How good are US dermatologists at discriminating skin cancers? A number-needed-to-treat analysis

Abstract

Background: Identification of skin cancer requires discrimination of malignant lesions from benign lesions. The number of biopsies performed to yield one cancer diagnosis can be presented as a number needed to treat (NNT), and provides an assessment of the efficiency of skin cancer detection. Objective: To assess the clinical accuracy of US dermatologists screening for skin cancer, the NNT for both melanoma and non-melanoma skin cancer was examined. Methods: Pathology reports from 2021 biopsies performed at the Wake Forest University Department of Dermatology were reviewed, including the physician's differential diagnosis and final pathological diagnosis. The NNT was calculated for melanoma, non-melanoma skin cancer, and all skin cancer diagnosed. Results: Of 1240 biopsies suspicious for skin cancer, 559 cancers were diagnosed, yielding a NNT of 2.22 for any cancer. The NNT specifically for non-melanoma skin cancer was 1.6, while the NNT for melanoma was 15. Patient age, anatomical location, sex and physician all significantly impacted on NNT values. Conclusions: The NNT for melanoma in our study was lower compared to recently published values obtained from general practitioners in Australian skin cancer clinics (NNT of 30). Variability amongst institutions, practice settings and physicians supports the need to establish a benchmark NNT.