
This review discusses the classification, applications, and projected future directions of AI within the care continuum of HCC. We cover work on risk stratification, early detection, diagnosis, prognostication, treatment selection, and response prediction. Multiple ML models have outperformed traditional regression models predicting HCC development, with superiority sustained on external validation. DL models have achieved AUROC above 0.90 across screening, diagnosis, and locoregional therapy response prediction. Pathomics-based DL models outperformed clinical models in predicting survival, metastasis, and local recurrence. LLM-driven systematic review workflows screened HCC trial reports with 96
Endohepatology integrates advanced endoscopic techniques with hepatology to enhance diagnosis and management of liver-related diseases, notably through EUS and ERCP. This review summarizes recent advances across the diagnostic and therapeutic applications of endohepatology, including parenchymal assessment, liver biopsy, portal pressure gradient measurement, and management of the complications of portal hypertension. EUS enables detailed visualization of the hepatic parenchyma, lesions, vasculature, and ascites, and facilitates tissue sampling with EUS-guided liver biopsy (EUS-LBx). Recent advancements, such as shear wave elastography (EUS-SWE), outperform transabdominal methods, particularly in obese patients, with comparable or superior fibrosis detection. EUS also allows for portal pressure gradient assessment (EUS-PPG), showing high correlation with hepatic venous pressure gradient (HVPG), aiding in portal hypertension evaluation. Management approaches focus on gastric varices, where EUS-guided coil and glue injections have demonstrated reduced rebleeding rates and enhanced safety over traditional techniques. Additionally, EUS assists in ascites sampling, differentiation of malignancy, and minimally invasive embolization of spontaneous portosystemic shunts, improving treatment outcomes. The future of endohepatology includes EUS-guided TIPS placement, portal venous sampling for tumor markers, and hepatic malignancy ablation, promising further minimally invasive options. These innovations facilitate comprehensive care, potentially improving patient outcomes through precise diagnosis, monitoring, and targeted therapy. Ongoing research aims to standardize techniques, validate efficacy, and expand applications, positioning endohepatology as a pivotal component in hepatology practice.
Alcohol-related liver disease (ALD) is increasingly recognised as a dynamic condition in which removal of the causal driver may permit hepatic recompensation. This review examines recompensation in ALD and MetALD, focusing on alcohol use disorder (AUD) treatment, phosphatidylethanol (PEth)-based abstinence monitoring, and cardiometabolic risk modification. Using Baveno VII criteria, multicentre studies report recompensation in approximately one-quarter to one-third of patients with decompensated ALD following sustained abstinence, with improved transplant-free survival. Early abstinence preserved hepatic reserve, and the absence of further decompensation predicts recovery. In MetALD and MASLD, weight loss and glycaemic control provide additional benefit, while PEth improves assessment of alcohol exposure and outcomes. Recompensation requires integrated addiction care, objective alcohol monitoring, cardiometabolic optimisation, and continued surveillance for portal hypertension and hepatocellular carcinoma. It represents substantial prognostic improvement but not a cure.
Alcohol-associated liver disease is now a leading indication for liver transplantation worldwide and in India, where living-donor transplantation (LDLT) predominates. Because transplantation treats the organ but not the addiction, this review examines how alcohol relapse affects the graft, immunosuppression and survival, with particular attention to the Indian LDLT setting. Relapse is best understood as a spectrum: occasional use has little measurable impact, whereas sustained harmful drinking accelerates graft steatohepatitis and fibrosis, often operates through immunosuppression non-adherence and rejection, and roughly triples long-term mortality. The excess late mortality is predominantly cardiovascular and malignant rather than hepatic. Reported relapse rates are markedly lower and reproducible in Indian cohorts (approximately 9–10
Liver fibrosis in alcohol-related liver disease (ALD) closely links to the development of adverse outcomes. However, underreporting of alcohol consumption and subclinical presentation hinder timely assessment. Various non-invasive tests (NITs) may mitigate the challenges. This review summarises the role of NITs in ALD. Objective biomarkers (e.g., phosphatidyethanol) quantify alcohol consumption for the diagnosis of ALD. Other readily available blood-based (e.g., Fibrosis-4 index) and imaging-based (e.g., vibration-controlled transient elastography) biomarkers help rule out and/or rule in advanced liver fibrosis with prognosis value, though the results can be confounded by ongoing alcohol consumption. Novel biomarkers and artificial intelligence (AI) models are emerging to refine risk assessment. While NITs help with the unmet needs in the field, careful adoption is needed by taking into consideration of the development of each NIT and the potential confounders to the test results. With the emerging evidence in specialised biomarkers and AI models, there is a future for achieving personalised management in ALD.