Clinical significance and predictors of oncologic outcome after radical prostatectomy for invisible prostate cancer on multiparametric MRI

Abstract

Background: The objective of our study was to evaluate the clinical significance of invisible prostate cancer (iPCa) on multiparametric magnetic resonance imaging (mpMRI) by analyzing clinical parameters and oncologic outcomes. Methods: We retrospectively reviewed the records of patients treated with radical prostatectomy (RP) from 2010 to 2015 at our institution. Before RP, all patients were confirmed to have prostate cancer based on prostate biopsy. We excluded patients who underwent neoadjuvant therapy. Additionally, we excluded patients who had incomplete mpMRI based on PI-RADS (Prostate Imaging Reporting and Data System). iPCa was defined as having no grade 3 or higher region of interests using a scoring system established by PI-RADS without limitations on interpretation from mpMRI by radiologists. We selected patients with iPCa using this protocol. We analyzed data using univariate and multivariate cox regression analysis, logistic analysis, Kaplan-Meier curves, and receiver operator characteristic curves to predict biochemical recurrence (BCR). Results: A total of 213 patients with iPCa were selected according to the patient selection protocol. Among them, pathological findings showed that Gleason score (GS) G6, G7 and ≥ G8 were present in 115 cases (54.0%), 78 cases (36.6%), and 20 cases (9.4%), respectively. Further, extracapsular extension (ECE), positive surgical margins (PSM), and lymphovascular invasion (LVI) were present in 28 (13.1%), 18 (8.5%), and 3 cases (1.4%), respectively. Seminal vesicle invasion (SVI) was observed in one case (0.5%). During a median follow-up time of 51 months, BCR was observed 29 cases. Adverse pathology (AP) was defined as GS ≥8, ECE, SVI and LVI. AP and prostate specific antigen (PSA) were significantly associated with BCR. Moreover, PSA > 6.2 ng/ml was suggested as a cut-off value for predicting BCR. Conclusions: In our results, cases of iPCa had clinically significant PCa, and AP and poor prognosis were also observed in some. Additionally, we found that PSA is the most clinically reliable predictor of oncologic outcome. We suggest that active treatment and diagnosis should be considered for patients with iPCa with PSA > 6.2 ng/ml.