RESUMO
OBJECTIVE: To analyse the detection ability of a multiparametric 3T MRI with phased-array coil in comparison with the pathological data provided by the prostatectomy specimens. METHODS: Prospective study of 30 months, including 74 patients for whom a diagnosis of prostate cancer had been made on randomized prostate biopsies, and all eligible to a radical prostatectomy. They all underwent multiparametric 3T MRI with pelvic phased-array coil including T2-weighted imaging (T2W), dynamic contrast-enhanced (DCE) and diffusion-weighted imaging (DWI) with an ADC mapping. Each gland was divided in octants. Three specific criteria have been sought (detection ability, capsular contact [CC] and extracapsular extension [ECE]), in comparison with the pathological data provided by the prostatectomy specimens. RESULTS: Five hundred and ninety-two octants were considered with 124 significant tumors (volume ≥ 0.1cm(3)). The general ability of tumor detection had a sensitivity, specificity, PPV and NPV respectively to 72.3%, 87.4%, 83.2% and 78.5%. The estimate of the CC and ECE had a high negative predictive power with specificities and VPN respectively to 96.4% and 95.4% for CC, and 97.5 and 97.7% for ECE. CONCLUSIONS: Multiparametric 3T MRI with pelvic phased-array coil appeared to be a reliable imaging technique in clinical and routine practice for the detection of localized prostate cancer. Estimation of the CC and millimeter ECE remains to be clarified, even if the negative predictive power for these parameters seems encouraging.
