Int J Cancer 48:182–188, 1991. 75. Van Noord PAH, Collette HJA, Maas MJ, de Waard F. Selenium levels in nails of premenopausal breast cancer patients assessed prediagnostically in a cohort-nested case-referent study among women screened in the DOM project. Int J Epidemiol 16 (Suppl.):318–322, 1987. 76. Yoshizawa K, Willett WC, Morris SJ, et al. Study of the prediagnostic selenium levels in toenails and the risk of advanced prostate cancer. J Natl Cancer Inst 90:1219–1224, 1998.
of Longitudinal 107 Measurement Nail Growth Jeffrey S. Roth and Richard K. Scher Department of Dermatology, College of Physicians and Surgeons, Columbia University, New York, New York
CONTENTS 107.1 Introduction ..........................................................................................................................................................919 107.2 Historical Overview .............................................................................................................................................919 107.3 Measurement of Nail Growth ..............................................................................................................................920 107.3.1 Fixed Landmarks ...................................................................................................................................920 107.3.2 Distal Landmarks...................................................................................................................................921 107.3.3 Technique of Measurement ...................................................................................................................921 107.4 Summary...............................................................................................................................................................921 References .......................................................................................................................................................................921
107.1 INTRODUCTION Initial thoughts on the longitudinal measurement of nail growth turn to the simple and intuitive application of a ruler to the nail and recording of differences in length over time. While this is sound in principle, precision in nail growth measurement demands a more rigorous technique and has been the subject of numerous articles over many decades. Achieving consistency, reproducibility, and accuracy of results requires that there be consensus about the definition of stable landmarks, avoiding sources of variability such as nail plate wear. In addition, nail growth has been reported to vary with season, time of day, digit, sex, and state of health,1 making it important to establish a biologically meaningful interval over which nail growth should be measured. Though a technique may claim to allow measurement of nail growth over a 15-min period,2 this may not be meaningful biologically. Who might be interested in measuring nail growth? Research applications would include the response of the onychomycotic nail to novel antifungals under investigation or to quantify the response of psoriatic nails to treatment. Such applications would require relative precision to allow meaningful statistical evaluation of study data. Clinical applications might be the measurement of Beau’s lines to time recent trauma to the nail matrix from disease or chemotherapy, the quantification of nail growth in a patient who complains of nails that grow poorly, or the establishing of efficacy of specific therapy for nail disease
in the clinic setting. Clinical and research applications may impact differently on the choice of technique. Research techniques should be precise, may involve specialized equipment, need not be quick, and need not (though should) be inexpensive. Clinical methods should be rapid, inexpensive, need no specialized equipment, and can be relatively less precise. Both research and clinical applications should do no harm to the patient (an intuitive “given” that is not always adhered to). The following will take into account the history of the treatment of this issue and a review of the major techniques that have been described.
107.2 HISTORICAL OVERVIEW The development of methods for measuring nail growth spans many decades. Interest in the growth of nails initially stemmed from the notion that the health of the nail and the vigor of its growth closely parallel the general health. While this notion remains generally sound, it is now appreciated that although the nail is not a direct reflection of the general health, it may offer clues to underlying systemic illness. Early reports of methods of measurement and of patterns of nail growth were flawed by poor scientific rigor and anecdotal experience. Thus, Berthold3 measured only the growth of his own fingernails (using the lunula as a fixed landmark). Similarly, Sharpley-Schafer4 measured only his own hand. Bloch5 criticizes the 919
920
Handbook of Non-Invasive Methods and the Skin, Second Edition
previous pioneering work of Dufour,6 published in 1872, which measures the distance between the cuticle and a permanent mark made in the nail plate using silver nitrate, stating that the cuticle is a constant landmark only in the well groomed. (He offers in place of Dufour’s methodology two techniques, one in which a permanent mark in the nail plate is measured against a fixed india ink tattoo on the dorsal finger and another in which the fixed point on the nail plate is measured against the knuckle of the finger bent at a fixed angle.) LeGros Clark and Buxton7 criticize the measurement of nail growth using the proximal nail fold as a landmark. Their methodology in turn is derided as too vague in subsequent studies. Internal inconsistencies abound as well. For example, they use as their fixed landmark a point on the nail “about 2 mm from the margin of the lunula,” yet they go on to give measurement of change in “micromillimeters” (microns). William Bennett Bean,8,9 perhaps the most indefatiguable nail watcher of all time, measured his own fingernails over a 35-year period both longitudinally and by weighing his nail clippings. His perseverance allowed him to observe the deceleration of nail growth with advancing years, from a rate of 0.123 mm/day at the age of 32 to 0.095 mm/day at age 67. Intuitively, measuring nail growth requires no sophisticated instrumentation; nevertheless, it has been subject to sometimes faddish application of technology as it became available. The modern age brought with it a flurry of new techniques: photography,10,11 magnifying devices such as the “biomicrometer” of Basler18 and the tool of LeGros Clark and Buxton,7 the “split image rangefinder adapted to a trinocular microscope” described by Orentreich et al.,2 and time-lapse photography.12 Authorities have opined on the suitability of anatomic structures for use as landmarks from which to measure the growing nail. Some favor the proximal nail fold.1 Others prefer the lunula, though it is conceded that on some fingers the lunular margin is in fact blurred under magnification, and therefore unsuitable for submillimeter measurements.7 Furthermore, many individuals lack a visible lunula. Some favor the use of bony landmarks by xray or physical examination, relying on the intimate relationship between the distal phalanx and the nail plate.10,12,14 The results of animal studies may be difficult to extrapolate to humans, owing to fundamental differences in the shape of the nail in other species, such as the rat.17 In summary, such a simple issue as how to measure the growing nail is beset with good-natured controversy. We intend to endorse no single technique, but to advise a common-sense approach: simplicity, reliability, ease, inexpensiveness, and, above all, avoidance of harm to the patient.
107.3 MEASUREMENT OF NAIL GROWTH We will approach the measurement of nail growth by presenting techniques and points of view in each of several aspects of the problem.
107.3.1 FIXED LANDMARKS The issue of finding a stable landmark as a point of reference from which to measure nail growth is important since if this point varies between determinations, the measurements will be rendered meaningless. Several proximal (fixed) landmarks have been proposed: 1. The cuticle. This landmark appears to be relatively stable in patients who do not manipulate or cut back the cuticle, especially if submillimeter measurements are not needed. The cuticle is obviously invalid as a point of reference if pushed back or cut. 2. The proximal nail fold. This is similar to the cuticle in its ability to be pushed back, though not as easily or permanently as the cuticle. Whether the edge of the proximal nail fold can be precisely defined for very fine measures is unclear. This has been shown to have a small interobserver error rate13 and may be the most suitable landmark. 3. The distal interphalangeal (DIP) joint. This is among the more precise landmarks. It can be used in two ways: one in which a mold or rigid brace is constructed so that the DIP joint is flexed at an identical angle at each determination and the other as a radiographic landmark (used with a radiopaque marker cemented to the nail plate). The first technique requires a simple tool and may be unsuitable for submillimeter measurements. The second technique, though perhaps the most precise, involves exposure to ionizing radiation and is therefore suitable only in research settings with informed consent, if ever. 4. The lunula. Using this structure has several disadvantages. First, not everyone has a visible lunula, especially on the index, third, ring, and small fingers. Second, while the lunula appears well defined, under magnification its edge is somewhat vague and would therefore introduce unnecessary error into measurement. 5. A structure cemented to the skin of the dorsal distal phalanx. This has the advantage of being relatively stable and easy to see, but may come loose during a meaningful interval of nail
Measurement of Longitudinal Nail Growth
growth. This technique was cited more than 50 years ago12 and later echoed.2
921
physician to follow the dictum primum non nocere, “first, do no harm.”
