Comparison of tumor regression grading system in locally advanced esophageal squamous cell carcinoma after preoperative radio-chemotherapy to determine the most accurate system predicting prognosis

Background: Nowadays, preoperative radio-chemotherapy is a standard treatment for locally advanced esophageal squamous cell carcinoma (ESCC). Tumor regression grade (TRG), referring to a classification of cancer response to preoperative treatment, can predict a prognosis of survival. Many TRG systems are proposed for use in esophageal cancer, but none of them has become standard grading system. This research compared five TRG systems, including Mandard system, Chirieac system, Schneider system, Hermann system, and Japan Esophageal Society (JES) system, to find the most accurately predictive system. Methods: We recruited 37 participants with locally advanced ESCC from 2006 to 2014. All of them were treated with radio-chemotherapy followed by esophagectomy. The resection specimens were evaluated microscopically for percentage of viable residual tumor comparing with tumor bed, number of positive lymph nodes and, consequently, assigned TRG grade according to each TRG system. Kaplan-Meier (KM) graphs were used to describe the median survival time. Log-rank tests and cox proportional hazard regression models were used in assessing associations between TRG systems and survival. Proportional hazard assumptions were evaluated on the basis of Schoenfeld and log-log plot. Akaike information criterion (AIC) values and pseudo R-squared values assessed model fit. All statistical tests were two-sided. Results: The KM graphs displayed overlapped curves in all TRG systems. The log-rank tests revealed that Schneider, JES and Mandard systems were statistically associated with overall-survival (P< 0.05). Only the multivariate cox regression analysis of Schneider system showed the statistically significant hazard ratio (P=0.037). Schneider system also had the best AIC and pseudo R-squared values. Conclusions: Schneider system might be the best predictive system. However, the overlapped KM curve opposed. This study had limitation due to small number of participants. More participants were needed to confirm our findings.