Gastroenterology Insights — Clinical Updates, Guides & Resources — Page 4

5 min readJul 23, 202621reads

CKM Syndrome Guidelines 2026: What They Mean for MASLD Screening in Gastroenterology

A 51-year-old woman is referred to hepatology for elevated ALT and hepatic steatosis on ultrasound. Her BMI is 29, her waist circumference is 91 cm, and nobody has ever ordered an eGFR or checked her blood pressure trend. Under the framework hepatologists have used for years, she's a routine MASLD workup. Under the guidelines released this month, she may already be a stage 2 or 3 cardiovascular-kidney-metabolic (CKM) syndrome patient - and that changes what happens next. The first CKM syndrome guidelines, jointly issued by the American Heart Association, American College of Cardiology, American Society of Nephrology, and American Diabetes Association, formalize something gastroenterologists have watched unfold clinically for a decade: metabolic dysfunction–associated steatotic liver disease rarely travels alone. As JAMA's coverage of the guideline explains, the framework helps clinicians diagnose, stage, treat, and monitor CKM syndrome based on the interconnections among metabolic risk factors, chronic kidney disease, and cardiovascular disease, replacing the 2013 obesity guidelines entirely. For CKM syndrome and MASLD screening in particular, the practical question for GI clinicians isn't whether the liver matters - it's whether hepatology's workup now needs to include kidney and cardiovascular staging it has traditionally left to other specialties. How CKM Staging Reframes the MASLD Workup The guideline defines five stages, and MASLD patients cluster heavily in the middle of them. Stage 0 requires normal BMI, waist circumference, glucose, blood pressure, and lipids with no CKD or cardiovascular disease; stage 1 is excess or dysfunctional adiposity alone; stage 2 adds metabolic risk factors or CKD; stage 3 adds subclinical cardiovascular disease or very high-risk CKD; and stage 4 is established clinical cardiovascular disease. A patient with MASLD and even mild hypertension already clears the bar for stage 2 - most of the panel a gastroenterologist would order for steatosis anyway. What's new is the mandate to close the loop on kidney function. The guidelines call for regular eGFR testing in all adults, with urine albumin-to-creatinine ratio (UACR) added for anyone at stage 2 or higher, and albuminuria is treated as a modifiable risk factor on par with blood pressure. Most MASLD patients already meet the threshold for UACR testing by virtue of their metabolic profile alone, yet it's rarely part of a standard hepatology intake. The PREVENT equations are the second addition: for stage 0–3 patients, they estimate 10- and 30-year risk of total cardiovascular disease, heart failure, and atherosclerotic disease, and factor in BMI and eGFR directly - a natural extension of data hepatology is already collecting, and a useful companion to the fibrosis-risk models covered in our post on machine learning for liver fibrosis prediction in MASLD . Case in point The patient above returns for her follow-up. Her transaminases have improved slightly on lifestyle counseling, and her ultrasound still shows steatosis without obvious cirrhotic features. This time, her workup includes eGFR (78 mL/min/1.73m²), UACR (42 mg/g - mildly elevated), and a blood pressure average of 134/86. Run through the PREVENT calculator, her 10-year cardiovascular risk lands at 9%. That crosses the 7.5% threshold the guidelines use to inform whether to initiate a GLP-1-based therapy or an SGLT2 inhibitor for CKM syndrome - not as a liver-directed decision, but as a cardiometabolic one that happens to also help her steatosis. She's staged as CKM 2 going on 3, and that stage, not her ALT trend, is now the number driving the conversation about escalation. Screening and Referral Triggers Gastroenterologists Should Adopt Three additions to a standard MASLD intake bring a hepatology visit in line with the new guideline, without requiring a cardiology consult for every patient. First, waist circumference belongs in the chart alongside BMI. The guidelines set abdominal obesity thresholds at 88 cm or greater for women and 102 cm or greater for men, with lower thresholds for patients of Asian ancestry. Second, eGFR and, where indicated, UACR should be added to the standard MASLD panel rather than deferred to primary care - kidney disease often goes unrecognized until later stages despite now being treatable earlier. Third, a PREVENT score above 7.5% or 20% (the stage 3 threshold) is a referral trigger, not just a documentation exercise - it's the point at which cardiology or a structured GLP-1/SGLT2 pathway should enter the conversation. None of this requires the gastroenterologist to manage cardiovascular risk directly. The full CKM Health guideline in Circulation explicitly favors team-based care with a coordinating point person, noting that clinicians already treating these patients can often fill that role. For a MASLD patient already in a hepatology clinic every six months, that coordinator can reasonably be you. A frequently overlooked point The instinct in hepatology has been to treat MASLD as the index diagnosis and everything else as a comorbidity list. The CKM framework inverts that: obesity and its metabolic consequences are the root process, and MASLD is one downstream expression of it, not the anchor. A patient staged purely on liver fibrosis risk (say, a low FIB-4) can still be CKM stage 3 on the strength of subclinical coronary calcium or a high PREVENT score - numbers a liver-focused workup will never surface. This distinction matters even more for the lean phenotype discussed in our post on cryptogenic steatotic liver disease in lean patients , where the absence of classic cardiometabolic risk factors can mask real CKM-relevant risk rather than rule it out. Fibrosis staging and CKM staging are two different axes measuring two different futures for the same patient. Bottom line for clinical practice Add eGFR to every MASLD workup, and add UACR once a patient has any second metabolic risk factor - most MASLD patients qualify. Calculate a PREVENT score before deciding on GLP-1 or SGLT2 escalation; a 10-year CVD risk ≥7.5% is the guideline's own threshold for starting therapy. Document waist circumference at every visit, using the lower Asian-ancestry thresholds where applicable - BMI alone under-detects visceral risk. Treat a PREVENT-CVD score ≥20% as a stage 3 flag warranting cardiology or nephrology referral, not just a note in the chart. Don't assume a reassuring fibrosis score means reassuring CKM staging - check both independently, especially in lean or atypical presentations. Next time a MASLD case in your clinic raises a CKM staging question - whether a PREVENT score changes the therapy conversation, or whether a UACR result should trigger referral - walk GastroAGI through the specifics. It returns a guideline-anchored answer in seconds, built for exactly this kind of cross-specialty decision

GastroAGI TeamRead more

11 min readJul 22, 202628reads

CYP2C19 Genotype and PPI Response in Pediatric EoE: Does Dose Matter?

