Dynamic TurboFLASH subtraction technique for contrast-enhanced MR imaging of the prostate: Correlation with histopathologic results

PURPOSE: To assess whether a TurboFLASH (fast low-angle shot) magnetic resonance (MR) sequence can improve the accuracy of fast spin-echo (SE) endorectal coil MR imaging in the staging and localization of prostate cancer. MATERIALS AND METHODS: In 57 patients with prostate cancer, MR imaging was performed with the following sequences: T1-weighted SE, T2-weighted fast SE, single-section gadolinium-enhanced dynamic subtracted TurboFLASH (one image every 1.25 or 2.5 seconds), and late-phase gadolinium-enhanced T1-weighted SE. Retrospectively, two blinded independent readers graded onset and steepest slope of enhancement and assessed tumor involvement and capsular penetration. MR findings were correlated with histopathologic results. RESULTS: On TurboFLASH images, prostate cancer was characterized by early and rapidly accelerating enhancement compared with that of surrounding tissues. Average sensitivity, specificity, and accuracy for detection of tumor involvement for the two readers with TurboFLASH images were 73.5%, 81.0%, and 77.5%. These values with fast SE images were 57.5%, 80.5%, and 72.0%. Depiction of capsular penetration and delineation and staging of tumor were better when TurboFLASH images were included with fast SE images. Differences between the two sequences, however, were not statistically significant. CONCLUSION: Because the TurboFLASH sequence did not statistically significantly improve tumor localization and staging results, routine use is not recommended. The technique may be useful for selected patients with equivocal evidence of capsular penetration.