A Review of Selected Presentations From the ACG 2026 Annual Scientific Meeting
October 9-14, 2026 • Nashville, Tennessee
Machine Learning Models of Noninvasive Tests to Predict MASH and Fibrosis Stage Based on MAESTRO-NAFLD-1 and MAESTRO-NASH Liver Biopsies
Metabolic dysfunction-associated steatotic disease (MASLD) can progress to metabolic dysfunction-associated steatohepatitis (MASH), a progressive liver disease associated with hepatic inflammation, hepatocyte ballooning, and varying degrees of fibrosis.1 In patients with MASH, advanced fibrosis is associated with a higher risk of adverse clinical outcomes, including hepatocellular carcinoma.2 As a result, there is a need to accurately assess the extent of fibrosis in patients with MASH.
The gold standard approach for assessing hepatic fibrosis is liver biopsy; however, biopsy is associated with limitations, including high costs, susceptibility to sampling error, variability in pathologic interpretation, and a risk of procedure-related adverse outcomes.1 There has been significant interest in the development of noninvasive biomarker tests as alternatives to serial liver biopsies.3
Loomba and colleagues developed machine learning methods to predict MASH and/or the extent of fibrosis using various noninvasive tests in patients with MASLD.3 The authors retrospectively analyzed data from 1970 biopsy-confirmed patients in the phase 3 MAESTRO-NASH and MAESTRO-NAFLD-1 trials who met the criteria of having 3 of 4 defined metabolic risk factors (diabetes/insulin resistance, hypertension, obesity/overweight, and dyslipidemia) and with evidence of hepatic steatosis and fibrosis from recent assessments.
Two outcomes were accessed: fibrosis stage, categorized as F0/F1, F2/F3, or F4; and the presence of MASH with F2/3 fibrosis, or at-risk MASH, which was defined as a baseline nonalcoholic fatty liver disease activity score (NAS) of 4 or higher and moderate to advanced fibrosis without cirrhosis. Models were developed using the random forest method for the full predictor model that incorporated 37 demographic and clinical variables and a lean model that included 20 to 24 routinely available parameters.
Among the 1970 patients included, 71.8% had moderate to advanced fibrosis and 74.7% had MASH (baseline NAS ≥4). Investigators reported that the machine learning models accurately predicted fibrosis stage and MASH with F2/F3 fibrosis. Using the full data model, the cross-validation mean area under the curve (AUC) for predicting the 3 fibrosis categories was 0.76 for predicting F0/F1, 0.75 for predicting F2/F3, and 0.94 for predicting F4. The investigators noted the particularly high mean AUC of 0.94 for discriminating patients with cirrhosis (F4). Using the best-performing lean model (Figure 1), the mean AUCs for predicting F0/F1, F2/F3, and F4 were 0.83, 0.78, and 0.89, respectively.
The mean AUC for predicting MASH with significant (stage F2/F3) liver fibrosis was 0.75 using the full model and 0.79 using the lean model. The sensitivity of the full model for predicting F4 fibrosis was 82.3% and the negative predictive value (NPV) was 98.1%. The lean model 1 was found to perform as well as the full model for sensitivity (72.5% vs 65.3%), specificity (73.7% vs 71.1%), positive predictive value (PPV) (78.3% vs 74.6%), and NPV (67.4% vs 61.2%). The factors that contributed most to predicting fibrosis stage were liver stiffness measures from imaging with magnetic resonance elastography (MRE) or vibration-controlled transient elastography (VCTE) and MASH composite algorithms such as FAST score (FibroScan measurements with aspartate aminotransferase levels), SAFE (Steatosis-Associated Fibrosis Estimator), and FIB-4 (Fibrosis-4) index.
The findings support potential use of machine learning models of noninvasive biomarkers to identifying patients with MASH and moderate to advanced liver fibrosis for clinical trial enrollment. However, the authors cautioned that there is a need for additional external and prospective validation of these tools before implementing them in clinical practice.
References
1. Ribeiro TCR, de Paula Boechat Soares V, Campos Fabri J, de Aragão Ramos LE. Noninvasive liver fibrosis assessment in metabolic dysfunction-associated steatotic liver disease. World J Hepatol. 2026;18(1):111534.
