| Title: | Genetic Information Enhances a Novel Non-Invasive Steatosis Index and Outperforms Existing Steatosis Indices |
| Journal: | Liver International |
| Published: | 11 Jan 2026 |
| Pubmed: | https://pubmed.ncbi.nlm.nih.gov/41521699/ |
| DOI: | https://doi.org/10.1111/liv.70410 |
Publication 18652
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
Abstract
BACKGROUND AND AIMS: Early detection of individuals with metabolic dysfunction-associated steatotic liver disease (MASLD) is important as interventions to reduce steatosis can prevent progression to advanced liver disease. Identification of individuals with hepatic steatosis relies on imaging techniques, which are costly and often unavailable in large populations. Identifying hepatic steatosis through blood tests and demographics may serve as a cost-effective alternative to identifying hepatic steatosis. Here, we aim to identify demographic, serum and genetic variables that can help predict hepatic steatosis in two large population-based cohorts.</p>
METHODS: We analysed data from 32 008 participants in the UK Biobank (UKBB) with liver fat quantified using magnetic resonance proton density fat fraction (MRI-PDFF). We created non-invasive models to predict steatosis, the Enhanced Steatosis Indexes (ESIs), using logistic regression. Candidate predictors included variables from the literature and genetic markers. ESI-1 variables included demographic and serum values. ESI-2 included the ESI-1 variables plus the genetic markers. We trained and tested these models in UKBB and validated them in the Framingham Heart Study (FHS).</p>
RESULTS: In UKBB, ESI-1 and ESI-2 predicted ≥ 5% PDFF with AUC 0.844 and 0.858, respectively, outperforming DSI, HSI, FSI and FLI (AUC 0.818-0.834 vs. ESI-1 and AUC 0.830-0.844 vs. ESI-2). In FHS, ESI-1 and ESI-2 predicted CT-measured liver steatosis with AUC 0.803 and 0.808, respectively.</p>
CONCLUSIONS: We developed a non-invasive model to diagnose steatosis that outperforms existing steatosis models. Including genetic information improved the performance of ESI-2. Deployment of our models can facilitate non-invasive screening of steatosis so that, with early intervention, disease progression can be prevented.</p>
IMPACT AND IMPLICATIONS: We developed a non-invasive model based on demographic, serum laboratory and genetic predictors that outperforms existing steatosis indices in identifying individuals with hepatic steatosis. This Enhanced Steatosis Index (ESI) enables more accurate and personalised early detection of hepatic steatosis, a highly prevalent yet underdiagnosed condition. By leveraging readily available clinical measures and genetic data, both ESI-1 and ESI-2 improve risk stratification and support large-scale screening and early intervention efforts. These advances have the potential to provide more affordable, accurate and targeted screening, facilitating earlier treatment, reducing the burden of metabolic liver disease and helping to address disparities in the diagnosis and management of hepatic steatosis.</p>
13 Keywords
- Adult
- Aged
- Fatty Liver
- Female
- Genetic Markers
- Humans
- Liver
- Logistic Models
- Magnetic Resonance Imaging
- Male
- Middle Aged
- Non-alcoholic Fatty Liver Disease
- United Kingdom
6 Authors
- Yousif Alyousifi
- Chinmay Raut
- Vincent L. Chen
- Antonino Oliveri
- Yanhua Chen
- Elizabeth K. Speliotes
1 Application
| Application ID | Title |
|---|---|
| 18120 | Identifying causes of obesity and liver disease using population genetics |
Enabling scientific discoveries that improve human health