Interobserver agreement of the dermoscopic diagnosis of 129 small melanocytic skin lesions

Abstract

Aim and background: Dermoscopic diagnosis of pigmented skin lesions is based on the evaluation of dermoscopic criteria (classical pattern analysis) and on alternative diagnostic methods, such as the ABCD (A, asymmetry; B, border; C, color; D, differential structures) rule based on the total dermatoscopic score. The aim of the study was to investigate the interobserver agreement of standard dermoscopic criteria between two observers and the diagnostic validity of dermoscopic diagnosis by pattern analysis and by the ABCD rule.

Study design: The study included a total of 129 small (< or = 5 mm) melanocytic skin lesions selected from all lesions observed in consecutive patients between April 1996 and September 1998. Before surgery, each lesion was photographed with a Dermaphot. Dermoscopic images were examined independently by two observers to evaluate the presence or absence of standard dermoscopic criteria and to establish the dermoscopic diagnosis by pattern analysis and by the ABCD rule.

Results: Interobserver agreement for dermoscopic criteria varied from moderately good to good, with the highest agreement for radial streaks (k = 0.96) and the lowest for pseudopods (k = 0.49). Interobserver agreement was moderately good in dermoscopic diagnosis by pattern analysis (k = 0.48) and by the total dermatoscopic score (k = 0.44). The sensitivity and specificity of dermoscopic diagnosis by pattern analysis were 40% and 99%, respectively, for both observers. As regards the total dermatoscopic score (a cutoff score of < or = 5.45 vs > 5.45), sensitivity ranged from 80% to 100% and specificity from 48% to 59%.

Conclusions: The study showed that the pattern analyses as well as the ABCD rule give a poor discrimination between benign and malignant lesions and do not add relevant information for management decision in small melanocytic lesions. However, close follow-up examinations of small equivocal melanocytic lesions using digital equipment allow evaluation of their dermoscopic features during progression and whether their rather commonly found atypical dermoscopic features are lost during their natural course of growth.