2. Sanyal AJ, Harrison SA, Ratziu V, et al. The natural history of advanced fibrosis due to nonalcoholic steatohepatitis: data from the simtuzumab trials. Hepatology. 2019;70(6):1913-1927.
3. Loomba R, Taub R, Labriola D, Priel J, Padmanabhan K, Kim WR. Machine learning models of noninvasive tests to predict MASH and fibrosis stage based on MAESTRO-NAFLD-1 and MAESTRO-NASH liver biopsies. Abstract P6074. Presented at: ACG 2026 Annual Scientific Meeting; October 9-14, 2026; Nashville, Tennessee.
Efruxifermin Induced a Rapid and Direct Antifibrotic Response That Was Sustained Through 96 Weeks in Phase 2b Trials of Participants With MASH and F2/F3 Fibrosis or Cirrhosis
Efruxifermin, an investigational fibroblast growth factor 21 analog, has been evaluated in two randomized phase 2b trials of patients with MASH. In the HARMONY trial in patients with moderate or severe (F2/F3) fibrosis, efruxifermin was associated with a significant improvement in fibrosis at Week 24 (the primary outcome) and at Week 96.1 In the SYMMETRY trial in patients with compensated cirrhosis (F4c), efruxifermin was not associated with a significant improvement in fibrosis at Week 36 (the primary outcome) but was associated with a significant improvement at Week 96.2
Sanyal and colleagues evaluated the time course of fibrosis improvement with efruxifermin in patients enrolled in the HARMONY and SYMMETRY trials.3
The HARMONY trial enrolled 126 patients with F2/F3 MASH who were randomly assigned to efruxifermin 28 mg (n=40), efruxifermin 50 mg (n=43), or placebo (n=43).1 The mean age of enrolled patients was 55 years; 62% of patients were female, 70% had type 2 diabetes, 16% were receiving a glucagon-like peptide-1 receptor agonist (GLP-1 RA), and the mean body mass index (BMI) was 38 kg/m2. Week 96 liver biopsies were available for 88 patients.
In the HARMONY trial, at Week 96, rates of MASH resolution without fibrosis worsening were 62% and 57% in the efruxifermin 28 mg and 50 mg arms and 24% in the placebo arm (P<.001 for each efruxifermin group vs placebo).3 Improvement in fibrosis of at least 1 stage without MASH worsening occurred in 46% of patients receiving efruxifermin 28 mg (12/26), including 10 patients (38%) with a sustained response at Weeks 24 and 96 and 2 patients (8%) with a new response at Week 96. In the efruxifermin 50-mg arm, 75% of patients (21/28) attained an improvement in fibrosis of at least 1 stage without MASH worsening (P<.001 vs placebo), including 11 patients (42%) with a sustained response at Weeks 24 and 96 and 10 patients (38%) with a new response at Week 96. In the placebo arm, 24% of patients had an improvement in fibrosis of at least 1 stage without MASH worsening at Week 96 (Figure 2, A).
The SYMMETRY trial enrolled 181 patients with F4c MASH who were randomly assigned to efruxifermin 28 mg (n=57), efruxifermin 50 mg (n=63), or placebo (n=61).2 The mean age of enrolled patients was 61 years; 67% were female, 80% had type 2 diabetes, 35% were receiving a GLP-1 RA, and the mean BMI was 36 mg/m2. Week 96 liver biopsies were available for 134 patients.
In the SYMMETRY trial, at Week 96, rates of MASH resolution without fibrosis worsening were 59% and 55% with efruxifermin 28 mg and 50 mg, respectively, and 18% in the placebo arm (P<.001 and P<.01, respectively). Improvement in fibrosis of at least 1 stage without MASH worsening occurred in 29% of patients receiving efruxifermin 28 mg (12/41), including 6 patients (14.6%) with a sustained response at Weeks 36 and 96 and 6 patients (14.6%) with a new response at Week 96. In the efruxifermin 50-mg arm, 39% of patients (18/46) attained an improvement in fibrosis of at least 1 stage without MASH worsening (P<.01 vs placebo), including 9 patients (19.5%) with a sustained response at Weeks 36 and 96 and 9 patients (19.5%) with a new response at Week 96. In the placebo arm, 15% of patients had an improvement in fibrosis of at least 1 stage without MASH worsening at Week 96 (Figure 2, B).
