Development and validation nomograms to predict long-term survival in patients with locally advanced rectal adenocarcinoma after surgery: a population-based study

Abstract

Background: This study aimed to develop predictive nomograms for long-term cancer-specific survival (CSS) and overall survival (OS) in patients diagnosed with locally advanced rectal adenocarcinoma (LARA). Methods: Patients diagnosed with LARA between 2004 and 2015 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. The patients were randomly divided into training and validation cohorts. Optimal cutoff values for age, lymph node ratio (LNR), and tumor size were determined using X-tile. Univariate and multivariate Cox regression analyses were also conducted to identify independent factors associated with CSS and OS, and these factors were used to construct the nomograms. The performance of nomogram was assessed using the concordance index (C-index), calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA). External validation was performed at a single center in China. Furthermore, the predictive performance of the nomograms was compared with that of the 8th edition of the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system. Results: This study included a total of 4900 patients diagnosed with LARA from the SEER database, with an additional 116 patients composing the external validation cohort from a single institution in China. The determined optimal cutoff values for age, LNR, and tumor size were 67, 17.1%, and 62, respectively. Cox regression analyses revealed age, race, T and N stage, carcinoembryonic antigen (CEA) levels, tumor size and differentiation, chemoradiotherapy, perineural invasion, and the LNR as independent prognostic factors for both CSS and OS. The C-indexes of the long-term survival nomograms in the training, internal validation, and external validation sets were 0.713 (0.676-0.750), 0.707 (0.670-0.744), and 0.702 (0.600-0.804) for CSS, and 0.700 (0.669-0.731), 0.700 (0.651-0.749), and 0.705 (0.631-0.779) for OS, respectively. The predictive performance of the nomograms was superior to that of the 8th edition of the AJCC TNM staging system. Conclusions: We established and validated novel nomograms for more precise prediction of CSS and OS in patients with LARA, and the predictive power could guide prognostic prediction and therapeutic decisions.