RESUMO
PURPOSE: To determine whether quantitative dynamic contrast-enhanced MRI improves the performance of T2W-MRI for the localisation of non-palpable prostate cancer (PCa) and for the estimation of tumor volume. MATERIALS AND METHODS: Twenty-three patients (PSA: 8.91+/-6.2ng/m) with a non-palpable cancer underwent endorectal MRI with T2W and dynamic contrast enhanced (DCE) imaging before radical prostatectomy. Each level of evaluation (apex, mid-portion, base) was divided in eight areas (24 areas per prostate and 552 areas for the 23 patients). Localisation and volume of tumor foci greater than 0,2cc present on the radical prostatectmoy specimens were retrospectively correlated to their MR appearance on the 552 evaluated areas. The dynamic parameters included capillary permeability (K(trans)), maximum concentration of gadolinium after 60s of perfusion ([Gd]) and wash-out (K(ep)). Uni- and multivariate analysis were performed to determine which parameters were predictive of PCa. RESULTS: Mean values of K(trans), K(ep) and [Gd] were significantly higher in the 58 tumor foci greater than 0,2 cm(3) of the PZ and the TZ (all p<0.05). Logistic regression for each zone provides provided a value of the area under the ROC curve of 0.83 for the PZ and 0.81 for the TZ (0.7 and 0.75, respectively, for the T2W imaging), only significant for the PZ (p<0.002). Sensitivity and specificity were 79 and 77% for the PZ, 62.5 and 94% for the TZ. Above 0,2 cm(3), tumor volume on dynamic MR showed a mean difference of 51+/-100% (range: -145 to +248%). CONCLUSIONS: Quantitative dynamic MRI is more accurate than T2W imaging for tumor localisation of non-palpable cancer greater than 0,2 cm(3), but the difference is only significant for the PZ. Above this volume, correlation between tumor volume measured on dynamic MRI and that on the specimen is poor.