Why do some children with active eosinophilic esophagitis respond histologically to proton pump inhibitor therapy, while others do not? For gastroenterologists and pediatric GI clinicians, this question is clinically familiar. A child is diagnosed with active histologic eosinophilic esophagitis, treated with a proton pump inhibitor, and then reassessed. In some patients, esophageal eosinophilia improves. In others, it persists despite therapy. The decision that follows may involve continuing treatment, changing therapy, adding another approach, or rethinking whether the dose was adequate for that individual patient. A new article in The American Journal of Gastroenterology brings pharmacogenetics into this discussion. The verified title is “Dose May Matter: CYP2C19 Genotype and PPI Response in Pediatric Eosinophilic Esophagitis.” The article was published in July 2026 with DOI 10.14309/ajg.0000000000004117 . The available source identifies it as an article by Paroma Bose, Chizoba N. Umeweni, Anthony Perkins, Elizabeth Jensen, and additional authors. The key question is focused: does proton pump inhibitor metabolism, as reflected by CYP2C19 genotype, influence treatment response in pediatric EoE? The authors state that little is known about the effect of PPI metabolism based on CYP2C19 genotype on treatment response in pediatric eosinophilic esophagitis. This is an important but early piece of evidence. It is not a guideline, not a randomized trial, and not a mandate for routine genetic testing. It is a cohort study of pediatric subjects with active histologic EoE treated with PPI therapy, in whom whole blood samples were tested for CYP2C19 genotype. A precision-medicine question hiding inside a common clinical pathway The clinical appeal of this study is easy to understand. PPI therapy is already part of the therapeutic landscape for EoE, but response is variable. The source does not provide a complete background review of EoE treatment, nor does it specify response rates, patient numbers, dosing duration, or endoscopic findings. What it does clearly establish is that the investigators were interested in whether PPI metabolism , defined through CYP2C19 genotype , helps explain histologic response in pediatric EoE. That framing matters. Many EoE discussions focus on disease phenotype, allergic inflammation, diet, topical steroid therapy, biologics, or fibrostenotic complications. This article asks a different but practical question: could some apparent PPI non-response reflect pharmacologic exposure rather than only disease biology? The title itself— “Dose May Matter” —is appropriately cautious. It does not claim that genotype fully determines response. It does not claim that a specific genotype-based dosing algorithm is established. It suggests that dose and metabolism may interact in a clinically meaningful way. For clinicians, this is exactly the kind of hypothesis that needs careful handling. It is plausible and potentially useful, but the source supports association with histologic response, not proof that genotype-guided dosing improves outcomes. What the study investigated The study design was a cohort study . The population was pediatric subjects with active histologic eosinophilic esophagitis . The intervention or treatment exposure was PPI therapy . The pharmacogenetic exposure was CYP2C19 genotype , assessed through testing of whole blood samples . The outcome highlighted in the abstract was histologic response to PPI therapy . The available source does not provide sample size, age distribution, sex distribution, PPI type, treatment duration, exact histologic response threshold, number of subjects in each genotype group, or the statistical model used. Because those details are not available in the provided source, they should not be inferred. What can be verified is the core structure: Children had active histologic EoE. They were treated with PPI therapy. Their CYP2C19 genotype was assessed using whole blood. The investigators then evaluated how metabolizer status related to histologic response, particularly in the context of high-dose PPI exposure. This is clinically relevant because histologic response is a tissue-level endpoint. The source does not report symptom response, endoscopic response, quality of life, stricture outcomes, food impaction rates, or long-term remission. Therefore, the blog interpretation must remain centered on histology. The key finding: response differed by metabolizer status at higher PPI dose The central finding reported in the abstract is that among subjects prescribed a high PPI dose greater than 1 mg/kg/day , intermediate or slow PPI metabolizers were significantly more likely to have histologic response to PPI therapy compared with normal and rapid PPI metabolizers . The abstract also states that no ultra-rapid metabolizers responded to PPI therapy. This is the clinically provocative signal. The finding suggests that PPI metabolism may influence whether a given dose achieves sufficient pharmacologic effect in some children with EoE. Intermediate or slow metabolizers may have greater exposure to PPI therapy at the same or similar dose, while normal, rapid, or ultra-rapid metabolizers may have lower effective exposure. The source supports the observed association between metabolizer category and histologic response; it does not prove the pharmacokinetic mechanism directly unless drug levels or acid-suppression metrics were measured, and those details are not provided in the accessible source. The dose qualifier is also important. The result is described specifically among subjects prescribed high PPI dose >1 mg/kg/day . This means the finding should not be generalized to all PPI doses, all dosing regimens, or all pediatric EoE patients without additional data. Why the phrase “dose may matter” is clinically meaningful The article’s title is well chosen because it points to an interaction clinicians already recognize in practice: the same prescribed drug dose does not necessarily produce the same biologic effect in every patient. In this study, the relevant variable is CYP2C19 genotype, used as a marker of PPI metabolism. The authors report that metabolizer status was associated with histologic response among children receiving high-dose PPI therapy. This raises a practical question: when a child with EoE does not respond histologically to a PPI, is the issue that the disease is PPI-nonresponsive, or that the patient’s metabolism leads to inadequate drug effect at the prescribed dose? The source does not answer that question definitively. It does, however, support the idea that CYP2C19 genotype may help explain variability in response. The discussion statement in the abstract is that knowing CYP2C19 genotype may optimize PPI dose for EoE treatment in individual patients. That wording is important. “May optimize” is not the same as “should routinely guide.” It signals a potential future direction rather than established clinical guidance. Practical interpretation for gastroenterologists For clinicians, the finding should be interpreted as a pharmacogenetic association with histologic response, not as a new standard of care. A reasonable interpretation is that CYP2C19 metabolizer status may identify pediatric EoE patients in whom a standard or even high PPI dose is less likely to produce histologic response. The most striking reported observation is that no ultra-rapid metabolizers responded to PPI therapy in the cohort. However, without knowing the number of ultra-rapid metabolizers, confidence intervals, absolute response rates, PPI agents used, adherence assessment, treatment duration, and adjustment for confounders, clinicians should avoid overinterpreting this finding. A result can be statistically significant and still require replication before it changes practice. The study is especially relevant for pediatric gastroenterologists and fellows because it reframes PPI response as potentially influenced by host pharmacogenetics. It also fits within a broader movement toward individualized therapy in EoE. But the evidence provided here does not establish how to adjust the dose, which PPI should be selected for different genotypes, whether split dosing matters, or whether genotype-guided therapy improves outcomes compared with usual care. What clinicians