Background: The aim of this study was to evaluate genomic risk of patients with persistent prostate specific antigen (PSA) using mRNA expression analysis and a validated prognostic genomic-risk classifier.
Methods: Monocentric retrospective study including all patients who underwent radical prostatectomy (RP) by one surgeon and Decipher Test from October 2013 to December 2018. PSA persistent population was defined as all patients with two consecutive PSA>0.1 ng/mL at follow-up after the surgery. Neurovascular Structure-adjacent Frozen-section Examination (NeuroSAFE) was performed intraoperatively for research of positive surgical margins. Multivariate analysis was performed for persistent PSA (pPSA) predictors. A specific localized, organ-confined, and negative margins sub-population with PSA persistence was compared to a similar sub-population without PSA persistence for genomic differential expression analyses.
Results: A total of 564 patients were included and 61 of them had pPSA. Preoperative PSA was higher in the PSA persistent group (11.6 [6.4, 21.2] vs. 6.2 [4.7, 9.2] P=0.00010), as well as PSA density (PSAd) (0.3 [0.2, 0.5] vs. 0.2 [0.1, 0.3] P=0.0001). Postoperative characteristics, Gleason Score, and positive surgical margins were significantly higher in the PSA persistent population. 31 patients had pPSA in our specific subpopulation and were compared to 217 patients with no pPSA. On multivariate analysis, only Decipher Score (OR=5.64 [1.28; 24.89], P=0.022) and preoperative PSA (OR=1.06, [1.02; 1.09], P=0.001) were significant predictors for PSA persistence. We found two genes to be significantly upregulated with a 2.5-fold change in our specific subpopulation (SERPINB11 and PDE11A).
Conclusions: We found unique genomic features of patients with pPSA, whilst confirming previous clinical findings that this condition behaves to a worse prognosis. Given this high genomic risk, further imaging studies should be performed to select patients for early treatment intensification.