[Magnetic resonance with endorectal coil in the local staging of prostatic carcinoma. Comparison with histologic macrosections in 40 cases]. / La Risonanza Magnetica con bobina endorettale nella stadiazione locale del carcinoma prostatico. Confronto con macrosezioni istologiche in 40 casi.

ABSTRACT

PURPOSE: Endorectal coil MRI is widely used in the diagnostic workup of prostate cancer, but diagnostic accuracy rates reported in the literature are quite variable. We report our personal experience with endorectal coil MRI in the local staging of prostate carcinoma. MATERIAL AND METHODS: Forty consecutive patients with histologically proved prostate carcinoma were examined with endorectal coil MRI at high field strength (1.5 T). All patients underwent a sagittal T1-weighted SE location sequence (TR 400, TE 20), an axial T1-weighted SE (TR 400, TE 20), two axial T2-weighted FSE sequences (TR 3000, TE 102, ETL 8) with and without fat suppression, and a coronal T2-weighted FSE sequence (TR 3000, TE 102, ETL 8); an axial Fast Multiplanar Spoiled Gradient Recalled (FMSPGR) dynamic sequence after Gd-DTPA injection was also performed in 18 patients. MR staging of local tumor spread was done according to the current literature criteria. All patients were submitted to radical prostatectomy, and histologic macrosections on the same plane as MR images were obtained from surgical specimens. MR and histologic staging were compared to assess MR accuracy in detecting capsular infiltration, seminal vesicles and apex involvement. The diagnostic yield of Gd-DTPA was also investigated. RESULTS: MRI correctly staged 31 of 40 cases (77.5%). MR accuracy was 80% in detecting capsular infiltration (85.7% sensitivity and 73.6% specificity), 90% in seminal vesicle involvement (91.6% sensitivity, 89.2% specificity) and 72.5% in apex involvement (79.1% sensitivity, 62.5% specificity). Dynamic studies with Gd-DTPA did not improve staging accuracy in any case. DISCUSSION AND CONCLUSIONS: In agreement with most of the current literature, MRI showed moderate overall accuracy in the local staging of prostate carcinoma. Particularly, MRI had good accuracy in detecting seminal vesicle involvement but moderate sensitivity and specificity in demonstrating capsular infiltration and apex involvement. Due to its high cost, MRI should not be routinely used in prostate cancer staging but should be reserved to the patients whose clinical and serological data suggest extraprostatic tumor spread, whose preoperative demonstration could avoid noncurative surgery.