Machine learning predicts hepatocellular carcinoma risk from routine clinical data: a large population-based multicentric study - PubMed
5 hours ago
- #risk stratification
- #hepatocellular carcinoma
- #machine learning
- Machine learning framework developed for hepatocellular carcinoma (HCC) risk stratification using routine clinical data.
- Study utilizes data from over 900,000 individuals and 983 HCC cases across two large cohorts: UK Biobank and All of Us Research Program.
- Assesses contributions from demographics, lifestyle, health records, blood, genomics, and metabolomics.
- Random-forest-based models outperform existing risk-scores on internal and external test sets.
- Robust performance across ethnic subgroups with comprehensive interpretability.
- Code, model weights, and a web-calculator released for external validation and integration.
- Introduces PRE-Screen-HCC, a tool for HCC risk stratification and early detection.