Modified 3-Point MRI-Based Tumor Regression Grade Incorporating DWI for Locally Advanced Rectal Cancer

The purpose of this study was to evaluate the prognostic relevance of a modified 3-point MRI-based tumor regression grading system incorporating DWI for patients with locally advanced rectal cancer after preoperative chemoradiotherapy (CRT).

Between March 2012 and September 2013, 118 consecutively registered patients with middle or lower locally advanced rectal cancer who underwent CRT followed by surgery were enrolled in this retrospective study. Two radiologists in consensus assessed MRI tumor regression grade (mrTRG) based on T2-weighted images and high b value DW images (0 and 1000 s/mm²) using the following grades: 0, complete regression (no obvious tumor); 1, intermediate regression (dominant fibrosis, regression > 50%); 2, poor regression (dominant tumor, regression ≤ 50%). Multivariate analysis with a Cox regression model was performed to evaluate the association between modified mrTRG and 3-year disease-free survival (DFS) rate. A Kaplan-Meier method with a log-rank test was used to compare the DFS rate between responder (grades 0 and 1) and nonresponder (grade 2) groups.

Both the accuracy (72.9% vs 38.1%; p < 0.001) and the interreader agreement (κ = 0.580 vs 0.338; p < 0.001) of modified 3-point mrTRG were improved over the established 5-point mrTRG. Modified mrTRG (adjusted hazard ratio, 2.505; 95% CI, 1.231-5.100) was independently associated with 3-year DFS rate (p = 0.011). There was also a significant difference in the 3-year DFS rate between responders (73.8%; 95% CI, 64.2-81.3%) and nonresponders (41.7%; 95% CI, 10.9-70.8%) (p = 0.028).

In patients with middle or lower locally advanced rectal cancer, the modified 3-point mrTRG incorporating DWI was independently associated with the 3-year DFS rate after CRT followed by surgery. The grading scale may be used as a surrogate for expected prognosis of preoperative CRT. Further prospective trials are warranted.