Skip to content

Shain AH, et al. The Genetic Evolution of Melanoma from Precursor Lesions. The New England journal of medicine. 2015;373:1926-36. PMID 26559571

One-paragraph summary

The pathogenic mutations of melanoma had been catalogued but their order had not. The authors sequenced 293 cancer-relevant genes across 150 microdissected areas from 37 primary melanomas and their histologically adjacent precursor lesions, spanning unequivocally benign lesions, intermediate lesions, and intraepidermal or invasive melanoma. The result is an ordered account of melanoma genesis in which specific alterations mark specific histological stages, and in which the point-mutation burden rises monotonically from benign through intermediate to invasive disease with a strong ultraviolet signature detectable at every stage.

Key findings

  • Unequivocally benign lesions harboured BRAF V600E mutations exclusively.
  • Intermediate lesions were enriched for NRAS mutations and additional driver mutations; more than one pathogenic alteration defines the intermediate category.
  • 77% of intermediate lesions and melanomas in situ harboured TERT promoter mutations — selected at an unexpectedly early stage of neoplastic progression.
  • Biallelic CDKN2A inactivation emerged exclusively in invasive melanoma; PTEN and TP53 mutations only in advanced primary melanoma.
  • Copy-number alterations became prevalent only at invasion; tumour heterogeneity emerged as genetically distinct subpopulations as melanomas progressed.
  • The companion analysis of 230 areas from 82 patients ordered the pathway disruptions — MAPK activation, telomerase up-regulation, chromatin modulation, G1/S override, MAPK ramp-up, p53 disruption, PI3K activation — and found no mutation specifically associated with metastatic progression (Shain 2018, PMID 29990500).

Limitations

  • Targeted panel of 293 genes rather than whole genomes, so structural variation and non-coding events outside the panel were not assessed.
  • Precursor lesions were identified as histologically adjacent to a melanoma, so the sample is enriched for precursors that did progress; it cannot estimate the probability that a given benign or intermediate lesion progresses.
  • Histological classification of "intermediate" lesions is itself poorly reproducible — the diagnostic class with 25–40% accuracy against expert consensus (Elmore 2017, PMID 28659278) — so the stage assignments inherit that uncertainty.
  • Small number of patients (37 in the primary analysis).

Why it matters

This paper supplies the molecular basis for the two hardest problems in melanoma. First, it explains why diagnosis is irreproducible in the middle of the spectrum: the genetic events that pathologists are asked to distinguish are shared across the benign–intermediate–malignant boundary, with BRAF V600E present in unequivocally benign lesions and TERT promoter mutation present in 77% of intermediate lesions. Second, it therefore underwrites the overdiagnosis argument at a mechanistic level — a lesion carrying melanoma's early driver mutations is not thereby a lesion that would have killed anyone (see screening and overdiagnosis). The WHO responded by proposing melanocytoma for intermediate tumours with increased but still low progression probability (Elder 2020, PMID 32057276). No marker separating indolent from progressive early melanocytic lesions has emerged in the decade since, and its absence remains the single most consequential gap in the field.

Cited by wiki pages

  • molecular subtypes and genomics
  • screening and overdiagnosis
  • histopathology and prognostic factors
  • uveal melanoma
  • overview