Endorectal magnetic resonance imaging has limited clinical ability to preoperatively predict pT3 prostate cancer

OBJECTIVE To assess the clinical value of preoperative knowledge of the presence of extracapsular extension (ECE) or seminal vesicle invasion (SVI) in the planning for prostatectomy. MATERIALS AND METHODS An institutional database of 1161 robotic-assisted laparoscopic prostatectomies (RALP) performed by a single surgeon (D. B. S.) was queried for those who underwent endorectal coil magnetic resonance imaging (erMRI) before robotic-assisted laparoscopic prostatectomy. erMRI reports were dichotomized into positive or negative and compared with the final histopathology. The erMRIs performed at academic centres were compared with those performed in non-academic settings. A sub-group of high-risk patients was also analyzed for erMRI accuracy. RESULTS The 179 patients who underwent erMRI had significantly worse disease compared to the 982 patients without imaging. Of the 110 patients with histopathologically organ-confined disease, 81 (74%) were correctly diagnosed as such on erMRI, whereas 29 (26%) were felt to have cT3 disease and constituted false-positives. Among the 69 patients with pT3 disease, erMRI correctly predicted 30 (43%), whereas 39 (57%) were incorrectly considered organ-confined. The overall sensitivity and specificity for diagnosing pT3 disease was 43% and 73%. When stratified by pT3a and pT3b, the sensitivity and specificity of erMRI to accurately diagnose ECE is 33% and 81%, respectively. In evaluating SVI, erMRI has a sensitivity and specificity of 33% and 89%, respectively. The positive predictive value of erMRI to assess for ECE and SVI is 50% in both, with a negative predictive value of 61% and 63%, respectively. erMRIs performed at academic centres compared to non-academic locations demonstrated similar rates of sensitivity at 67% vs 77% and specificity at 39% vs 54%, respectively (P = 0.33). CONCLUSIONS In the setting of the present study, which was designed to be more reflective of current practice patterns in the USA, erMRI has limited clinical value in preoperatively detecting ECE and SVI. The accuracy of detecting T3 disease did not improve in academic centres or in high-risk patients.