Objective: The purpose of this study was to assess the accuracy of 3-T MRI in the preoperative diagnosis, staging, and planning of surgical management of rectal carcinoma.
Subjects and methods: Thirty-eight patients (23 men, 15 women) with clinically suspected rectal carcinoma underwent 3-T MRI. Coronal, axial, and sagittal T2-weighted sequences with and without fat suppression; axial T1-weighted spin-echo sequences; axial T1-weighted gradient-echo sequences with and without fat suppression; oblique 2D MR hydrography; and 3D fat-suppressed dynamic contrast-enhanced MRI were performed. Image quality with these sequences was evaluated by three radiologists experienced in body MRI. The significance of difference in results with the sequences was tested. The manner in which MRI staging and feasibility of sphincter-sparing surgery agreed with operative and pathologic findings was evaluated with kappa statistics.
Results: Rectal carcinoma was identified on MRI and confirmed histologically in all 38 patients. MRI findings were correctly predictive of T category in 35 cases (accuracy, 92.1%). In 31 (96.9%) of 32 resectable cases,sphincter-sparing surgical approaches were accurately chosen on the basis of MRI findings. Among the 11 sequences, 3D fat-suppressed dynamic contrast-enhanced MRI best delineated tumor margins. Coronal and axial T2-weighted images also well depicted tumor margins with minimal artifact. T2-weighted images were superior to unenhanced T1-weighted images.
Conclusion: MRI of rectal cancer at 3 T is accurate for prediction of T category and the feasibility of sphincter-sparing surgery. The best images were obtained with coronal, sagittal, and axial T2-weighted sequences and 3D fat-suppressed dynamic contrast-enhanced MRI.