Using Large Language Models to Predict Advanced Liver Fibrosis in Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD): A Proof-of-Concept Analysis - PubMed
2 hours ago
- #Liver Fibrosis
- #GPT-4
- #MASLD
- Study evaluates GPT-4 and GPT-3.5 for predicting advanced liver fibrosis in MASLD patients using NHANES data.
- GPT-4 achieved an AUROC of 0.91, outperforming traditional risk scores like FIB-4.
- GPT-3.5 showed strong performance with an AUROC of 0.90, slightly lower than GPT-4.
- Both models demonstrated high specificity and calibration, with GPT-4 being superior.
- Findings suggest LLMs could serve as scalable, interpretable tools for clinical risk prediction in MASLD.
- Further validation across diverse populations is needed before clinical adoption.