107.3.2 DISTAL LANDMARKS Once the issue of proximal landmarks is settled, it becomes relatively easy to choose a way of marking the growing portion of the nail. It makes no difference if the nail is etched, drilled, or if a radiopaque band is cemented to the nail plate. The only specific requirement is that the marking be permanent over the course of measurement.
107.3.3 TECHNIQUE
OF
MEASUREMENT
Several options exist in this regard: 1. A straight rule with at least millimeter (and preferably submillimeter) increments indicated. Such a device should be used with a loupe of 8× magnification or greater. 2. A caliper similar to those used in electrocardiography,16 which is then applied to the straight rule. This is preferable, as the direct application of the straight rule to the curved nail plate may introduce inaccuracy. 3. Devices such as the “split image rangefinder”2 have only very specialized applications, as does time-lapse photography.12
107.4 SUMMARY The technique employed to measure the longitudinal growth of nails should satisfy several criteria: it should be simple, disfigure as little as possible, be safe, and be biologically relevant. The extremes of precision that many methods aim to achieve are only of value in research settings. Except in specialized situations when very short term growth rates are measured, the rate of nail plate growth is so variable between day and night and season15 in an individual as to make overly precise or overly frequent measurements meaningless. No dermatologic diagnosis rests on accurate measurement of longitudinal growth of the nail. In pursuing his interest in this parameter, therefore, it behooves the
REFERENCES 1. Dawber, R. and Baran, R., Nail growth, Cutis, 39, 99, 1987. 2. Orentreich, N., Markofsky, J., and Vogelman, J.H., The effect of aging on the rate of linear nail growth, J. Invest. Dermatol., 73, 126, 1979. 3. Berthold, Beobachtungen uber das quantitative Verhaltniss der Nagel: und Haarbildung beim Menschen, Mullers Arch., 156, 1850. 4. Sharpley-Schafer, E., Relative growth of nails on right hand and left hand respectively: on seasonal variations in rate; and on influences of nerve section upon it, Proc. R. Soc. Edinburgh, 51(1), 8. 5. Bloch, A.M., Etude de la croissance des ongles, C. R. Soc. Biol., 58, 253, 1905. 6. Dufour, 1872. 7. LeGros Clark, W.E. and Buxton, L.H.D., Studies in nail growth, Br. J. Dermatol. Syph., 50, 221, 1938. 8. Bean, W.B., A note on fingernail growth, J. Invest. Dermatol., 20, 2, 1953. 9. Bean, W.B., Nail growth. Thirty five years of observation, Arch. Int. Med., 140, 73, 1980. 10. Babcock, M.J., Methods of measuring fingernail growth rates in nutritional studies, J. Nutr., 55, 323, 1955. 11. Sibinga, M.S., Observations on growth of fingernails in health and disease, Pediatrics, 24, 225, 1959. 12. Morton, R., Visual assessment of nail growth, J. Audiovis. Media Med., 14(1), 31, 1991. 13. Dawber, R., Fingernail growth in normal and psoriatic subjects, Br. J. Dermatol., 82, 454, 1970. 14. Kandil, E., Accurate measurement of nail growth, Int. J. Dermatol., 11(1), 54, 1972. 15. Scher, R.K. and Daniel, R.C., Nails: Therapy, Diagnosis, Surgery, W.B. Saunders, Philadelphia, 1991. 16. Hillman, R.W., Fingernail growth in the human subject. Rates and variations in 300 individuals, Hum. Biol., 27, 255, 1955. 17. Godwin, K.O., An experimental study of nail growth, J. Nutr., 69, 121, 1959. 18. Basler, A., Growth processes in fully developed organisms, Med. Klin., 33, 1664, 1937.
108 Measurement of Nail Thickness Gregor B.E. Jemec Department of Dermatology, Roskilde Hospital, University of Copenhagen, Roskilde, Denmark
CONTENTS 108.1 Variable.................................................................................................................................................................923 108.2 Methods ................................................................................................................................................................923 108.3 Correlation between the Methods........................................................................................................................924 108.4 Practical Recommendations .................................................................................................................................924 References .......................................................................................................................................................................924
The nail is a well-defined keratin structure. The larger part of it is clearly visible and immediately accessible for studies. Yet comparatively few studies have been made of the nail compared with, e.g., hair. In these, mostly nail longitudinal growth has been measured, although a few studies have also paid attention to the structure of the nail plate and volume of nail growth. Studies of nail thickness may be relevant to quantification of nail matrix output, to penetration studies, and to studies of the nails as markers of nondermatological disease.1
108.1 VARIABLE The variable studied is simply the thickness of the nail, defined as the distance between the superficial and profound surfaces of the nail. The nail plate is of uneven thickness, being thinner at the nail matrix where it is formed and gradually thickening toward the free edge.2 All measurements of the thickness should therefore be made at a defined point along the nail plate, most often at the free edge or immediately proximal to this. The thickness is most appropriately expressed in millimeters. Normal values range between 0.3 and 0.9 mm, with the first finger having a thicker nail than the fifth finger. Toenails are usually thicker than fingernails, and men appear to have thicker nails than women.
108.2 METHODS Several noninvasive methods are available for the measurement of nail thickness: simple callipers or micrometers, high-frequency ultrasound, and optical coherence tomography. However, only the first two are discussed, as the necessary apparatus is more readily available.
Simple callipers or micrometers are readily available and have sufficient precision to measure nails. The nails can be measured in vivo provided that a sufficient free length of nail is available for the instrument to grip at the distal end of the nail plate. The instrument should also preferably exert a standardized pressure on the gripped tissue, although this is less important in hard tissues such as nail than in, e.g., skin. The positioning should always be perpendicular to the nail plate surface, and the shortest measurements are the most correct.3 In one published study the coefficient for such measurements was 5.3% (SD = 2.4%). The main disadvantages of the method are that it can be difficult to get a grip on the free edge of the nail if it is cut short, and that it is only the thickness at the free edge that is measured. The nail progressively increases in thickness toward the free edge, and the calliper measurements therefore represent a maximum value of nail thickness rather than, e.g., an average or a minimum value. High-frequency ultrasound offers the advantage of being able to measure the thickness of the nail plate anywhere along the nail, but also here repeated measurements are necessary to identify the correct (smallest) thickness of the plate. It should be specified as well as possible where along the nail plate the measurements have been made.4 Both A- and B-scans give the thickness, but the B-scan provides additional information (Figure 108.1). Optical coherence tomography gives images similar to Bscans. The speed of sound within the nail has been found to be 2459 m/sec. This can be used for exact calculations of distance. Nail hydration may lead to slowing of the speed of sound within the nail, and hence to an overestimation of 923
924
Handbook of Non-Invasive Methods and the Skin, Second Edition
108.4 PRACTICAL RECOMMENDATIONS B
A D C
FIGURE 108.1 Ultrasound image with A-scan superimposed on B-scan showing the thickness of the normal human fingernail, marked between the arrows. A: Ultrasound membrane (machine). B: Proximal nail fold. C: Dorsal nail plate. D: Ventral nail plate.
actual distances. The coefficient of variation for ultrasound measurements was 4% (SD = 1.3%) in one study.