In both trials, efruxifermin was associated with rapid reductions in liver stiffness by VCTE that improved over time. In the HARMONY trial, the mean change in liver stiffness from baseline to Week 96 was +1% with placebo, -20% with efruxifermin 28 mg (P<.05 vs placebo), and -39% with efruxifermin 50 mg (P<.001 vs placebo). In the SYMMETRY trial, the mean change in liver stiffness from baseline to Week 96 was -10% with placebo, -19% with efruxifermin 28 mg, and -25% with efruxifermin 50 mg (P<.05 vs placebo).
Liver fibrosis as assessed using the Enhanced Liver Fibrosis (ELF) score also improved significantly and rapidly with efruxifermin. In both trials, improvements from baseline in mean ELF score were significantly greater with either efruxifermin dose compared with placebo at Week 24 and these differences were sustained at Week 96.
The most common adverse events across both trials were gastrointestinal, including diarrhea and nausea. Investigators noted that these were mild to moderate, transient, and generally occurred early in the treatment course. There were reports of increased appetite, but mean body weight did not generally change. Small but statistically significant reductions in bone mineral density at Week 96 were also reported, although investigators noted that the clinical relevance of these is to be determined, and the incidence of fractures was not increased. Markers of liver function and hemostasis were stable.
Efruxifermin was associated with improvements in MASH and fibrosis in patients with MASH and F2/F3 fibrosis or compensated cirrhosis. Most early responses were maintained for 96 weeks, and some patients without an early response gained a response by Week 96. Ongoing phase 3 trials evaluating efruxifermin in patients with MASH include the SYNCHRONY Histology trial in patients with F2/F3 fibrosis (NCT06215716) and the SYNCHRONY Outcomes trial in patients with compensated cirrhosis (NCT06528314).
References
1. Noureddin M, Frias JP, Neff GW, et al. Safety and efficacy of once-weekly efruxifermin versus placebo in metabolic dysfunction-associated steatohepatitis (HARMONY): 96-week results from a multicentre, randomised, double-blind, placebo-controlled, phase 2b trial. Lancet. 2025;406(10504):719-730.
2. Noureddin M, Rinella ME, Chalasani NP, et al. Efruxifermin in compensated liver cirrhosis caused by MASH. N Engl J Med. 2025;392(24):2413-2424.
3. Sanyal AJ, Behling C, Bedossa P, et al. Efruxifermin induced a rapid and direct antifibrotic response that was sustained through 96 weeks in phase 2b trials of participants with MASH and F2/F3 fibrosis or cirrhosis. Abstract 28. Presented at: ACG 2026 Annual Scientific Meeting; October 9-14, 2026; Nashville, Tennessee.
Reducing Cardiovascular Risk in Patients With MASH Independent of Baseline Based on Lipoprotein(a) and Low-Density Lipoprotein Lowering by Resmetirom
MASH is associated with a high prevalence of cardiovascular (CV) disease driven partially by atherogenic dyslipidemia.1,2 Resmetirom, a thyroid hormone receptor-β agonist, is indicated in conjunction with diet and exercise for treatment of adults with MASH with moderate to advanced liver scarring without cirrhosis.3 In the randomized phase 3 MAESTRO-NASH and MAESTRO-NAFLD-1 trials, resmetirom was associated with significant reductions in atherogenic lipid and lipoprotein (Lp) levels, including low-density lipoprotein cholesterol (LDL-C), apolipoprotein B (ApoB), and Lp(a), in patients with MASH.4,5
Bansal and colleagues presented the results of an analysis of the effects of resmetirom on lipid-related parameters and safety based on use of statins at baseline among patients in the MAESTRO-NASH trial.6 The analysis included patients with fibrosis stages F1B, F2, and F3 who were randomly assigned to resmetirom 80 mg, resmetirom 100 mg, or placebo.
