The melanocytic state carried an unexpected survival penalty.
Digital spatial RNA profiling of in-transit metastatic melanoma from 105 patients resolved axes corresponding to differentiation, immune infiltration, stromal/neural-crest-like features, and a melanocytic program. In treatment-naive disease, melanocytic-state-high tumors had median melanoma-specific survival of 5.16 years versus 12.88 years for melanocytic-state-low tumors, a 7.72-year difference. The association persisted in multivariable analysis and external datasets; acral melanomas showed higher melanocytic-state expression.
A biologically defined cell state—not only an immune score—may identify a prognostic and therapeutically distinct metastatic subtype.
The retrospective 1990–2020 cohort spans major diagnostic and treatment eras, and only 30% received checkpoint inhibitors.
Is the poor-prognosis signal cell-autonomous, subtype-linked, or a spatial proxy for a permissive niche?