108.3 CORRELATION BETWEEN THE METHODS The two methods correlate well (r = 0.79, p < 0.001).3
Position your instrument of choice so as to be perpendicular to the surface of the nail. The smallest thickness is the correct value. If you are using ultrasound, mark where you measure, and do not soak the nail.
REFERENCES 1. Wollina, U., Berger, M., and Karte, K., Calculation of nail plate and nail matrix parameters by 20 MHz ultrasound in healthy volunteers and patients with skin disease, Skin Res. Technol. 7, 60, 2001. 2. Johnson, M. and Shuster, S., Continuous formation of nail along the bed, Br. J. Dermatol. 128, 277, 1993. 3. Finlay, A.Y., Western, B., and Edwards, C., Ultrasound velocity in human fingernail and effects of hydration: validation of in vivo thickness measurement techniques, Br. J. Dermatol. 123, 365, 1990. 4. Jemec, G.B.E. and Serup, J., Ultrasound structure of the human nail plate, Arch. Dermatol. 125, 643, 1989.
109 Image Analysis of the Nail Surface Claudine Piérard-Franchimont and Gérald E. Piérard Department of Dermatopathology, University Hospital Sart Tilman, Liège, Belgium
CONTENTS 109.1 Introduction ..........................................................................................................................................................925 109.2 Nail Surface Modifications ..................................................................................................................................925 109.2.1 Longitudinal Striations ..........................................................................................................................925 109.2.2 Herringbone Nail ...................................................................................................................................925 109.2.3 Beau’s Lines...........................................................................................................................................926 109.2.4 Pitting and Rippling...............................................................................................................................926 109.2.5 Trachyonychia........................................................................................................................................926 109.2.6 Ridging of the Nail Underface ..............................................................................................................926 109.3 Instrumental Assessments ....................................................................................................................................926 109.3.1 Static Microtopography of the Nail Surface .........................................................................................926 109.3.2 Dynamic Microtopography of the Nail Surface....................................................................................926 109.3.3 Nail Microindentation............................................................................................................................926 109.3.4 Nail Sclerometry ....................................................................................................................................927 109.4 Conclusion............................................................................................................................................................927 References .......................................................................................................................................................................926
109.1 INTRODUCTION Nail is a hard but flexible structure growing continuously and potentially submitted to many types of microtraumatisms. It may grossly look smooth, but closer examination shows it is not. The structure of the nail surface has attracted so far little interest from researchers. Descriptive reports are rarely supported by quantification of the nail plate microrelief. However, several patterns of nail surface abnormalities are well identified by clinical inspection. They result from endogenous dermatosis or from trauma and weathering.
109.2 NAIL SURFACE MODIFICATIONS 109.2.1 LONGITUDINAL STRIATIONS Longitudinal striations at the nail surface present as indented grooves separated by projecting ridges. This condition may be considered a physiological feature when presenting as shallow depressions, usually parallel, and separated by low projecting ridges. They become more prominent with age and in some particular conditions described hereafter.
Onychorrhexis consists of a series of narrow, longitudinal, parallel superficial striations with the appearance of having been scratched by an awl. Sometimes dust particles are ingrained into the nail surface. Splitting of the free edge of the nail is common. The small rectilinear projections extend from the proximal nail fold to the free edge of the nail. They may also stop short or be interrupted at regular intervals, giving rise to a beaded appearance. In some patients a wide, longitudinal median ridge has the appearance, in cross section, of a circumflex accent. Median nail dystrophy is an uncommon condition consisting of a longitudinal groove in the mid-line or just off center of the thumbnails, starting at the cuticle and growing out of the free edge. Tumors nearby the nail matrix may exert pressure and produce a single wide, deep, longitudinal groove or canal. This aspect disappears when the cause is removed.
109.2.2 HERRINGBONE NAIL Nail ridging, with oblique lines pointing centrally to meet in the mid-line, is an uncommon pattern occurring in childhood. 925
926
Handbook of Non-Invasive Methods and the Skin, Second Edition
109.2.3 BEAU’S LINES Beau’s lines are transverse grooves extending from one lateral edge of the nail to the other. They affect all nails at corresponding levels. The width of the transverse groove relates to the duration of the process that has altered the nail matrix activity. An abrupt distal limit of the groove indicates a sudden outbreak of disease. A sloping aspect suggests a more progressive onset. The proximal limit of the depression may also be abrupt or sloped.
109.2.4 PITTING
AND
RIPPLING
The eponym Rosenau’s depressions refers to nail pitting and rippling. Pits develop as a result of defective nail formation in punctate areas located in the proximal portion of the nail matrix. The surface of the nail plate may be studded in a buckshot pattern with small punctate depressions that vary in number, size, depth, and shape. It is usual to accept five pits as an arbitrary number for the physiological condition. The depth and width of the pits relate to the extent of the matrix involved. Their length is determined by the duration of the damage. Deep and irregularly shaped pits often suggest psoriasis, but they are not pathognomonic for any disease. Pits may be randomly distributed or evenly patterned in series along one or several longitudinal lines. They are sometimes arranged in a crisscross pattern, and may thus resemble the external surface of a thimble. Regular pitting may convert to rippling or ridging, and these conditions may correspond to variants of uniform pitting.
109.2.5 TRACHYONYCHIA Trachyonychia refers to a spectrum of alterations resulting in severe nail roughness, as if the surface had been rubbed with sandpaper.
109.2.6 RIDGING
OF THE
NAIL UNDERFACE
The undersurface of the nail disclosed after avulsion exhibits a topographical aspect unrelated to the outer surface of the same nail. Deep longitudinal striations are present. They deepen with age.
109.3
INSTRUMENTAL ASSESSMENTS
The distinctive features of the nail microrelief can be studied using various procedures. Clinical examination allows qualitative assessment of the gross microtopography. Low-magnification photographs under carefully controlled and repeatable conditions can be used to document the nail microrelief. More precise quantitative examinations can be made on the outer portion of the nail in vivo, or after its avulsion, or after making a copy of negative
replicas. The rigorous use of optical profilometric methods or any other microtopographic assessment1 on this material brings quantitative information. The diverse typical alterations are clearly evidenced.2–5 In addition, due to the continuous nail growth, onychochronobiology can be studied by the same means.6,7 The same methods are suitable for assessing the mechanical properties of nail in combination with the microindentation and the sclerometry methods.