Because the blinded MAESTRO-NASH trial is still ongoing, a pharmacokinetic analysis was not completed on this population. Instead, pharmacokinetics were analyzed for patients in MAESTRO-NAFLD-1, a completed 52-week trial that enrolled a similar population of patients with MASLD without cirrhosis.5
Of the 473 patients enrolled in MAESTRO-NASH, 49% were receiving statins at baseline—most commonly atorvastatin—and 26.9% were receiving high-intensity statin therapy. Compared with patients not receiving statins, those receiving statins were more likely to have diabetes (79.5% vs 55.0%), dyslipidemia (100% vs 43.8%), and advanced fibrosis (F3: 64.3% vs 56.6%).
In a modified intention-to-treat analysis, the dual primary endpoints from MAESTRO-NASH at Week 52 of MASH resolution (defined as a reduction in NAS by ≥2 points) and improvement in fibrosis (by ≥1 stage with no worsening of NAS) were similar regardless of baseline statin use, with both outcomes showing a nominally significant improvement with resmetirom at 80-mg and 100-mg doses vs placebo. Specifically, the Week 52 MASH resolution rates were 29% in patients receiving a statin and 35% in patients not receiving a statin in the resmetirom 80-mg arm; 34% and 43%, respectively, in the resmetirom 100-mg arm; and 10% and 12%, respectively, in the placebo arm. Week 52 fibrosis improvement rates were 26% in patients receiving a statin vs 33% in patients not receiving a statin in the resmetirom 80-mg arm, 28% and 39%, respectively, in the resmetirom 100-mg arm, and 11% and 22%, respectively, in the placebo arm (Figure 3).
Resmetirom was also associated with numerically greater reductions in liver fat assessed by magnetic resonance imaging–proton density fat fraction compared with placebo in both dose groups at Week 52, and reductions in LDL-C, ApoB, triglycerides, and Lp(a) at Week 24, regardless of baseline statin use. Overall, 44.4% of patients on statins at baseline with an LDL-C level of 70 mg/dL or higher who were assigned to resmetirom 100 mg attained an LDL-C reduction of less than 70 mg/dL; 36.3% and 37.5% of patients with an elevated Lp(a) level at baseline shifted to a lower level (≥30 or ≥50 mg/dL to <30 or <50 mg/dL).
Among patients in the highest quartile of CV risk based on Lp(a) at baseline (>18.75 mg/dL), resmetirom was associated with a significant reduction in Lp(a) regardless of baseline statin use. Investigators reported no significant statin-related adverse events in the trial and concluded that the findings support concomitant use of resmetirom and statins in these patients.
References
1. Sanyal AJ, Husain M, Diab C, et al. Cardiovascular disease in patients with metabolic dysfunction-associated steatohepatitis compared with metabolic dysfunction-associated steatotic liver disease and other liver diseases: a systematic review. Am Heart J Plus. 2024;41:100386.
2. Rinella ME, Neuschwander-Tetri BA, Siddiqui MS, et al. AASLD practice guidance on the clinical assessment and management of nonalcoholic fatty liver disease. Hepatology. 2023;77(5):1797-1835.
3. Rezdiffra [package insert]. West Conshohocken, PA: Madrigal Pharmaceuticals Inc; 2026.
4. Harrison SA, Bedossa P, Guy CD, et al; MAESTRO-NASH Investigators. A phase 3, randomized, controlled trial of resmetirom in NASH with liver fibrosis. N Engl J Med. 2024;390(6):497-509.
5. Harrison SA, Taub R, Neff GW, et al. Resmetirom for nonalcoholic fatty liver disease: a randomized, double-blind, placebo-controlled phase 3 trial. Nat Med. 2023;29(11):2919-2928.
6. Bansal M, Taub R, Labriola D, et al. Reducing cardiovascular risk in patients with MASH independent of baseline based on lipoprotein(a) and low-density lipoprotein lowering by resmetirom. Abstract P6267. Presented at: ACG 2026 Annual Scientific Meeting; October 9-14, 2026; Nashville, Tennessee.
