Integrating static and dynamic features of melanoma: The DynaMel algorithm

BackgroundSequential digital dermatoscopy identifies dynamic changes in melanocytic lesions. However, no algorithm exists that systematically weights dynamic changes regarding their association with melanoma.ObjectiveWe sought to identify relevant dynamic changes and to integrate these into a novel diagnostic algorithm.MethodsDuring follow-up (mean 44.28 months) of 688 patients at high risk, 675 pigmented lesions with prospectively documented dynamic changes were excised. The association between specific changes and melanoma was assessed.ResultsWe detected 61 melanomas (38 invasive, median thickness 0.42 mm) with dynamic changes. Multivariate logistic regression analyses revealed a significant association between the diagnosis of melanoma and 5 dynamic criteria. According to the observed odds ratios we defined two dynamic major criteria (2 points each: asymmetric-multifocal enlargement and architectural change) and 3 dynamic minor criteria (1 point each: focal increase in pigmentation, focal decrease in pigmentation, and overall decrease in pigmentation when not accompanied by a lighter pigmentation of the adjacent skin). The DynaMel score was generated by addition of dynamic and 7-point checklist scores with a threshold for excision of 3 or more points. Including information about dynamic changes increased the sensitivity of the 7-point checklist from 47.5% (29 of 61 melanomas detected) to 77.1% (47 of 61 melanomas detected). The specificity slightly decreased from 99.0% to 98.1%.LimitationsBefore broad application the DynaMel algorithm needs to be validated using data from a different prospective study.ConclusionsThe DynaMel algorithm integrates a scoring system for dynamic dermatoscopic changes into the 7-point checklist for dermatoscopy and thereby increased the sensitivity of melanoma detection.