Establishment and validation of a prognostic survival prediction model for colorectal cancer patients with liver metastasis:based on SEER database

CHEN Chen; MU Conghui; YUAN Shuai

doi:10.12173/j.issn.1004-5511.202411053

Abstract：Objective To explore the relevant factors affecting patients with liver metastasis of colorectal cancer,and to establish a nomogram prediction model to provide evidence for individualized diagnosis and treatment.Methods To retrospectively analyze the case information of colorectal cancer liver metastasis patients in the SEER database from January 2010 to December 2019.Lasso regression and Cox regression model were used to analyze independent risk factors affecting patient prognosis and establish the prediction nomogram model.The discrimination and consistency of the model were evaluated based on the receiver operating characteristic curve and its area under the curve(AUC),calibration curve and decision curve analysis(DCA).Results A total of 5,121 patients with colorectal cancer liver metastases were included,comprising 3,631 cases in the modelling group and 1,490 cases in the validation group.Undifferentiated tumour grade[HR=4.29,95%CI(2.74,6.71)],positive neural infiltration tumour[HR=1.32,95%CI(1.16,1.49)],lymph node ratio(LNR)>0.6[HR=1.70,95%CI(1.28,2.28)]were risk factors for patients with colorectal cancer liver metastases.Age>60 years old[HR=0.85,95%CI(0.75,0.95)]and surgery for metastatic lesions[HR=0.85,95%CI(0.75,0.96)]were protective factors for the prognosis of colorectal cancer patients with liver metastasis.A survival prediction model was established by nomogram.The AUC values for the modeling group and validation group were 0.722 and 0.732,respectively.The calibration curve approximates the ideal curve,and the DCA curve demonstrates that the model possesses good clinical validity.Conclusion This study is based on the SEER database to construct a survival prediction nomogram model,which has high discrimination,consistency,and clinical effectiveness,demonstrating the practicality of the prediction model in clinical practice.