Aims and background: Dermoscopy (dermatoscopy, skin surface microscopy, epiluminescence microscopy) has been increasingly employed in recent years for the preoperative detection of cutaneous melanoma, and dermatoscopic features of pigmented skin lesions have been previously defined using histopathology (HP) as the "key to the code". The aim of the present study was to evaluate the interobserver agreement on the HP diagnosis in a series of epiluminescence microscopy equivocal melanocytic skin lesions.
Study design: Ten melanocytic skin lesions were selected on the basis of diagnostic disagreement of at least 2 out of 9 epiluminescence microscopy observers. The histologic specimens from the 10 lesions were examined by 9 HP observers. The agreement of the HP diagnoses was calculated by means of Fleiss' k statistics.
Results: The overall HP agreement was less than excellent (k = 0.5). When considering the prevailing epiluminescence microscopic and HP diagnoses, 2 cases were shown to be epiluminescence microscopy false-negative melanomas. Virtually no agreement was found among epiluminescence microscopy observers in 4 cases (40%) or among HP observers in 3 cases (30%). However, only one pigmented skin lesion remained unclassifiable on epiluminescence microscopy as well as HP.
Conclusions: When at least 2 epiluminescence microscopy experts disagree in the evaluation of a given melanocytic skin lesion, even HP consultations may give equivocal results. The need to establish more reliable epiluminescence microscopic and HP criteria by performing an improved and meticulous clinicopathologic correlation, e.g. by using telecommunication via Internet, is emphasized.