should conclude—and what they should not Clinicians can conclude that this AJG article reports a cohort study in pediatric active histologic EoE treated with PPI therapy, with CYP2C19 genotype assessed from whole blood. They can also conclude that, among children prescribed high-dose PPI therapy greater than 1 mg/kg/day, intermediate or slow PPI metabolizers were significantly more likely to show histologic response than normal or rapid metabolizers, and that no ultra-rapid metabolizers responded in the reported cohort. Clinicians should not conclude that CYP2C19 testing is now required for every child with EoE. The available abstract does not provide a clinical algorithm. It does not compare genotype-guided dosing with standard dosing in a randomized design. It does not report prospective outcomes after genotype-based dose adjustment. It does not show whether changing dose based on genotype improves histologic remission, symptoms, endoscopic outcomes, or long-term disease course. Clinicians should also not conclude that PPI non-response is fully explained by genotype. EoE response is likely multifactorial, but the provided source only supports the specific association between CYP2C19 metabolizer category, high-dose PPI therapy, and histologic response. Association is not causation This point deserves emphasis. The study reports an association between CYP2C19 metabolizer status and histologic response to PPI therapy in pediatric EoE. It does not prove that CYP2C19 genotype caused response or non-response. It also does not prove that increasing the PPI dose in rapid or ultra-rapid metabolizers would produce response. The causal pathway is plausible, but it remains unproven by the accessible source. To establish causation and clinical utility, future studies would need to test whether genotype-guided dosing changes outcomes compared with non-genotype-guided care. For example, an implementation study could assign children to genotype-guided PPI dosing versus usual PPI dosing and then compare histologic remission, symptom change, endoscopic change, adverse events, adherence, need for therapy escalation, and cost-effectiveness. The current source does not report such a trial. Therefore, the safest interpretation is that CYP2C19 genotype may be a clinically relevant predictor or modifier of histologic PPI response, particularly at high-dose exposure, but this remains early evidence. Strengths of the evidence The study has several strengths based on the available source. First, it addresses a clinically important problem: variability in response to PPI therapy in pediatric EoE. The authors explicitly state that little is known about the effect of PPI metabolism based on CYP2C19 genotype in this setting. Second, the population is clinically specific: pediatric subjects with active histologic EoE treated with PPI therapy. This avoids extrapolating from adult acid-mediated disease or from non-EoE indications. Third, the exposure is biologically coherent: CYP2C19 genotype as a marker of PPI metabolism. The study tested whole blood samples, which suggests direct genotyping rather than inferred metabolizer status from clinical response alone. Fourth, the outcome highlighted is histologic response, which is central to EoE assessment. The source does not define the threshold used, but it clearly identifies histologic response as the reported endpoint. Finally, the result is clinically interpretable: intermediate or slow metabolizers were more likely to respond than normal and rapid metabolizers among those receiving high-dose PPI, while ultra-rapid metabolizers did not respond. Limitations and evidence gaps The accessible source is concise and does not provide enough detail for complete critical appraisal. The most important missing information is sample size. Without knowing how many children were included overall and how many fell into each metabolizer group, it is difficult to judge precision. The statement that no ultra-rapid metabolizers responded is important, but its clinical weight depends heavily on how many ultra-rapid metabolizers were studied. The source also does not specify the PPI agent, dosing schedule, treatment duration, adherence assessment, baseline disease severity, distribution of eosinophil counts, or whether other therapies were excluded or controlled. These factors could influence histologic response. The abstract does not report symptom outcomes. That matters because EoE management often requires integrating histology, symptoms, and endoscopic findings. A histologic signal is important, but clinicians should not infer symptom benefit unless reported. The source also does not report safety outcomes. If genotype-guided dose escalation is considered in future studies, safety and tolerability will be essential, particularly in children. Another limitation is that the study is observational. Cohort studies can identify clinically useful associations, but they cannot fully exclude confounding. Adherence, disease phenotype, timing of follow-up endoscopy, PPI selection, and clinical decision-making may all influence observed response. How this may influence future research This study creates a clear research agenda. The next step should be larger prospective studies that report genotype distribution, PPI exposure, dosing schedule, adherence, histologic thresholds, endoscopic outcomes, symptoms, and safety. These studies should also clarify whether genotype has predictive value beyond standard clinical variables. A more practice-oriented question is whether CYP2C19-guided PPI therapy improves outcomes. That requires an interventional design. It is not enough to show that genotype is associated with response; clinicians need to know whether acting on the genotype changes care in a beneficial way. Researchers should also determine whether the effect differs by PPI type. The provided source does not specify the PPI agents used, so no conclusions can be made about one PPI versus another. Implementation research will also matter. Even if genotype-guided dosing proves useful, clinicians will need practical answers: when should testing be ordered, how quickly must results return, which patients benefit most, what dose changes are appropriate, and how should nonresponse be managed? Clinical Takeaway “Dose May Matter: CYP2C19 Genotype and PPI Response in Pediatric Eosinophilic Esophagitis” is a clinically relevant AJG cohort study suggesting that CYP2C19 metabolizer status may influence histologic response to PPI therapy in pediatric EoE. Among children prescribed high-dose PPI therapy greater than 1 mg/kg/day, intermediate or slow metabolizers were significantly more likely to have histologic response than normal or rapid metabolizers, while no ultra-rapid metabolizers responded in the reported cohort. The study supports a promising precision-medicine hypothesis: pharmacogenetics may help explain variable PPI response and could eventually help individualize dosing. But the current evidence is not yet practice-changing. It should be interpreted as observational, association-based evidence that requires larger validation and prospective genotype-guided treatment studies before routine clinical implementation. Five key clinical takeaways The verified article is “Dose May Matter: CYP2C19 Genotype and PPI Response in Pediatric Eosinophilic Esophagitis,” published in The American Journal of Gastroenterology in July 2026 . The study design was a cohort study of pediatric subjects with active histologic EoE treated with PPI therapy . Whole blood samples were tested for CYP2C19 genotype , linking pharmacogenetic metabolizer status to PPI treatment response. Among subjects prescribed high PPI dose >1 mg/kg/day , intermediate or slow metabolizers were significantly more likely to have histologic response than normal or rapid metabolizers; no ultra-rapid metabolizers responded. The evidence is observational and association-based . It suggests genotype may help optimize PPI dosing in the future, but it does not establish routine CYP2C19-guided EoE management. Source reference and link Bose P, Umeweni CN, Perkins A, Jensen E, et al. Dose May Matter: CYP2C19 Genotype and PPI Response in Pediatric Eosinophilic Esophagitis. The American Journal of Gastroenterology. July 2026. DOI: 10.14309/ajg.0000000000004117 .

