Evaluation of locally recurrent pelvic malignancy: performance of T2- and diffusion-weighted MRI with image fusion

PURPOSE: To evaluate the performance of T2- and diffusion-weighted magnetic resonance imaging (MRI) with image fusion for detection of locally recurrent pelvic malignancy. MATERIALS AND METHODS: The study group consisted of 28 patients (27 female, 1 male) who underwent pelvic MRI at 1.5 T after treatment of pelvic malignancy. MR images were reviewed independently by three blinded readers. The performance of the four sequences for detecting local recurrence was evaluated using receiver operating characteristic analysis: T2-weighted fast spin-echo (FSE), diffusion-weighted echo-planar imaging (DWI), dynamic contrast-enhanced (DCE) fat-suppressed T1-weighted spoiled gradient echo (SPGR), and T2-DWI with image fusion, the latter created using OsiriX Medical Imaging Software. RESULTS: Local recurrence was confirmed at biopsy in 16 patients. Twelve patients showed no evidence of recurrence on two consecutive MRI studies. The Az value for T2-DWI with image fusion (0.949) was statistically greater than that for T2-weighted FSE (0.849) (P<0.05). The sensitivity and specificity was 87.5% and 47.2%, respectively, for T2-weighted FSE, 100.0% and 50.0% for DWI, 95.8% and 58.3% for DCE fat-suppressed T1-weighted SPGR, and 93.8% and 72.2% for T2-DWI with image fusion. CONCLUSION: For depicting locally recurrent pelvic malignancy, T2-DWI with image fusion outperforms standard T2-weighted FSE and DWI and is comparable to DCE fat-suppressed T1-weighted SPGR.