Proteomic signatures for fibrosis in MASLD: a biopsy-proven dual-cohort study
Scandinavian Journal of Gastroenterology, 2025
Blomdahl J., Åberg M., Fridén M., Ahlström H., Hockings P., Hulthe J., Eriksson N., Gabrysch K., Nasr P., Risérus U., Kechagias S., Rorsman F., Ekstedt M., Vessby J.
| Disease area | Application area | Sample type | Products |
|---|---|---|---|
Metabolic Diseases Hepatology | Patient Stratification | Plasma |  Olink Target 96 |
Abstract
Objectives
Predicting disease progression in metabolic dysfunction-associated steatotic liver disease (MASLD) is challenging, and current non-invasive tests (NITs) lack the precision to replace liver biopsy. This study aimed to identify plasma biomarkers for different stages of fibrosis using affinity-based proteomics in two biopsy-proven cohorts. The primary objective was to identify biomarkers capable of distinguishing between low-to-no fibrosis (F0-1) and significant fibrosis (F2-4) in MASLD.
Materials and methods
Participants in the discovery cohort were recruited from Uppsala University Hospital and Swedish CArdioPulmonary bioImage Study (SCAPIS), while the validation cohort was included from Linköping University Hospital. All participants diagnosed with MASLD underwent liver biopsy and were categorized by fibrosis stage (F0-1 or F2-4). A total of 276 plasma proteins were analyzed using Olink® panels, with biomarkers identified through ordinal logistic regression, random forest (RF) analysis and the Boruta algorithm.
Results
The discovery cohort included 60 participants, with 60% having fibrosis stage F0–1 and 40% having F2-4. The validation cohort had 59 participants, of whom 35 had fibrosis stage F0-1 (59.3%) and 24 had stage F2-4 (40.7%). Five biomarkers were significantly associated with fibrosis stage in the discovery cohort, with four confirmed in the validation cohort. A model combining angiotensin converting enzyme-2 (ACE2), hepatocyte growth factor (HGF) and insulin-like growth factor-binding protein-7 (IGFBP-7) demonstrated strong predictive performance for significant fibrosis (c-statistics 0.82–0.83), outperforming fibrosis-4 (FIB-4) (c-statistics 0.61–0.72).
Conclusions
A biomarker model including ACE2, HGF and IGFBP7 shows promise in distinguishing between low-stage and significant fibrosis.