Chemical shift effect predicting lymph node status in rectal cancer using high-resolution MR imaging with node-for-node matched histopathological validation.

Objectives: To evaluate the value of the chemical shift effect (CSE) as well as other criteria for the prediction of lymph node status.

Materials And Methods: Twenty-nine patients who underwent radical surgery of rectal cancers were studied with pre- and postoperative specimen MRI. Lymph nodes were harvested from transverse whole-mount specimens and compared with in vivo and ex vivo images to obtain a precise slice-for-section match. Preoperative MR characteristics including CSE, as well as other predictors, were evaluated by two readers independently between benign and metastatic nodes.

Results: A total of 255 benign and 35 metastatic nodes were obtained; 71.4% and 69.4% of benign nodes were detected with regular CSE for two readers, whereas 80.0% and 74.3% of metastatic nodes with absence of CSE. The CSE rendered areas under the ROC curve (AUC) of 0.879 and 0.845 for predicting nodal status for two readers. The criteria of nodal location, border, signal intensity and minimum distance to the rectal wall were also useful but with AUCs (0.629-0.743) lower than those of CSE.

Conclusions: CSE is a reliable predictor for differentiating benign from metastatic nodes. Additional criteria should be taken into account when it is difficult to determine the nodal status by using only a single predictor.

Key Points: • CSE is good for predicting nodal status with high confidence. • Nodal border and signal intensity are useful for assessing nodal status. • Location of mesorectal nodes could facilitate the prediction of nodal status. • Primary tumour stage could be used as reference for nodal staging.