Morphologic predictors of lymph node status in rectal cancer with use of high-spatial-resolution MR imaging with histopathologic comparison

PURPOSE: To evaluate signal intensity and border characteristics of lymph nodes at high-spatial-resolution magnetic resonance (MR) imaging in patients with rectal cancer and to compare these findings with size in prediction of nodal status. MATERIALS AND METHODS: Forty-two patients who underwent total mesorectal excision of the rectum to determine if they had rectal carcinoma were studied with preoperative thin-section MR imaging. Lymph nodes were harvested from 42 transversely sectioned surgical specimens. The slice of each lymph node was carefully matched with its location on the corresponding MR images. Nodal size, border contour, and signal intensity on MR images were characterized and related to histologic involvement with metastases. Differences in sensitivity and specificity with border or signal intensity were calculated with Cls by using method 10 of Newcombe. RESULTS: Of the 437 nodes harvested, 102 were too small (<3 mm) to be depicted on MR images, and only two of these contained metastases. In 15 (68%) of 22 patients with nodal metastases, the size of normal or reactive nodes was equal to or greater than that of positive nodes in the same specimen. Fifty-one nodes were above the area imaged, and seven of these contained metastases. The diameter of benign and malignant nodes was similar; therefore, size was a poor predictor of nodal status. If a node was defined as suspicious because of an irregular border or mixed signal intensity, a superior accuracy was obtained and resulted in a sensitivity of 51 (85%) of 60 (95% CI: 74%, 92%) and a specificity of 216 (97%) of 221 (95% CI: 95%, 99%). CONCLUSION: Prediction of nodal involvement in rectal cancer with MR imaging is improved by using the border contour and signal intensity characteristics of lymph nodes instead of size criteria.