Machine Learning Enables Single-Score Assessment of MASLD Presence and Severity
Chen, R.; Petrazzini, B. O.; Nadkarni, G. O.; Rocheleau, G.; Bansal, M.; Do, R.
Show abstract
Metabolic dysfunction-associated steatotic liver disease (MASLD) affects 30% of the global population but is often underdiagnosed. To fill this diagnostic gap, we developed a digital score reflecting presence and severity of MASLD. We fitted a machine learning model to electronic health records from 37,212 UK Biobank participants with proton density fat fraction measurements and/or a MASLD diagnosis to generate a "MASLD score". In holdout testing, our model achieved areas under the receiver-operating curve of 0.83-0.84 for MASLD diagnosis and 0.90-0.91 for identifying MASLD-associated advanced fibrosis. MASLD score was significantly associated with MASLD risk factors, progression to cirrhosis, and mortality. External testing in 252,725 diverse American participants demonstrated consistent results, and hepatologist chart review showed MASLD score identified probable MASLD underdiagnosis. The MASLD score could improve early diagnosis and intervention of chronic liver disease by providing a non-invasive, low-cost method for population-wide screening of MASLD.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Organ damage proteomic signature identifies patients with MASLD at-risk of systemic complications 93%
- CD4+ T cells promote fibrosis during metabolic dysfunction-associated steatohepatitis 92%
- Breakthrough SARS-CoV-2 Infection Outcomes in Vaccinated Patients with Chronic Liver Disease and Cirrhosis: A National COVID Cohort Collaborative Study 92%
Similar papers in this journal
- Unannotated small RNA clusters in circulating extracellular vesicles detect early stage liver cancer 92%
- Low coverage whole genome sequencing of low-grade dysplasia strongly predicts colorectal cancer risk in ulcerative colitis 91%
- Multiple indicators of gut dysbiosis predict all-cause and cause-specific mortality in solid organ transplant recipients 91%
Similar papers in this journal
- Exome-by-phenome-wide rare variant gene burden association with electronic health record phenotypes 92%
- Genome-wide polygenic score with APOL1 risk genotypes predicts chronic kidney disease across major continental ancestries 92%
- Actionable druggable genome-wide Mendelian randomization identifies repurposing opportunities for COVID-19 91%
Similar papers in this journal
- Rifaximin does not increase the rate of 30-day mortality in patients with cirrhosis and daptomycin in two National US-based cohorts 94%
- Outcomes of SARS-CoV-2 Infection in Patients with Chronic Liver Disease and Cirrhosis: a N3C Study 94%
- Machine Learning-Based Prediction of Pediatric Ulcerative Colitis Treatment Response using Diagnostic Histopathology 91%
Similar papers in this journal
- A systematic analysis of the contribution of genetics to multimorbidity and comparisons with primary care data 91%
- Multi-ancestry omic Mendelian randomization revealing putative drug targets of COVID-19 severity 91%
- Sedentary lifestyle, Physical Activity, and Gastrointestinal Diseases: Evidence from Mendelian Randomization Analysis 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.