Semaglutide and Tirzepatide Are Associated With Lower Risk of Cirrhosis and Liver Decompensation in MASH
Several recent randomized clinical trials have demonstrated histologic benefit with incretin-based therapy in patients with MASH and moderate or advanced (F2 or F3) fibrosis.1,2 In the ESSENCE trial, semaglutide was associated with significant improvements over placebo at Week 72 in multiple liver histologic outcomes, including rates of resolution of steatohepatitis without worsening of fibrosis (62.9% vs 34.3%; P<.001), reduction in liver fibrosis without worsening of steatohepatitis (36.8% vs 22.4%; P<.001), and a combined outcome of resolution of steatohepatitis and reduction in liver fibrosis (32.7% vs 16.1%; P<.001).1 In the SYNERGY-NASH trial, tirzepatide was associated with significant improvements over placebo at Week 52 in rates of MASH resolution without worsening of fibrosis (44% with tirzepatide 5 mg, 56% with tirzepatide 10 mg, and 62% with tirzepatide 15 mg, compared with 10% with placebo; P<.001 for all 3 comparisons) and higher rates of improvement of at least 1 fibrosis stage without worsening of MASH (55%, 51%, and 51%, respectively, with tirzepatide, compared with 30% with placebo).2 In both trials, gastrointestinal adverse events were more common with incretin-based therapy vs placebo.
Nyancho and colleagues evaluated the association between incretin-based therapy and liver outcomes in a real-world cohort of patients with MASH by using electronic health record data.3 The cohort included 12,017 adults who had received semaglutide or tirzepatide and a control group of 47,343 adults who had not received either agent. In a 1:1 propensity matched analysis controlled for demographics, clinical variables, and treatment, outcomes were compared in 3558 patients receiving incretin-based therapy with semaglutide or tirzepatide and a control group of 3358 patients receiving neither agent.
The primary outcome, a composite of cirrhosis or hepatic decompensation, occurred at a significantly lower rate in the incretin-based therapy group vs the control group (hazard ratio [HR], 0.53; 95% CI, 0.44-0.64) (Figure 4). In terms of individual outcomes, incretin therapy was associated with significant reductions over the control group in rates of incident cirrhosis (4.0% vs 12.8%; odds ratio [OR], 0.29; 95% CI, 0.24-0.35) and liver decompensation (1.7% vs 9.8%; OR, 0.16; 95% CI, 0.12-0.22).
The researchers cautioned that there are limitations to this study, including its retrospective, observational design, residual confounding factors, an incomplete data capture, adherence bias considerations, and the presence of duplicate patients. However, they concluded these real-world findings do provide further support to the role of incretin-based therapies in the management of MASH.
References
1. Sanyal AJ, Newsome PN, Kliers I, et al; ESSENCE Study Group. Phase 3 trial of semaglutide in metabolic dysfunction-associated steatohepatitis. N Engl J Med. 2025;392(21):2089-2099.
2. Loomba R, Hartman ML, Lawitz EJ, et al; SYNERGY-NASH Investigators. Tirzepatide for metabolic dysfunction-associated steatohepatitis with liver fibrosis. N Engl J Med. 2024;391(4):299-310.
3. Nyancho D, Pfeir KA, Greer K. Semaglutide and tirzepatide are associated with lower risk of cirrhosis and liver decompensation in MASH. Abstract P6080. Presented at: ACG 2026 Annual Scientific Meeting; October 9-14, 2026; Nashville, Tennessee.
Efruxifermin in MASH With F2/F3 Fibrosis: Post Hoc Analysis Indicated Early Histologic Features of Regression Were Associated With Week 96 Fibrosis Improvement
The randomized, double-blind, phase 2b HARMONY trial evaluated the efficacy and safety of efruxifermin compared with placebo in 747 patients with MASH and F2/F3 fibrosis.1 The trial met its primary endpoint, showing improvement in liver fibrosis (decrease in ≥1 fibrosis stage) without worsening of MASH at Week 24 in a significant proportion of patients in the efruxifermin 28-mg group (39%; risk ratio, 2.3; P=.025) and efruxifermin 50-mg group (41%; risk ratio, 2.2; P=.036) over patients in the placebo group (20%).
