DAY 1: Clinical Trial Updates Presented at the ACG President’s Plenary

Dr. Evan S. Dellon, from the University of North Carolina, Chapel Hill, NC, presented the 24-week results of the REMODEL trial, a phase 4 randomized, placebo-controlled trial of dupilumab, a dual inhibitor of IL-4 and IL-13, in a population with severe eosinophilic esophagitis. Patients were randomized 2:1 to dupilumab (n=46) or placebo (n=23). The primary outcome was absolute change from baseline in esophageal distensibility plateau by FLIP. Dupilumab improved esophageal distensibility by 1.28 mm, with no change in the placebo group. Patients with a baseline distensibility plateau below 17 mm responded more strongly: 60% improved with dupilumab versus 15% with placebo, compared with 40% versus 10% in the overall population. Endoscopic inflammatory and fibrostenotic scores and histologic scores also fell significantly with dupilumab and were unchanged with placebo; these reductions were similar regardless of baseline diameter. Several post-hoc analyses explored thresholds of disease progression, and safety data were presented. The full results are expected to be published shortly in the American Journal of Gastroenterology. An open-label follow-up phase is ongoing.

Dr. Jodie A. Barkin, from the University of Miami Miller School of Medicine in Miami, FL, presented the results from the PACT-CP Registry of adults with exocrine pancreatic insufficiency. This 5-year prospective observational registry across 18 US sites characterized symptoms, pancreatic enzyme replacement therapy (PERT) use, quality of life, and treatment satisfaction in 190 patients with chronic pancreatitis or recurrent acute pancreatitis, with a median follow-up of 24 months; at least three-quarters of patients reached 18 months of follow-up. The cohort was predominantly non-Hispanic White, with a median age of 60 years, and most had idiopathic or toxic-metabolic disease that was most often diagnosed based on 2 or more symptoms or test results. Diagnosis was often delayed, averaging 3 years, and more than 10% of patients waited 5 years or longer. Chronic abdominal pain was common, and more than half had been on PERT for at least a year at enrollment. PERT use remained stable over time, but only 40% of patients were taking PERT as recommended by guidelines. Despite this, 80% of patients reported symptom improvement and treatment satisfaction held steady at about 70%. Only 25% had to restrict their diet all the time. Quality of life, measured with the SF-12, was lower than normal but stable, and anxiety declined modestly with treatment. Healthcare provider visits stayed stable while the number of acute care visits fell. The investigators identified several priorities for future research: understanding the reasons for diagnostic delays, improving PERT use, improving both physical and mental quality of life, and addressing ongoing abdominal pain and psychosocial burden.

Dr. Mitchell Shiffman from the Liver Institute of Virginia, in Richmond, VA, presented results from the VISTAS trial of volixibat, an ileal bile acid transporter (IBAT) inhibitor, in patients with primary sclerosing cholangitis (PSC) and moderate-to-severe pruritus. Blocking IBAT interrupts bile acid reabsorption, which has been shown to increase fecal bile acid excretion, lower serum bile acids and cholesterol, and relieve symptoms including pruritus. The trial randomized patients 12 years of age and older to volixibat 20 mg (n=54) or placebo (n=57) for 28 weeks, followed by an open-label extension. The primary endpoint was change in pruritus from baseline, and secondary endpoints included quality of life, change in serum bile acids and alkaline phosphatase (ALP), and safety. Mean age was 45 years and mean time since diagnosis was 8–9 years. Baseline pruritus scores were similar between arms, and volixibat met the primary endpoint, reducing pruritis scores by 2.72 units versus 1.08 units with placebo, a difference that was both statistically and clinically significant. The response was rapid, beginning within 2 weeks, plateauing at 4 weeks, and remaining stable thereafter. Fatigue and sleep disturbance also improved at week 28. Changes in serum bile acids and ALP were not significant, with high variability. Treatment-emergent adverse events were slightly more frequent with volixibat but were not severe, and no serious adverse events occurred in the treatment arm. Discontinuations were higher with volixibat, mainly because of diarrhea. The most common adverse events were diarrhea, abdominal pain, nausea, and vitamin D deficiency, all known effects of IBAT inhibition. One death occurred in the placebo group. VISTAS is the largest study of an IBAT inhibitor in PSC and the first randomized controlled trial to assess symptom burden in this population.

