Nomogram for predicting the overall survival for young-onset colorectal cancer with liver metastases: a SEER database analysis

Abstract

Abstract Background: This study aimed to identify independent clinicopathologic factors associated with overall survival (OS) of Young-onset colorectal cancer with liver metastases (YO-CRCLM) and then construct and validate a prognostic nomogram for patients with YO-CRCLM. Methods We rigorously screened eligible YO-CRCLM patients diagnosed from January 2010 to December 2018 from the Surveillance, Epidemiology, and End Results (SEER) database and then randomly divided them into a training and a validation cohort. The prognostic indicators were screened by univariate and multivariate cox regression analysis. The nomogram predicted the 1-, 3- and 5-year OS based on the selected prognostic factors. The model was internally validated in testing cohort. The accuracy, reliability, and clinical utility of the model were assessed using the operating characteristic curve (ROC), the area under the curve (AUC), calibration curve, and decision curve analysis (DCA), respectively. Different risk subgroups of prognosis were determined based on the nomogram scores in training cohort of YO-CRCLM. Results 2976 patients with YO-CRCLM from the SEER database were eventually included in the study. The nomogram model incorporated ten independent prognostic risk factors to predict 1-, 3-, and 5-year OS based on univariate and multivariate cox regression. The ROC curve analysis showed that the area under the curve (AUC) of the training and validation cohorts exceeded 0.7 at 1, 3, and 5 years. Internal validation of calibration curves and DCA showed good stability and clinical utility values. Conclusion Nomogram predicting OS for YO-CRCLM patients was built and had an excellent performance, which will help clinical decision-making and formulation of the individualized treatment for YO-CRCLM patients.