Dynamic endorectal magnetic resonance imaging for local staging and detection of neurovascular bundle involvement of prostate cancer: Correlation with histopathologic results

Objectives. To assess the staging accuracy and detection of neurovascular b undle involvement by dynamic subtraction contrast-enhanced endorectal magne tic resonance imaging (MRI) in patients with localized prostate cancer. Methods. In 38 patients with biopsy-proven prostate cancer, endorectal MRI was performed on a 1.5-Tesla magnetic resonance system using the dynamic te chnique with gadolinium-diethylenetriaminepentaacetic acid bolus enhancemen t. Two radiologists prospectively assessed the tumor involvement, localizat ion, capsular penetration, seminal vesicle invasion, and neurovascular bund le involvement. All patients subsequently underwent radical prostatectomy, and the MRI findings were correlated with the histopathologic results. Results. The overall accuracy of detecting cancer localization in the prost ate was 72%. The detection of involvement in the peripheral zone had an 80% accuracy rate, but for lesions in the transition zone, the rate was 63%. T he sensitivity and specificity of tumor detection was 81% and 79% for perip heral zone cancers and 37% and 97% for transition zone cancers, respectivel y. The accuracy rate, was 84%, 97%, and 97% for the detection of capsular p enetration, seminal vesicle invasion, and neurovascular bundle involvement, respectively. Conclusions. Prostatic MRI with an endorectal surface coil using the dynami c technique more accurately detected tumor localization, capsular penetrati on, seminal vesicle invasion, and neurovascular bundle involvement than pre viously reported methods. The detection of tumor localization was more accu rate in the peripheral zone than in the transition zone. This technique may be useful for the selection of patients for radical prostatectomy and, par ticularly, for identifying candidates for nerve-sparing surgery. UROLOGY 57 : 721-726, 2001. (C) 2001, Elsevier Science Inc.