Portal de Periódicos da CAPES

Three‐dimensional imaging of vitiligo

2015; Wiley; Volume: 24; Issue: 11 Linguagem: Inglês

10.1111/exd.12791

1600-0625

Indermeet Kohli, Prescilia Isedeh, Mohammed Aljamal, Diego Ruiz Dasilva, Amanda Batson, Douglas Canfield, Nikiforos Kollias, Henry W. Lim, Iltefat Hamzavi,

Atherosclerosis and Cardiovascular Diseases

Vitiligo is a skin disorder that results from the loss of pigmentation (S8) due to the destruction of melanocytes. It affects 1–2% of the world population and affects all skin types 1. It can be a very psychologically distressing disease that can negatively affect the subject's quality of life 2. Unfortunately, although there have been many qualitative and semiquantitative outcome measures 2-7 (brief discussion in e-supplement), there is no standardized outcome method to determine the treatment efficacy for vitiligo. In this letter, we describe this emerging technique of taking whole-body 3D images followed by 3D image analysis to quantify the vitiligo-affected body surface area (BSA) of the subjects. Other research groups 8, 9 (S9–S11) have validated the accuracy and utility of 3D imaging and analysis for a range of clinical applications, and we have extended this utility to evaluate vitiligo. The objective of this work was to accurately quantify the vitiligo-affected BSA by analysing 3D images of vitiligo subjects and evaluating data for reliability measures, where reliability refers to consistent repeated measurements. Further comparisons were made to compare lesion area calculated by 2D and 3D image analyses. The use of high-resolution Canfield Vectra Whole Body 360 (WB360) allowed for taking of 3D images of the entire body of the vitiligo subject at once. The physical dimensions of the Canfield Vectra WB 360 system (Fig. 1a), comprised of 46 cameras, are 2.8 m × 2.4 m × 2.4 m, for the x, y and z dimension, respectively. Subjects were positioned at the centre of the imaging system. Whole-body imaging was performed for six subjects, skin type 1-VI, with generalized vitiligo diagnosed by a dermatologist. Subject consent was obtained using commercial photographic release prior to imaging at Canfield Scientific Inc. located in Fairfield, New Jersey USA. 3D image acquisition was followed by pixel-based image analysis using Vectra 3D VAM (Visualization, Analysis, Measurement) application to accurately measure the area of each vitiligo lesion in square cm as well as total body surface area (BSA) (Fig. 1b). The lesion area was selected manually using a multipoint polygon region-selection tool in VAM designed for precise planimetry applications. With alternate image processing, using blue channel isolation, we were able to clearly visualize the contrast between the vitiligo-affected skin and normal skin (Fig. 1b). This processing was especially helpful for identifying the subtle lesion borders in lighter skin subjects. The percentage BSA affected by vitiligo was thus calculated. Time required for 3D image acquisition and for analysing total vitiligo-affected BSA was evaluated for each subject. Data reliability was analysed by accounting for inter-rater and intra-rater measurement reliability using intraclass correlation coefficient (ICC) with a one-way random-effects model for absolute agreement. All analyses were carried out using SAS 9.4 (SAS Institute Inc., Cary, North Carolina, USA). The inter-rater reliability was assessed by having three raters of different skill levels, who independently measure the percentage vitiligo-affected BSA. The intra-rater reliability was assessed by having all three raters that perform three repeated measures of the percentage vitiligo-affected BSA (Fig. 1c). To compare lesion area calculated by 2D and 3D image analyses, a circle of diameter 5 cm was carefully applied to the curved forearm of a subject to simulate area coverage of a generic lesion using a known surface area reference. 3D as well as 2D images of this generic lesion were taken, and corresponding area was calculated using Vectra 3D VAM (Canfield Scientific Inc., Fairfield, New Jersey, USA) application. 3D image acquisition by Canfield Vectra WB 360 system took <3 ms per subject, and download time was 0.95. For the known surface area reference, 2D image analysis resulted in lesion area of 14.98 cm2, whereas 3D image analysis resulted in a lesion area of 19.50 cm2. Comparing the calculated areas to the actual area of the circle, 19.63 cm2, it can be concluded that 2D image analysis of contoured surfaces can underestimate lesion area by more than 20%, consistent with that reported earlier 7, whereas 3D image analysis accounts for the changes in relative contour and has an associated error of <1%. The exact estimation of the area of a contoured surface, typical surface of human skin, thus requires the use of 3D imaging because it provides a very accurate, objective measurement which is time-efficient and reproducible. One limitation of this new system is the manual selection of the lesion boundaries. Future work will involve the development of analysing software that can assist in fully automated selection of the lesion area. Another limitation is the equipment cost as well as the space needed. In conclusion, to the authors’ knowledge, this is the first study of its kind to use 3D imaging to assess vitiligo. The high ICCs indicate the high reliability and thus rater-independent nature of measurements associated with this imaging system. It provides an objective element to account for the amount of de/re-pigmentation along with ease to use and time efficiency. The associated low uncertainty indicates the accuracy of 3D image analysis in determining vitiligo-affected BSA. This 3D imaging analysis will be helpful in precisely documenting a lesion in terms of its location, extent and pattern. This in turn will aid in better evaluation of response to therapy, worsening of the subject's vitiligo or stabilization of the disease and would allow the vitiligo researcher to precisely quantify the changes in pigmentation without the variability inherent in any subjective measure. Indermeet Kohli and Prescilia Isedeh wrote the manuscript. Indermeet Kohli, Dennis DaSilva, Amanda Batson, Douglas Canfield, Nikiforos Kollias, Henry W. Lim and Iltefat Hamzavi made substantial contributions to research design. Prescilia Isedeh, Mohammed Al-Jamal and Dennis DaSilva performed major data acquisition. Indermeet kohli and Dennis DaSilva made significant contribution to data analysis and interpretation. All the authors contributed to the critical revision and approval of the final version of the manuscript. None. Indermeet Kohli is a subinvestigator for Estee Lauder and Ferndale Laboratories. Prescilia Isedeh is a subinvestigator for Estee Lauder, Ferndale Laboratories and Clinuvel. Mohammed Al-Jamal is a subinvestigator for Ferndale Laboratories. Dennis DaSilva, Douglas Canfield and Amanda Batson are full-time employees of Canfield. Scientific Inc. Nikiforos Kollias is a consultant for Canfield and Johnson & Johnson. Henry W. Lim is a consultant for Ferndale Laboratories, Uriage, Sanofi and Johnson & Johnson and has received grant support from Estee Lauder, Ferndale Laboratories and Clinuvel. Iltefat Hamzavi is an investigator for Estee Lauder, Ferndale Laboratories and Clinuvel. Data S1. Discussion of vitiligo assessment methods. Data S2. References. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.