Prediction of bile duct injury after transarterial chemoembolization for hepatocellular carcinoma: Model establishment and verification.

Objective:This study aimed to establish and validate a predictive model for bile duct injury in patients with hepatocellular carcinoma (HCC) after drug-eluting bead transarterial chemoembolization (DEB-TACE). Methods:We retrospectively analyzed 284 patients with HCC treated with DEB-TACE at our hospital between January 2017 and December 2021, of whom 63 patients experienced postoperative bile duct injuries. Univariate and logistic multivariate regression analyses were performed to identify the risk factors for bile duct injury, as well as establish and internally validate the nomogram model. The area under the curve (AUC) of the receiver operating characteristic (ROC) curve, calibration curve, Hosmer-Lemeshow goodness of fit test, decision curve analysis (DCA), and clinical impact curve (CIC) were used to assess the predictive power, clinical value, and practicability of the nomogram model. Results:The incidence of bile duct injuries after DEB-TACE was 22.18% (63/284), with one injury occurring in every 2.86 sessions of DEB-TACE treatment. Univariate and logistic multivariate regression analyses indicated that a history of hepatectomy (odds ratio [OR]=2.285; 95% confidence interval [CI]=1.066-4.898; P<0.05), subjective angiographic chemoembolization endpoint level (OR=1.832; 95% CI=1.258-2.667; P<0.05), alkaline phosphatase (OR=1.005; 95% CI=1.001-1.010; P<0.05), and platelet count (OR=1.005; 95% CI=1.001-1.009; P<0.05) were independent risk factors for bile duct injury after DEB-TACE among patients with HCC. The risk nomogram model based on the above four variables was validated using the bootstrap method, showing consistency between the predicted and experimental values. Furthermore, the model performed well in the Hosmer-Lemeshow goodness-of-fit test (2=3.648; P=0.887). The AUC of this model was 0.749 (95% CI=0.682-0.817), with an overall accuracy of 69.01%, a positive predictive value of 73.02%, a negative predictive value of 67.87%, a sensitivity of 73.0%, and a specificity of 67.90%, suggesting that the nomogram model had good accuracy and discrimination. In addition, DCA and CIC revealed a high clinical value and practicability of the model. Conclusion:Bile duct injury in patients with HCC treated with DEB-TACE is caused by multiple factors rather than a single factor. The nomogram prediction model used in this study had a good fitting degree and prediction efficacy, with high clinical value and practicability.