Affinity proteomics-based non-invasive detection of clinically significant liver disease
Balasubramani, S.; Karl, A. S.; Borchert, J. A.; Schrader, C.; Fromme, M.; Kayatekin, C.; Zhang, B.; Levit, M.; Krishnaswami, P.; van Eekeren, L. E.; Joosten, L. A. B.; Tomanova, P.; Remih, K.; Strnad, P.
Show abstract
BackgroundSince liver disease is often clinically unapparent, non-invasive biomarkers predicting future major adverse liver outcomes (MALO) are urgently needed. Therefore, we assessed the usefulness of a novel, proximity extension assay (PEA)-based high-throughput targeted proteomics method to predict MALOs. MethodsPEA plasma proteomic data (>2900 proteins) and clinical information were accessed from the population-based UK Biobank (UKB) cohort, including >53000 individuals with a median follow-up of >10 years and its subcohorts of obese (>12900) and diabetic (>1600) participants. The validation cohorts comprised 287 subjects with severe alpha1-antitrypsin deficiency (AATD), Pi*ZZ genotype, and 960 people living with HIV (PLHIV), who underwent liver stiffness measurement (LSM) via transient elastography. Selected PEA parameters were compared to routine measurements. Bayes-moderated linear models (age and sex as covariates) assessed the differential abundance. Logistic regression was used to identify and validate a novel prognostic score. ResultsRoutine gamma-glutamyltransferase (GGT) and aspartate aminotransferase (AST) levels strongly correlated with PEA-based measurements (r=0.91 and r=0.68, respectively). Similarly, PEA-based thrombospondin-2 levels strongly correlated with immunoassay-based values (r=0.85). Twenty proteins were consistently associated with future MALOs/increased LSM in all cohorts. UKB cohort was used to develop a novel five-component PEA score that demonstrated superior predictive power (AUROC=0.84) compared to established indices/scores, including AST-to-platelet-ratio index (APRI, AUROC=0.73) and Fibrosis-4 index (FIB4, AUROC=0.72) and its attractive predictive power was sustained in diabetic/obese subcohorts. In PLHIV and AATD validation cohorts, all five components showed gradual increases across fibrosis stages, and PEA score numerically outperformed APRI/FIB4 in predicting significant liver disease. ConclusionOur study identifies a new PEA score consisting of epithelial/hepatic stellate cell markers that demonstrates an attractive discriminative ability in several independent cohorts and different liver disease aetiologies.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Identification and functional characterisation of a rare MTTP variant underlying hereditary non-alcoholic fatty liver disease 94%
- Liver-specification of human iPSC-derived endothelial cells transplanted into mouse liver 93%
- A single-cell fixed RNA profiling of liver fibrosis progression and regression reveals SEMA4D and LMCD1 as key mediators of fibrogenesis 93%
Similar papers in this journal
- Hepatocyte Dedifferentiation Profiling In Alcohol-Related Liver Disease Identifies CXCR4 As A Driver Of Cell Reprogramming 94%
- Liver disease is a significant risk factor for cardiovascular outcomes - a UK Biobank study 94%
- Molecular signatures of chronic antibody-mediated rejection in human liver transplants 93%
Similar papers in this journal
- Clinical course and risk factors for mortality of COVID-19 patients with pre-existing cirrhosis: A multicenter cohort study 92%
- Faecal virome transplantation decrease symptoms of type-2-diabetes and obesity in a murine model 90%
- GATA4 and GATA6 loss-of-expression is associated with extinction of the classical programme and poor outcome in pancreatic ductal adenocarcinoma 90%
Similar papers in this journal
- Hepatic MCPIP1 protein levels are reduced in NAFLD patients and are predominantly expressed in cholangiocytes and liver endothelium 93%
- Circulating cell-free messenger RNA secretome characterization of primary sclerosing cholangitis 93%
- A Single-Cell Atlas Of Human Pediatric Liver Reveals Age-Related Hepatic Gene Signatures 93%
"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.