The influence of clinical information in the histopathologic diagnosis of melanocytic skin neoplasms.

Abstract

BACKGROUND: We tested the relevance of clinical information in the histopathologic evaluation of melanocytic skin neoplasm (MSN). METHODS: Histopathologic specimens from 99 clinically atypical MSN were circulated among ten histopathologists; each case had clinical information available in a database with a five-step procedure (no information; age/sex/location; clinical diagnosis; clinical image; dermoscopic image); each step had a histopathologic diagnosis (D1 through D5); each diagnostic step had a level of diagnostic confidence (LDC) ranging from 1 (no diagnostic certainty) to 5 (absolute diagnostic certainty). The comparison of the LDC was employed with an analysis of variance (ANOVA) for repeated measures. FINDINGS: In D1 (no information), 36/99 cases (36.3%) had unanimous diagnosis; in D5 (full information available), 51/99 cases (51.5%) had unanimous diagnosis (p for difference between proportions <0.001). The observer agreement expressed as kappa increased significantly from D1 to D5. The mean LDC linearly increased for each observer from D1 through D5 (p for linear trend <0.001). On average, each histopathologist changed his initial diagnosis in 7 cases (range: 2-23). Most diagnostic changes were in D2 (age/sex/location). INTERPRETATION: The histopathologic criteria for the diagnosis of MSN can work as such, but the final histopathologic diagnosis is a clinically-aided interpretation. Clinical data sometimes reverse the initial histopathologic evaluation.