肝细胞癌合并酒精性肝病患者移植后乙型肝炎病毒再激活与生存率研究
Renyi Su , Huigang Li , Xuanyu Zhang , Linping Cao , Zhe Yang , Jinyan Chen , Shusen Zheng , Xiao Xu , Di Lu , Xuyong Wei
Engineering ›› 2025, Vol. 49 ›› Issue (6) : 304 -313.
肝细胞癌合并酒精性肝病患者移植后乙型肝炎病毒再激活与生存率研究
Post-Transplant Hepatitis B Virus Reactivation and Survival in Hepatocellular Carcinoma Patients with Alcoholic Liver Disease
Alcohol consumption poses an escalating public health challenge. However, the impact of alcoholic liver disease (ALD) on post-transplant hepatitis B virus (HBV) reactivation and surgical outcomes remains inadequately characterized. Herein, we retrospectively analyzed our cohort (NCT06114251) comprising 453 patients with an HBV background. Propensity score matching (PSM) and sensitivity analyses were employed to assess the influence of ALD on surgical outcomes. Benchmark analysis compared the predictive performance of 21 models for post-transplant HBV reactivation. The Shapley additive explanation (SHAP) algorithm facilitated feature ranking and model interpretation. Patients were stratified into three subgroups based on the alcohol-modified HBV reactivation index (AMBRI). Among the cohort, 113 patients (24.9%) had concurrent pre-transplant diagnoses of ALD and HBV infection, while 340 (75.1%) had HBV infection alone. The presence of ALD was associated with an elevated risk of HBV reactivation and liver metastasis. PSM and sensitivity analyses revealed significantly lower five-year HBV reactivation-free survival (74.9% vs 85.4%), overall survival (OS, 56.2% vs 70.5%), and tumor recurrence-free survival (RFS, 47.8% vs 63.3%) in the ALD cohort. In recipients without HBV reactivation, hepatocellular carcinomas (HCCs) arising from both ALD and HBV exhibited inferior RFS (log-rank P = 0.026) and OS beyond one year (landmark P = 0.032) compared to HBV-related HCC alone. Benchmark analysis identified the surv.cforest model as the optimal predictor, achieving an area under the receiver operating characteristic (AUC) curve of 0.914 in internal validation and 0.884 in external validation, outperforming the published Cox model (AUC = 0.78). AMBRI-based stratification delineated three distinct risk subgroups, with the intermediate- and high-risk groups exhibiting significantly worse OS and RFS than the low-risk group. In this study, stratification by AMBRI identified intermediate- and high-risk groups with poorer post-transplant outcomes, underscoring the necessity for intensified surveillance and enhanced HBV treatment regimens, particularly in recipients with pre-transplant ALD.
Liver transplantation / Hepatitis B virus reactivation / Alcoholic liver disease / Machine learning
/
〈 |
|
〉 |