Exploring the accuracy of multiparametric MRI in prostate cancer staging: An institutional retrospective study.

Abstract

e17056 Background: Multiparametric Magnetic Resonance Imaging (mpMRI) has been increasingly utilized in prostate cancer (PCa) diagnosis and staging. While there is Level 1 data supporting MRI utility in identifying clinically significant PCa and guiding PCa diagnosis, there is little data on its ability to predict surgical outcomes and its utility as a staging study. We aimed to evaluate the accuracy of mpMRI in predicting common surgical pathology outcomes in patients who underwent radical prostatectomy (RP). Methods: Men who underwent either open radical prostatectomy (ORP) or robotic assisted laparoscopic prostatectomy (RALP) for prostate adenocarcinoma from January-December 2021 at a single tertiary level care academic medical center were identified. Chart review for relevant patient demographics, mpMRI related variables and final surgical pathology was completed. In patients who had pre-operative mpMRI, we evaluated the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of mpMRI in predicting relevant surgical outcomes, including presence of pT2N0 organ confined disease (OCD), extracapsular extension (ECE), seminal vesicle invasion (SVI), lymph node involvement (LNI), and bladder neck invasion (BNI). Results: 168 eligible patients were identified in a 12-month period. The mean age was 63.5±6.24 years and mean Prostate Specific Antigen (PSA) was 11.4±23.7, with 166 (98.8%) patients undergoing RALP and 115 (68.5%) having pre-operative mpMRI. Median GGG was 2 in both MRI and CT subsets (p = 0.580), and patients who had pre-op MRI were more likely to have higher PSA (12.7 ±28.1 vs 8.38± 6.32, p = 0.073) and clinically node positive disease (p < 0.001) than those with CT. However, there was no significant difference in final surgical pathology or positive surgical margin rates between these two groups. On subset analysis of the MRI subset, Table summarizes the sensitivity, specificity, PPV, and NPV of pre-op MRI to predict OCD, ECE, SVI, LNI, and BNI. While specificity of pre-op MRI was adequate for all outcomes (89.1-100%), sensitivity (2.9-49.2%), PPV (40-100%), and NPV (56.3-94.3%) were poor. Conclusions: At present, pre-op MRI of the prostate does not appear to be accurate in its ability to predict important pathologic outcomes at the time of radical prostatectomy and should be used cautiously as a local staging tool. More work is needed before MRI can be used as a reliable staging tool for PCa.[Table: see text]