Dr. Eden Koo from the University of Michigan, Ann Arbor, MI, reported 4-week results of a prospective, single-arm survey-based observations study of an oral enzyme blend targeting fructans, galactooligosaccharides, and lactose in patients with irritable bowel syndrome (IBS). The blend is taken as roughly 0.5 g of powder sprinkled on food. A total of 442 patients were assessed at baseline and 4 weeks. Participants were a mean of 50 years age and predominantly female, represented a range of IBS phenotypes, and nearly half were already following a low-FODMAP diet. The primary outcome was change in the IBS Symptom Severity Score (IBS-SSS; range 0–500, with higher scores indicating more severe symptoms, across 5 domains). Secondary outcomes included PROMIS T-scores for belly pain and bloating, which are scaled to a population mean of 50 with a standard deviation of 10, and dietary satisfaction on a 5-point scale. IBS-SSS fell by a mean of 102.4 points, with significant decreases across all 5 domains, and more than two-thirds of patients achieved a clinically meaningful response, defined as a reduction of more than 90 points. PROMIS T-scores for belly pain and bloating decreased by about 7 points, with roughly 40% of patients reporting improvement, and dietary satisfaction improved significantly, by 0.28 points. Prespecified sensitivity analyses showed consistent improvements regardless of IBS subtype and, among patients on a low-FODMAP diet, regardless of diet phase. Strengths include a large, heterogeneous cohort, while limitations include the absence of a placebo comparator, short follow-up, and reliance on survey data. Targeted enzyme supplementation may offer a less restrictive adjunct for IBS management, potentially allowing patients to liberalize their diet while maintaining symptom control. Randomized controlled trials are needed to confirm these findings.

Dr. Eugenia Shmidt from the University of Minnesota in Minneapolis, MN, presented a prospective study of fertility treatment use and outcomes in women with inflammatory bowel disease (IBD) from the WIsDoM cohort. Fertility in IBD is thought to be impaired by pelvic adhesions, including fallopian tube adhesions and scarring, and by surgery, particularly during periods of active disease. The study asked how fertility treatments are being used in this population, how successful they are, and which factors are associated with outcomes. Investigators examined women undergoing intrauterine insemination (IUI), egg retrieval alone, or IVF with embryo transfer; pregnancy outcomes at 12 weeks were measured only in those undergoing IUI or IVF with embryo transfer, and results were analyzed at both the patient and cycle levels. Of 67 women who underwent fertility treatment, 52 actively attempted pregnancy by IUI or IVF with embryo transfer and 15 underwent egg retrieval only. Women using assisted reproductive technology were about 33 years old and, compared with other cohort participants, had attempted pregnancy for longer, were more likely to have had prior fertility treatment and pelvic surgery, and more often had active disease within 3 months of treatment. Across 148 treatment events in the 52 women attempting pregnancy, 42% required multiple attempts and 73% ultimately achieved pregnancy. Success varied sharply by modality: 38% of women undergoing IUI alone became pregnant, compared with 85% of those undergoing embryo transfer. Per attempt, 10% of IUI cycles and 73% of embryo transfers resulted in pregnancy, compared with 10–15% and 45%, respectively, in the general population. After adjustment for surgical history and other covariates, the odds of pregnancy doubled, and prior IUI was not associated with success in subsequent embryo transfer. The investigators hypothesize that pelvic adhesions in women with IBD may impair IUI while leaving IVF relatively unaffected, making embryo transfer the more effective approach; they plan to collect hysterosalpingography data to test this hypothesis.

Dr. Mikin Patel of the Joint & Vascular Institute in Libertyville, IL, presented preliminary results from the prospective HemorrhoIdaL EmboliSation (PILES) registry. Hemorrhoidal artery embolization, first performed in 1995 and grounded in the vascular theory of hemorrhoid development, was initially limited by post-procedural pain; the modern interventional radiology approach, known as the emborrhoid technique, uses 80-micron spheres and coils for embolization. It is an outpatient procedure with no sedation, no food or medication holds, and no transrectal intervention. Prospective data remain sparse, outcomes vary across studies, and the largest retrospective series includes only 170 patients; the single-center PILES registry was designed to address this gap. Patients with symptomatic bleeding from internal hemorrhoids underwent a modified emborrhoid procedure, with a primary outcome defined of greater than 50% reduction in bleeding. Over 9 months of enrollment, 136 patients were treated with 97% technical success, 65% clinical success, adverse events in 1.5%, and 70% patient satisfaction, with outcomes improving over the course of the registry. Clinical success fell within the range reported in prior studies despite stringent inclusion criteria. Limitations include the single-center design, lack of technique standardization, and the need to assess durability and recurrence over time. Planned analyses include subgroup comparisons, effects on non-bleeding symptoms, and comparison with rubber band ligation. Although well established in vascular practice, the emborrhoid approach remains novel in gastroenterology.

Dr. Kira L. Newman of the University of Michigan, Ann Arbor, MI, presented her work demonstrating that inulin restores colonization resistance against Clostridioides difficile in the microbiomes of patients with ulcerative colitis (UC). Microbiome dysfunction in UC manifests clinically as an increased burden of C. difficile infection and persists despite anti-inflammatory treatment. Inulin, a prebiotic fiber derived from chicory root, has been shown to reduce C. difficile in infants, and the study aimed to identify microbiome features associated with susceptibility and test whether inulin could reverse them. Investigators used a miniature bioreactor array, or "minicolon," consisting of anaerobic chambers that cultivate patient microbiomes over time with an 8-hour turnover of media and waste, using media low in fermentable carbohydrates and supplemented with bile acids and hemin to mimic the colonic environment. In this system, healthy microbiomes cleared C. difficile after challenge more effectively than UC microbiomes, with differences in beta diversity involving taxa responsible for short-chain fatty acid and bile acid metabolism and corresponding metabolomic differences, including in 3-hydroxybutyrate and deoxycholic acid. When high-dose and standard-dose inulin were compared, the higher dose produced microbiomes resistant to C. difficile colonization, with increased short-chain fatty acid levels, particularly butyrate, along with notable changes in bile acids. Further work in the bioreactor system indicated that butyrate acts on vegetative growth of C. difficile, and the investigators noted that differences in inulin chain length, such as shorter chicory-derived versus longer agave-derived inulin, may have differential effects.

