Use of Preoperative MRI to Select Candidates for Local Excision of MRI-Staged T1 and T2 Rectal Cancer: Can MRI Select Patients With N0 Tumors?

Abstract

BACKGROUND: To minimize the recurrence rate after local excision of rectal cancer, the false-negative rate of nodal staging should be minimized.

OBJECTIVE: The purpose of this study was to develop a set of criteria using preoperative MRI that would minimize the false-negative rate for the diagnosis of regional lymph node metastasis.

DESIGN: A prospectively maintained colorectal cancer database and MRI images were retrospectively reviewed.

SETTINGS: This study was conducted at a multidisciplinary tertiary center.

PATIENTS: A total of 246 consecutive patients who underwent MRI and curative-intent surgery for MRI-staged T1/T2 rectal cancer from January 2008 to July 2012 were included.

MAIN OUTCOME MEASURES: MRI features significantly associated with lymph node metastasis were identified using a χ test. Five diagnostic criteria for lymph node metastasis were created based on these predictive MRI features, and their false-negative rates were compared using the generalized estimating equation method.

RESULTS: Small size/homogeneity of lymph nodes and no visible tumor/partially involved muscular layer were significantly associated with lower risks of lymph node metastasis. When tumor invasion depth was not considered, the false-negative rate did not decrease below 10%, even when the strictest criterion for morphologic evaluation of lymph nodes (not visible or <3 mm) was used. Adding invasion depth to the diagnostic criteria significantly decreased the false-negative rate as low as 1.8%.

LIMITATIONS: This study is limited by its small sample size and retrospective nature. CONCLUSIONS: Assessing both the depth of tumor invasion and lymph node morphology may reduce the false-negative rate and can be helpful to better identify candidates suitable for local excision of early stage rectal cancer. However, strict MRI criteria for oncologic safety might result in considerable false-positive cases and limit the application of local excision.