Can We Diagnose Early Melanoma by Genomics?

Melanoma incidence has increased approximately 6-fold over the past 40 years, driven primarily by melanoma in situ and thin invasive melanomas, which are among the most subjective lesions to diagnose on histopathology. This review examines whether genomics clarifies the "gray zone" between benign and malignant pigmented lesions, or instead recasts existing histopathologic ambiguity in molecular terms, while also considering whether genomic classification shifts overdiagnosis upstream instead of resolving underlying diagnostic uncertainty. Across cytogenetic, mutation-based, and gene-expression platforms, current assays perform best in unequivocal melanoma-versus-nevus comparisons and lose accuracy in diagnostically ambiguous lesions where clarification is most needed. Gene expression profile (GEP) tests can be useful diagnostic aids in tissue-rich equivocal lesions, including atypical Spitz tumors, MELTUMPs, and cellular blue nevus/blue nevus-like melanocytoma differentials. One clinicopathologic-genomic model, the Merlin CP-GEP test, has been prospectively validated for predicting sentinel lymph node status and is now recognized as supporting sentinel lymph node biopsy decisions in T1b/T2a melanomas. Importantly, a low-risk prognostic GEP class assignment should not displace the standard sentinel lymph node biopsy discussion. GEP tests are least useful for small, thin, equivocal junctional proliferations, precisely the lesions that are most difficult to diagnose. Pre-biopsy non-invasive assays carry the additional risk of lowering, rather than raising, the biopsy threshold. Overall, current genomic tools should be viewed as an adjunct to, rather than a replacement for, histopathology and clinical context.