The Predictive Value of Albumin-bilirubin Score Combined with Liver Function Index and CEA for Liver Metastasis of Colorectal Cancer

Objective To investigate the predictive value of albumin-bilirubin (ALBI) score combined with liver function index and carcinoembryonic antigen (CEA) for liver metastasis of colorectal cancer. Methods The clinical data of patients with colorectal cancer who underwent surgical treatment in the Second Hospital of Lanzhou University from January 2016 to July 2021 and were followed up for 24 months were retrospectively analyzed. According to the follow-up results, the enrolled patients were divided into liver metastasis group and non-liver metastasis group, and were randomly divided into modeling group and validation group according to the ratio of 2:1. The risk factors of liver metastasis in patients with colorectal cancer were analyzed. Lasso-Logistic regression was used to construct the prediction model. Bootstrap method was used for internal verification. Receiver operating characteristic curve, calibration curve and clinical decision curve were used to evaluate the reliability of the prediction model. Finally, a nomogram was drawn to show the prediction results. Results A total of 195 patients who met the inclusion and exclusion criteria were enrolled, including 130 in the modeling group and 65 in the validation group. Through Lasso regression variable screening and Logistic regression analysis, the results showed that ALBI (OR=8.062,95% CI:2.545~25.540), alanine transaminase (ALT) (OR=1.037,95% CI:1.004~1.071) and CEA (OR=1.025,95% CI:1.008~1.043) were independent predictors of liver metastasis in colorectal cancer. The area under curve (AUC) of the combined prediction of liver metastasis of colorectal cancer in the modeling group was 0.921, the sensitivity was 78%, the specificity was 95%, the C-index was 0.921, the H-L fitting curve X²=0.851, P=0.654, and the slope of the calibration curve was close to 1, suggesting that the accuracy of the model was high, and the DCA curve showed that the model had good clinical application value. For the data of the modeling group, the Bootstrap method was used for internal verification of 1000 resamplings. The accuracy was 0.869, the kappa consistency was 0.709, and the AUC was 0.913. When ALBI, ALT and CEA were used to diagnose liver metastasis of colorectal cancer alone, the AUC of CEA was the largest (0.897), and the combination of the three had the highest efficacy in the diagnosis of liver metastasis of colorectal cancer. In the validation group, the AUC, sensitivity, specificity, Cindex and H-L curve of the combined prediction of liver metastasis of colorectal cancer were 0.918, 85%, 95.6%, 0.918, X²=0.586, P=0.746, respectively. Conclusion ALBI, ALT and CEA have certain predictive value for liver metastasis of colorectal cancer. The combined diagnosis of liver metastasis of colorectal cancer has high efficacy, and the risk prediction model constructed by them has a good predictive effect.