GastroAGI TeamRead more

11 min readJul 22, 202640reads

Risk-Based Pathology Reporting After ESD for Early GI Cancer: Why Standardisation Matters

After endoscopic submucosal dissection for early gastrointestinal cancer, the endoscopy report tells only part of the story. The resection may appear technically successful, the lesion may have been removed en bloc, and the patient may leave the unit believing that cancer treatment is complete. But the next major clinical decision often depends on the pathology report: was the resection curative, or does the patient need further treatment? That decision is rarely based on one histological feature alone. It depends on a risk-based synthesis of invasion depth, invasion breadth, margin status, lymphovascular invasion, tumour budding, differentiation, histological subtype, perineural invasion, specimen handling, and the reliability of the measurements reported. When these elements are reported inconsistently, multidisciplinary teams may struggle to determine whether surveillance is appropriate or whether additional therapy should be considered. A new international consensus article in Gut , titled “Risk-based pathology reporting after endoscopic submucosal dissection for early gastrointestinal cancer: international consensus standards,” directly addresses this problem. It was published online ahead of print on 10 July 2026 and carries the DOI 10.1136/gutjnl-2025-337567 . The work was led by Kareem Khalaf and colleagues and developed through an international modified Delphi consensus process. A pathology report that determines the next clinical step Endoscopic submucosal dissection, or ESD, enables en bloc resection of early gastrointestinal cancers and provides specimens that can be examined for detailed pathological risk assessment. The clinical strength of ESD is not only that it can remove selected lesions endoscopically, but that it preserves specimen architecture in a way that allows pathologists to assess features linked to residual disease risk, lymph node metastasis risk, and the need for additional treatment. The problem identified by the consensus group is variability. The abstract specifically notes that reporting remains variable for key parameters that determine curative resection and the need for additional treatment. These include submucosal invasion depth and breadth, margin status, lymphovascular invasion, tumour budding, differentiation, and the use of ancillary stains. For clinicians, this variability is not a technical inconvenience. It is a clinical problem. If one report describes invasion depth imprecisely, another uses different margin terminology, and a third omits tumour budding or lymphovascular invasion assessment, decision-making becomes less reliable. The same patient could be discussed differently depending on local pathology conventions rather than tumour biology. This consensus article is therefore clinically relevant because it focuses on the interface between endoscopy, pathology, surgery, oncology, and surveillance planning. It is not a trial of ESD technique. It is not a drug study. It is a consensus standards paper designed to make post-ESD pathology reporting more reproducible and clinically meaningful. What the consensus process investigated The objective was to develop practical international standards for the pathology assessment and reporting of invasive carcinoma in ESD specimens. The study design was an international modified Delphi consensus process involving 42 experts from 15 countries . The panel included 28 gastrointestinal pathologists and 14 therapeutic endoscopists , which is important because the reporting standards were developed across the two specialties most directly involved in post-ESD interpretation and decision-making. The consensus statements addressed several domains: measurement of invasion, margin assessment, staining, specimen handling, prognostic histological features, and clinically relevant reporting. The process resulted in 56 recommendations reaching consensus across seven domains . The population here should be understood correctly. This was not a patient cohort study. There was no intervention group, no control group, and no patient-level clinical outcome analysis reported in the accessible abstract. The “population” for the consensus process was the international expert panel. The clinical material under consideration was ESD specimens containing invasive carcinoma from early gastrointestinal cancer. The “outcome” was consensus on practical pathology reporting standards, not survival, recurrence, lymph node metastasis, or cost-effectiveness. That distinction matters. These standards may influence clinical practice by improving reporting consistency, but the abstract does not show that implementing them improves patient outcomes. It provides consensus-based reporting criteria intended to support clinical decision-making and future validation of risk models. Measuring submucosal invasion: precision over habit One of the most clinically important areas addressed is submucosal invasion. In early gastrointestinal cancers treated by ESD, the depth of invasion into the submucosa is a key element in risk assessment. The consensus panel recommends using the Sm1–Sm3 subclassification only when the muscularis propria is present . If the muscularis propria is not present, the depth of submucosal invasion should be reported in micrometres , rounded to the nearest 100 µm . This recommendation is practical because the anatomical landmark used for subclassification may not always be present in an ESD specimen. If a classification system depends on a structure that is absent, reporting can become inconsistent or misleading. Reporting invasion depth in micrometres provides a more explicit measurement and reduces ambiguity. The panel also recommends reporting submucosal invasion breadth in millimetres as an adjunct metric for future validation. This is a subtle but forward-looking point. The consensus group is not presenting breadth as a fully validated standalone decision rule in the accessible abstract. Instead, it identifies it as a metric that should be reported so future datasets can test its prognostic value more consistently. For clinicians, the implication is clear: pathology reporting should not only describe that invasion is present, but define how it was measured and express it in reproducible terms. That does not automatically determine management by itself, but it improves the reliability of multidisciplinary interpretation. Margin status: standard terminology for a high-stakes variable Margin assessment is another central domain. The panel recommends that margin positivity should be defined as direct tumour contact with the inked surface . This definition is supported by standardised specimen pinning, inking, complete embedding, and parallel sectioning. This matters because “positive,” “close,” “involved,” and “uncertain” margins can be used differently across institutions. In a post-ESD setting, margin status may influence whether a lesion is considered completely resected and whether additional treatment, repeat endoscopic therapy, surgery, or close surveillance is discussed. A precise definition helps reduce interpretive drift. The consensus also links the margin definition to specimen handling. This is important because the quality of margin interpretation depends on what happens before the slide is read. Pinning, inking, embedding, and sectioning are not clerical details; they determine whether the pathologist can confidently assess the relationship between tumour and resection surface. For endoscopists, this reinforces the need for close communication with pathology teams. A high-quality ESD specimen can still produce a less useful report if handling and orientation are inconsistent. Conversely, standardised pathology workflows can enhance the clinical value of an en bloc resection. Ancillary stains: selective use, not automatic escalation The consensus standards retain H&E as the baseline stain . Selective immunohistochemistry or elastic stains are recommended for situations such as equivocal lymphovascular invasion, distorted architecture, or difficult margin interpretation. This is a balanced approach. It avoids implying that every ESD specimen requires extensive ancillary staining. At the same time, it recognises that certain clinically important features may be difficult to assess on routine staining alone. Lymphovascular invasion is particularly relevant because it can influence risk stratification after endoscopic resection. The consensus does not state that ancillary stains should replace conventional histology. Rather, it supports selective use when interpretation is uncertain. This distinction is important for clinicians reviewing reports. The absence of ancillary stains does not necessarily mean inadequate reporting if the case is straightforward. But when lymphovascular invasion or margins are equivocal, selective stains may improve interpretive confidence. Tumour budding and composite risk: individual features need context The panel recommends that tumour budding should be reported according to International Tumour Budding Consensus Conference criteria . It also states that differentiation, histological subtype, lymphovascular invasion, perineural invasion, and margin status should be integrated into composite risk assessment. This is perhaps the core conceptual shift: pathology after ESD should not be a disconnected list of microscopic observations. It should support clinically relevant risk assessment. Composite risk does not mean inventing a new score without validation. It means recognising that individual features interact in clinical decision-making. For example, a report that separately lists differentiation, invasion depth, lymphovascular invasion, and margins may be complete in a descriptive sense, but the clinician still needs to understand whether the combined profile suggests low-risk or higher-risk pathology. The consensus standards aim to make these elements “synoptic-ready,” meaning suitable for structured reporting formats. Synoptic reporting is especially useful in multidisciplinary care because it reduces omission, improves clarity, and allows data aggregation for research and quality improvement. Why this matters for multidisciplinary teams The conclusion of the article states that these standards provide immediately implementable, synoptic-ready pathology reporting criteria after ESD. The authors also state that standardising measurement landmarks, margin terminology, ancillary stain use, and reporting of adverse histological features aims to reduce interinstitutional variability, improve multidisciplinary decision-making, and support future validation of risk models in early gastrointestinal cancer. That conclusion is highly relevant to tumor boards and post-resection pathways. After ESD, the decision is often not binary. Some patients clearly meet criteria for endoscopic cure. Others clearly require further treatment. But many cases sit in the difficult middle: a close or uncertain margin, borderline invasion depth, equivocal lymphovascular invasion, or histological features that raise concern without providing a single definitive answer. Standardised reporting does not eliminate clinical judgment. It improves the substrate on which clinical judgment is based. For gastroenterologists and hepatologists involved in upper GI, colorectal, and early cancer pathways, the most practical value may be in reducing ambiguity at the moment when endoscopic therapy transitions into longitudinal cancer-risk management. For fellows and trainees, the paper also provides a useful reminder that pathology reports are not passive documents. They are active clinical tools. What clinicians should conclude Clinicians can conclude that an international expert panel reached consensus on 56 recommendations for pathology assessment and reporting after ESD for early gastrointestinal cancer. The standards address invasion measurement, margin definition, staining, specimen handling, histological risk features, and clinically relevant reporting. They can also conclude that the paper supports more structured, risk-based, synoptic-ready reporting after ESD. This is especially relevant where endoscopic resection is being used for early invasive carcinoma and post-resection decisions depend heavily on histological risk features. Clinicians should not conclude that this article proves improved survival, reduced recurrence, or reduced need for surgery. The accessible source does not report patient outcomes after implementation of these standards. It also does not validate a new risk model. Instead, it provides consensus criteria intended to improve reporting consistency and support future validation. What remains uncertain Several important questions remain. First, implementation may vary by pathology resources. Complete embedding, standardised inking, parallel sectioning, and selective ancillary staining require workflow alignment. Institutions with high ESD volume may adopt these standards more readily than low-volume centers. Second, the clinical effect of implementation still needs study. Future research should evaluate whether synoptic, risk-based reports reduce reporting variability, improve agreement in multidisciplinary recommendations, reduce unnecessary surgery, identify patients needing additional treatment more accurately, or improve long-term outcomes. Third, the role of submucosal invasion breadth remains investigational in the accessible abstract. The panel recommends reporting it as an adjunct metric for future validation, not as an established independent determinant of management. Fourth, the abstract does not provide organ-specific algorithms for every GI site. Early esophageal, gastric, colorectal, and other gastrointestinal cancers may have different risk frameworks. The value of common reporting standards is clear, but clinical application still needs site-specific interpretation within local and international guidelines. Clinical Takeaway The new Gut international consensus standards mark an important step toward more consistent, clinically actionable pathology reporting after ESD for early gastrointestinal cancer. The key message is not that pathology should become more complex; it is that the clinically decisive elements should be measured, defined, and reported in a standardised way. For gastroenterologists and endoscopists, the report after ESD should answer more than “was cancer present?” It should help determine whether resection was likely curative, whether adverse histological features are present, whether margins are truly positive, and whether the case requires further multidisciplinary discussion. This is consensus-based guidance, not outcome-proven evidence of improved survival or recurrence reduction. Its immediate value lies in standardisation. Its future value will depend on whether these reporting standards improve risk prediction, reduce variability, and support better patient-level decisions after endoscopic resection. Five key clinical takeaways The verified article is “Risk-based pathology reporting after endoscopic submucosal dissection for early gastrointestinal cancer: international consensus standards,” published online in Gut on 10 July 2026 . The study design was an international modified Delphi consensus process involving 42 experts from 15 countries , including 28 gastrointestinal pathologists and 14 therapeutic endoscopists . The consensus produced 56 recommendations across seven domains , including invasion measurement, margin assessment, staining, specimen handling, prognostic histological features, and clinically relevant reporting. Key recommendations include reporting submucosal invasion depth in micrometres when muscularis propria is absent, defining margin positivity as direct tumour contact with the inked surface, and using ancillary stains selectively. These standards are intended to reduce reporting variability and improve multidisciplinary decision-making, but the accessible source does not show patient-outcome benefits after implementation. Source reference and link Khalaf K, Li H, Iwaya M, Orr CE, Schneider M, Iwaya Y, Yuan Y, Saito Y, Shimamura Y, Messmann H, Jacques J, Hassan C, Repici A, von Renteln D, Pellisé M, Elkholy S, Anderson JT, Cai M, Pouw RE, Yang D, Chiu PWY, Lauwers GY, Kumarasinghe MP, Ushiku T, Streutker CJ, Wang T, Hurlbut D, Grin A, Bellizzi A, Kim KM, Charissoux A, Fenouil T, Terris B, de Hertogh G, Jansen M, Meijer SL, Vieth M, Nakanishi Y, Kawachi H, Xu C, Abd El-Kareem D, Ohashi K, Brown I, Kirsch R, Singh C, Knight K, Montgomery EA, Cuatrecasas M, Saez de Gordoa K, Bechara R. Risk-based pathology reporting after endoscopic submucosal dissection for early gastrointestinal cancer: international consensus standards. Gut. Published online 10 July 2026. DOI: 10.1136/gutjnl-2025-337567 .

