The role of magnetic resonance tumour regression grade in the prediction of regression and survival of rectal adenocarcinoma after long-course chemoradiotherapy: a cohort study.

Abstract

Between 2011 and 2016, patients with rectal cancer who underwent LCCRT and had post-LCCRT MRI were included in the study. Both mrTRG and pTRG were dichotomised into good responders (mrTRG 1-3 and pTRG 0-1) and poor responders (mrTRG 4-5 and pTRG 2-3). Correlation between mrTRG and pTRG was assessed with Cohen κ analysis. Survival analysis was performed with Kaplan-Meier test and Cox proportional hazard models. There were 59 patients included in this study. There were significant reductions in anal sphincter and circumferential resection margin involvement in post-LCCRT MRI. Fair agreement was found between mrTRG and pTRG (κ=0.345). Sensitivity, specificity and accuracy of mrTRG 1-3 to predict good pathological response were 100%, 46.3% and 62.7%, respectively. On survival analysis, mrTRG 1-3 was not associated with improved overall survival and recurrence-free survival. While there is fair agreement in correlation between mrTRG and pTRG, MRI remains an objective, noninvasive assessment of tumour response. Further studies are required to improve the ability of mrTRG to predict good responders to LCCRT and evaluate its role as a prognostic marker for survival.