Serum protein risk stratification score for diagnostic evaluation of metabolic dysfunction-associated steatohepatitis

Abstract

Background: Reliable, noninvasive tools to diagnose at-risk metabolic dysfunction-associated steatohepatitis (MASH) are urgently needed to improve management. We developed a risk stratification score incorporating proteomics-derived serum markers with clinical variables to identify high-risk patients with MASH (NAFLD activity score >4 and fibrosis score >2).

Methods: In this 3-phase proteomic study of biopsy-proven metabolic dysfunction-associated steatotic fatty liver disease, we first developed a multi-protein predictor for discriminating NAFLD activity score >4 based on SOMAscan proteomics quantifying 1305 serum proteins from 57 US patients. Four key predictor proteins were verified by ELISA in the expanded US cohort (N = 168) and enhanced by adding clinical variables to create the 9-feature MASH Dx score, which predicted MASH and also high-risk MASH (F2+). The MASH Dx score was validated in 2 independent, external cohorts from Germany (N = 139) and Brazil (N = 177).

Results: The discovery phase identified a 6-protein classifier that achieved an AUC of 0.93 for identifying MASH. Significant elevation of 4 proteins (THBS2, GDF15, SELE, and IGFBP7) was verified by ELISA in the expanded discovery and independently in the 2 external cohorts. MASH Dx score incorporated these proteins with established MASH risk factors (age, body mass index, ALT, diabetes, and hypertension) to achieve good discrimination between MASH and metabolic dysfunction-associated steatotic fatty liver disease without MASH (AUC: 0.87-discovery; 0.83-pooled external validation cohorts), with similar performance when evaluating high-risk MASH F2-4 (vs. MASH F0-1 and metabolic dysfunction-associated steatotic fatty liver disease without MASH).

Conclusions: The MASH Dx score offers the first reliable noninvasive approach combining novel, biologically plausible ELISA-based fibrosis markers and clinical parameters to detect high-risk MASH in patient cohorts from the United States, Brazil, and Europe.

Graphical abstract

FIGURE 1

Discovery proteomics in MASLD identifies serum proteins associated with MASH and a multi-protein MASH predictor. (A, B) Discrimination between MASH and MASLD without MASH patient samples for the most significant 20 proteins by SOMAscan. Two different unstructured learning methods, HC (A) or PCA (B), were applied to the 20 most significant differentially expressed proteins (Benjamini-Hochberg corrected p < 0.01), detected by SOMAscan in 57 patients, demonstrating separation between MASH and MASLD without MASH. (A) In the HC colormap, red denotes upregulation, and green denotes downregulation. The * indicates the wrong classification. (B) PCA plot using the 20 proteins demonstrates good separation into 2 clusters with the SVMs separating line based on a linear kernel. Blue circles, MASH; red circles, MASLD without MASH. (C) HC using the 6 proteins from the MASH predictor demonstrates discrimination between MASH and MASLD without MASH. In the HC colormap, red denotes upregulation, and green denotes downregulation. The * indicates the wrong classification. (D) ROC curve analysis of the 6-protein MASH predictor. The 6-protein serum MASH predictor discriminates MASH subjects from MASLD without MASH subjects with high AUC. Abbreviations: HC, hierarchical clustering; MASH, metabolic dysfunction–associated steatohepatitis; MASLD, metabolic dysfunction–associated steatotic fatty liver disease; PCA, principal components analysis; ROC, receiver-operating characteristic; SVM, support vector machines.

FIGURE 2

Pathway analysis of 73 serum proteins associated with MASH. (A–C) Systems biology analysis of the 73 significant differentially expressed proteins was conducted using ingenuity pathway analysis. See Supplement for details, http://links.lww.com/HC9/B95. (A) Biological functions that are significantly enriched by the 73 input protein list. Level of significance: p < 10⁻⁸. (B) Upstream regulators that best explain the observed expression changes in the input 73 protein list as their targets. Level of significance: p < 10⁻⁵. (C) Downstream targets of TGFB1 (ie, proteins whose expression is affected by TGFB1) from among the 73-protein list. Red indicates upregulation and green denotes downregulation in MASH. Proteins are coded by shape; square: cytokine, vertical rhombus: enzyme, horizontal rhombus: peptidase, trapezoid: transporter, ellipse: transmembrane receptor, circle: other. Links are color-coded as red: leads to activation, blue: leads to inhibition, yellow: findings inconsistent with the state of the downstream protein, and black: effect not predicted. Red arrow highlights proteins incorporated in the MASH predictor. (D) STRING analysis and visualization of interaction clusters (k-means = 11 clusters indicated by node color) formed by serum proteins and labeled on related functional categories. Solid line represents within-cluster, dashed gray line represents between-cluster interactions. Line thickness indicates the strength of data support. Blue box indicates the major hub node. Red boxes indicate proteins included in the 6-protein MASH predictor. Abbreviation: MASH, metabolic dysfunction–associated steatohepatitis.

FIGURE 3

Distribution of ELISA levels for each of the 4 MASH predictor serum proteins in the US training cohort as visualized by box and whisker plots (A) and scattergram (B) of median protein levels. All 4 proteins had significant p value <0.001. Abbreviation: MASH, metabolic dysfunction–associated steatohepatitis.

FIGURE 4

Predictive performance of a multi-component risk classifier, MASH Dx score, in the training and independent validation cohorts. (A) Diagnostic performance of the 9-feature MASH Dx score, for the diagnosis of MASH (NAS ≥4 + F ≥0) versus MASLD without MASH in the training and external validation cohorts. The risk scores for the cutoffs for MASH, gray zone, and MASLD without MASH are indicated. The table provides the detailed performance criteria of the MASH Dx score for the US training cohort and the external, independent German and Brazilian validation cohorts as well as the pooled validation cohort. The AUC of the ROC curve, patient numbers (n), MASH prevalence, test outcome, PPV, NPV, sensitivity (SEN), and specificity (SPE) are shown. (B) ROC curves for the MASH Dx score. ROC curves and associated AUCs are shown for the US training cohort (TR), the German validation cohort (G), the Brazilian validation cohort (B), and the pooled validation cohort (VAL). (C-E) Predictive performance of a Multi-Component Risk Classifier, MASH Dx score, to discriminate between at risk MASH F2-4 and MASH F0-1 and MASH F2-4 and MASH F0-1 + MASLD without MASH in the training and independent validation cohorts. ROC curves for the MASH Dx score on at risk MASH F2-4 and MASH F0-1 (C) and at risk MASH F2-4 and MASH F0-1 + MASLD without MASH (D). ROC curves for FIB-4 on at risk MASH F2-4 and MASH F0-1 + MASLD without MASH (E). ROC curves and associated AUCs are shown for the US Training cohort (TR) and the pooled validation cohort (VAL). Abbreviations: MASH, metabolic dysfunction–associated steatohepatitis; MASLD, metabolic dysfunction–associated steatotic fatty liver disease; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver-operating characteristic.

FIGURE 5

Algorithm for clinical pathway in MASH evaluation and management. All patients with suspicion of MASLD should be screened using the MASH Dx score for fibrosis. If the score is <42.6 (low risk ), counseling should be continued on lifestyle modifications. If the score is >42.6, current guidelines for progression should be followed to closely monitor fibrosis staging. Moreover, therapeutics and clinical trials should be considered. Abbreviations: MASH, metabolic dysfunction–associated steatohepatitis; MASLD, metabolic dysfunction–associated steatotic fatty liver disease.