GastroAGI TeamRead more

11 min readJul 22, 202612reads

Cryptogenic Steatotic Liver Disease in Lean Patients: A Hidden Risk Signal for Hepatology Practice

What should clinicians do when hepatic steatosis is present in a lean patient who has no recorded cardiometabolic risk factors? That question sits at the edge of the current steatotic liver disease framework. In many clinics, steatosis is interpreted through a metabolic lens. When obesity, diabetes, dyslipidemia, hypertension, or other cardiometabolic risk factors are present, the diagnostic pathway is relatively intuitive. But a smaller group of patients do not fit that pattern. They have steatotic liver disease, yet they are lean and have no documented cardiometabolic risk factors. The clinical temptation is to consider them lower risk, atypical, or simply “cryptogenic.” A recent Gut article challenges that assumption. The verified article title is “Cryptogenic steatotic liver disease: a lean phenotype associated with increased liver-related mortality.” It was published online ahead of print on July 16, 2026 , with DOI 10.1136/gutjnl-2026-339142 . The accessible source does not support the alternative wording “distinct gut microbiome features”; microbiome findings are not reported in the PubMed-indexed abstract. The article instead focuses on the clinical phenotype and outcomes of operationally defined cryptogenic steatotic liver disease. A phenotype that does not fit the usual metabolic script The study addresses a definitional and clinical gap: cryptogenic steatotic liver disease remains undefined . To make this group measurable, the investigators operationalised cryptogenic SLD as lean SLD without recorded cardiometabolic risk factors . This matters because modern SLD classification depends heavily on metabolic and alcohol-related categories, yet some patients with hepatic steatosis do not clearly meet those categories. This was not an interventional study. There was no drug, procedural exposure, or lifestyle intervention being tested. The main exposure of interest was a clinical phenotype : lean steatotic liver disease without recorded cardiometabolic risk factors. The investigators compared this phenotype with other SLD and non-SLD groups, assessed liver-related features, and examined longitudinal outcomes. The population was drawn from several large cohorts. Clinical characteristics were evaluated in the UK Biobank magnetic resonance proton density fat fraction cohort , where SLD was defined as PDFF ≥5% . Findings were then assessed in a hepatic steatosis index-based UK Biobank cohort and externally validated in three cross-sectional cohorts , including the National Health and Nutrition Examination Survey and the Korean National Health Insurance Service cohorts. Longitudinal outcomes were examined using Cox regression in two cohorts: the UK Biobank hepatic steatosis index cohort and the Korean National Health Insurance Service cohort. What the investigators found In the UK Biobank MR-PDFF cohort, there were 30,847 participants . Among them, 1,195 participants , or 3.4% , had non-obese SLD, defined by body mass index below 25 kg/m² . Within this non-obese SLD group, 13.7% had operationally defined cryptogenic SLD, while 86.3% had lean metabolic dysfunction-associated SLD. The key message is that cryptogenic SLD was not simply a benign “lean fatty liver” category. Across cohorts, cryptogenic SLD showed less favourable metabolic and liver-related profiles than participants without SLD and without cardiometabolic risk factors. This wording is important. It indicates an observed association, not proof that cryptogenic SLD itself causes worse outcomes. The investigators also reported cohort-specific signals that suggest biological heterogeneity. In the UK Biobank MR-PDFF cohort, cryptogenic SLD showed higher contrast-enhanced T1-weighted image values and an increased prevalence of PNPLA3 and TM6SF2 risk variants. In the magnetic resonance elastography cohort, higher fibrosis rates were observed. These findings suggest that the phenotype may include patients with measurable liver injury or fibrotic risk despite the absence of recorded cardiometabolic risk factors. The outcome signal was clinically important. In longitudinal analyses, cryptogenic SLD was associated with liver-related death in the Korean National Health Insurance Service cohort, with a hazard ratio of 2.5 and 95% confidence interval 1.4 to 4.3 . A similar but less precise association was observed in the UK Biobank hepatic steatosis index cohort, with a hazard ratio of 13.2 and 95% confidence interval 1.9 to 92.4 . The Korean cohort association persisted after stricter alcohol exclusion. Why this is clinically relevant For gastroenterologists and hepatologists, the practical relevance is straightforward: absence of obesity and recorded cardiometabolic risk factors should not automatically reassure clinicians when steatosis is present. The study does not say that every lean patient with steatosis is high risk. It does not provide a new management algorithm. It does not establish that cryptogenic SLD should be treated like advanced MASLD. But it does suggest that a subset of lean patients with apparently unexplained steatosis may have liver injury markers, genetic risk enrichment, higher fibrosis rates in one cohort, and increased liver-related mortality in longitudinal analyses. This is especially relevant in real-world practice because lean patients with steatosis may be under-prioritized. Clinicians often focus fibrosis assessment on patients with diabetes, obesity, metabolic syndrome, or abnormal liver enzymes. A lean patient without obvious cardiometabolic disease may not be triaged with the same urgency. The current study supports a more cautious approach: cryptogenic SLD should not be dismissed merely because the metabolic phenotype is not obvious. The evidence should still be interpreted within its design. This was observational research using cohort data. The association with liver-related death is important, but association is not causation. The phenotype may be a marker for unmeasured risk factors, genetic susceptibility, unrecorded metabolic dysfunction, alcohol exposure not fully captured, medication-related steatosis, or other causes not identified in the available datasets. The study’s operational definition is useful for research, but it is not the same as a complete etiologic diagnosis in an individual patient. A useful reminder about “recorded” cardiometabolic risk A subtle but important phrase in the abstract is “without recorded cardiometabolic risk factors.” That does not necessarily mean that cardiometabolic risk factors were biologically absent. It means they were not recorded within the datasets according to the study’s operational criteria. This distinction matters clinically. In practice, a patient may lack a formal diagnosis of diabetes or hypertension but still have insulin resistance, visceral adiposity, sarcopenic obesity, evolving dyslipidemia, or other metabolic abnormalities not captured by standard diagnostic labels. The study does not provide enough detail in the accessible abstract to determine how completely these possibilities were assessed. Therefore, clinicians should avoid interpreting cryptogenic SLD as a purely non-metabolic disease entity. The finding that cryptogenic SLD had less favourable metabolic and liver-related profiles than non-SLD/no-CMRF participants also complicates the idea of a completely “metabolically healthy” group. The available source does not specify which metabolic measures drove that less favourable profile, but the direction of the finding suggests that absence of recorded cardiometabolic risk factors may not equal absence of relevant risk. Genetic signals: intriguing, not deterministic The reported enrichment of PNPLA3 and TM6SF2 risk variants in the UK Biobank MR-PDFF cohort is one of the more interesting aspects of the study. These variants are named in the abstract as being more prevalent in cryptogenic SLD. For clinicians, the correct interpretation is cautious. The finding supports the possibility that genetic susceptibility may contribute to steatosis or liver injury in some lean patients without recorded cardiometabolic risk factors. However, the abstract does not establish that these variants explain the phenotype, nor does it report whether genetic testing should be used clinically. It also does not provide management recommendations based on genotype. The practical implication is not routine