In our previous blog, we highlighted LiverExplore,* PathAI’s digital pathology tool that provides comprehensive, granular quantifications of fibrosis directly from routine, H&E-stained liver biopsies.1 The quantitative power of LiverExplore provides researchers with the opportunity to measure treatment-induced changes in fibrosis.
PathAI scientists recently returned from AASLD’s The Liver Meeting, where we witnessed how LiverExplore can enable meaningful analyses in MASH clinical trials. Through collaborative research with our partners, results were presented highlighting the ability of LiverExplore to detect disease-relevant changes in fibrosis in the treatment arms of two Phase 2 clinical trials.
Here, we will dive deeper into these exciting results.
Reduction of Liver Fibrosis by AI-Based Digital Pathology Analysis: Results from the Pemvidutide Phase 2b IMPACT Trial (collaboration with Altimmune)
Pemvidutide is a balanced 1:1 glucagon/GLP-1 dual receptor agonist. In MASH, the activation of glucagon receptors results in direct effects on the liver, including reductions in liver fat, inflammation and fibrosis, while the GLP-1 receptors mediate metabolic effects such as appetite suppression and weight loss. In the IMPACT study, (NCT05989711), a phase 2b trial of patients with biopsy-confirmed F2 or F3 MASH, pemvidutide achieved MASH resolution in 55% of patients within 24 weeks.2
In collaboration with PathAI, investigators from Altimmune, Inc. performed a retrospective analysis of the IMPACT study using LiverExplore to assess changes in fibrosis across the study arms. Here, digitized H&E-stained biopsies from baseline and week 24 were assessed for the steatosis-adjusted proportionate areas of total, early-stage, and advanced fibrosis.
As shown in Figure 1, LiverExplore quantified treatment-specific effects on fibrosis that were not seen through manual pathologist evaluation using MASH Clinical Research Network (CRN) scoring:
These results demonstrate that, compared to placebo, pemvidutide treatment can reduce the areas of total, early, and advanced pathologic liver fibrosis in a dose-dependent manner. Furthermore, these results highlight LiverExplore’s ability to detect disease-relevant changes in fibrosis beyond what can be seen through manual pathologist evaluation using MASH CRN scoring.
Read the full poster on our IMPACT analysis here.
Figure 1. LiverExplore-quantified effects of pemvidutide on total, early-stage, and advanced fibrosis area compared to placebo at 24-weeks.
AI-powered histology analysis of HARMONY reveals efruxifermin-driven changes in the liver microarchitecture in F2/F3 MASH (collaboration with Akero Therapeutics)
FGF21 is known to regulate lipid and glucose homeostasis, and preclinical work has demonstrated that this hormone can reduce features of MASH, including steatosis and fibrosis. In Phase 2b trials for MASH, efruxifermin (a long-acting FGF21 analog administered once weekly) improved fibrosis compared with placebo after 96 weeks of treatment in participants with either F2-F3 disease (HARMONY)3 or compensated cirrhosis (F4c) due to MASH (SYMMETRY).4
Together with Akero Therapeutics, we performed a retrospective analysis of the HARMONY study (NCT04767529) using LiverExplore. In this trial, participants received either placebo or efruxifermin (28 or 50 mg) for 96 weeks, and biopsies were taken at baseline as well as after 24 weeks and 96 weeks of treatment.
As shown in Figure 2, quantification of fibrosis features by LiverExplore in these biopsies yielded results that expanded the previously reported effects of efruxifermin on fibrosis:
These results reveal that efruxifermin-driven reductions in fibrosis are dose- and duration-dependent: earlier-stage fibrosis was reduced by week 24, whereas reductions in advanced fibrosis occurred at week 96 in the 50 mg efruxifermin group. The analysis underlines both the utility of LiverExplore to dissect treatment-induced anti-fibrotic effects as well as the potential efficacy of efruxifermin for modulating patterns of advanced fibrosis.
Figure 2. Improvements in LiverExplore-quantified fibrosis features with efruxifermin.
(left) LiverExplore fibrosis overlays from representative biopsies across study arms and timepoints.
(right) Fold change from baseline of LiverExplore-quantified fibrosis features across study arms at weeks 24 and 96.
In this poster, through LiverExplore feature analysis, we also demonstrated efruxifermin’s effects on other aspects of the liver microarchitecture – particularly, tissue and cell composition:
(left) LiverExplore tissue overlays from representative biopsies across study arms and timepoints.
(right) Fold change from baseline of LiverExplore-quantified relative tissue area features (out of total usable tissue) across study arms at weeks 24 and 96.
The reduction in LiverExplore-quantified hepatocellular ballooning, steatosis, and lobular inflammation aligns with the previously reported conventional pathology results from HARMONY that efruxifermin led to MASH resolution with no worsening of fibrosis. Furthermore, the reduction in immune cells and corresponding increase in normal hepatocytes that were measured by LiverExplore suggests that efruxifermin treatment resulted in improvements to overall liver health.
Figure 4. Increase in LiverExplore-quantified normal hepatocytes and reduction in immune cells with efruxifermin.
(left) LiverExplore cell overlays from representative biopsies across study arms and timepoints.
(right) Fold change from baseline of LiverExplore-quantified cell proportion features (relative to total cell count) across study arms at weeks 24 and 96.
Read the full poster on our HARMONY analysis here.
Conclusions
LiverExplore is a suite of digital pathology models that provide comprehensive, granular quantifications of fibrosis features from routine, H&E-stained liver biopsies. Through its ability to quantify disease-relevant fibrosis subtype-specific features, LiverExplore has the unique potential to dissect therapeutic-induced changes in fibrosis using digitized images of preexisting H&E-stained trial biopsies, thus yielding critical evidence of fibrotic regression in these trials.
The results that we presented with our partners at AASLD showcase the ability of LiverExplore to detect treatment-induced changes in fibrosis beyond categorical MASH CRN staging. In both the HARMONY and IMPACT trials, therapeutic-driven reductions in early-stage fibrosis occurred at 24 weeks, regardless of dose or therapeutic intervention. Longer treatment with one of the interventions, efruxifermin, was also associated with reductions in advanced fibrosis. We also demonstrated that LiverExplore-quantified cell and tissue features associated with MASH pathology were decreased with efruxifermin treatment in the HARMONY trial, and that these changes were accompanied by an increase in normal hepatocytes, suggesting that the treatment reduced inflammation and improved liver health.
The granular data provided by LiverExplore on these clinical trial cohorts demonstrates the tool’s functionality beyond a simple measurement tool to be a key strategic asset in drug development. By moving beyond categorical staging, LiverExplore unlocks a new level of understanding about how experimental therapeutics truly impact the liver microarchitecture. Ultimately, the future of MASH clinical trials depends on this precision. After seeing the insights revealed by LiverExplore in the AASLD presentations highlighted here, Jörn Schattenberg, M.D. of Saarland University Medical Center (UKS), a key opinion leader and renowned expert in MASH biology, noted:
“Leveraging LiverExplore to quantify fibrosis along with other cell and tissue features lends to transformational insights that could impact the field in multiple ways, including furthering our understanding of the mechanisms and pathophysiology of MASH, and supporting drug development through its ability to dissect response patterns. There is a key place for LiverExplore in clinical trials and research.”
This makes LiverExplore an essential tool for accelerating the discovery and validation of effective MASH treatments, uniquely positioning it to enhance the field’s understanding of biomarkers, combination therapy approaches, and potential mechanism(s) of action of experimental therapeutics in hepatology, from discovery through clinical trials and beyond.
Learn more about LiverExplore here.
References