Liver cancer is associated with a few factors, such as viruses and alcohol consumption, and hepatectomy is an important treatment for patients with liver cancer. However, post-hepatectomy liver failure (PHLF) is the most serious complication and has a high mortality rate. Effective prediction of PHLF allows for the adjustment of clinical treatment strategies and is critical to the long-term prognosis of patients. Many factors have been associated with the development of PHLF, so there is an increasing interest in the development of predictive models for PHLF, such as nomograms that integrate intra-operative factors, imaging and biochemical characteristics of the patient. Ultrasound, as a simple and important examination method, plays an important role in predicting PHLF, especially the Nomogram established based on ultrasound measurements of liver stiffness and spleen area provides a more convenient way to predict the occurrence of PHLF.

To investigate the value of contrast-enhanced portal vein imaging at the hepatobiliary phase obtained with gadobenate dimeglumine for predicting clinical outcomes in patients with chronic liver disease (CLD).

Three hundred and fourteen CLD patients who underwent gadobenate dimeglumine-enhanced hepatic magnetic resonance imaging were stratified into three groups: nonadvanced CLD (n = 116), compensated advanced CLD (n = 120), and decompensated advanced CLD (n = 78) groups. The liver-to-portal vein contrast ratio (LPC) and liver-spleen contrast ratio (LSC) at the hepatobiliary phase were measured. The value of LPC for predicting hepatic decompensation and transplant-free survival was assessed using Cox regression analysis and Kaplan-Meier analysis.

The diagnostic performance of LPC was significantly better than LSC in evaluating the severity of CLD. During a median follow-up period of 53.0 months, the LPC was a significant predictor for hepatic decompensation (p < 0.001) in patients with compensated advanced CLD. The predictive performance of LPC was higher than that of the model for end-stage liver disease score (p = 0.006). With the optimal cut-off value, patients with LPC ≤ 0.98 had a higher cumulative incidence of hepatic decompensation than patients with LPC > 0.98 (p < 0.001). The LPC was also a significant predictive factor for transplant-free survival in patients with compensated advanced CLD (p = 0.007) and those with decompensated advanced CLD (p = 0.002).

Contrast-enhanced portal vein imaging at the hepatobiliary phase obtained with gadobenate dimeglumine is a valuable imaging biomarker for predicting hepatic decompensation and transplant-free survival in CLD patients.

• The liver-to-portal vein contrast ratio (LPC) significantly outperformed liver-spleen contrast ratio in evaluating the severity of chronic liver disease. • The LPC was a significant predictor for hepatic decompensation in patients with compensated advanced chronic liver disease. • The LPC was a significant predictor for transplant-free survival in patients with compensated and those with decompensated advanced chronic liver disease.

Post-hepatectomy liver failure (PHLF) is one of the most serious complications and causes of death in patients with hepatocellular carcinoma (HCC) after hepatectomy. This study aimed to develop a novel machine learning (ML) model based on the light gradient boosting machines (LightGBM) algorithm for predicting PHLF.

A total of 875 patients with HCC who underwent hepatectomy were randomized into a training cohort (n=612), a validation cohort (n=88), and a testing cohort (n=175). Shapley additive explanation (SHAP) was performed to determine the importance of individual variables. By combining these independent risk factors, an ML model for predicting PHLF was established. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, and decision curve analyses (DCA) were used to evaluate the accuracy of the ML model and compare it to that of other noninvasive models.

The AUCs of the ML model for predicting PHLF in the training cohort, validation cohort, and testing cohort were 0.944, 0.870, and 0.822, respectively. The ML model had a higher AUC for predicting PHLF than did other non-invasive models. The ML model for predicting PHLF was found to be more valuable than other noninvasive models.

A novel ML model for the prediction of PHLF using common clinical parameters was constructed and validated. The novel ML model performed better than did existing noninvasive models for the prediction of PHLF.