genetic testing for every lean patient with steatosis. Rather, the finding should encourage clinicians and researchers to think beyond body mass index and conventional metabolic labels when evaluating steatotic liver disease. It also supports the need for future studies that integrate genetics, imaging, metabolic phenotyping, alcohol exposure assessment, and longitudinal outcomes. Fibrosis and mortality: the signal that should change our level of attention The most clinically meaningful part of the abstract is not simply that cryptogenic SLD exists. It is that this phenotype was associated with liver injury markers and liver-related mortality . The authors’ conclusion states that operationally defined cryptogenic SLD accounted for a substantial proportion of lean SLD and, despite no recorded cardiometabolic risk factors, was associated with liver injury markers and liver-related mortality. That conclusion should shift the clinician’s level of attention, but not yet the standard of care. A reasonable response is heightened clinical vigilance: confirm the presence and degree of steatosis, assess fibrosis risk, review alcohol intake carefully, exclude secondary causes where appropriate, and monitor according to clinical context. However, those actions are general clinical reasoning rather than specific new recommendations from this abstract. The abstract itself does not prescribe a diagnostic or management algorithm. The mortality association also needs careful interpretation. The Korean National Health Insurance Service cohort produced a more precise association than the UK Biobank hepatic steatosis index cohort. The UK Biobank estimate had a wide confidence interval, indicating uncertainty around the magnitude of risk. The direction was similar, but the precision was limited. This is a key point for evidence-based communication. The study suggests a risk signal; it does not define the exact individual-level risk for a patient sitting in clinic. Strengths of the study Several strengths make this article worth attention for GastroAGI readers. First, the investigators used large population-based resources rather than a single specialty clinic cohort. The UK Biobank MR-PDFF cohort provided imaging-defined steatosis, while additional cohorts were used for assessment and validation. Second, the study did not stop at cross-sectional characterization. It examined longitudinal outcomes using Cox regression in two cohorts. That is important because the clinical relevance of any SLD phenotype ultimately depends on outcomes, not merely imaging or biochemical differences. Third, the phenotype was examined across multiple settings, including UK Biobank, NHANES, and Korean National Health Insurance Service cohorts. External validation across different datasets strengthens the credibility of the signal, although it does not eliminate concerns about residual confounding or measurement differences. Fourth, the study incorporated more than one type of liver-related assessment. The abstract mentions MR-PDFF, contrast-enhanced T1-weighted imaging values, magnetic resonance elastography, genetic variants, and liver-related mortality. This multidimensional approach is useful because cryptogenic SLD is unlikely to be explained by one marker alone. Limitations clinicians should keep in mind The accessible source is an abstract-level summary, so some important details are not available. It does not provide the complete phenotype definitions, variable lists, missing-data handling, alcohol thresholds, medication exclusions, liver enzyme values, fibrosis thresholds, cause-of-death adjudication methods, or full adjustment models. The operational definition also has limitations. “Lean SLD without recorded cardiometabolic risk factors” may group together biologically different patients. Some may have unmeasured metabolic dysfunction. Some may have genetic predisposition. Some may have alcohol exposure below documented thresholds. Others may have secondary causes that were not fully captured. The study’s conclusion acknowledges the phenotype but does not resolve its underlying cause. Another limitation is that the study should not be interpreted as proving causality. The association with liver-related mortality may reflect the phenotype itself, associated genetic factors, unmeasured confounders, diagnostic misclassification, or other factors. Observational cohort design can identify risk signals, but it cannot by itself prove that cryptogenic SLD causes liver-related death. Finally, because the source does not report gut microbiome data, clinicians should not use this article to make microbiome-based claims about cryptogenic SLD. Any discussion of microbiome mechanisms would require a different source. What clinicians should and should not conclude Clinicians should conclude that operationally defined cryptogenic SLD is a measurable phenotype within lean SLD and that, in this study, it was associated with less favourable liver-related profiles and liver-related mortality. They should also recognize that a lack of recorded cardiometabolic risk factors does not necessarily mean a patient with steatosis is low risk. Clinicians should not conclude that the study creates a new treatment pathway, proves causality, or justifies genotype-based management. They should not conclude that cryptogenic SLD is microbiome-driven based on this source. They should also not assume that all lean patients with steatosis have the same prognosis. The phenotype is heterogeneous, and the available abstract does not provide enough detail for individualized risk prediction. For fellows and researchers, the study is a strong reminder that nomenclature is not just semantics. How we classify SLD affects who is studied, who is monitored, and who may be missed. Cryptogenic SLD may represent a diagnostic blind spot rather than a benign residual category. Where future research should go The next research step is clearer phenotyping. Future studies should determine whether cryptogenic SLD represents unmeasured metabolic dysfunction, genetic susceptibility, environmental exposures, alcohol misclassification, medication effects, sarcopenia-related risk, or multiple overlapping pathways. Prospective studies are also needed. A clinically useful framework would show how to identify these patients in routine practice, which fibrosis assessment strategy is most appropriate, and whether earlier recognition changes outcomes. The current study identifies an association with liver-related mortality, but it does not show that intervention based on this phenotype reduces that risk. More granular reporting would also help. Clinicians need to know how risk varies by age, sex, ethnicity, liver enzyme profile, fibrosis markers, imaging features, and genotype. Without that information, cryptogenic SLD remains clinically important but incompletely actionable. Clinical Takeaway This Gut study suggests that cryptogenic steatotic liver disease, operationalised as lean SLD without recorded cardiometabolic risk factors, should not be dismissed as a benign phenotype. In large cohort analyses, this group represented a meaningful subset of non-obese SLD and was associated with liver injury markers and liver-related mortality. The evidence is clinically relevant but observational. It should increase awareness and encourage careful fibrosis-oriented assessment in lean patients with steatosis, while avoiding premature causal claims or unsupported management recommendations. Most importantly, the verified source supports a mortality-risk phenotype—not a microbiome-based conclusion. Five key clinical takeaways The verified article title is “Cryptogenic steatotic liver disease: a lean phenotype associated with increased liver-related mortality,” published in Gut on July 16, 2026 . The study operationalised cryptogenic SLD as lean steatotic liver disease without recorded cardiometabolic risk factors . In the UK Biobank MR-PDFF cohort, cryptogenic SLD accounted for 13.7% of non-obese SLD . Cryptogenic SLD was associated with liver injury markers, higher fibrosis rates in an MRE cohort, and liver-related mortality in longitudinal analyses. This is observational evidence and should be interpreted as a risk signal, not proof of causation or a new clinical guideline. Source reference and link Yoon EL, Lee HY, Lee J, et al. Cryptogenic steatotic liver disease: a lean phenotype associated with increased liver-related mortality. Gut. Published online July 16, 2026. DOI: 10.1136/gutjnl-2026-339142.

