Unraveling the Plasma Extracellular Vesicles Proteome: Predictors of Resistance in Immunotherapy of Advanced Melanoma.

Despite substantial clinical benefit from immune checkpoint inhibitors (ICI), advanced melanoma remains challenging due to frequent treatment resistance. Resistance may be intrinsic (primary) or emerge over time (secondary). Biomarkers predicting distinct resistance phenotypes before therapy are lacking. As key mediators of cellular communication, extracellular vesicles (EVs) represent promising biomarkers. This study aimed to identify baseline EV proteome-derived pathways and biomarkers associated with overall, primary, and secondary resistance to ICI in advanced melanoma and to derive biomarker signatures predictive of progression-free survival (PFS). EVs were isolated from pretreatment plasma samples of 46 patients with advanced melanoma using size exclusion chromatography and ultracentrifugation. Proteomic profiling was performed by liquid chromatography-mass spectrometry using DIA-NN. Pathway enrichment and network analyses were conducted using Reactome, Metascape, Cytoscape, and DAVID. Resistance-associated proteins were integrated into composite biomarker signatures and evaluated for association with PFS. Overall resistance was characterized by enrichment of platelet- and complement-associated pathways. Primary resistance was associated with enhanced Fc gamma receptor (FCGR) signaling and downregulation of KSRP-associated post-transcriptional regulatory processes. In contrast, secondary resistance was preceded by distinct baseline EV proteomic patterns involving complement activation and reduced hemostasis- and platelet-related pathways. EV-derived biomarker signatures for overall, primary, and secondary resistance independently discriminated patients according to PFS. Baseline plasma EV proteomics reveals distinct systemic biological programs associated with different resistance phenotypes to ICI in advanced melanoma. EV-derived biomarker signatures enable stratification by PFS and warrant validation in larger, multicentric cohorts.