关键词: 直肠癌, 局部进展期, 新辅助放化疗, 肿瘤退缩预测

Abstract:

Background and purpose: The standard therapy for locally advanced rectal cancer (LARC) is neoadjuvant chemoradiotherapy (nCRT) followed by surgery. NCRT can make the tumor regress and downstage, and increase the R0 resection rate. However, individual differences in rectal cancer are large, and some patients respond poorly to nCRT and cannot benefit from nCRT. Therefore, it is necessary to establish effective screening measures to identify patients with poor response to nCRT. This study aimed to analyze the influencing factors of nCRT for LARC and construct the tumor regression prediction model. Methods: Data of 158 LARC patients who underwent total mesenteric resection after receiving nCRT at the First Hospital Affiliated to Air Force Medical University from January 2016 to December 2020 were collected. Baseline clinical indicators before nCRT were collected, including laboratory examination, tumor markers and magnetic resonance imaging (MRI). According to the tumor size reported by MRI before and after nCRT, Response Evaluation Criteria in Solid Tumors (RECIST) was used to evaluate the extent of tumor regression after nCRT. After receiver operating characteristic (ROC) curve was used to standardize the clinical baseline indicators, logistic regression analysis was carried out to screen the factors affecting the tumor regression. The tumor regression prediction model was constructed by logistic regression, and the performance of the model was evaluated based on decision curve analysis (DCA) and the calibration curve. The accuracy of the model was tested by 10-fold cross-validation. Results: This retrospective cohort study enrolled 158 patients, in which, 98 patients achieved complete response (CR) or partial response (PR). The objective response rate was 62%. Sixty patients had poor response to nCRT, either stable disease (SD) or progressive disease (PD). Multivariate logistic regression analysis showed that tumor diameter before treatment (P<0.001), time to surgery after nCRT (P = 0.006), D-dimer (P = 0.010), prognostic nutrition index (PNI) (P = 0.035), carcinoembryonic antigen (CEA) (P = 0.004) and extramural vascular invasion (EMVI) (P = 0.026) were significantly related to tumor regression after nCRT. The area under ROC curve (AUC) of tumor regression after nCRT prediction model for LARC was 0.84 (95% CI: 0.780-0.899), sensitivity was 85.0%, and specificity was 72.4%. In the calibration curve, the predicted results were in good agreement with the actual results, and the prediction accuracy was good. The DCA showed that the tumor regression prediction model could bring clinical net benefit to diagnosis. Conclusion: Tumor diameter before treatment, time to surgery after nCRT, D-dimer, PNI, CEA and EMVI are independent risk factors for the tumor regression after nCRT in LARC patients. The tumor regression prediction model based on the above factors has good predictive efficacy for the tumor regression after nCRT in LARC patients.

Key words: Rectal cancer, Local advanced, Neoadjuvant chemoradiotherapy, Tumor regression prediction