This Week in Gastroenterology — Sep 29, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Gastroenterology. This week we're covering 10 notable papers spanning liver disease — from the first joint European guidance on drug therapy for steatohepatitis to a head-to-head trial of a blood test against ultrasound for liver cancer — plus positioning of advanced therapies in inflammatory bowel disease, and a set of papers on cancer surveillance and post-treatment symptom burden. Let's dive in.

We start in the liver, where the therapeutic landscape has shifted faster than practice has. In the Journal of Hepatology, a joint guidance from the European liver, diabetes and obesity societies sets out how resmetirom and semaglutide should be positioned in metabolic dysfunction-associated steatohepatitis [1]. The framework is built on non-invasive tests rather than biopsy — transient elastography, magnetic resonance elastography, or the enhanced liver fibrosis score can identify the patients at high risk of major adverse liver outcomes in whom pharmacotherapy may be considered. The recommendation is for adults with non-cirrhotic steatohepatitis and fibrosis stage two or three, always alongside structured lifestyle intervention and comorbidity management rather than instead of it. The choice between agents is framed around the cardio-renal-metabolic profile: the panel suggests semaglutide at two point four milligrams weekly may be preferred where there is obesity, type two diabetes, or established cardiovascular disease, while resmetirom may be prioritised in people without obesity or those already on, or intolerant of, a GLP-1 receptor agonist. Two caveats are worth carrying forward. The guidance asks for nutritional and sarcopenia assessment during semaglutide therapy, and it explicitly states that in adults with cirrhosis from steatohepatitis, both drugs appear safe but should not be used as liver-targeted treatment until data on liver-related benefit exist. Outcome trials for both agents are still running.

Staying in the liver, two papers take on the weakest link in cirrhosis care — finding cancer early. Also in the Journal of Hepatology, the ALTUS study, reported by John and colleagues, prospectively compared a multi-target blood test against standard-of-care ultrasound in adults with cirrhosis or chronic hepatitis B, with cross-sectional imaging as the reference standard [3]. Among more than two thousand evaluable participants, thirty-seven cancers were found, twenty-seven of them early stage. The blood test picked up two thirds of early-stage cancers; ultrasound picked up about a fifth. That paired difference of roughly forty-four percentage points is large, and the test met its non-inferiority margin and then some. But the second co-primary outcome was missed: specificity came in just under eighty-two percent, short of the prespecified target, against ultrasound specificity of close to ninety-nine percent. In a screening population that is a real cost — a substantial number of false positives feeding into diagnostic imaging. The authors frame this as supporting further evaluation of blood-based screening, not as a replacement, and that reading is the honest one. A narrative review in Alimentary Pharmacology and Therapeutics by Park and colleagues arrives at a complementary and more conservative position for hepatitis B specifically [4]. It argues for separating risk prediction from early detection, notes that patients with an estimated annual cancer risk below zero point two percent generally do not require routine surveillance, and holds that guideline-eligible patients should continue six-monthly ultrasound with or without alpha-fetoprotein. Crucially, it points out that much of the evidence for GALAD, multi-target blood tests, and abbreviated magnetic resonance imaging comes from mixed-aetiology or non-hepatitis-B cohorts, so hepatitis-B-specific prospective validation is still missing and these tools remain adjunctive or investigational.

A third liver paper addresses something far more mundane and probably more immediately actionable. In Clinical Gastroenterology and Hepatology, Mahmud and colleagues analysed more than sixty thousand outpatient paracenteses in over four thousand patients with cirrhosis across the Veterans Health Administration [5]. Diagnostic fluid studies for spontaneous bacterial peritonitis were sent in only about a third of procedures, with wide variation by region, provider type and facility. After weighting, patients whose fluid was tested at every paracentesis had roughly a third lower risk of seven-day hospitalisation compared with those tested about half the time, along with lower thirty-day mortality, and the association was strongest in those with higher Child-Turcotte-Pugh scores. They were also more likely to receive outpatient antibiotics for peritonitis. The authors are appropriately careful: this is observational, and differences in patient selection and care delivery could explain part of the signal. Interestingly, never testing also looked better than intermediate testing, which hints at unmeasured selection. It supports evaluating standardised, risk-adapted testing protocols rather than settling the question.

