Researchers Developed New Liver Fibrosis Screening Model
A new diagnostic tool leveraging serum protein levels identifies liver fibrosis more effectively in early trials.
Updated on Sept. 20, 2026 in Life Sciences
Live Poll
Do you trust noninvasive screening models to accurately identify early signs of potential health conditions?
Researchers have developed a noninvasive liver fibrosis screening model that incorporates serum Golgi protein 73 (GP73). This new research-stage method aims to improve detection compared to existing clinical standards.
Why it matters
The study addresses the need for more accurate, noninvasive methods to detect liver fibrosis in patients without a prior diagnosis. By identifying GP73 as a biomarker, this approach potentially offers a more precise tool for early intervention.
The logistic regression model achieved an AUC of 0.788 in a cohort of 758 participants, while the Random Forest and XGBoost models reached 0.789 and 0.773, respectively. The model defines fibrosis as a liver stiffness measurement of at least 7.3 kPa.
The details
The model uses Least Absolute Shrinkage and Selection Operator (LASSO) regression—a statistical method that improves prediction accuracy by shrinking coefficients—to identify four independent predictors: age, aspartate aminotransferase (AST) levels, BMI, and GP73. These predictors are synthesized into a nomogram, a visual representation of a mathematical formula, to output a probability of fibrosis. Researchers evaluated this performance against the established FIB-4 (a fibrosis index based on four clinical factors) and APRI (AST to platelet ratio index) tools using a 7:3 ratio for training and testing.
Timeline
September 20, 2026: The peer-reviewed research article was published.
The Tech Race
This research seeks to improve upon long-standing noninvasive diagnostic indices like FIB-4 and APRI by incorporating GP73 as a novel biomarker. It represents a broader effort in digital medicine to refine diagnostic accuracy using machine learning models on serum data.
This screening model is currently in the research stage and not yet available for clinical use. It will require further validation and regulatory review before it can be adopted into standard medical diagnostic workflows.
The takeaway
This study demonstrates that incorporating GP73 into clinical models can improve the predictive accuracy of liver fibrosis detection. Researchers and clinicians should monitor for future prospective trials that test this nomogram against larger, varied patient cohorts.
Further reading
Explore more advancements in diagnostic technologies on our Life Sciences page.
More information
View the complete results in the peer-reviewed research article.
Source note: This article includes information reported by Nature.
Live Poll
Do you trust noninvasive screening models to accurately identify early signs of potential health conditions?