GastroAGI TeamRead more

5 min readJul 21, 202646reads

GVAX Pancreatic Cancer Vaccine: What the 2026 Trial Data Means for PDAC Care

A patient two weeks out from a Whipple for resected pancreatic adenocarcinoma sends you a news link about a "cancer vaccine" and asks if she should be getting it instead of chemotherapy. The article is accurate but incomplete, and the gap between what it says and what a GI oncology team actually does with this data is exactly where these conversations go wrong. Here's what the vaccine is, what the evidence supports, and how to answer her. The Clinical Problem With Pancreatic Cancer Vaccine Coverage The GVAX pancreatic cancer vaccine is not new, but it keeps resurfacing in patient-facing media every time a new trial reports out, and each cycle of coverage tends to flatten years of incremental data into a single hopeful headline. Developed at Johns Hopkins by Elizabeth Jaffee and Daniel Laheru, GVAX is a whole-cell, GM-CSF–secreting allogeneic vaccine designed to convert pancreatic ductal adenocarcinoma from an immunologically "cold" tumor into one immune cells will actually enter and attack. That mechanism is real and well-documented. What's harder to communicate is that GVAX remains investigational, is never used as monotherapy in place of chemotherapy, and its clinical benefit so far is measured mostly in immune correlates and pathologic response rather than confirmed survival gains in randomized trials. Patients need that distinction before they build expectations around it. How GVAX Primes an Anti-Tumor Immune Response GVAX is manufactured from irradiated allogeneic pancreatic cancer cell lines genetically modified to secrete granulocyte-macrophage colony-stimulating factor. The irradiation halts proliferation without destroying the cells' antigenic surface proteins. Injected intradermally, the GM-CSF acts as a chemoattractant, drawing dendritic cells and other antigen-presenting cells to the injection site, where they take up tumor antigens — including mesothelin, a target expressed on most PDAC cells — and traffic to regional lymph nodes to prime CD8+ T cells. This matters clinically because PDAC is characteristically immune-excluded: dense desmoplastic stroma and an immunosuppressive microenvironment keep effector T cells from infiltrating the tumor, which is also why single-agent PD-1/PD-L1 blockade has performed poorly in this disease. The rationale for GVAX is that priming a mesothelin-specific T-cell response systemically, then pairing it with a second hit — cyclophosphamide to blunt regulatory T cells, or stereotactic body radiation to release additional antigen and remodel the stroma — creates a combination that checkpoint inhibitors alone cannot. A 2026 Nature Communications study formalized this logic in locally advanced disease, sequencing induction chemotherapy, then cyclophosphamide, pembrolizumab, GVAX, and SBRT before reassessing patients for resection. Case in Point A 61-year-old man presents with locally advanced pancreatic adenocarcinoma of the body, encasing the celiac axis on staging CT, with no distant metastases. He completes eight cycles of FOLFIRINOX with a partial response but remains borderline unresectable on restaging imaging. His oncology team enrolls him on a neoadjuvant protocol combining cyclophosphamide, pembrolizumab, GVAX, and SBRT prior to surgical reassessment. Three months later, restaging shows no new lesions and reduced vascular encasement, and he proceeds to a margin-negative distal pancreatectomy. His surgical pathology shows increased CD8+ T-cell infiltration compared to historical FOLFIRINOX-and-SBRT-only specimens — the biologic signal the trial was actually designed to detect, distinct from a guarantee of long-term survival benefit. This is the honest framing to give patients: converting an unresectable tumor into a resectable one is a meaningful, discussable outcome — a confirmed survival advantage from the vaccine component is not yet established. What the Trial Data Actually Shows The data supporting GVAX has accumulated in stages, and none of it currently supports monotherapy or first-line use. The original Johns Hopkins phase II adjuvant trial in 60 resected, mostly node-positive patients reported a median overall survival of 24.8 to 26.8 months across reported analyses, compared with a historical range of 17 to 22 months for surgery plus adjuvant chemoradiation alone — encouraging, but a single-arm comparison against historical controls, not a randomized result. A later phase II study combining GVAX with ipilimumab as maintenance therapy in metastatic disease, published in Clinical Cancer Research , showed the addition of checkpoint blockade did not clearly improve outcomes over GVAX alone in that setting. The more relevant recent work is neoadjuvant: the borderline-resectable GVAX-cyclophosphamide-nivolumab-SBRT trial and the 2026 locally advanced pembrolizumab-based protocol both used the vaccine to convert marginal candidates into surgical candidates while tracking immune infiltration as the primary endpoint, not survival. That distinction — pathologic and immunologic endpoints versus confirmed overall survival — is the single most important thing to communicate when a patient brings in coverage of this vaccine. A Frequently Overlooked Point The detail patients and even some referring clinicians miss is that GVAX's value proposition was never "better chemotherapy" — it was a strategy to overcome PDAC's specific resistance to checkpoint inhibition by manufacturing the T-cell response those drugs need to work with. Every combination trial since the original adjuvant study has been chasing that same problem: checkpoint blockade alone fails in PDAC because there's nothing for it to unleash. Framing GVAX as a vaccine that "boosts immunity" in the generic sense undersells the actual mechanism and invites patients to weigh it against chemotherapy as if they were interchangeable options, when in every published protocol it has been given alongside standard therapy, never instead of it. Bottom Line for Clinical Practice GVAX is available only through clinical trials at select academic centers — it has no FDA approval and should never be presented to patients as a standard-of-care alternative to chemotherapy. Refer interested, eligible patients to ongoing trials (search NCT registries for active Johns Hopkins/Sidney Kimmel protocols) rather than describing it as generally accessible immunotherapy. In resected disease, the survival data behind GVAX comes from single-arm trials against historical controls — communicate this as promising, not proven. In locally advanced or borderline resectable disease, the current evidence supports GVAX-based regimens as a strategy to improve resectability and immune infiltration, not as a confirmed survival intervention. Correcting the "vaccine versus chemotherapy" framing before a patient walks into oncology consult saves that visit from starting on the wrong premise. Next time a patient forwards you a headline like this mid-workup, run the actual trial data and staging details through GastroAGI — it separates what's investigational from what's actionable and gives you a guideline-anchored answer before the follow-up visit. Read Next: Sphingolipid Metabolism and KRAS in Pancreatic Cancer: A New Translational Signal

