This Week in Gastroenterology — Sep 8, 2026

Welcome to This Week in Gastroenterology. This week we're covering 10 notable papers spanning colorectal cancer risk stratification and screening, the maturing science of non-invasive liver assessment, hepatobiliary and cancer risk in inflammatory bowel disease, and a group of papers on functional and nutritional disease, including a positive phase 3 trial in intestinal failure. Let's dive in.

We'll start with colorectal cancer, where three papers each attack a different point in the risk-stratification chain. In the American Journal of Gastroenterology, Chang and colleagues asked whether everyone we label high risk after an index colonoscopy really carries the same risk. Using a prospective Taiwanese multicentre cohort of 730 patients under surveillance, followed on average about two and a half years, roughly one in nine patients developed metachronous advanced neoplasia. The interesting signal is where that risk sat. A single high-risk adenoma on its own did not significantly raise risk compared with having three or four low-risk adenomas. Risk climbed only when that advanced lesion came with company, roughly two and a half times higher when accompanied by synchronous low-risk adenomas and close to four times higher when accompanied by other high-risk adenomas, with a clear dose-response across the groups. The practical read is that synchronous burden, not the single worst lesion, is what should drive whether you bring someone back at three years or stretch the interval, though the follow-up here is short and this needs replication before guidelines move. [1] Complementing that, The Lancet Gastroenterology and Hepatology published a systematic review and meta-analysis from Mannucci and colleagues on germline multigene panel testing in unselected colorectal cancer patients, pooling 21 studies and nearly 7,000 individuals. About 15 percent of patients carried at least one pathogenic or likely pathogenic variant, and about 8 percent carried a high-penetrance variant. In early-onset disease under 50, the high-penetrance yield was around 14 percent, but the key finding for practice is what happened at older ages: yield declined progressively with age, yet high-penetrance variants stayed above the 5 percent threshold up to about age 67. Heterogeneity between studies was high, so treat the pooled figures as approximate, but the direction supports offering panel testing well beyond the young-onset population we currently prioritise. [2]

The third colorectal paper, also in the American Journal of Gastroenterology, is a decade of real-world data on multi-target stool DNA screening from Garg and colleagues, covering nearly 150,000 tests in more than 110,000 individuals across a multi-site health system. Positivity was about 14 percent, and the diagnostic colonoscopies that followed were high quality, with adenoma detection above 50 percent and advanced adenoma found in about 31 percent of those examined and cancer in just over 1 percent. The problem was the care cascade around the test. Only about two thirds of patients with a positive result completed colonoscopy within twelve months, and delays beyond six months were associated with a higher rate of advanced adenoma. Even more striking, after an initial negative test only about a third of patients were retested within three years, and among those who did undergo colonoscopy within three years of a negative stool test, close to half OF THOSE PATIENTS had an adenoma or serrated lesion. Stool-based screening only works as a programme, not as a single test, and these data quantify exactly where the leaks are. [3]

Turning to inflammatory bowel disease, two papers reframe what we should be watching for. In Gastroenterology, Everhov and colleagues used Swedish nationwide registers to compare colorectal cancer incidence in more than 124,000 patients with inflammatory bowel disease against over 1.2 million matched comparators, with a median follow-up of eleven years. Baseline incidence was modestly higher in inflammatory bowel disease, at about 1.2 versus 0.9 cases per 1,000 person-years. But the family history findings are the practice-relevant part: the largest absolute excess came in patients with two or more affected first-degree relatives, at roughly 2.7 additional cases per 1,000 person-years, whereas heredity for early-onset colorectal cancer added only a small and statistically uncertain increment. Since guidelines currently single out family history of early-onset disease, the authors argue surveillance intensity might be better targeted at patients with multiple affected relatives. [4] Then in Clinical Gastroenterology and Hepatology, Attauabi and colleagues report from the prospective population-based Copenhagen inception cohort, in which newly diagnosed adults were offered magnetic resonance cholangiopancreatography at the time of diagnosis. Among 389 patients scanned, 8 percent had radiological primary sclerosing cholangitis, rising to nearly 10 percent when non-diagnostic sclerosing cholangitis-like lesions were included, higher in extensive colitis, and notably present in Crohn's disease as well as ulcerative colitis. Only about half OF THOSE PATIENTS with radiological disease had abnormal liver biochemistry, meaning liver enzymes would have missed the rest. Those patients also went on to more inflammatory bowel disease hospitalisations and more biologic therapy. The authors are appropriately cautious: with only two years of follow-up we don't know the long-term significance of these images, and cost-effectiveness work is needed before routine cholangiography at diagnosis can be recommended. [5]

