Evaluation of skin erythema severity by dermatoscopy in dogs with atopic dermatitis

Background: To estimate the extent and severity of atopic dermatitis (AD)-related skin lesions, clinical trials enrolling dogs with AD often use categorical scales such as the Canine Atopic Dermatitis Extent and Severity Index, 4^th iteration (CADESI-04) and Canine Atopic Dermatitis Lesion Index (CADLI). Despite recent progress in the standardization of these AD-grading scales, the evaluation of the severity of skin lesions (including erythema) remains subjective. Objectives: To validate an optical set-up with a smartphone and a dermatoscope for the objective estimation of skin erythema severity in atopic dogs. Animals: Forty-three dogs with AD. Methods and materials: An erythema index (EI) was calculated from calibrated skin images and compared to the dermatologist’s erythema severity estimate using the erythema grading scale used in the CADESI-04, as well as an ad hoc Visual Analog Scale (VAS) with a continuous palette of red shades. Results: We found a strong correlation based on the Spearman rank correlation coefficient between all erythema valuations: CADESI-04 and VAS: 0.93 [95% CI: (0.85, 0.96)]; CADESI-04 and EI: 0.85 (0.72, 0.92); VAS and EI: 0.82 (0.67, 0.91). There was a good agreement between the objective EI and CADESI-04-based estimates because 71% of samples were classified in the same erythema severity category. When comparing the EI and the VAS, the standard deviation of misestimates was 12% (maximum 100%). Conclusions and clinical relevance: The proposed optical set-up has the potential to make erythema severity estimation objective, thus leading to more reliable AD severity scales for the use in experimental canine AD models or in clinical trials enrolling atopic dogs.