109.3.1 STATIC MICROTOPOGRAPHY SURFACE
OF THE
NAIL
Quantitative assessments of the nail microtopography are usually performed by longitudinal scans. Transversal scans are less easy to interpret due to the natural curvature of the nail. When information must be obtained in this direction, it is recommended to examine sections 5 mm in length to minimize this pitfall. It should be kept in mind that native alterations are better revealed at the proximal part of the nail. Weathering and natural microabrasions may add their effects in a cumulative way when moving toward the distal part of the nail. Sources of variability such as nail plate wear should be discarded. Controlled positioning of the nail is of the upmost importance.
109.3.2 DYNAMIC MICROTOPOGRAPHY OF THE NAIL SURFACE Repeated controlled assessments over time give insight in onychochronobiology. The effects of therapies can thus be assessed. The speed of growth of the nail can be assessed simultaneously when a mark has been engraved at the initial examination. Thus, it is possible to evaluate the rate of improvement or degradation of the nail condition. A biologically meaningful interval should be respected between successive measurements. In this consideration the speed of nail growth must be taken into consideration. Indeed, there may be interdependence between the disclosed microtopography changes and variations in the nail growth rate. An example is given by Beau’s lines and beaded nail.1 Seasonal variations in the nail surface microtopography may vary from insignificant to quite obvious.6
109.3.3 NAIL MICROINDENTATION Experimental microindentation allows the study of some mechanical properties of the nail. A load is applied under controlled conditions on a small surface. The indentation is usually measured during the test procedure. If any residual plastic deformation persists after releasing the force, the imprint of the device can be observed by profilometry.
Image Analysis of the Nail Surface
109.3.4 NAIL SCLEROMETRY Sclerometry deals with the dynamic assessment of the response of an object during microabrasion. In addition to the classical abrasion parameters, profilometry may describe in another way the groove traced in the nail. The effects of nail-hardening products and nail protectors can be conveniently assessed by that way. Similarly, nail softening by xenobiotics or altered states of health can also be quantified.
109.4 CONCLUSION The nail microrelief is subjected to variability due to physiological parameters, weathering, external trauma, and pathological features altering the nail matrix. Some microtopographic alterations are linked to changes in the nail growth rate and in the nail hardness. Objective assessments of the nail surface topography have been seldom addressed in the literature. Presumably the methods developed for the skin surface microtopography can be applied to the nail apparatus. These methods could give insight into onychochronobiology, providing unique information about the physiopathological processes having affected the nail over the past weeks and months.
927
REFERENCES 1. Lévêque, J.L., EEMCO guidance for the assessment of skin topography, J. Eur. Acad. Venereol., 12, 103, 1999. 2. De Doncker, P. and Piérard, G.E., Acquired nail beading in patients receiving itraconazole: an indicator of faster nail growth? A study using optical profilometry, Clin. Exp. Dermatol., 19, 404, 1994. 3. Nikkels-Tassoudji, N., Piérard-Franchimont, C., De Doncker, P., and Piérard, G.E., Optical profilometry of nail dystrophies, Dermatology, 190, 301, 1995. 4. Piérard, G.E. and Piérard-Franchimont, C., Fractal microrelief of the skin and nail, Giorn. Int. Dermatol. Ped., 8, 75, 1996. 5. Piérard-Franchimont, C. and Piérard, G.E., Surface image analysis of nail alterations in juvenile pityriasis rubra pilaris, Skin Res. Technol., 4, 34, 1998. 6. Piérard, G.E. and Piérard-Franchimont, C., Dynamics of psoriatic trachyonychia during low dose cyclosporin A treatment. A pilot study on onychochronobiology using optical profilometry, Dermatology, 192, 116, 1996. 7. Piérard-Franchimont, C., Jebali, A., Ezzine, N., Letawe, C., and Piérard, G.E., Seasonal variations in polymorphic nail surface changes associated with diabetes mellitus, J. Eur. Acad. Dermatol. Venereol., 7, 182, 1996.
Section III Clinical Experimentation, Evaluation, and Quantification
Guidelines for Assessment 110 General of Skin Diseases Elisabeth A. Holm and Gregor B.E. Jemec Department of Dermatology, Roskilde Hospital, University of Copenhagen, Roskilde, Denmark
CONTENTS 110.1 110.2 110.3 110.4 110.5
Introduction .......................................................................................................................................................931 Validity...............................................................................................................................................................932 Reliability ..........................................................................................................................................................933 Sensitivity and Responsiveness.........................................................................................................................933 Generic Instruments ..........................................................................................................................................934 110.5.1 Sickness Impact Profile (SIP).............................................................................................................934 110.5.2 Nottingham Health Profile (NHP) ......................................................................................................934 110.5.3 Medical Outcomes Study (MOS) 36-Item Short Form (SF-36)........................................................934 110.6 Dermatology Quality of Life Instruments ........................................................................................................935 110.7 Generic Quality of Life Instruments for Skin Disorders .................................................................................935 110.7.1 Dermatology Life Quality Index (DLQI)...........................................................................................935 110.7.2 Children’s Dermatology Life Index (CDLQI) ...................................................................................935 110.7.3 Skindex................................................................................................................................................935 110.8 Disease-Specific Dermatology Quality of Life Instruments ............................................................................935 110.8.1 Atopic Dermatitis................................................................................................................................935 110.8.2 Psoriasis Index of Quality of Life (PSORIQoL) ...............................................................................936 110.9 Disease-Specific Assessment.............................................................................................................................936 110.9.1 Atopic Eczema ....................................................................................................................................936 110.9.2 Psoriasis...............................................................................................................................................938 110.9.3 Acne ....................................................................................................................................................938 110.10 Other Skin Disorders.........................................................................................................................................938 110.11 Conclusion .........................................................................................................................................................938 References .......................................................................................................................................................................940
110.1 INTRODUCTION The skin can be assessed by a diversity of methods as demonstrated by the breadth of topics covered in this book. The accurate and appropriate measurement of health outcome is an important aspect of clinical work and research, and forms the basis of good evidence-based practice. This is of particular importance in chronic recurrent diseases such as skin diseases, where the absolute endpoints in treatment are not death or survival but relative improvement. Assessing disease severity in dermatological disorders presents practical problems because laboratory tests of
disease severity often do not exist. Biopsy and blood test provide valuable information regarding diagnosis, but are often only supplementary to the assessment of disease activity. Measurement as quality of life (QoL) has grown to be an important — and sometimes the decisive — endpoint of disease severity considerations. In a busy dermatological department or practice, the assessment of the patient’s skin disease is often the quick clinical look. The outcome depends upon the doctor’s knowledge, experience, and how detailed the severity is described in the patient’s records. Comments such as “much better” or “fine” convey little to a doctor who did not see the patient previously, and probably also very little to the original author when reviewing the notes months later. A more 931
932
Handbook of Non-Invasive Methods and the Skin, Second Edition
objective and reproducible assessment would therefore contribute significantly to the quality of even routine clinical work. All disease severity assessments should satisfy basic requirements if they are to be clinically useful. These requirements are primarily validity, reliability, sensitivity, and responsiveness. These four properties are described in the following using examples from the assessment of atopic eczema.
