MR imaging of renal masses interpreted on CT to be suspicious

Abstract

Objective: Prior studies have shown that renal MR contrast enhancement improves the efficacy of mass and proximal vascular evaluation. This study assessed the usefulness of different sequences for characterization of masses that appeared suspicious on CT and for prediction of their potential for malignancy.

Subjects and methods: In a prospective manner 32 patients (age range, 26-78 years: average age, 54 years), each with at least one suspicious mass on CT, were examined with MR imaging. The following sequences were performed: conventional spin-echo with and without fat saturation, fast spin-echo, and dynamic gadopentetate dimeglumine-enhanced infusion using a 1.5-T superconducting magnet. Results were analyzed and compared with pathologic results after resection.

Results: A total of 65 renal masses of average size 2.6 cm (range, 1-10 cm) were detected with dynamic MR imaging. Seventeen of the 65 masses were malignant. Of the 17 malignant masses, three did not enhance on dynamic MR imaging (because of hemorrhage). Sixteen of the 17 malignant masses were heterogeneous on T2-weighted images. Three enhancing masses contained fat and all were angiomyolipomas. Thirty-five of the 65 masses (four with hemorrhage) did not show enhancement, all of which were homogeneous on T2-weighted images and were proven to be cysts. Five masses resulted from infections and had heterogeneous T2 appearance. The remaining masses were three hematomas with hemorrhage, one column of Bertin, and one aneurysm.

Conclusion: Renal masses that are interpreted as suspicious on CT may lack MR enhancement because of hemorrhage effects; heterogeneity of their T2 appearance is thus critical in differentiating malignancy from benign disease. Odds-ratio calculations give an adjusted estimate of a 3.36-fold increase (95% confidence interval, 1.8-6.27) in the likelihood of malignancy when masses are heterogeneous on T2-weighted images and a 29-fold increase (95% confidence interval, 3.67-241.8) for predicting malignancy when enhancement is present.