AI Insight
Researchers used deep learning to extract 200 liver MRI features from over 43,000 UK Biobank participants and linked these radiomic phenotypes to genetic variants through genome-wide association studies. They found that certain imaging signals predict chronic liver disease independently of steatosis and conventional risk factors, and that the genetic architecture underlying these features extends well beyond hepatic fat accumulation. The study also identified genetic overlaps between liver MRI traits and plasma proteins, metabolites, and cardiometabolic conditions, along with putative causal relationships between liver imaging markers and metabolic disease outcomes.
Why it matters
These findings could improve early detection of liver and cardiometabolic diseases by providing genetically grounded, imaging-based biomarkers that capture disease processes beyond what fat content alone can reveal. The identification of pathway-specific imaging markers also opens the door to more precise monitoring of therapeutics that act through the liver.
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
The liver plays a central role in systemic metabolism, yet large-scale genetic studies of quantitative liver imaging phenotypes remain limited. Here, we applied deep learning-based segmentation and radiomics extraction to derive 200 well-defined liver MRI features across multiple categories and imaging contrasts in 43,176 UK Biobank participants. Association analyses revealed steatosis-independent radiomic signals predicting incident chronic liver disease beyond conventional risk factors. We conducted genome-wide association studies in 37,725 individuals and identified multiple heritable liver MRI features; joint genetic structure and pleiotropy analyses demonstrated that these radiomic traits capture complex genetic architecture beyond the extent of hepatic steatosis. These MRI features showed widespread genetic overlap with plasma proteins, metabolites, and cardiometabolic traits through shared genetic loci and genetic correlations independent of adiposity. We identified putative causal links between liver MRI traits and cardiometabolic and liver-related outcomes, as well as evidence for pathway-specific imaging biomarkers to track activity of hepatically-influenced therapeutics.