Accuracy of magnetic resonance imaging in prediction of tumour-free resection margin in rectal cancer surgery

Abstract

Background: Incomplete surgical removal of the circumferential tumour spread is believed to be the main cause of local recurrence after resection of rectal cancer. This study assessed the accuracy of magnetic resonance imaging (MRI) with a phased-array coil for preoperative staging and prediction of the distance of the tumour from the circumferential resection margin in a total mesorectal excision.

Methods: 76 patients with primary rectal cancer were preoperatively assessed by MRI at 1.5 T, with a phased-array coil. Two observers independently scored, on two occasions, the tumour stage and measured the distance to the mesorectal fascia. Their findings were compared with the final histological findings.

Findings: The MRI tumour stage agreed with the histological stage in 63 (83%) of 76 patients (weighted kappa=0.77 [95% CI 0.66-0.89]) for observer 1, and in 51 (67%) patients (weighted kappa=0.52 [0.37-0.67]) for observer 2. The intraobserver agreement on the tumour stage was good (kappa=0.80 [0.69-0.91]) for observer 1 but moderate (kappa=0.49 [0.34-0.65]) for observer 2. The interobserver agreement was moderate (kappa=0.53 [0.38-0.69]). In 12 patients with an obvious T4 tumour, a margin of 0 mm was correctly predicted. Of 29 patients for whom the pathologist reported a distance of at least 10 mm without specifying the actual distance, a distance of at least 10 mm was predicted in 28 by observer 1 and 27 by observer 2. For the remaining 35 patients, a regression curve was constructed; from this, a histological distance of at least 1.0 mm can be predicted with high confidence when the measured distance on MRI is at least 5.0 mm.

Interpretation: MRI with a phased-array coil showed moderate accuracy and reproducibility for predicting the tumour stage of rectal cancers. The clinically more important circumferential resection margin can, however, be predicted with high accuracy and consistency, allowing preoperative identification of patients at risk of recurrence who will benefit from preoperative radiotherapy, more extensive surgery, or both.