X X X X X X X X X X X X X X X X X X X X
X X X X X X X X X X X X X X X X X X X X (a)
110.2 VALIDITY
X X X X X X X X X X
Validation of methods is the process of determining whether the method or instrument measures what it is intended to measure, and if it is useful for the intended purpose. For example, to what extent is it reasonable to claim that a clinical scoring system like SCORAD for atopic eczema (AE) really is assessing the severity of AE? Since we are attempting to measure an ill-defined variable (severity of AE), we can only infer that the instrument is valid insofar as it correlates with other observable behavior in AE. This validation process consists of a number of stages, and it is traditionally subdivided into three main aspects: (1) content, (2) criterion, and (3) construct validity. Content validity concerns the extent to which the symptoms we include in our new scoring system are sensible and reflect the intended domain (the specific disease) of interest. Are all relevant areas of interest covered in our new construct? For assessing the symptomatology of atopic eczema, the method should include items relating to all major symptoms of the disease. As an example, we design a scoring system that only includes symptoms like erythema, papulation, and edema, which are dominant features in the acute phase of the disease (Figure 110.1a). Such a scorings systems would have no content validity for the general population of patients with AE, in whom chronic lesions are prominent. We must therefore include symptoms like lichenification and skin dryness to ensure that it also covers patients in a stable phase, if we want a valid system for both the acute and chronic phases (Figure 110.1b). Methods of content validation are not amenable to formal statistical testing. To ensure that the instrument covers all the relevant issues, it is important in the construction phase to:
X X X X X X X X X X
• • •
Review literature, including published results from clinical trials Include input from specialists Collect information from relevant patient associations
After the scoring system has been constructed and before it is tested in a pilot study, it is useful to check
No content validity, since only one small area of interest is covered by our focus.
X X X X X X X X X X
Content validity, since items cover a wide aspect of the phenomena that we want to measure.
X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X (b)
FIGURE 110.1 See text for details.
whether the instrument covers the intended area by once more contacting experts in the field of interest and relevant patient groups. This critical review of an instrument after it has been constructed is called by some authors face validity. Criterion validity describes how well an instrument agrees with the true value, ideally a golden standard. For atopic eczema no golden standard exists; therefore, criterion validities for new instruments are often compared against one or more well-established instruments, e.g., SCORAD. This method may be problematic, if the motivation for creating a new tool is to compensate for inadequacies of the existing instrument. In this case, the comparison of new instruments against established instruments is of limited value. Comparing a new scoring system with an established one is, however, a good method if the motive is to develop either a shorter or simpler system. The statistical methods used for comparison of new and old are correlation, which is the method of analysis of the possible association between two continuous variables, and regression, if we want to describe more precisely the degree of association between the two methods.1 Construct validity is one of the most important properties of a measurement instrument. It is an assessment of the degree to which an instrument measures the construct that it was designed to measure. For example, does SCORAD really measure severity of atopic eczema? The process of analyzing construct validity involves several steps, and it is a lengthy and ongoing process. A hypothetical model is constructed and data are collected and analyzed. If the relationship between data and other well-known factors of the same construct relates well, the instrument appears to be valid. The greater the supporting evidence, the more confident we can be that the model is
General Guidelines for Assessment of Skin Diseases
an adequate representation of the construct that we want to measure, e.g., atopic eczema severity. Every study in which the new instrument is seen to correlate with other aspects of the disease supports the validity of the tool. In AE such supportive evidence could be, e.g., steroid hormone consumption, doctors visits, and time spent on treatment. In contrast, a single negative finding may lead to reconsideration of the whole theoretical background for the new instrument. Construct validity is the most applicable to numerical analysis of the three validation processes. For greater detail, see Fayers and Machin.2 Cases where our methods or instruments correlate positively with other aspects of our construct display convergent validity; e.g., an increase in objective SCORAD is normally followed by an increase of pruritus. If dimensions of the severity appear to be negatively correlated, this is called divergent validity. Correlation and factor analysis plays an important role in construct validation.2 One of the simpler forms of construct validation is known-groups validity. This is based on the principle that specified groups of patients are anticipated to score differently from each other, and the instrument should be able to distinguish this difference. Known-groups comparison is therefore a combination of test for validity and a form of sensitivity or responsiveness. An example of known-group validity in dermatology is Jemec and Wulf’s study,3 where two disease-specific scoring systems (PASI for psoriasis and SCORAD for AD) were compared for patients suffering from psoriasis and AD.
110.3 RELIABILITY Assessment of reliability is to determine if, e.g., a scoring system produces reproducible and consistent results. A reliable scoring system will give reproducible or similar values if it is used repeatedly on the same patient, when the patient’s condition is stable. This repeatability reliability is based upon analysis of correlations between repeated measurements. The measurements can be repeated over time (test–retest reliability), by different observers (interobserver reliability), or by the same observer on two or more occasions (intraobserver reliability). In many cases information about correlations gives insufficient data to assess the methods. If one is more interested in prediction or estimation, regression analysis should be used. The simplest method of assessing repeatability for binary data is the proportion of agreement, when the same instrument is applied on two occasions. For example, the same patients are assessed twice and the results are presented in a 2 × 2 table:
933
First Assessment Second Assessment Positive Negative Total
Positive A11 C21 AC
Negative B12 D22 BD
Total AB CD N
The number of agreements, that is, the number of patients who respond in the same way in both assessments, is A11 + D22, and the proportion of agreement is PAgree = (A11 + D22)/N Another widely used method for assessing repeatability for binary data is the kappa coefficient, κ. The kappa coefficient provides a better method than the above concept, because the kappa coefficient also reflects the agreement, which exists purely by chance. If there is perfect agreement, κ = 1. If the agreement is no better than by chance, κ = 0. If the agreement is less than would be expected by chance, κ = negative value. Pearson’s correlation is often used as a measure of reliability. This is not recommended since repeated measures may be highly correlated even though they can be systematically different. For example, if all patients in a second assessment score 20 points higher than the first assessment, and then data for both measures are plotted in a X–Y diagram, the slope for the two graphs will be exactly the same. This gives a correlation value of 1, indicating perfect association, while in reality the agreement between the two measurements is poor, as they are parallel displaced by 20. A method recommended for assessing reliability is analysis of variance (ANOVA). For details, see the Altman.1 The word reliability is rather confusingly used for another property of scale validation: internal reliability. Internal reliability assesses if the scores from different items correlate with each other and with the total scale score. Are all items related to the same latent variable? Cronbach’s coefficient α is one of the most widely used methods of assessing internal consistency. If items are uncorrelated, αCronbach = 0, and if all items are identical and have perfect correlation, αCronbach = 1. Cronbach’s coefficient α will increase with the number of items used in the scale.
