Pre-surgical immune-proteomic profiles in serum identify inflamed and suppressed immune states associated with biochemical recurrence and PSA persistence in prostate cancer

Abstract Background Prostate cancer progression is shaped by complex tumor-immune interactions and inflammatory processes. However, immune-related protein biomarkers associated with prognosis after radical prostatectomy remain limited. This study aimed to characterize pre-surgical serum immune-proteomic profiles associated with biochemical recurrence (Recurrence; prostate-specific antigen [PSA] ≥ 0.2 ng/mL) and PSA persistence (Persistence; PSA ≥ 0.1 ng/mL) after surgery, targeting proteins within immuno-oncology and inflammation. Method Pre-surgical serum from two independent prostate cancer cohorts (Trondheim, n = 136; Oslo, n = 87) was analyzed using the Olink Inflammation and Immuno-Oncology panels. The 34 proteins overlapping between panels were analyzed in the combined cohort. Group comparisons used one-way ANOVA with false discovery rate adjustment and multivariate modeling used partial least squares discriminant analysis with fivefold cross-validation and recursive feature elimination. Prognostic associations were evaluated using Cox proportional hazards for recurrence and logistic regression for PSA persistence. Result IL18 was consistently lower in patients with recurrence and PSA persistence compared to non-recurrent patients, suggesting reduced antitumor immune activity before surgery. PSA persistence group was characterized by lower TNFSF12, CCL19, and CD244 compared to non-recurrent patients. We observed a broader trend toward reduced protein levels across the panel in patients with PSA persistence in both cohorts, consistent with a more immune-deserted profile. The protein-based model discriminated recurrence and PSA persistence more accurately than the clinical-parameter model, which included age, PSA before surgery, biopsy Gleason Grade Group, clinical T stage, and Prostate Imaging-Reporting and Data System (PI-RADS) score (Trondheim: Recurrence AUC 0.73 vs. 0.64; Persistence AUC 0.83 vs. 0.81; Oslo: Recurrence AUC 0.77 vs. 0.54; Persistence AUC 0.89 vs. 0.63). Combining proteins and clinical parameters further improved discrimination (Trondheim: Recurrence AUC 0.75; Persistence AUC 0.91; Oslo: Recurrence AUC 0.75; Persistence AUC 0.88). Conclusion Pre-surgical serum immune-proteomic profiling revealed distinct immune states particularly associated with PSA persistence and recurrence. Serum protein signatures improved discrimination beyond presurgical clinical parameters, supporting their potential for presurgical risk stratification and treatment planning with further validation in larger independent cohorts.