Two hepatology papers this week address the same frustration from opposite directions, namely how to use non-invasive tests over time. In Hepatology, Loomba and colleagues tested the American Association for the Study of Liver Diseases suggestion that a 30 percent decline in liver stiffness on vibration-controlled transient elastography can serve as a treatment-response endpoint in metabolic dysfunction-associated steatohepatitis. Using 160 patients with biopsy-proven disease and stage 2 to 3 fibrosis from a phase 2b pegozafermin trial, with paired elastography and biopsy, a 30 percent stiffness decline was independently associated with roughly a four-fold increase in the odds of fibrosis regression, but discriminative accuracy was only modest, an area under the curve of about 0.68, and about 0.62 in a separate validation cohort from the United States and Singapore. Their conclusion is blunt: this threshold is not accurate enough as a stand-alone surrogate, and better response biomarkers are needed. [6] Clinical Gastroenterology and Hepatology offers a different use for the same measurement. Moreau and colleagues applied a joint latent class model to the international elastography prognosis cohort of over 13,600 patients with metabolic dysfunction-associated steatotic liver disease, in whom 238 liver-related events occurred over a median four years. By modelling the trajectory of liver stiffness together with fibrosis-4, age, sex and platelets, the model produced well-calibrated individual risk estimates with integrated discrimination in the high eighties to low nineties percent, and it flagged 61 to 80 percent of patients who went on to have an event as high risk, compared with roughly 40 to 55 percent using liver stiffness alone. Read together, these papers suggest the trajectory of non-invasive tests is more informative for predicting hard outcomes than for certifying histological change.

Finally, three papers on luminal and nutritional disease. Gut published a multicentre cross-sectional study from Chan and colleagues of 580 patients with acid exposure time above 6 percent off therapy across eleven international centres. Excessive supragastric belching was identified in about 19 percent of these patients with conclusive reflux disease, with roughly a quarter of total acid exposure attributable to belching episodes, and the finding was consistent across four world regions. Patients whose dominant complaint was heartburn or chest pain were less likely to have this phenotype, whereas a high total reflux episode count pointed towards it, with more than 115 total reflux episodes or more than 56 proximal episodes giving over 90 percent specificity. That's a usable signal to consider behavioural therapy rather than escalating acid suppression. [8] Also in Gut, the Rome Foundation paediatric global study led by Nurko surveyed caregivers of 8,000 children across China, Italy, Mexico and the United States using Rome V criteria, and found disorders of gut-brain interaction in 28 percent of children, driven mainly by defecation and anorectal disorders at about 20 percent. Prevalence was highest in Mexico and lowest in China, fell with age from roughly 41 percent in the under-threes to about 22 percent in adolescents, and was associated with anxiety, depression, functional disability, healthcare use and substantial parental caregiving time, particularly when upper and lower disorders overlapped. [9] And in Clinical Gastroenterology and Hepatology, the global phase 3 STARS trial from Joly and colleagues randomised 163 patients with short bowel syndrome and intestinal failure two-to-one to once-weekly subcutaneous apraglutide, a long-acting glucagon-like peptide-2 analogue, or placebo. At 24 weeks, weekly parenteral support volume fell by about a quarter with apraglutide versus about 12 percent with placebo, a significant difference, and this was also significant in the stoma subgroup. More patients came off at least one day of parenteral support per week, 43 percent versus 28 percent. Importantly, in the colon-in-continuity group at week 48, the differences in days off support and in enteral autonomy were not statistically significant. Tolerability was good. [10]

If you only have time for one paper this week, make it the decade-long multi-target stool DNA cohort in the American Journal of Gastroenterology [3]. It is the rare paper that tells you your colonoscopy quality is fine while showing that the system around your non-invasive screening test is where patients are being lost, and both of those gaps are fixable in your own practice this month.

Here are the key takeaways from this week in Gastroenterology. First, within the conventional high-risk post-polypectomy group, synchronous adenoma burden is what drives metachronous advanced neoplasia, so a single advanced lesion in isolation may not warrant the shortest interval. Second, germline panel testing in colorectal cancer yields clinically meaningful variants well past age 60, arguing against restricting testing to early-onset disease. Third, stool DNA screening fails at the follow-up step, not the detection step, so build systems for timely colonoscopy after a positive test and for re-screening after a negative one. Fourth, in inflammatory bowel disease, multiple affected relatives may matter more for colorectal cancer surveillance than early-onset heredity, and radiological sclerosing cholangitis is commoner at diagnosis than expected and often biochemically silent, though routine cholangiography is not yet justified. Fifth, in fatty liver disease, a 30 percent fall in liver stiffness is only a modest marker of histological improvement, but the trajectory of non-invasive tests over time predicts liver-related events well. And finally, consider supragastric belching in reflux patients with very high reflux episode counts and non-heartburn symptoms, and remember that apraglutide meaningfully reduces parenteral support in short bowel syndrome, with the colon-in-continuity secondary endpoints falling short of significance.

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 you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

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