110.4 SENSITIVITY AND RESPONSIVENESS Sensitivity and responsiveness are two closely related properties to repeatability reliability. Sensitivity is the instrument’s ability to detect differences between groups. This can, e.g., be the difference between patients with various degrees of disease severity or different treatment groups. Sensitivity is one of the most important properties of an instrument, since the usefulness
934
Handbook of Non-Invasive Methods and the Skin, Second Edition
of a measure is dependent upon its ability to detect clinically relevant differences. There are situations where an established sensitive method fails to detect differences. This occurs if the data only take up a small spectrum of the scale. For example, we have selected patients with either very mild (SCORAD 1 to 10) or very severe (SCORAD 93 to 103) atopic eczema. If a new instrument is only sensitive to differences, which correspond to a change in SCORAD of 10, it will not be able to detect any differences within the groups of mild or severe eczema. This phenomenon is called the floor-and-ceiling effect. An instrument that is valid and reliable in one study can therefore lose its properties in another study, depending upon the selection of sample size, type of study populations, types of interventions, etc. Sensitivity is usually assessed by cross-sectional comparisons of groups of patients in which differences are expected. In practice, sensitivity is closely related to known-groups validity. A highly sensitive scale will usually also be reliable and highly responsive. Responsiveness is closely related to sensitivity, but relates to within patients, where sensitivity relates to between patients. If a method is responsive, it is able to detect changes in a patient’s disease status over time. Responsiveness can also be regarded as providing additional evidence of our new instrument’s validity, since it confirms that the anticipated responses occur when the patient’s status changes. The construct of our hypothetical model is confirmed. Finally, besides being valid, reliable, and sensitive to changes, our instrument also needs to be acceptable for both patient and investigator.
110.5 GENERIC INSTRUMENTS Generic instruments have been constructed for general use, irrespective of the illness or condition of the patients, to compare data between different diseases, including skin disorders. The generic questionnaires may often also be applicable to healthy people, and are constructed to cover a spectrum from healthy people to very sick patients. Because of the width of the spectrum, individual sensitivity and responsiveness suffer. The first generic instruments were developed primarily with population surveys in mind, although they were later applied in clinical trials. They are commonly described as quality of life (QoL) questionnaires, but many of these only measure physical symptoms, and are therefore more appropriately termed health status instruments than QoL instruments. Many different generic assessment instruments exist. In this chapter, three different instruments will be described: Sickness Impact Profile (SIP), Nottingham Health Profile (NHP), and Medical Outcomes Study 36Item Short Form (SF-36).
110.5.1 SICKNESS IMPACT PROFILE (SIP)* Bergner et al.4 developed SIP; the final version was finished in 1981. The SIP is a questionnaire designed to measure sickness-related behavioral dysfunction and was developed for use as an outcome measure in the evaluation of health care. The full questionnaire (16 pages) consists of 136 items and takes about 30 minutes to complete. The items describe everyday activity and cover the impact upon health of activities, behavior, social functioning, and emotional well-being. All items are negatively worded, representing dysfunction. Twelve main areas of dysfunction are covered. A standard scoring method is used for each of these 12 dysfunctions. These 12 represent two higher-order dimensions, (1) physical and (2) psychosocial, which can be scored in a similar manner as the 12. SIP has been tested in respect to validity, reliability, responsiveness, and sensitivity. Test–retest reliability (r = 0.92) and internal consistency (r = 0.94) were high, and SIP is sensitive even to minor changes in morbidity.
110.5.2 NOTTINGHAM HEALTH PROFILE (NHP)** Hunt et al.5 developed the NHP in 1981. It is often used in population studies of general health assessment and in clinical trials. Although it is less sensitive to minor changes, it was mainly developed to assess whether there are any health problems. It measures health-related quality of life within the sections of energy, sleep, emotions, pain, mobility, and social isolation, as well as the frequency of health-related problems pertaining to paid employment, housework, hobbies, family life, social life, sex life, and holidays. It consists of 38 items, where the respondents are given a list of statements that they answer yes or no. As for SIP, all items are negatively worded, representing dysfunction. Compared to SIP, the NHP is short and easy to complete. The NHP is well documented with regard to reliability and validity, and it is useful in describing the impact of chronic disease.6
110.5.3 MEDICAL OUTCOMES STUDY (MOS) 36-ITEM SHORT FORM (SF-36)*** The most widely used of the general health status questionnaires is SF-36.7 Ware and Sherbourne7 developed SF36 nearly 10 years after both the SIP and NHP were constructed. SF-36 was constructed to survey health status * For permission to use contact: Health Services Research & Development Center, John Hopkins School of Hygiene and Public Health, 624 North Broadway, Baltimore, MD 21205-1901. ** For permission to use contact: Dr. Stephen McKenna, Galen Research, Enterprise House, Manchester Science Park, Lloyd Street North, Manchester M15 6SU, U.K. *** For permission to use contact: Dr. John Ware, Medical Outcomes Trust, 20 Park Plaza, Suite 1014, Boston, MA 02116-4313.
General Guidelines for Assessment of Skin Diseases
in the Medical Outcomes Study. The SF-36 was also designed for use in clinical practice and research, health policy evaluations, and general population surveys. As the name implies, there are 36 questions, where most refer to the past 4 weeks. The SF-36 includes one multi-item scale that assesses eight health concepts: (1) limitations in physical activities because of health problems, (2) limitations in social activities because of physical or emotional problems, (3) limitations in usual role activities because of physical health problems, (4) bodily pain, (5) general mental health (psychological distress and well-being), (6) limitations in usual role activities because of emotional problems, (7) vitality (energy and fatigue), and (8) general health perceptions. These eight can be scored separately, but also summarized into two measures: (1) physical health and (2) mental health. SF-36 is used in several dermatological studies8–10 and is tested for validity, reliability, and sensitiveness/responsiveness.11–13
110.6 DERMATOLOGY QUALITY OF LIFE INSTRUMENTS Generic instruments are constructed to cover a wide range of conditions and diseases. Therefore, they often lack the sensitivity and responsiveness to detect differences that arise as a consequence of treatments compared in, e.g., clinical trials. As a consequence, disease-specific questionnaires have been developed.
110.7 GENERIC QUALITY OF LIFE INSTRUMENTS FOR SKIN DISORDERS
935
110.7.2 CHILDREN’S DERMATOLOGY LIFE INDEX (CDLQI)** CDLQI is designed for use in children, i.e., patients from age 5 to age 16. It was developed by Lewis-Jones and Finlay in 199316 and is based upon the DLQI. Only a few questions have been changed, so the construct is more relevant for children. It is widely used, self-explanatory, and self-administered. It is usually completed in 1 to 2 min, and like the DLQI, it is tested for validity, reliability, and sensitivity to changes and is translated into several languages.
110.7.3 SKINDEX*** The Skindex is constructed to measure the effect of skin disease on patients’ quality of life. Initially a 61-item selfadministered survey instrument was developed, the socalled Skindex. The original Skindex had eight scales, each of which addressed a construct, or an abstract component, in a comprehensive conceptual framework: cognitive effects, social effects, depression, fear, embarrassment, anger, physical discomfort, and physical limitations. Item responses were standardized from 0 (no effect) to 100 (maximal effect).17,18 The 61-item instrument was tested for validity and reliability, but the acceptability by the respondents was weak, because it was time-consuming. As a consequence, the Skindex-2919 and later the Skindex-1620 were developed. These new and shorter versions have improved discriminative and evaluative capability and have decreased respondent burden compared to Skindex-61. The Skindex is commonly used and translated into several languages.
