T3 Subclassification With Emd/Mesorectum Ratio Predicts Neoadjuvant Chemoradiation Outcome In T3 Rectal Cancer Patients

Objective: To evaluate the feasibility of the EMD (extramural depth)/mesorectum ratio as a marker for T3 rectal cancer and its ability to predict tumour response to neoadjuvant chemoradiation and survival. Methods: From 2010 to 2016, 284 T3 rectal cancer patients who underwent high resolution MRI before neoadjuvant chemoradiation were enrolled. The EMD was defined as the distance from the outer edge of the muscularis propria to the outermost edge of the tumour. The measurement of the tumour EMD and mesorectum was in the same layer and their ratio was calculated. Receiver operating characteristic analysis and relative area under the curve statistics were used to choose the cut-off value. The association of the EMD/mesorectum ratio and other MRI or clinical factors with the tumour regression grade (TRG) was analysed. Cox regression analysis was used to estimate independent risk factors for disease-free survival (DFS) and overall survival (OS). Results: The mean EMD/mesorectum ratio was 0.47 +/- 0.3. We chose an EMD/mesorectum ratio of 0.5 in further analyses after receiver operating characteristic analysis. Of 284 patients, 177 (62.3%) had an EMD/mesorectum ratio <= 0.5. Patients with an EMD/mesorectum ratio <= 0.5 had a higher TRG 0-1 rate than patients with a ratio > 0.5 (53.1% vs 36.4%, p = 0.006). A multivariate analysis identified that an EMD/mesorectum ratio > 0.5 [hazard ratio (HR) 2.020; p = 0.028] and ypTNM II-III (HR 3.550; p = 0.017) were independent prognostic factors to indicate decreased DFS. For OS, only patients with TRG 2-3 had decreased OS compared with patients with TRG 0-1 (HR 2.959; p = 0.035). Conclusion: When the EMD/mesorectum ratio was applied to categorize T3 rectal cancer patients, the ratio of 0.5 can be used as a cut-off value for T3 rectal cancer. Patients with a ratio <= 0.5 had a higher response rate and better DFS. However, further validation is needed in a larger sample of patients.