Retrospective study of role of neoadjuvant rectal scores in evaluating the 10-year disease-free survival of patients with locally advanced rectal cancer undergoing neoadjuvant chemoradiotherapy followed by surgery

Abstract

Objective: To investigate the correlation between the neoadjuvant rectal (NAR) score and long-term survival in patients with locally advanced rectal cancer who have undergone neoadjuvant chemoradiotherapy. Methods: Clinical and pathological data of 487 patients diagnosed with rectal adenocarcinoma from October 2004 to April 2014 at Sun Yat-sen University Cancer Center who had received neoadjuvant chemoradiotherapy were retrospectively analyzed and the impact of NAR score on prognosis studied. Disease-free-survival (DFS) was calculated by the Kaplan-Meier method and survivals compared using the log-rank test. Cox models were used for univariate and multivariate analyses. Receiver operating characteristic curves were utilized to evaluate the predictive capability of NAR and tumor regression grade scores for the risk of 10-year postoperative recurrence and metastasis. The Delong test was employed to compare the diagnostic performance of the two scores. Results: Of the 487 patients included in the study, 166 were men (34.1%). The median age was 56 years (interquartile range [IQR]: 46-63). All patients completed adequate preoperative chemoradiotherapy and underwent R0 resection.The median interval between the end of chemoradiotherapy and surgery was 51 days (IQR: 44-58). Post-chemoradiotherapy downstaging occurred in 329 patients (67.6%). Tumor regression grades (TRGs) were 1-2 in 246 patients (50.5%) and 3-4 in 241 patients (49.5%). A total of 394 patients (80.9%) received postoperative chemotherapy. NAR scores were <8 in 182 patients (37.4%), 8-16 in 180 (37.0%), and >16 in 125 (25.6%). The median follow-up time was 111.5 months (IQR: 70.7-133.7 months). One hundred and thirteen patients died of rectal cancer, among whom 13 patients developed local recurrence, 88 patients developed distant metastasis, and 12 patients had unknown recurrence patterns. The 10-year DFS and overall survival rate of f the whole group were 68.9% and 71.5% respectively. The 10-year DFS rates for patients with NAR scores <8, 8-16, and >16 were 85.1%, 80.5%, and 66.4%, respectively (P<0.001). Multivariate analyses revealed that the Dixon operation (HR=0.606, 95%CI: 0.408-0.902, P=0.014), and >16 (HR=2.569, 95%CI: 1.559-4.233, P<0.001) were independent predictors of the 10-year DFS of patients with locally advanced rectal cancer (P<0.05 for all). In the entire patient cohort, the AUC of the receiver operating characteristic curve for NAR score predicting 10-year recurrence and metastasis was 0.67 (95%CI: 0.62-0.72), whereas the AUC for TRG score was 0.54 (95%CI: 0.49-0.60). The two scores differed significantly in accuracy (Z=-4.06, P<0.001), the NAR score being a significantly better predictor of risk of 10-year recurrence and metastasis than the TRG score. Conclusion: The NAR score is a reliable predictor of 10-year DFS in patients with locally advanced rectal cancer who have undergone neoadjuvant chemoradiotherapy followed by curative surgery.