Quantifying Cutaneous Dermatomyositis: A Novel 3D Image-based Approach.

Abstract

OBJECTIVE Visual examination of skin lesions has considerable subjectivity and inter-rater variability. This study assessed the feasibility of a 3D image-based assessment of cutaneous disease activity in dermatomyositis (DM). METHODS DM patients were evaluated in a prospective study at 2 time points for skin rash assessment using the Cutaneous Dermatomyositis Disease Area and Severity Index (CDASI) and 3D images. A "3D image disease activity score (3DAS)" was calculated based on the percentage of the rashes relative to the total body surface area, multiplied by the degree of rash redness. The construct validity and responsiveness of 3DAS were evaluated using the Spearman correlation coefficient (rsp) against standard CDASI and patient-reported outcome measures (PROMs). A generalized linear regression model assessed the relationship between the 3D image-derived rash area and redness with the CDASI score. RESULTS 27 DM patients (81.5% female, 96.3% White; median age 50.0 years) were enrolled. The median (IQR) CDASI score at baseline was 6.0 (0.0 - 17.0). For the construct validity, 3DAS correlated strongly with the CDASI (rsp = 0.83, p < 0.001) and PROMs. The generalized linear regression analysis identified the rash area and redness from 3D images as significant predictors of the CDASI score. Regarding responsiveness, absolute changes from baseline in the 3DAS correlated strongly with the CDASI score (rsp = 0.61, p = 0.004). CONCLUSION Our results demonstrate favorable validity and responsiveness of the 3D images for evaluating rash in DM patients. The 3D image-derived rash area and redness are significant predictors of CDASI scores.