Reliability of skin biopsy pathology.

Abstract

The diagnosis of skin disease by histologic examination is regarded as the reference standard upon which therapy and follow-up are determined. Our study investigated the reliability of skin biopsy diagnosis requested by family physicians and physicians' assistants. Biopsy diagnoses by a community-based pathology group were reinterpreted by our study dermatopathologist on a sample of 119 skin biopsies randomly selected from the 1844 biopsies performed by family physicians and physicians' assistants at a large Washington State health maintenance organization during a 4 1/2-year period. There were 107 exact matches and 3 mismatches of premalignant lesions and 6 mismatches of benign diagnoses. In addition, two melanomas diagnosed by the community-based pathologists were interpreted as benign by our study dermatopathologist. A third melanoma diagnosed by the community-based group was interpreted as a poorly differentiated squamous cell cancer by the university dermatopathologist. The weighted kappa, 0.83, indicated excellent interrater agreement. Although our study showed excellent interrater concordance of skin biopsy interpretation, there was disagreement about three melanomas between a community-based general pathology group and our study dermatopathologist. The melanoma disagreement is consistent with previous studies that found poor interrater agreement for early melanomas. The community-based pathologists were uncertain about two of these melanomas, and as part of their quality control and review procedures requested confirmation by an expert pathologist, who agreed with the melanoma diagnosis. Family physicians are justified in requesting a second opinion (if not automatically requested by a community laboratory) when the histopathologic diagnosis is not in concordance with the clinical history or impression or when the pathologist is unsure of the diagnosis.