Dr. Aasma Shaukat of NYU Langone Health in New York, NY, presented a cluster-randomized clinical trial of a multilevel intervention to improve colorectal cancer screening at federally qualified health centers (FQHCs) in New York City. Screening rates remain low among underserved populations receiving care at FQHCs, and this 5-year study tested tailored mailed fecal immunochemical test (FIT) outreach in an area of Brooklyn with high colorectal cancer rates. Patients at 5 FQHCs were cluster-randomized at the primary care provider and clinic level, with screening completion at 5 months after randomization as the primary outcome; the study also explored barriers and facilitators through interviews and assessed cost-effectiveness. Of 2007 patients, 1116 were assigned to usual care and 891 to mailed outreach, which included 2 letters with a FIT kit, a phone call or voicemail, text message reminders, and a final letter encouraging completion and referring patients to their primary care provider. Patients who completed testing received their results, and those with positive results received navigation and help with transportation to follow-up colonoscopy. Participants had a mean age of 57 years, were predominantly Black and Hispanic, were largely insured through Medicaid, and many were unemployed. Outreach increased screening completion from 23% to 35.9% and follow-up colonoscopy completion from 54% to 92%. Older age and Black race were associated with higher adherence, while unemployment was associated with lower screening rates independent of location and insurance status. Patients identified navigation as the most helpful component, particularly arranging for an escort rather than transportation alone. Because sustainability was a major focus, the investigators calculated the cost per additional person screened and downstream costs, with the goal of presenting these data to health system leadership to justify adoption. Dr. Shaukat noted that the dismantling of the Agency for Healthcare Research and Quality (AHRQ) has eliminated funding for continued research.

Dr. Vignesh Krishnan Nagesh, from Mayo Clinic in Rochester, MN, presented results from a retrospective study comparing a random forest machine learning model with a large language model (LLM) for automated detection of colorectal neoplasia across 5,000 colonoscopies. Adenoma and sessile serrated lesion detection rates (ADR and SSLDR) are key quality measures for precursor lesions but are not consistently reported, prompting interest in automated approaches that can calculate these rates, flag patients, and recommend next steps. LLMs, however, carry risks of hallucination and high computational cost. Investigators compared a random forest model, a supervised learning method built on decision trees, with an LLM trained to interpret and generate human language, applying both to the 5,000 colonoscopy results and manually adjudicating discordant cases. The two models disagreed in about 2.5% of cases, and on manual review, the random forest model had a higher false-positive rate than the LLM. A cost-effectiveness analysis is underway. The investigators noted that broader implementation may introduce model drift and that different models may be needed for different health systems and practices, as nonstandard pathology report language may affect LLM performance. A prospective study is needed to validate these findings.

Dr. Mai Khalaf from Baylor College of Medicine, Houston, TX, presented results from a multicenter randomized controlled trial comparing traction-assisted endoscopic submucosal dissection (T-ESD) with conventional ESD (C-ESD) for colorectal lesions. ESD allows en bloc resection but is technically challenging and available at a limited number of centers. Traction devices, such as the Sure-Trac system used in this study, can help maintain a clear view of the dissection plane. The trial asked whether traction makes ESD easier without extending procedure time. Patients at 3 centers were randomized 1:1 to C-ESD or T-ESD with Sure-Trac. All colorectal lesions were eligible except pedunculated lesions, appendiceal lesions, and previously resected lesions, and crossover was permitted when issues arose during dissection. A total of 150 patients were randomized, 75 to each arm; after 21 crossovers, 61 patients completed C-ESD and 68 completed T-ESD as assigned. Most lesions were in the right colon, and lesion size was comparable between groups. Dissection speed did not differ significantly between the two groups (C-ESD: 22 mm²/min vs T-ESD: 19.6 mm²/min), and total procedure time and curative resection rates were also comparable. T-ESD required fewer intraprocedural rescue maneuvers and was associated with less post-procedural pain, while other adverse events were similar. Of the 21 patients who crossed over, 14 switched from C-ESD to T-ESD because of visualization problems, and 7 switched from T-ESD to C-ESD because of concerns about en bloc removal or adverse events such as bleeding. The investigators concluded that T-ESD is safe and effective and improves visualization, with fewer patients needing to switch to C-ESD, although it did not reduce resection time or increase dissection speed.