Objective: The purpose of this study was to prospectively evaluate the diagnostic ability of diffusion-weighted MRI (DWI) for detecting upper urinary tract cancer.
Subjects and methods: Seventy-six patients (36 women, 40 men; median age, 70 years) suspected of having upper urinary tract cancer underwent T1- and T2-weighted imaging and DWI (b values, 0 and 800 s/mm(2)) with or without T1-weighted dynamic contrast-enhanced MRI (DCE-MRI). Two radiologists independently interpreted the images.
Results: Of the 76 patients suspected of having upper urinary tract cancer, 49 were diagnosed with upper urinary tract cancer and the remaining 27 were diagnosed as not having upper urinary tract cancer. The sensitivity, specificity, and accuracy of DWI interpretation for each reviewer was 92%, 96%, and 93% and 94%, 81%, and 89%, respectively. The sensitivity and accuracy of DWI were significantly higher than those of T1- and T2-weighted imaging (p < 0.01 and p = 0.03 for reviewer 1 and p < 0.01 for both values for reviewer 2), although the specificity did not change. The diagnostic abilities of DWI and DCE-MRI were not significantly different. The interobserver agreement of DWI between the two reviewers was excellent (κ score = 0.801). The apparent diffusion coefficient values of upper urinary tract cancer with grade 3 were significantly lower than those of upper urinary tract cancer with grades 2 and 1 (p < 0.028).
Conclusion: DWI provides accurate information for the diagnosis of upper urinary tract cancer in a noninvasive manner. The additional use of DWI to T1- and T2-weighted imaging increases the sensitivity of MRI in identifying upper urinary tract cancer with excellent interobserver agreement. Furthermore, DWI could be a useful adjunct to preoperative assessment of histologic grade.