Related Experiment Video
Updated: May 9, 2025
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
Development and validation of a prognostic model for post-surgical overall survival in Asian colon cancer patients: a
Cheng Liu1, Huaide Qiu2, Junqiang Wang3
1Department of Rehabilitation Medicine, Yixing No.2 People's Hospital (Yixing Prevention and Treatment Hospital for Occupational Diseases), Yixing, Jiangsu, China.
Objective:
This study aimed to identify the determinants of postoperative overall survival in Asian patients with colon cancer and to establish a prognostic nomogram model.
Methods:
The study included colon cancer cases diagnosed between 2010 and 2015, sourced from the SEER database as well as an external cohort from Yixing No.2 People's Hospital. Records with incomplete data on predetermined variables were excluded. The SEER dataset of eligible Asian postoperative colon cancer cases was split into a training set and a validation set with a 7:3 ratio. Prognostic factors affecting overall survival were identified using univariate and multivariate Cox regression analyses on the training set. A prognostic nomogram was developed with the R software package, and its predictive accuracy was evaluated in training, validation and external cohorts using ROC curves and calibration plots. Concordance index (C-index) and area under curves (AUCs) were also calculated, while decision curve analysis (DCA) was performed to examine the clinical utility.
Results:
Based on the criteria, 8738 cases from the SEER database were deemed suitable for analysis, and were divided into a training set (6118 cases) and a validation set (2620 cases) with a 7:3 ratio. An external cohort consisting of 73 cases with colon cancer was collected for external validation. The Cox regression analysis revealed that factors such as age, gender, marital status, histological type, grade classification, AJCC_T stage, AJCC_N stage, AJCC_M stage, CEA levels, and chemotherapy significantly influenced OS (P<0.05). These factors were incorporated into the nomogram, which demonstrated a C-index of 0.775 (95% CI: 0.766-0.784) for predicting OS in the training set, a C-index of 0.774 (95% CI: 0.760-0.787) in the validation set, and a C-index of 0.763 (95% CI: 0.698-0.828) in the external cohort. The nomogram was validated with good accuracy and clinical utility across three datasets.
Conclusion:
This study identified several independent prognostic factors influencing the postoperative overall survival of Asian colon cancer patients, including age, gender, marital status, histological type, grade classification, AJCC_T, AJCC_N, and AJCC_M stages, CEA levels, and chemotherapy. The constructed prognostic model showed good discrimination and accuracy, offering clinicians an individualized tool for survival predictions.
