Validation of MRI and Surgical Decision Making to Predict a Complete Resection in Pelvic Exenteration for Recurrent Rectal Cancer.

BACKGROUND: The main predictor of long-term survival in patients with recurrent rectal cancer is surgical resection with a clear resection margin. MRI plays a role in patient selection and surgical planning. OBJECTIVE: This study aimed to validate MRI in determining pelvic involvement by comparing MRI to histological outcomes, to assess the effect of MRI on surgical planning by comparing MRI findings with the surgical procedure, and to compare MRI anatomical involvement with resection outcome to assess if MRI can predict a clear resection margin. DESIGN: Retrospective study reviewing prepelvic exenteration MRI and correlating organ, involving an MRI with pathological involvement and surgical outcomes. SETTINGS: Single quaternary referral center with a special interest in pelvic exenteration. PATIENTS: The patients included 40 men and 22 women with median age of 60 years who had locally recurrent rectal cancer. MAIN OUTCOME MEASURES: The accuracy of MRI as measured using sensitivity and specificity by correlating MRI involvement with pathological involvement was the primary outcome measured. RESULTS: Recurrence in the anterior and central compartments was identified with accuracy on MRI and was likely to be associated with clear resection margins. MRI was less accurate at determining pelvic sidewall involvement. Lateral recurrence, high sacral, and nerve involvement were more likely to be associated with a positive resection margin. Sensitivity and specificity for pelvic sidewall structures was 46% and 91%. Involvement of nerve roots (60%-69%) and the upper sacrum (80%) on MRI was more likely to predict a positive resection margin than involvement of major pelvic viscera (22%). LIMITATIONS: This study was limited by its retrospective nature. CONCLUSIONS: MRI findings can be used to help predict resection margin. Prospective work with MRI interpretation and close correlation and involvement by pathologists is needed to address imaging and surgical limitations at the pelvic sidewall and high posterior margin.