Comparison of tumor regression grade and clinical stage based on MRI image as a selection criterion for non-radical management after concurrent chemoradiotherapy in locally advanced rectal cancer: a multicenter, retrospective, cross-sectional study.

ABSTRACT

PURPOSE: There has been no comparative study on the clinical value of magnetic resonance tumor regression grade (mrTRG)1-2 and ycT0-1N0 for the prediction of ypT0-1N0 after concurrent chemoradiotherapy (CCRT) for rectal cancer. We compared the diagnostic performance between mrTRG1-2 and ycT0-1N0 for predicting ypT0-1N0 as a selection criterion for non-radical management after CCRT in locally advanced rectal cancer. METHODS: This retrospective study enrolled 291 patients from three referral hospitals between January 2018 and March 2020. The diagnostic performance of ycT0-1N0 and mrTRG1-2 for the prediction of ypT0-1N0 was compared in terms of sensitivity, specificity, positive-predictive value, negative-predictive value, and area under the curve (AUC). RESULTS: Sixty-eight patients (23.4%) achieved ypT0-1N0. Nineteen patients (6.5%) had ycT0-1N0, and 91 patients (31.2%) had mrTRG1-2. For predicting ypT0-1N0, ycT0-1N0 had a sensitivity of 16.2% (95% confidence interval [CI] 8.36‒27.10) and positive-predictive value of 57.9% (95% CI 36.57‒76.63), while mrTRG1-2 had a sensitivity of 58.8% (95% CI 46.23‒70.63) and positive-predictive value of 44.0% (95% CI 36.46‒51.74). When predicting ypT0-1N0, mrTRG1-2 showed a higher AUC (0.680, 95% CI 0.604‒0.756) than ycT0-1N0 (0.563, 95% CI 0.481‒0.645) (P < 0.001). CONCLUSION: mrTRG1-2 might be a better indicator than ycT0-1N0 for the selection of non-radical management of advanced rectal cancer post-CCRT. However, additional diagnostic tools are required for predicting ypT0-1N0 because mrTRG1-2 or yc stage on MRI has insufficient evidence for diagnosing ypT0-1N0.