Validation of the prognostic value of a three-gene signature and clinical parameters-based risk score in prostate cancer patients

BackgroundThe study aimed to validate the prognostic value of the Prostatype (R) risk score (P-score), which includes a three-gene signature and conventional risk factors, in a retrospective cohort. MethodsAll 716 patients diagnosed with prostate cancer from 2008 to 2010 at Skane University Hospital, Sweden, were included. After excluding patients based on pathological and clinical eligibility criteria, RNA quality, and presence of metastases at diagnosis, a final cohort comprising 316 patients was further analyzed. Expression levels of three genes (IGFBP3, F3, and VGLL3) were measured in archived formalin-fixed paraffin-embedded core needle biopsies. The gene expression data were combined with clinical parameters (Gleason score, prostate-specific antigen, and clinical tumor stage) to calculate the P-score for each patient. Predictive performance of the P-score in terms of prostate cancer-specific mortality (PCSM), distant metastasis and adverse pathological outcomes were investigated. ResultsThe P-score predicted both PCSM (hazard ratio [HR] = 1.6) and metastasis (HR = 1.46). The P-score had an area under curve (AUC) of 0.93 when predicting the PCSM risk at 10 years (95% confidence interval [CI]: 0.89-0.98), which was significantly better than both D'Amico (AUC: 0.81, 95% CI: 0.72-0.90, p < 0.001) and UCSF-CAPRA (AUC: 0.88, 95% CI: 0.80-0.96, p < 0.05). Decision curve analysis showed a higher net benefit of the P-score compared to both D'Amico and CAPRA. All three risk scores performed similarly in the prediction of distant metastases. For patients who underwent radical prostatectomy (RP), a higher P-score correlated with adverse pathological features such as pathologic tumor stage T3-4 (p < 0.0001) and >= International Society of Urological Pathology grade group 3 (p < 0.0001). ConclusionsOur findings provide evidence for the prognostic value of the P-score. The P-score predicted the risk for PCSM more accurately than the D'Amico and CAPRA scores. Performance was similar when predicting the risk for development of distant metastases within 10 years. Moreover, the P-score correlated with adverse pathological outcomes in RP specimens. Thus, the P-score could provide useful information for patients and their doctors to make informed decisions at the time of diagnosis.