Objective: The purpose of this article is to assess the detection and negative prediction rate of upper urinary tract tumors in nonopacified urinary tracts on portal venous phase MDCT.
Materials and methods: This retrospective case-control study included 20 patients with upper urinary tract tumors and 40 age- and sex-matched control subjects. All studies were assessed independently by two reviewers. Reviewers determined whether each of four segments of the upper urinary tract could be fully visualized and whether tumor was present or absent. For each tumor, reviewers characterized its morphologic features (i.e., infiltrative or polypoid mass, urothelial thickening, and associated hydroureter or hydronephrosis).
Results: The detection rate of the proximal two upper urinary tract segments was significantly higher than that for the distal segments (p < 0.001). For each upper urinary tract, the sensitivity, specificity, and negative predictive value of portal venous phase MDCT for detecting tumors were 95%, 97%, and 100%, respectively. The positive predictive value for an estimated population prevalence of 0.0005-0.004 was 0.6-4.8%. The morphologic features significant for the presence of tumor were urothelial thickening and the presence of a discrete polypoid mass. Interobserver agreement for all features was good or very good, except for moderate agreement on urothelial thickening involving the ureter (κ = 0.60).
Conclusion: The detection rate of upper urinary tract tumors on nonopacified portal venous phase is high. Furthermore, in the absence of morphologic features suggestive of urothelial malignancy, a normal-appearing ureter may be reassuring.