Additional follow-up continued to show efficacy improvements with efruxifermin over placebo at 96 weeks. Among 88 patients with biopsies at Week 96, the proportion of patients with fibrosis improvement (≥1-stage) without MASH worsening was 46% in the efruxifermin 28-mg group (n=26), 75% in the efruxifermin 50-mg group (n=28), and 24% in the placebo group (n=34).2 Mild to moderate gastrointestinal adverse events, including diarrhea and nausea, were reported with efruxifermin more than placebo.
Behling and colleagues performed a post hoc analysis of the HARMONY trial to identify detailed histologic features of fibrosis regression.3 Included in the analysis were patients with liver biopsies at baseline and at Week 24. For the analysis, 2 pathologists independently reassessed biopsy images and generated a combined score for 6 predetermined features of regression at baseline and Week 24. Features of regression included: regenerative changes, interrupted septa, isolated thick collagen fibers, thick and thin collagen, large lipogranulomas, and isolated arteries. A change in combined score from baseline to Week 24 was calculated for each patient. An increase in combined score indicated early improvement for that feature at Week 24. Subsequently, this score was analyzed in patients with histologic responses at Week 96. In this subset of patients, the number of features with early improvement was numerically higher in patients receiving efruxifermin 28 mg or 50 mg, with a median of 3 features with early improvement, than in patients receiving placebo, with a median of 1 feature with early improvement. The most frequently observed changes at Week 24 among patients with a Week 96 histologic response were changes related to collagen morphology.
In a combined analysis of 33 patients receiving either dose of efruxifermin with a Week 96 histology response, there was a strong correlation between improvements in interrupted septa and improvements in thick/thin collagen at Week 24. In contrast, improvement in large lipogranulomas was not associated with improvements in other features assessed.
Investigators also reported an association between the number of early features with improvement at Week 24 and Week 96 response in all patients. In all measurements, greater improvements at Week 96 were noted in the patients receiving efruxifermin (28 mg or 50 mg) who showed early improvement in more features than in patients receiving placebo. Greater improvement was noted in those with more features (4-6 features vs 0-3 features). Measurements included change in NAS at Week 24 and at Week 96 and change in markers of fibrosis, such as ELF, liver stiffness (Figure 5), and Pro-C3, and markers of liver injury (alanine transaminase [ALT] and aspartate transaminase [AST]) at Week 96 in patients receiving efruxifermin vs placebo.
Clinical trials are continuing to evaluate efruxifermin in different patient populations, including the SYNCHRONY Histology trial4 in patients with MASH and F2/F3 fibrosis and the SYNCHRONY Outcomes trial5 in patients with compensated cirrhosis.
References
1. Harrison SA, Frias JP, Neff G, et al; HARMONY Study Group. Safety and efficacy of once-weekly efruxifermin versus placebo in non-alcoholic steatohepatitis (HARMONY): a multicentre, randomised, double-blind, placebo-controlled, phase 2b trial. Lancet Gastroenterol Hepatol. 2023;8(12):1080-1093.
2. Noureddin M, Frias JP, Neff GW, et al. Safety and efficacy of once-weekly efruxifermin versus placebo in metabolic dysfunction-associated steatohepatitis (HARMONY): 96-week results from a multicentre, randomised, double-blind, placebo-controlled, phase 2b trial. Lancet. 2025;406(10504):719-730.
3. Behling C, Bedossa P, Rinella ME, et al. Efruxifermin in MASH with F2/F3 fibrosis: post hoc analysis indicated early histologic features of regression were associated with week 96 fibrosis improvement. Abstract P1675. Presented at: ACG 2026 Annual Scientific Meeting; October 9-14, 2026; Nashville, Tennessee.
4. ClinicalTrials.gov. A study evaluating efruxifermin in subjects with non-cirrhotic nonalcoholic steatohepatitis (NASH)/metabolic dysfunction-associated steatohepatitis (MASH) and fibrosis. https://clinicaltrials.gov/study/NCT06215716. Accessed October 23, 2026.
5. ClinicalTrials.gov. A study evaluating efruxifermin in subjects with compensated cirrhosis due to NASH/MASH. https://clinicaltrials.gov/study/NCT06528314. Accessed October 23, 2026.