Five-protein model predicts MASLD with ‘strong accuracy’ up to 16 years before onset
Key takeaways:
- Participants with higher baseline levels of five specific plasma proteins had increased risk for MASLD.
- The combination of these proteins, along with other clinical predictors, enhanced predictive accuracy.
Five plasma proteins may predict a patient’s risk for metabolic dysfunction-associated steatotic liver disease up to 16 years before symptoms develop, according to research presented at Digestive Disease Week.
“Among more than 2,700 proteins, five stood out,” Shiyi Yu, MD, resident physician in the department of gastroenterology at Guangdong Provincial People’s Hospital in China, said during a presentation. “People with higher levels of these proteins were significantly more likely to develop MASLD, even 16 years before they were diagnosed. Our five-protein model predicted liver disease with strong accuracy, not just in the short-term, but over many years.”
The global burden of MASLD is surging, according to researchers, with data showing the Middle East and East Asia emerging as two regions with the fastest annual percent change in adult MASLD prevalence from 2010 to 2019.
Early identification of at-risk individuals is critical, yet long-term predictive tools remain limited. Yu and colleagues conducted an observational study to identify effective biomarkers and create predictive models that could forecast early onset of disease in the short- and long-term.
The researchers analyzed proteomics data from 52,952 UK Biobank participants who did not have MASLD at baseline. They selected 2,737 proteins and developed predictive models to compare cumulative MASLD incidences across baseline protein concentration quintiles.
Five proteins — CDHR2 (area under the curve = 0.825), FUOM (AUC = 0.815), KRT18 (AUC = 0.810), ACY1 (AUC = 0.803), and GGT1 (AUC = 0.797) — served as robust predictive biomarkers for MASLD incidence within 5 years. The proteins demonstrated “significant deviations” in plasma concentrations up to 16 years before onset, the researchers noted.
Participants with higher baseline levels of these proteins appeared at higher risk for MASLD ( HRs = 7.05-9.81).
Additionally, data showed that the combination of the five proteins showed predictive accuracy for MASLD at different time points over 16 years: 5-year (AUC = 0.857), 10-year (AUC = 0.775), over 10-year (AUC = 0.739) and all-time (AUC = 0.758).
The combined levels of the 5 specific proteins achieved 83.8% accuracy at predicting disease five years from onset and 75.6% accuracy at predicting 16 years ahead of diagnosis, according to a press release from DDW.
The addition of clinical biomarkers, such as BMI and daily exercise, improved 5-year accuracy to 89.4% and 16-year accuracy to 82.2%.
“These findings could be a game-changer for how we screen and intervene for liver disease,” Yu said. “Instead of waiting for symptoms or irreversible damage, we could flag high-risk individuals early and take steps to prevent MASLD from developing. This is especially important because MASLD often progresses silently until advance stages.
“These proteins don’t diagnose disease, but they signal increased risk, giving clinicians more information to act on earlier,” she added. “While more work is needed to understand the biology behind these biomarkers, this is a big step towards personalized prevention.”
References:
- Yu S, et al. Abstract 323. Presented at: Digestive Disease Week; May 3-6, 2025. San Diego.
- Blood proteins can predict liver disease up to 16 years before symptoms (Press release). Available at: https://ddw.org/blood-proteins-can-predict-liver-disease-up-to-16-years-before-symptoms/. Published April 25, 2025. Accessed April 25, 2025.
Collapse
Yu S, et al. Plasma proteomic profiles predict metabolic dysfunction-associated steatotic liver disease up to 16 years before onset. Presented at: Digestive Disease Week; May 3-6, 2025. San Diego.