110.7.1 DERMATOLOGY LIFE QUALITY INDEX (DLQI)*
110.8 DISEASE-SPECIFIC DERMATOLOGY QUALITY OF LIFE INSTRUMENTS
DLQI is a simple practical questionnaire technique for routine clinical use for any skin disease. It was developed by Drs. Finlay and Khan in 199214,15 and is today the most widely used general life quality questionnaire in dermatology. It covers different aspects of life impairment, e.g., symptoms, feelings, daily activities, work or school, personal relationships, and treatment. It consists of 10 questions, where each question has four alternative responses and scores from 0 to 3. The DLQI is calculated by summing the score of each question, resulting in a score range from 0 to 30. The higher the score, the greater the impairment of quality of life. The DLQI is easy and quick to use. It is tested according to validity, reliability, and sensitivity to changes and is translated into many different languages.
110.8.1 ATOPIC DERMATITIS****
* For permission to use contact: A.Y. Finlay, Head of Department of Dermatology, University of Wales College of Medicine, Heath Park, Cardiff, CF 14 4XN, Wales, U.K.
The Dermatitis Family Impact Questionnaire (DFI)21 and the Infants’ Dermatitis Quality of Life Index (IDQOL)22 are two simple questionnaires that are designed to assess the impact on quality of life of infants with atopic dermatitis and their families. The authors of CDLQI developed both questionnaires, and the two instruments therefore have the same basic construction and scoring system as ** For permission to use contact: Ms. Lewis-Jones or A.Y. Finlay, Head of Department of Dermatology, University of Wales College of Medicine, Heath Park, Cardiff, CF 14 4XN, Wales, U.K. *** For permission to use contact: M.M. Chren, R.J. Lasek, S.A. Flocke, and S.J. Zyzanski, Dermatology Service, Cleveland Veterans Affairs Medical Center, Cleveland, Ohio. **** For permission to use contact: Ms. Lewis-Jones or A.Y. Finlay, Head of Department of Dermatology, University of Wales College of Medicine, Heath Park, Cardiff, CF 14 4XN, Wales, U.K.
936
Handbook of Non-Invasive Methods and the Skin, Second Edition
TABLE 110.1 Disease-Specific Quality of Life Instruments for Dermatological Diseases Type of Disease
Name of Instrument
Reference
Acne Atopic dermatitis
The Acne-Specific Quality of Life Questionnaire (Acne-QoL) The Family Dermatitis Impact Questionnaire (FDI) The Infants’ Dermatitis Quality of Life Index (IDQOL) Hairdex Melasma Quality of Life Scale (MELASQOL) Psoriasis Disability Index Psoriasis Life Stress Inventory Psoriasis Index of Quality of Life (PSORIQoL) Scalpdex Hornheide questionnaire for psychosocial support
24 21 22 25 26 27 28 23 29 30
Hair disorder Melasma Psoriasis
Scalp, dermatitis Skin cancer
the CDLQI and DLQI. In addition to the 10 items, the IDQOL also has a global severity question. Both the DFI and the IDQOL are tested for validity, reliability, and sensitivity to changes and are translated into several languages.
110.8.2 PSORIASIS INDEX (PSORIQOL)
OF
QUALITY
OF
LIFE
The PSORIQoL is a new (2003) psoriasis-specific measure of QoL. It consists of 25 dichotomous items, making it short and practical for use in clinical studies. It differs from existing patient-reported outcome measures used in dermatology in the way that items do not directly assess impairment or disability, but rather the impact of these and other influences on the QoL of the patient.23 For other disease-specific QoL instruments, see Table 110.1.
110.9 DISEASE-SPECIFIC ASSESSMENT In lieu of objective measures in dermatology, many scoring systems have been developed. These methods generally describe a few clinical parameters: area involved, assessment of severity by scoring of three to six relevant symptoms on a scale from 0 = no involvement to 3 = severe, in the inflamed area. Some are designed for overall assessment, e.g., PASI and SCORAD, and some for target lesions, such as ADSI for AE. Measurements of extent have been claimed to be an “impossible task” for many dermatological diseases, in particular AE.31 Therefore, many different assessment methods such as the rule of hand, color coding,32 a system of tick boxes,33 computer-assisted body surface area34 image analysis system,35 and the rule of nines in, e.g., SCORAD,36 have been developed.
9%
Chest 18% 9%
Back 18%
9%
1% 18%
18%
FIGURE 110.2 Body surface area by Rules of Nines.
110.9.1 ATOPIC ECZEMA For atopic eczema at least 15 different disease-specific objective skin examination scales exist.37 In this section some of the most commonly used, like SCORAD and EASI, are presented, while other scoring instruments for AD are listed in Table 110.2. SCORAD (Scoring Atopic Dermatitis) was developed by the European Task Force on Atopic Dermatitis in 199336 and is the most extensively tested of all existing AE severity indices. It is an overall assessment and includes scoring of extent, intensity, and subjective symptoms (pruritus and sleep loss). The subjective symptoms can be left out in what is then termed an objective SCORAD. It consists of a combination of: 1. Assessment of extent (the rule of nines). Be aware of the differences in body area size
General Guidelines for Assessment of Skin Diseases
937
TABLE 110.2 Disease-Specific Instruments for Dermatological Diseases Type of Disease Acne Atopic dermatitis
Dry skin
Dyshidrotic eczema Hand dermatitis Inflammatory disease Lipodystrophy Mastocytosis Melasma Morphea Pemphigus Psoriasis
Rosacea Scleroderma
Name of Instrument
Reference
Leeds Acne Grading Scale SCORAD Eczema Area and Severity Index (EASI) Self-administered EASI (SA-EASI) Assessment Measure for Atopic Dermatitis (ADAM) Atopic Dermatitis Area and Severity Index (ADASI) Atopic Dermatits Severity Index (ADSI) Basic Clinical Scoring System (BCSS) Costa’s Simple Scoring System (Costa’s SSS) Leicester score Nottingham Eczema Severity Score Rajka and Langeland Six-Area, Six-Sign Atopic Dermatitis Severity Index (SASSAD) Skin Intensity Score (SIS) Six-Area, Total Body Severity Assessment (TBSA) Objective Severity Assessment of Atopic Dermatitis Score (OSAAD) Overall Dry Skin Score (ODS) Dry Skin Area and Severity Index (DASI) Specified Symptom Sum Score (SRRC) Dyshidrotic Eczema Area and Severity Index Work Productivity and Activity Impairment–Chronic Hand Dermatitis Questionnaire in chronic hand dermatitis Dermatology Index of Disease Severity (DIDS) Objective lipodystrophy severity grading scale Scoring Index of Mastocytosis (SCORMA) Melasma Area and Severity Index (MASI) Two methods to assess morphea: skin scoring and the use of a durometer Pemphigus Area and Activity Score (PAAS) A new grading system for oral pemphigus PASI Self-administered Psoriasis Area and Severity Index (SAPASI) National Psoriasis Foundation Psoriasis Score (NPF-PS) Rosacea staging Standard classification of rosacea European Scleroderma Study Group (EscSG) activity indices for systemic sclerosis Scleroderma visual analog scales U.K. Scleroderma Functional Score (UKFS) Arthritis Hand Function Test in adults with systemic sclerosis (scleroderma) Self-administered Systemic Sclerosis Questionnaire (SySQ)
42 36 38 39 43 32 44 45 46 47 33 48 49 50 51 34 52
between infants of 0.05 >0.05 >0.05
Mean ± SD. Compared by Student’s paired t-test.
