Pre-treatment T-cell Transcriptional Signatures Predict Immunotherapy Outcomes in Melanoma

Immune checkpoint inhibitors (ICIs) have improved outcomes for patients with melanoma and are now the standard of care for high-risk and advanced disease. However, long-term benefits are observed in only around 25% of patients, with significant risk for immune-related adverse events, highlighting the need for predictive biomarkers. To develop a minimally invasive, pre-treatment biomarker strategy, we profiled functional and subset-specific transcripts in peripheral blood T lymphocytes (PBTLs) and applied machine learning to identify predictive signatures. Patients were enrolled prior to receiving ICI monotherapy in the adjuvant (Exploratory n=61, Validation=78) or metastatic (Exploratory n=48, Validation=46) settings. Following feature selection, random forest models were trained and benchmarked against empirical null models. In the adjuvant setting, CD160 and GZMB predicted recurrence (95th percentile), while treatment-limiting toxicity was predicted by a signature comprising TNFRSF18, VTCN1, TIGIT, CCR4, and AHR (97th percentile). In the metastatic setting, baseline CD45RB, a marker of T-cell differentiation, most strongly predicted progression within one year (93.9th percentile). Distinct signatures in the adjuvant and metastatic settings suggest differences in T-cell programs associated with patient outcomes. These findings support further evaluation of pre-treatment circulating T-cell transcriptional profiles as predictors of ICI response and toxicity in melanoma.