This Week in Gastroenterology — Sep 15, 2026

Welcome to This Week in Gastroenterology. This week we're covering 10 notable papers spanning early aggressive therapy and imaging in Crohn's disease, the fast-moving field of fatty liver diagnosis and drug treatment, and a cluster of endoscopy papers on resection technique, surveillance intervals and ampullary lesions. Let's dive in.

We start with Crohn's disease, where the biggest news is long-term follow-up of the PROFILE trial, published in The Lancet Gastroenterology and Hepatology. Noor and colleagues randomised 386 patients with newly diagnosed Crohn's disease across 40 United Kingdom hospitals to either top-down infliximab plus an immunomodulator from the moment of diagnosis, or a conventional step-up strategy, with protocolised treatment for the first 48 weeks and then a return to local standard of care. They then tracked objective outcomes for up to five years. Ninety-three percent of the original cohort had records available, and the numbers are striking: there were 28 Crohn's-related abdominal operations among 26 step-up patients, compared with just six operations in six top-down patients, translating to a roughly fivefold shorter time to surgery in the step-up arm. Crohn's-related hospital admissions were about twice as common with step-up, affecting around one in five step-up patients versus one in nine treated top-down, and progression to stricturing or penetrating disease was likewise roughly doubled, seen in about 17 percent of step-up patients against 7 percent of those treated early. Serious infections and malignancies were no different between the groups. The key point is that by five years almost everyone in both arms had received a biologic or immunomodulator — so this is not a comparison of treated versus untreated, it is a comparison of timing, and timing appears to modify the disease course. That is about as close as we have come to a disease-modification signal in Crohn's disease.

Alongside that, Alimentary Pharmacology and Therapeutics published the Italian Group for the Study of Inflammatory Bowel Diseases UPGRADE-CD study, a prospective real-world cohort of upadacitinib in 391 advanced-therapy-experienced Crohn's patients across 38 centres, led by Barberio [2]. Nearly half of the patients had already failed more than two advanced therapies, and yet clinical remission was achieved in just over half of them at week 12 and sustained at week 24. Steroid-free clinical response rose modestly over that interval to just under two thirds of patients. Among the 40 percent with active extraintestinal manifestations at baseline, about two thirds had complete resolution by six months, which is clinically useful information when you are choosing between mechanisms for a patient with arthralgia or skin disease. Discontinuation by week 24 was around 11 percent, mostly for treatment failure, and there were no major cardiovascular or thromboembolic events and no malignancies — reassuring, though six months is a short window for those particular signals. A higher baseline Harvey-Bradshaw index was the strongest predictor of failing to reach remission, which again argues for acting before the disease burden is high. And on the imaging side, Clinical Gastroenterology and Hepatology published work from Takenaka and colleagues developing and temporally validating a magnetic resonance enterography framework for predicting which Crohn's strictures will not improve on advanced therapy [3]. They split severity into an activity score — bowel wall thickness over 6 millimetres, ulceration, and the comb sign — and a structural persistence score — stricture length over 3 centimetres, prestenotic dilation over 3 centimetres, and absence of intramural oedema. Each score independently predicted non-improvement, and the summed score discriminated better than segmental MaRIA alone, with an area under the curve of 0.90, and also separated patients by surgery-free survival. The authors are careful to frame this as baseline risk stratification rather than a fibrotic-versus-inflammatory classifier, but it is a practical way to counsel a patient about whether medical therapy is likely to open that stricture.

Moving to the liver, three papers address the gap between what we can now treat and what we can reliably measure. Gut published an international expert panel review from Mironova, Ratziu, Rinella and colleagues on the practical use of resmetirom and semaglutide in metabolic dysfunction-associated steatohepatitis with significant fibrosis [4]. With both agents now conditionally approved by the United States Food and Drug Administration for F2 to F3 disease without cirrhosis, this consolidates phase three data, label language and early clinical experience into one care pathway — non-invasive diagnosis, selection between a liver-directed thyroid hormone receptor beta agonist and a GLP-1 receptor agonist by patient phenotype, what to do for the patient already on a GLP-1 for weight or diabetes, on-treatment monitoring, and how to define response, non-response and when to switch or combine. If you have been improvising a MASH treatment algorithm, this is the document to read. The catch is that the whole pathway depends on non-invasive tests, and two papers this week probe their limits. In Clinical Gastroenterology and Hepatology, Nammi and colleagues asked what to do about the well-known age problem with FIB-4, comparing standard FIB-4 against an older-age cutoff of 2.0, an age-capped version, a log-transformed version, and FIB-3 — simply removing age from the formula — across two MASLD cohorts totalling 938 patients with magnetic resonance elastography or biopsy as reference, plus a national survey population [5]. All strategies performed essentially equivalently overall, with areas under the curve clustered in the high 0.70s, and differences in sensitivity and specificity did not translate into meaningful differences in predictive value. What stood out was that the age-independent FIB-3 gave the most balanced trade-off at the extremes — under 35 and 65 and over — which is exactly where FIB-4 misleads us. It is not yet a practice change, but it is a strong argument for prospective evaluation in primary care. Meanwhile Gastrointestinal Endoscopy reported a single-centre prospective trial from Kohli and colleagues of endosonographic shear wave measurement in 109 patients undergoing endoscopic ultrasound-guided liver biopsy [6]. Liver stiffness values from both lobes correlated strongly with histological fibrosis, with the right lobe giving more reliable signal than the left, and excess transducer pressure only distorted left-lobe readings. Splenic stiffness performed superbly for cirrhosis, with an area under the curve of 0.96, but added nothing for intermediate fibrosis, where liver measurement was better. Overall the technique beat APRI and FIB-4 for detecting cirrhosis, and operator experience did not affect reliability. For the endosonographer already in there taking a liver biopsy, this is a low-cost add-on worth knowing about.