Instrumental and Computer-Based Methods for Measurement of Surface Area Afflicted with Disease
112.5.2 ASSESSMENT OF INVOLVED AREA IN ATOPIC DERMATITIS (COMPUTER IMAGE ANALYSIS WITH CIAD AND CIAT) In comparison to psoriatic lesions, atopic dermatitis gives rise to ill-defined, diffuse-natured individual lesions. So, it is preferable to use the digital camera system to improve the quality of the photographs in the case of clear cutaneous lesions. The skin lesions of atopic dermatitis patients were estimated by using a visual scoring method, computer image analysis using a digital camera (CIAD), and computer image analysis of direct tracings of the lesions (CIAT). The comparison between the visual scoring method and the CIAT showed that, on the whole, the four dermatologists overestimated the involved area, and that, for each dermatologist, the differences between the results obtained using the visual scoring method and those obtained using the CIAT on each individual patient were statistically significant. Moreover, the mean overestimate was 3.5-fold the results of the CIAT (Figure 112.5). There were highly significant differences between the assessments of the individual dermatologists. On the other hand, three of the four dermatologists showed little day-to-day variation; only one dermatologist’s estimation differed significantly between days 1 and 2. This lack of difference is attributed to the interval between estimations being limited to 24 hours, and the fact that the observers were relatively well-trained dermatologists (Figure 112.6). However, there was some correlation between the visual scoring method and CIAT in that those patients having high scores on the CIAT system also had high scores on
961
the visual scoring method. In general, the visual scoring system overestimated the surface area involved (Figure 112.7). This may have a significant effect on the SCORAD index in atopic dermatitis, which is similar to the PASI score. The reliability of CIAD is tested with CIAT, which was presumed to precisely measure the involved area, although admittedly no method used in this study was absolutely accurate. On comparing CIAT with CIAD, no significant difference was found between the two methods. In terms of the area measurements, the agreement was high, especially on the trunk and arm areas. However, the leg areas of CIAD slightly exceeded the corresponding CIAT areas (Figure 112.8). It should be noted that in a photograph, a three-dimensional object is transformed into a two-dimensional photograph. As a consequence, there is some loss of lateral surface area, especially for cylindrical objects.9,10 In addition, the lesions of patients with atopic dermatitis are located mainly on the anterior or posterior areas of the limbs, rather than the lateral areas. Therefore, the percentage areas (the involved area/total surface area) as assessed by CIAD were overestimated compared to the corresponding values obtained using CIAT. The CIAT and CIAD methods might seem to show more variation in their scores between individual patients than the visual scoring method; however, this is not because these systems are more sensitive, but because of the wide variation in the estimates of the involved area. In 3 of the 23 patients these estimates were greater than 34% of the total body surface area, compared with the remaining 20 patients, in which the involved areas were lower than 13%.
Observer 1 80.00 P < 0.01 70.00
Area(%)
60.00
Visual grading method Computer image analysis of direct tracings
50.00 40.00 30.00 20.00 10.00 0.00
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Patients
FIGURE 112.5 Comparison of estimates between the visual scoring made using the rule of nines and the computer image analysis of direct tracings (CIAT) of the lesions. The results show that, on the whole, the four dermatologists overestimated the involved area compared to the values obtained using CIAT, and that, in each case, the differences between the values obtained by the visual scoring and those obtained using CIAT were statistically significant (p < 0.01, Student’s paired t-test).
962
Handbook of Non-Invasive Methods and the Skin, Second Edition
Observer 2 80.00 P < 0.01 70.00
Area(%)
60.00
Visual grading method Computer image analysis of direct tracings
50.00 40.00 30.00 20.00 10.00 0.00
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Patients
Observer 3 80.00 P < 0.01 70.00
Area(%)
60.00
Visual grading method Computer image analysis of direct tracings
50.00 40.00 30.00 20.00 10.00 0.00
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Patients
Observer 4 80.00 P < 0.01 70.00
Area(%)
60.00
Computer image analysis of direct tracings
50.00 40.00 30.00 20.00 10.00 0.00
Visual grading method
1
2
3
FIGURE 112.5 (Continued.)
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 Patients
Instrumental and Computer-Based Methods for Measurement of Surface Area Afflicted with Disease
963
80 70
Area (%)
60 50 40 30 20 10 0
Day 1 Day 2
Day 1 Day 2
Day 1* Day 2*
Day 1 Day 2
Dermatologist 1 Dermatologist 2 Dermatologist 3** Dermatologist 4 * P < 0.05 comparison between consecutive days by 3rd dermatologist (Student’s t-test) ** P < 0.05 comparison between four dermatologists (ANOVA)
FIGURE 112.6 Comparison of the results obtained by the four dermatologists on consecutive days using the visual scoring method. 80 n = 23
n = 23
n = 23
Area (%)
60
40
20
0
CIAD
Visual scoring method
CIAT
Visual scoring method
CIAT
CIAD
FIGURE 112.7 Comparisons of results between the computer image analysis of tracing (CIAT), the computer image analysis with the digital camera (CIAD), and the visual scoring method.
The visual scoring method cannot accurately estimate the total involved surface area in patients with skin diseases, whereas the CIAD method shows accuracy similar to that obtained by the CIAT. In addition, the CIAD is much less time-consuming than CIAT. Therefore, the CIAD can be recommended for use as an objective and accurate method for measurement of surface area afflicted with skin disease.
112.5.3 ASSESSMENT OF INVOLVED AREA IN PSORIASIS (COMPUTER IMAGE ANALYSIS WITH ANALOG CAMERA) The areas of psoriatic lesions were measured by the visual scoring method and computerized image analysis with
analog camera, by dividing the surface area of the body into eight parts: the head and anterior and posterior parts of the trunk, upper limbs, and lower limbs. The percentages of the involved areas measured by the visual scoring method were significantly higher than those obtained from the computerized image analysis with analog camera on the trunk, upper and lower limbs, and total involved body area (Table 112.3). The measurement of the involved area in the case of psoriasis showed the necessity for an objective assessment that could overcome the difference between the observers, which was statistically significant in this study. On the other hand, the measurements of the head area were not statistically different, showing that it was difficult to observe the involved area present on the head and to accurately estimate these
964
Handbook of Non-Invasive Methods and the Skin, Second Edition
25 CIAT 20 CIAD
%
15
10
5
0
Trunk Arm ∗Compared by student’s paired t-test
Leg∗
FIGURE 112.8 Comparisons of estimates of involved area in each anatomic region between the computer image analysis of tracing (CIAT) and the computer image analysis with digital camera (CIAD).
TABLE 112.3 The Average Area Percentages of Psoriatic Lesions in Four Body Regions Visual Scoring (%) Head and neck Upper extremity Trunk Lower extremity Total corrected %b
1.00 8.25 11.33 11.58 9.78
± ± ± ± ±
0.89 3.99 5.95 4.78 3.39
Image Analysis with Analog Camera (%) 0.28 2.79 4.45 5.52 4.13
± ± ± ± ±
0.27 1.78 3.01 3.95 2.50
p Valuea >0.05