Comparison of diagnostic accuracy for cutaneous malignant melanoma between general dermatology, plastic surgery and...

Since the 1980s there have been dedicated pigmented lesion clinics (PLCs) in the U.K. Important considerations when comparing the efficacy of the PLC with other referral clinics include diagnostic accuracy.

To compare the false-negative rate of clinical diagnosis (FNR) in the PLC with that in the other clinics of primary referral of malignant melanoma (MM) in the same geographical area. We have previously shown that certain clinical features are risk factors for diagnostic failure of MM. A further aim of this study was to correct for any differences in frequency of these factors in the melanoma populations between clinics and to estimate the false-positive diagnostic rate (FPR) in the PLC.

To compare the FNR between clinics, the case notes of all patients presenting with histologically proven cutaneous MM in Leicestershire between 1987 and 1997 were examined retrospectively. A false-negative diagnosis was defined as documentation of another diagnosis and/or evidence in the case notes that the diagnosis was not considered to be MM. The FNR was estimated as the number of false-negative clinical diagnoses/number of true-positive histological diagnoses. To estimate the diagnostic FPR, which was defined as the number of false-positive clinical diagnoses of MM/total number of positive clinical diagnoses, in the PLC, the outcome of 500 consecutive patients attending the PLC was surveyed.

The case notes of 731 patients were available, of whom approximately two-thirds initially attended the PLC, one-fifth the General Dermatology clinics (D) and the remainder were divided approximately equally (one-twentieth each) between Plastic Surgery clinics (P), other clinics (O) and the surgery of the general practitioner (GP). The last was regarded as the primary referral clinic if the lesion were excised there prior to any referral. The FNR was lowest for the PLC, at 10%, compared with 29% (D), 19% (P), 55% (O) and 54% (GP) (P < 0.0001). Lesions with risk factors for diagnostic failure were under-represented in the PLC (P < 0.0001), the mean frequencies of the risk factors being 20% (PLC), 25% (D), 22% (P), 31% (O) and 30% (GP). Differences were not large but still could partially explain the lower FNR of the PLC. However, when the FNR was estimated for lesions exhibiting each of these risk factors, the PLC was found to have the lowest rate in every case (PLC vs. all clinics combined, P = 0.04 to P < 0.0001). The mean FNR for the risk factors combined was 18% (PLC), 45% (D), 50% (P), 68% (O) and 71% (GP). Also on logistic multivariable analysis of the PLC vs. all the other clinics on FNR and the above factors, the higher FNR of the other clinics retained significance (odds ratio 5.9, P < 0.0001). In the 500 patients surveyed separately in the PLC, the MM pick-up rate on biopsy was 32% and the diagnostic FPR was 41%.

The FNR of MM was lower in the PLC than in the other clinics, while the pick-up rate for MM on biopsy and the FPR were acceptably low.