Prognostic scoring system using inflammation- and nutrition-related biomarkers to predict prognosis in stage I-III colorectal cancer patients
Ke-Jin Li, Ziyi Zhang, Kuan Wang, Subinur Sulayman, Xiangyue Zeng, Juan Liu, Yi Chen, Zeliang Zhao
Abstract
BACKGROUND Colorectal cancer (CRC) is a common malignancy that has become a global burden. The prognostic prediction of CRC patients on the basis of inflammatory biomarkers and nutritional biomarkers has shown some potential but has not been fully explored. AIM To develop and validate a prognostic model for CRC based on inflammation and nutrition-related biomarkers and to evaluate its predictive value for patient outcomes. METHODS Patients were randomized at a 3:2 ratio into a training cohort (n = 282) or a validation cohort (n = 188). To identify the optimal prognostic factors for constructing the risk score (RS), LASSO Cox regression analysis was conducted. The association between the RS and overall survival (OS) was evaluated using receiver operating characteristic (ROC) curves and Kaplan-Meier (K-M) survival analysis. Independent risk factors were screened by multivariate Cox regression analysis. Nomograms were constructed and validated on the basis of these factors. RESULTS In the training cohort, univariate analysis of all the inflammatory and nutritional biomarkers demonstrated some predictive value. A LASSO-Cox analysis included four biomarkers and constructed an RS. Through ROC analysis, the area under the prognostic curve was 0.795. K-M survival curve analyses revealed that the five-year OS was significantly greater in the Low-RS group than in the High-RS group (P < 0.001). Multivariate analysis demonstrated that the degree of differentiation (P = 0.001), degree of nerve invasion (P = 0.022), and RS (P < 0.001) were independent risk factors. We constructed a nomogram to predict the OS of CRC patients and validated it in a separate cohort. The calibration curve showed high accuracy. Additionally, decision curve analysis for 1-year, 3-year, and 5-year survival probabilities indicated significant clinical utility in predicting survival outcomes. CONCLUSION This study developed a nomogram based on the RS to predict the OS of CRC patients. This nomogram can guide treatment decisions and enable the formulation of personalized follow-up strategies on the basis of predicted recurrence risk, aiming to improve long-term prognosis.