Clinical characteristics and prognostic model of colorectal cancer based on SEER database / 国际外科学杂志

ABSTRACT

Objective:To investigate the important risk factors for the prognosis of patients with colorectal cancer (CRC) based on SEER database, and to construct a line chart prognostic model of 1-, 3-and 5-year survival rates for CRC patients.Methods:The clinical data of 52814 patients with CRC diagnosed pathologically from 2010 to 2015 in SEER database were collected. Cox proportional hazard model was used for univariate and multivariate survival analysis to determine the prognostic risk factors. Stepwise regression was used to screen the clinical factors that had the greatest impact on prognosis. Calibration curve was used to evaluate the accuracy of line chart prediction model for predicting 1-, 3-and 5-year OS in patients with CRC.Results:Surgical method, age, LNR, AJCC stage, T stage, N stage, M stage, vascular tumor thrombus and nerve invasion were independent risk factors for prognosis. Stepwise regression analysis showed that age, T stage, M stage, LNR and mode of operation were the most associated with the prognosis of CRC. A line chart model was constructed based on these five factors. The consistency index of age/T staging/M staging/LNR/ operation on the training set and verification set was the highest, which were 0.762 and 0.756 respectively. ROC curve analysis showed that in the training set and verification set, the AUC of age/T stage/M stage/LNR/ operation prognosis model was more than 0.7, and the model had high diagnostic value.Conclusion:The CRC prognostic line chart model based on age/T staging/M staging/LNR/ operation has a better predictive effect. The constructed prognosis OS diagram is convenient for clinical oncologists to make a more accurate assessment of the prognosis of patients, and provides a theoretical basis for individual diagnosis and treatment of CRC patients.