Measurement of body surface area in atopic dermatitis using specific PC software (ScoradCard)

In skin diseases, evaluation of involved surface area is a crucial factor in grading the degree of severity. We examined the reliability of body surface area assessment and relative inter-observer and intra-observer variability using new software (ScoraCard), specifically designed to evaluate automatically the extension of the involved area in the SCORAD index. Twenty pediatricians, untrained in the evaluation of skin disease, estimated the percentage of surface area involved in photo-tests of two children with artificial well-delimited lesions, at first by sight and then through software. As "gold standard" the exact amount of pixels was counted for the whole body surface of the children, for the different body zones and for the painted artificial lesions, expressed as percentage of the respective zone. For photo 1, gold standard was 38.06% and median percentage was 43.44% (95% CI 40.7-46.21) by sight (p = 0.002) and 37.99% (95% CI 36.04-39.94) by ScoradCard (p = 0.79). For photo 2, gold standard was 27.84%, median percentage was 30.44% (95% CI 28.25-32.63) by sight (p = 0.047) and 27.8% (95% CI 26.55-29.04) by ScoradCard (p = 0.79). The level of agreement (kappa statistic), cumulative for the two photo tests, was 0.38 (fair agreement) by sight method and 0.67 (good agreement) by ScoradCard. Among the 10 pediatricians who repeated the computer aided evaluation 3 months apart, the intra-observer variability was not significantly different: the median percentage was 31.5% (95% CI 27.0-49.4) at time 0 and 29.0% (95% CI 26.7-47.2) 3 months later (p = 0.76). This new software could be a useful tool in evaluating skin lesions extension, minimizing inter- and intra-observer variability, which is an important goal in multi-centre studies.