Diagnosing melanoma: how do we assess how good we are?

Abstract

Evaluating and improving diagnostic accuracy in identification of melanomas is important for both conservation of healthcare resources and reduction in patient morbidity. Useful indicators in assessing this accuracy include the number needed to treat (NNT) and the benign:malignant (B:M) ratio. Both of these methods lack sensitivity, as they do not account for the ability to detect early or in situ melanomas. To assess the NNT and B:M ratio for a busy hospital serving a population of 650,000 over a 5-year period, and to assess a new ratio of diagnostic accuracy by calculating the ratio of invasive (malignant) melanomas to melanoma in situ (MM:MMIS) as a marker of sensitivity. This was a retrospective analysis of data on all melanocytic lesions excised during two separate years (2006 and 2011) with a 5-year interval between them. The lesions were divided into benign naevi (BN), dysplastic naevi (DN), MMIS and MM. In 2006, 650 melanocytic lesions were excised (462 BN/DN, 45 MMIS, 143 MM). The NNT was 3.46, the B:M ratio was 2.46 and the MM:MMIS ratio was 3.18. In 2011, 730 melanocytic lesions were excised (464 BN/DN, 99 MMIS, 167 MM). The NNT was 2.74, the B:M ratio was 1.74 and the MM:MMIS ratio was 1.69. The NNT and B:M ratios from our study compare favourably with those in the published literature. The fall in the MM:MMIS and B:M ratios over this 5-year study appears to be an indicator of the ability to detect early disease and is probably secondary to the changes to our skin cancer service. This study may encourage physicians to aim not only for low B:M ratios but also low MM:MMIS ratios.