Related Experiment Video
Updated: Jul 17, 2025
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
A novel risk model for predicting peritoneal metastasis in colorectal cancer based on the SEER database
Li Yao1, Huan Shao1, Xinyi Zhang1
1Department of Gastroenterology, The First Affiliated Hospital of Zhejiang Chinese Medical University, Number 54, Youdian Road, Shangcheng District, Hangzhou, 310006, Zhejiang Province, China.
Background:
Early detection and intervention could significantly improve the prognosis of patients with peritoneal metastasis (PM). Our main purpose was to develop a model to predict the risk of PM in patients with colorectal cancer (CRC).
Methods:
Patients from the Surveillance, Epidemiology, and End Results (SEER) database with CRC classified according to the AJCC 8th TNM staging system were selected for the study. After data pre-processing, the dataset was divided into a training set and a validation set. In the training set, univariate logistic analysis and stepwise multivariate logistic regression analysis were utilized to screen clinical features and construct a risk prediction model. Then, we validated the model using the confusion matrix, receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves to examine its performance.
Results:
The model constructed using stepwise multivariate logistic regression analysis incorporated the following eight clinical features: age, tumor location, histological type, T stage, carcinoembryonic antigen (CEA) level, tumor deposits (TDs), log odds (LODDS) of metastatic lymph nodes, and extraperitoneal metastasis (EM). The areas under the curve (AUCs) of the model in the training and validation sets were 0.924 and 0.912, respectively. The accuracy and the recall ratio were higher than 0.8 in both cohorts. DCA and the calibration curves also confirmed its excellent predictive power.
Conclusions:
Our model can effectively predict the risk of PM in CRC patients, which is of great significance for the timely identification of patients at high risk of PM and further clinical decision-making.
