Metabolomic Machine Learning Predictor for Adequate and Deep Response to UDCA in Non-Cirrhotic PBC
Fan Yang1, Jin Lin1, Yujie Zhang1
1Department of Gastroenterology and Hepatology and Laboratory of Gastrointestinal Cancer and Liver Disease, West China Hospital of Sichuan University, Chengdu, China.
Summary
This study reveals metabolic differences in primary biliary cholangitis (PBC) patients responding to ursodeoxycholic acid (UDCA). Integrating metabolomics with clinical data improves prediction of UDCA response and patient outcomes.
Area of Science:
- Hepatology
- Metabolomics
- Personalized Medicine
Background:
- Primary biliary cholangitis (PBC) patients show varied responses to ursodeoxycholic acid (UDCA).
- Incomplete UDCA responders face high progression risks.
- Current prognostic tools lack crucial metabolic insights for personalized PBC management.
Purpose of the Study:
- To characterize the metabolic landscape associated with UDCA response in PBC.
- To develop predictive models integrating clinical and metabolomic data for UDCA response.
- To enhance risk stratification for improved personalized PBC management.
Main Methods:
- Analyzed plasma metabolomics in 347 non-cirrhotic PBC patients.
- Classified patients into complete responders (n=214) and non-responders (n=133).
- Developed machine learning models using clinical and metabolomic features for response prediction.
Main Results:
- Identified suppressed bile acid biosynthesis and altered glutathione metabolism in UDCA responders.
- Found significant differences in sphingolipid signaling and phospholipase D pathways between adequate and deep responders.
- Developed an integrated prediction model with high performance (AUC=0.91 for response, AUC=0.87 for deep response stratification).
Conclusions:
- Established a metabolomics-informed framework for PBC management.
- Identified key metabolic predictors for UDCA treatment response.
- Demonstrated the potential for improved risk stratification and personalized management in PBC.
