Proteomics and clinical data integration for the identification of plasma protein biomarkers in metastatic castration-resistant prostate cancer

Metastatic castration-resistant prostate cancer remains an incurable disease despite the availability of novel therapeutic options and is associated with limited survival. The lack of reliable prognostic and predictive biomarkers continues to limit the implementation of personalized treatment strategies. Currently, prostate-specific antigen is the main circulating biomarker used in clinical practice; however, its ability to accurately monitor disease progression and predict outcomes is limited. This study aimed to identify circulating protein biomarkers with prognostic and predictive value to improve patient stratification and support therapeutic decision-making. Plasma samples from 48 independent patients with mCRPC were analyzed by liquid chromatography–tandem mass spectrometry. Associations between protein abundance and clinical variables were evaluated using MaAsLin2. Candidate selection for circulating tumor cell (CTC) and androgen receptor splice variant 7 (AR-V7) status was further refined using sparse partial least squares discriminant analysis (sPLS-DA) with repeated stratified five-fold cross-validation, balanced error rate optimization, and feature-selection stability assessment. Discriminative performance was evaluated by receiver operating characteristic (ROC) analysis with internal bootstrap validation. Selected proteins were also compared with baseline PSA, and survival associations were assessed using Kaplan–Meier analysis and univariable and multivariable Cox proportional hazards models adjusted for treatment line, baseline PSA, and age. Nine candidate plasma proteins were identified. HP, SERPINA3, LRG1, and ITIH3 were associated with CTC status, whereas ENO1, HSP90AA1, S100A9, SPP1, and VCL were associated with AR-V7 status. Several proteins exhibited high discrimination of the corresponding clinical phenotype, and endpoint-specific multiprotein models demonstrated strong discriminatory performance and were significantly associated with overall survival. Direct comparison with baseline PSA suggested a complementary contribution of the proteomic panels, particularly for CTC status. Survival analyses further supported the prognostic relevance of the identified proteins. This study identified nine circulating proteins with clinical relevance in metastatic castration-resistant prostate cancer and showed that their combination into compact endpoint-specific panels may provide novel and important information to clinical practice. These biomarkers may improve patient stratification, facilitate treatment selection, and enhance disease monitoring through a minimally invasive liquid biopsy approach.