Predictive value of biopsy specimens suspicious for melanoma: Support for 6-mm criterion in the ABCD rule

Abstract

Clinical detection of melanoma can be challenging. The number of biopsy specimens performed to diagnose 1 melanoma is a measure of efficiency of skin cancer detection, but few data are available to describe this measure from US health care. We studied the diagnosis of melanoma among biopsy specimens of clinically concerning pigmented lesions at an academic dermatology department. We searched for all biopsy specimens that were performed because of clinical suspicion of melanoma in 2013. Characteristics of the patient, lesion, and clinician performing the biopsy, and the final pathology diagnosis were recorded. A total of 2643 biopsy specimens from 2213 patients submitted by 43 providers were included. Melanoma was diagnosed in 165 cases (positive predictive value 6.4%, 95% confidence interval 5.5%-7.4%). Older age (P < .001), male gender (P = .045), and nontrunk location (P < .001) were predictors of higher probability of melanoma detection. Lesions larger than 6 mm in size had higher positive predictive value 11.5% (8.8%-14.1%) than smaller lesions 2.6% (1.6%-3.6%). Factors influencing the decision to biopsy a lesion may be difficult to evaluate retrospectively. At an academic medical center, 16 clinically concerning lesions were biopsied to diagnose 1 melanoma. Biopsy specimens of clinically concerning pigmented lesions larger than 6 mm on older men had the highest yield.