Melanoma detection rate and concordance between self-skin examination and clinical evaluation in patients attending a pigmented lesion clinic in Italy

Background The early diagnosis of melanoma is based on the collaboration between dermatologists and family doctors, who filter subjects to be referred to a pigmented lesion clinic (PLC). Following growing media coverage, there is increasing concern in the general population about the risk of the 'changing mole', resulting in a progressively increased workload in PLCs. Aim and methods We investigated the causes of referral to a PLC in a series of 193 attendees seen consecutively at the PLC of the University of Florence. Because the number of naevi is the major risk factor for melanoma in Mediterranean populations, the concordance between self-counting of naevi and the clinical evaluation of a PLC dermatologist in order to classify high-risk individuals was also investigated. Results Detection of a clinically suspicious lesion at dermatological examination occurred in 13 of 193 subjects referred by general practitioners (6.7%), with three melanomas confirmed histologically (overall detection rate: three of 193, 1.6%). The positive predictive value of the 'presence of a suspicious lesion', the cause of referral in 39.9% of subjects, was 9.1% when based on the gold standard criterion represented by the clinical detection of a suspicious lesion by the dermatologist and 3.8% based on the histological diagnosis of melanoma; the negative predictive value was 94.8% (100% when based on the histological diagnosis of melanoma), suggesting that the clinical detection of a suspicious lesion in subjects with different causes of referral (such as risk factors for melanoma, or the need to be reassured about moles) is unlikely. There was poor agreement between self-evaluation based on the presence of multiple naevi and the dermatological examination (gold standard) for both common and atypical naevi. The highest concordance (kappa = 0.32, 95% confidence interval 0.20-0.43) was associated with a dichotomized count of naevi as up to 50 or more than 50 naevi. Conclusions In order to reduce the PLC workload, the filtering role of the family doctor needs to be improved, so that only subjects with a specific suspicious lesion are referred to the PLC. The self-assessment of melanoma risk based on the presence of multiple naevi was not reliable.