Lentigo maligna of the face: A quantitative simple method to identify individual patient risk probability on dermoscopy.

ABSTRACT

BACKGROUND/OBJECTIVES: The clinical and dermoscopic differential diagnosis of flat pigmented facial lesions represents a great challenge for the clinicians. Our aim was to report a quantitative method based on dermoscopic features to better classify pigmented facial lesions. METHODS: This is a retrospective case-series study that analysed the dermoscopic features of 582 pigmented facial lesions. RESULTS: The individual patient probability of lentigo maligna (LM) was predicted by a multivariate model, with an accuracy of 0.72. According to the odds ratio at the multivariate analysis, an individual scoring index was assigned to each criterion, and a value of 4.56 was identified as optimal cut-off point. Up to a score of 2.5, the probability that a lesion is an LM is 0. The probability increases from 10 to 50% for a score ranging between 4.5 and 6. It is about 90% for a score of 7. CONCLUSION: The optimal cut-off point obtained and the curve that identifies the probability of a patient having a LM could improve the classification and the management strategies of equivocal pigmented facial lesions.