Rectal carcinoma: prospective comparison of conventional and gadopentetate dimeglumine enhanced fat-suppressed MR imaging.

The purpose of this study is to compare the usefulness of conventional MR imaging and gadopentetate dimeglumine enhanced fat-suppressed MR imaging for the depiction and staging of rectal carcinoma. Thirty-two patients were prospectively evaluated by MR imaging using a 1.5-T unit. Based on the results of a barium study and/or digital examination, a balloon catheter was inserted to the level of the lesion before examination. Both conventional T1- and T2-weighted images and gadopentetate dimeglumine enhanced fat-suppressed T1-weighted images were obtained for all patients. The kappa statistics were performed for the evaluation of interobserver agreement and the McNemar test was performed for the analysis of staging accuracy. When only T1- and T2-weighted images were used, 5 of 32 tumors were not detected and the extent of 18 of 32 tumors were unclear. However, when gadopentetate dimeglumine enhanced fat-suppressed imaging was added, 24 of 32 tumors were well defined and only one tumor was not detected. In determining the depth of invasion, the staging accuracy was 72% for conventional imaging and 68% for all images combined. There was no significant difference between with gadopentetate dimeglumine fat-suppressed imaging and conventional imaging (P > .05). Use of gadopentetate dimeglumine (fat-suppressed imaging) resulted in overestimation of muscular invasion, perirectal fat invasion, and adjacent organ invasion in 12 patients, whereas nine patients were overestimated without the use of gadopentetate dimeglumine. In the detection of metastatic lymph nodes, gadopentetate dimeglumine enhanced fat-suppressed imaging also was not useful. Tumor detection was excellent using gadopentetate dimeglumine enhanced fat-suppressed images. However, the accuracy of staging was not improved by obtaining such images.