Machine Learning in Liver Fibrosis Diagnosis and Progression - EMJ

Machine Learning in Liver Fibrosis Diagnosis and Progression

LIVER fibrosis is a chronic and progressive condition caused by various factors, including alcohol-associated liver disease, metabolic dysfunction-associated steatohepatitis (MASH), autoimmune diseases, and viral hepatitis. Hepatitis B (HBV) and hepatitis C (HCV) are significant global contributors, leading to chronic hepatocyte injury, fibrous tissue deposition, nodular regeneration, and, ultimately, cirrhosis. Cirrhosis, the final stage of liver fibrosis, significantly increases the risk of hepatocellular carcinoma and complications such as ascites, variceal bleeding, and hepatic encephalopathy. 

Liver biopsy remains the gold standard for diagnosing fibrosis and assessing disease severity. However, its invasiveness, cost, and susceptibility to sampling errors have led to the development of non-invasive biomarkers for liver fibrosis estimation. Despite these advancements, the natural progression of liver fibrosis and the sequence of biomarker changes remain unclear. 

Recent developments in machine learning, particularly disease progression modelling, have allowed researchers to reconstruct long-term disease patterns from cross-sectional data. The ‘event-based model’ paved the way for methods such as ‘Subtype and Stage Inference’ (SuStaIn). SuStaIn integrates disease progression modelling with clustering to identify disease subgroups and progression patterns. 

Applying SuStaIn to liver fibrosis progression in HBV and HCV patients, researchers identified a probable sequence of biomarker conversion: APRI, FIB-4, CEI, Q-LSC, and RV/TV. These findings suggest that APRI may be the earliest biomarker for liver fibrosis screening. Additionally, serum biomarker conversion appears to precede imaging biomarker changes, aligning with previous studies showing that APRI and FIB-4 effectively predict fibrosis stages. 

However, the study had limitations, including a small sample size, lack of direct fibrosis severity validation, and absence of data on viral loads and treatment history. Future studies incorporating a larger cohort and additional imaging techniques are necessary for validation. 

Understanding liver fibrosis progression through machine learning provides valuable insights into disease mechanisms, improves patient stratification, and enhances clinical trial design. Identifying biomarker sequences may enable earlier diagnosis and targeted interventions, ultimately improving patient outcomes. 

Katie Wright, EMJ 

Reference 

Suzuki A et al. Modeling liver fibrosis progression in patients with viral hepatitis using the machine learning tool Subtype and Stage Inference (SuStaIn). Cureus. 2025;17(2):e78744. 

Rate this content's potential impact on patient outcomes

Average rating / 5. Vote count:

No votes so far! Be the first to rate this content.