[Preoperative endorectal MRI in prostate cancer: a monocentric retrospective cohort]

The aim of the study was to assess the added value of a prostatic MRI performed in a non-expert center before radical prostatectomy for prostate cancer.

All patients considered for a radical prostatectomy for a clinically localised prostate cancer in our institution between June 2006 and April 2011 were analysed. They underwent a systematic endorectal 1.5 T MRI [eMRI] (T2 and diffusion-weighted imaging). The procedure was performed at least 8 weeks after transrectal ultrasound guided prostate biopsies. They were analyzed by a single experimented genitourinary radiologist. The preoperative characteristics and biopsy data were collected, as the eMRI interpretation using a systematic scale. Correlation between eMRI and final histopathology has been analyzed (standardized pathological report using the Stanford procedure). We considered the following points: extracapsular extension, seminal vesicle invasion and cancer localisation. Cancer localization was analyzed both by halves (left/right) and on sextant-basis (base, middle, apex, left and right respectively). For each data, sensitivity, specificity, positive (PPV) and negative predictive value (NPV) were calculated. Finally, we compared the results with those from the Partin's table and the Kattan's nomogram for T stage.

Two hundred and fifteen patients underwent a radical prostatectomy, having a mean age of 63 ± 6 years, a mean PSA: 7.8 ± 3.8 ng/mL. One hundred and fifty-nine patients had organ confined disease. Fifty-six had pT3 disease (50 pT3a and 6 pT3b). Sensibilities and specificities were the following: for extracapsular extension 26% and 96%; for seminal vesicle invasion 17% and 98%; for cancer localisation 64% and 74%; for prediction of left or right side 81% and 48%.

MRI showed equal performances compared to the Partin's table and the Kattan's nomogram for T stage. It showed lower performance compared to biopsy results for cancer localization (sensitivity 88%, specificity 70%). Literature provides wide ranges of results for eMRI. The results of this study were at the low limit of these ranges but reflect everyday practice. When performed, as it was in our study, eMRI did not seem to be accurate for staging. It had at best a very limited added value compared to the existing tools.