Machine learning‑based programmed cell death‑related index to predict prognosis and immunotherapy response in skin cutaneous melanoma

Skin cutaneous melanoma (SKCM) is a highly aggressive malignancy with heterogeneous outcomes and a variable response to immunotherapy. Programmed cell death (PCD) serves a key role in tumor progression and immune regulation, but its prognostic and therapeutic relevance in SKCM remains to be elucidated. An integrative machine learning strategy encompassing 77 algorithm combinations was employed to construct a PCD-related index (PCDI). Associations between the PCDI and immune microenvironment, immunotherapy response and drug sensitivity were also investigated. Subsequently, a 13-gene risk signature was identified via the Lasso algorithm and used to calculate a PCDI-based risk score. Functional enrichment and experimental validation were conducted to explore potential biological functions of stratifin (SFN), which had the highest positive coefficient in the formula used to calculate the PCDI-based risk score. The PCDI robustly stratified patients into high- and low-risk groups with markedly different overall survival across The Cancer Genome Atlas and Gene Expression Omnibus cohorts. The risk score outperformed traditional clinical parameters in predicting prognosis and served as an independent prognostic factor. Low-risk patients exhibited higher immune cell infiltration, immune checkpoint expression and tumor immunogenicity, as well as lower tumor immune dysfunction and exclusion, and immune escape scores. The model predicted improved immunotherapy responses in low-risk groups, which was further validated in three independent immunotherapy cohorts. By contrast, high-risk patients were more sensitive to chemotherapeutic and targeted agents. Using the Cell Counting Kit-8 assay, SFN, one of the key genes in the signature, was experimentally validated as an oncogenic driver in SKCM. The present study developed and validated a machine learning-based PCDI that effectively predicts prognosis and immunotherapy response in SKCM. This PCDI provides novel insights into PCD-mediated tumor-immune interactions and demonstrates potential for personalized therapeutic decision-making in melanoma.