Turning to inflammatory bowel disease, where the question is no longer whether advanced therapies work but how to rank them. In Clinical Gastroenterology and Hepatology, Hupé and colleagues used a nationally representative United States claims database covering more than seven thousand patients with ulcerative colitis and over nine thousand treatment episodes, comparing infliximab, adalimumab, vedolizumab, ustekinumab, tofacitinib and upadacitinib [2]. After propensity-score weighting, upadacitinib was associated with roughly forty percent lower risk of treatment failure — a composite of disease-related hospitalisation, surgery, or a new corticosteroid prescription — compared with every other agent tested, and it had the highest proportion of quarters in corticosteroid-free disease stability, at about seventy percent. Vedolizumab modestly outperformed both anti-TNF agents. The findings held in patients with and without prior advanced therapy exposure. The caveat is size and timing: only one hundred and fifty-four patients started upadacitinib, the newest drug in the cohort, and claims data cannot capture endoscopic activity or disease severity directly. This is supportive real-world evidence, not a head-to-head trial. Alongside it, a review in Inflammatory Bowel Diseases by Gisbert and Chaparro addresses the malignancy question that inevitably follows [8]. Their reading of the evidence is broadly reassuring: anti-TNF monotherapy does not appear to increase malignancy risk, with residual uncertainty for lymphoma and melanoma, and the clearest excess lymphoma risk attaching to concomitant thiopurines. Vedolizumab and ustekinumab look favourable, and early data on selective anti-interleukin-23 agents are encouraging. For JAK inhibitors and S1P receptor modulators the evidence remains limited, which the authors say supports a cautious approach in older or higher-risk patients and those with prior cancer — a note that sits directly against the upadacitinib effectiveness signal and is worth holding in mind together.

On the oncology side, two papers deal with who benefits and who gets missed. In the Journal of Hepatology, Rimini and colleagues assembled more than thirteen hundred patients with advanced biliary tract cancer treated with first-line cisplatin, gemcitabine and durvalumab across fifty-five centres in twelve countries [7]. Molecular testing was done in most patients and found actionable tier-one alterations in about twenty-two percent — but only about a third of those patients ever received a matched agent. Among those treated after progression, matched targeted therapy was associated with median overall survival of nearly twenty-five months versus about seventeen, roughly halving the risk of death, with the benefit concentrated in second line; no significant benefit was demonstrated in third line. This is retrospective with small molecular subgroups, so the authors explicitly call it hypothesis-generating, but the access gap it documents is hard to argue with. Meanwhile, in Clinical Gastroenterology and Hepatology, Ausina-Poüs and colleagues pooled thirty-five studies and more than thirty-three thousand patients with Lynch syndrome [10]. Pooled incidence was about one gastric cancer per thousand person-years and about two duodenal cancers per thousand, with risk varying by mismatch repair gene — highest gastric risk with MSH2, very low with MSH6. Upper endoscopic surveillance was associated with lower observed gastric cancer risk but no difference in duodenal cancer, though that comparison rests on only two observational studies per outcome. The authors are explicit that this does not establish reduced cancer incidence or a survival benefit.

Finally, two papers on the symptoms that persist after the disease is treated. In The Lancet Gastroenterology and Hepatology, Olesen and colleagues review pain in chronic pancreatitis, which affects eighty to ninety percent of patients and arises from overlapping nociceptive, neuropathic and nociplastic mechanisms [6]. Opioid use remains highly prevalent despite its harms; among non-opioid options, gabapentinoids carry the strongest evidence, while cognitive behavioural therapy, antioxidants and dietary strategies have thinner support. Notably, they flag that endoscopic and surgical interventions relieve obstructive disease but the literature lacks sham-controlled trials, which makes true effect size hard to judge. And in Gut, the EUROSTAR consensus from Barman and colleagues, developed through a modified Delphi process with experts from ten European countries, produced eight statements and thirteen recommendations on symptom monitoring and service provision after major upper gastrointestinal cancer surgery, along with diagnostic criteria and management pathways for common postoperative conditions and sixteen research priorities [9]. It is expert consensus rather than trial evidence, but it fills a genuine standardisation gap in survivorship care.

If you only have time for one paper this week, make it the joint European guidance on resmetirom and semaglutide in steatohepatitis [1]. It is the first coordinated attempt to answer the question every hepatology and metabolic clinic is now facing — which patient, which drug, and when to stop — and it does so using non-invasive fibrosis testing rather than biopsy as the entry point.

Here is what this week's evidence adds up to in gastroenterology. First, drug therapy for steatohepatitis now has a society-endorsed framework anchored on fibrosis stage two to three by non-invasive testing, with explicit advice against using either agent as liver-targeted therapy in cirrhosis until outcome data exist. Second, blood-based liver cancer detection substantially outperformed ultrasound on early-stage sensitivity in a prospective head-to-head study, but missed its specificity target, and the hepatitis B literature still lacks disease-specific prospective validation — so imaging-based surveillance remains the standard for now. Third, in ulcerative colitis, real-world claims data place upadacitinib at the lowest risk of treatment failure among advanced therapies, though the upadacitinib group was small and the long-term malignancy evidence for JAK inhibitors remains limited. Fourth, in advanced biliary tract cancer, matched targeted therapy after chemoimmunotherapy failure was associated with markedly longer survival in second line but not third, while most patients with actionable alterations never received a matched drug — an access problem as much as a biology one. And fifth, across both outpatient ascitic fluid testing and upper endoscopic surveillance in Lynch syndrome, the signals favouring more systematic testing are observational and deserve prospective evaluation before they harden into standards.

That's your roundup for This Week in Gastroenterology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.