In endoscopy, Gut also published an individual patient data meta-analysis from Steinbrück and colleagues pooling four randomised trials of cold versus hot snare resection for large non-pedunculated colorectal polyps — 1509 polyps in 1423 patients from 50 centres, mean lesion size just over 3 centimetres [7]. The safety advantage of cold resection was clear and large: major adverse events, meaning deep mural injury, perforation or delayed bleeding, occurred in under 2 percent of cold cases versus about 6 percent of hot cases, and cold snare was the only independent protective factor. But residual or recurrent adenoma was more than doubled after cold resection, about 27 percent versus 13 percent, and the penalty was worst in lesions with high-grade dysplasia, where recurrence approached 41 percent against 18 percent. The narrowest gap was in sessile serrated lesions, where the difference did not reach significance. So the practical reading is that cold resection is not a blanket replacement for hot endoscopic mucosal resection — it is reasonable for sessile serrated lesions between 2 and 4 centimetres and for selected anticoagulated or multimorbid patients, but conventional adenomas, particularly dysplastic ones, still pay a real recurrence price. Staying with the colon, Clinical Gastroenterology and Hepatology published a cross-sectional study by Hendel and colleagues within Kaiser Permanente Northern California examining adherence to the 2020 post-polypectomy surveillance intervals [8]. For patients with three or four non-advanced adenomas, where the recommendation is three to five years, adherence exceeded 90 percent consistently. For patients with just one or two non-advanced adenomas, where the recommendation is now seven to ten years, adherence was only around 68 percent in 2021, improving to about 72 percent by 2024. Endoscopists also gravitated to the bounds rather than the middle of the range — a ten-year interval was assigned in roughly 46 to 47 percent of the one-to-two adenoma group. The implication is that over-surveillance is concentrated in the lowest-risk patients, which is precisely where capacity is wasted. And rounding out the endoscopy theme, Clinical Gastroenterology and Hepatology published the American Gastroenterological Association clinical practice update on ampullary neoplasms from Barakat and colleagues [9]. Practical points worth remembering: assess with a side-viewing duodenoscope, not a gastroscope; because 20 to 40 percent of ampullary adenomas harbour malignancy, take at least six biopsies, targeting ulcerated or indurated areas and avoiding the pancreatic orifice; and get endoscopic ultrasound for staging and intraductal extension in anything you might resect endoscopically, other than sub-centimetre lesions with no worrisome features. Lesions suited to papillectomy are well-demarcated, under 4 to 5 centimetres, soft, non-friable, with intraductal extension of a centimetre or less. Beyond a centimetre of intraductal extension, invasive cancer past the mucosa, or metastatic disease, send the surgically fit patient to surgery.

Finally, one pharmacovigilance signal. Alimentary Pharmacology and Therapeutics published a propensity-matched analysis by Aburumman and colleagues from a large United States federated network, comparing over 158,000 cancer patients exposed to immune checkpoint inhibitors with matched unexposed controls [10]. Incident coeliac disease diagnosed at least 90 days after the index event was nearly four times more common after checkpoint inhibitor exposure, with the signal consistent across both sexes and highest in those over 60. The absolute risk, though, is tiny — roughly 0.15 percent versus 0.07 percent, a number needed to harm of about 1200. So this is not a screening mandate; it is a reason to keep coeliac disease on the differential when a patient on immunotherapy develops chronic diarrhoea or malabsorption rather than reflexively labelling it checkpoint inhibitor colitis.

If you only have time for one paper this week, make it the five-year PROFILE follow-up in The Lancet Gastroenterology and Hepatology [1]. It is the strongest evidence we have that when you start advanced therapy in Crohn's disease changes the trajectory of the disease, not just the symptoms — and it should reshape how you counsel a newly diagnosed patient next week.

Here are the key takeaways from this week in Gastroenterology. First, in newly diagnosed Crohn's disease, top-down anti-TNF from the outset cut abdominal surgery, hospitalisation and progression to stricturing or penetrating disease at five years, with no safety penalty. Second, upadacitinib delivers remission in just over half of heavily pretreated Crohn's patients in real-world practice, and resolves extraintestinal manifestations in about two thirds. Third, for MASH with F2 to F3 fibrosis we now have a consolidated pathway for resmetirom and semaglutide, while FIB-4's age problem may be best solved by simply dropping age from the formula. Fourth, cold snare resection of large non-pedunculated polyps buys a real safety advantage but more than doubles residual adenoma, so reserve it for sessile serrated lesions and higher-risk patients rather than applying it universally. And fifth, remember that one or two non-advanced adenomas earns a seven-to-ten-year interval — that is where over-surveillance is still happening.

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 a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

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