GastroAGI TeamRead more

11 min readJul 21, 202625reads

Machine Learning for Liver Fibrosis Prediction in MASLD: Promising Tool or Premature Clinical Leap?

Can machine learning help gastroenterologists identify clinically important fibrosis in MASLD more accurately than the non-invasive tools already used in practice? That is the central question raised by the conference paper “Machine Learning Prediction of Liver Fibrosis in Patients with Metabolic Dysfunction Associated Steatotic Liver Disease,” published in Digestive and Liver Disease , Volume 58, Supplement 1, February 2026, page S130, as abstract F-80 . The work is authored by Salvatore Petta, Grazia Pennisi, Ciro Celsa, Sofiane Messaoudi, Emmanuel Tsochatzis, Elisabetta Bugianesi, Masato Yoneda, Ming-Hua Zheng, Hannes Hagström, Jérôme Boursier, José Luis Calleja, George Boon-Bee Goh, Wah-Kheong Chan, Rocio Gallego-Durán, Arun J. Sanyal, and collaborators. The DOI listed is 10.1016/j.dld.2026.01.204 . This is not a therapeutic trial and not a guideline. It is a machine-learning prediction-model development and validation study , reported as a conference paper/abstract. Its clinical focus is highly relevant: identifying liver fibrosis stage in patients with metabolic dysfunction-associated steatotic liver disease , or MASLD. The source explicitly frames the unmet need: accurate fibrosis staging is crucial; liver biopsy is invasive; and commonly used non-invasive tests such as FIB-4, liver stiffness measurement, and Agile 3+ may not always provide sufficient accuracy. The clinical problem: fibrosis risk, not steatosis alone For clinicians managing MASLD, the practical challenge is not simply detecting steatosis. The harder question is determining which patients have clinically meaningful fibrosis and therefore need closer assessment, follow-up, or specialist prioritization. The abstract starts from this point. Liver biopsy remains a reference standard for staging fibrosis, but it is invasive and not suited to broad, repeated, routine risk stratification. In daily practice, clinicians rely on non-invasive tests. FIB-4 is simple and accessible. Liver stiffness measurement provides imaging-based assessment. Agile 3+ combines selected parameters to estimate advanced fibrosis risk. Yet each test leaves room for uncertainty, especially when patients fall into indeterminate zones or when results are discordant. Machine learning is attractive in this setting because fibrosis risk may be influenced by multiple clinical and laboratory variables that interact in ways not easily captured by a fixed linear formula. The study investigated whether models trained on routinely collected clinical data could improve fibrosis assessment in biopsy-proven MASLD. This is the right type of question for AI in hepatology: not replacing clinicians, but potentially improving triage and staging where current tools are imperfect. What the investigators studied The investigators developed machine-learning models using a training cohort of biopsy-proven MASLD patients . The models incorporated 22 clinically relevant variables , including liver stiffness measurement. Internal validation used a 25% random test split across five MICE-imputed datasets , and external validation was performed in three independent cohorts from different centers . Several model types were evaluated. These included FT-Transformer , TabNet with and without liver stiffness measurement , and two ordinal models: MLP-CORAL and CoralTabNet . The binary prediction tasks focused on detecting fibrosis thresholds of F≥3 , F≥2 , and F4 . The authors also evaluated multiclass staging across F0 to F4 . This design matters clinically because it does not only ask whether a model can identify cirrhosis. It also asks whether machine learning can distinguish fibrosis thresholds that may influence monitoring intensity, referral decisions, and staging confidence. The use of biopsy-proven MASLD patients anchors the model against histological staging, while external validation across three cohorts is an important strength for assessing generalizability. At the same time, the accessible source is an abstract. It does not provide all details needed for full methodological appraisal, such as exact sample size, inclusion and exclusion criteria, the complete list of variables, missing data proportions, model calibration, decision-curve analysis, subgroup performance, or how the three external cohorts differed. These gaps are important when considering clinical adoption. Key findings: better performance, but not uniformly superior For detection of F≥3 fibrosis , FT-Transformer and TabNet achieved AUCs of 0.860 and 0.855 , respectively. The reported grey zones were small: 8.2% and 8.4% . These models outperformed FIB-4 , which had an AUC of 0.756 , and liver stiffness measurement alone, reported with an AUC of 0.837 . However, they did not significantly outperform Agile 3+ , which had an AUC of 0.847 with a reported p value of 0.12 . For F≥2 fibrosis , FT-Transformer and TabNet reached AUCs of 0.800 and 0.794 . These exceeded FIB-4, reported at 0.713 , and liver stiffness measurement, reported at 0.775 . Again, the models did not meaningfully outperform Agile 3+, which had an AUC of 0.794 and a p value of 0.548 . For F4 fibrosis , FT-Transformer and TabNet achieved AUCs of 0.834 and 0.831 , with the abstract stating that they outperformed conventional non-invasive tests. In multiclass staging from F0 to F4, MLP-CORAL showed the best agreement with biopsy, with a quadratic weighted kappa of 0.616 . Models without liver stiffness measurement performed consistently worse across tasks. The authors report that findings were confirmed across all three external validation cohorts, with stable AUROC values and minimal grey zones. These findings suggest that machine-learning models, especially FT-Transformer and TabNet, may improve discrimination compared with some established non-invasive tools. The most clinically relevant nuance is that superiority was not universal. For important thresholds such as F≥3 and F≥2, the models outperformed FIB-4 and liver stiffness measurement in the reported results, but did not significantly outperform Agile 3+. That distinction matters. It prevents overinterpreting the study as showing that machine learning is clearly better than all existing tools. Why liver stiffness measurement still matters One of the most clinically useful findings is that models without liver stiffness measurement performed worse across all tasks. This has two implications. First, machine learning did not appear to eliminate the value of elastography-derived information. Instead, liver stiffness measurement remained an important component of model performance. Second, a model requiring liver stiffness measurement may be less immediately scalable in settings where elastography access is limited. That does not make the model unhelpful. It simply clarifies the use case. A machine-learning tool that integrates liver stiffness measurement may be best positioned as an enhanced interpretation layer after elastography, not necessarily as a universal first-line screening test based only on routine blood work. For busy hepatology and gastroenterology services, this distinction is practical. If the model requires liver stiffness measurement, it may improve staging confidence after a patient has already entered a liver assessment pathway. If future versions can perform well without liver stiffness measurement, they may be more useful in primary-care triage or large population-level MASLD screening. The abstract reports that models without liver stiffness measurement performed worse, so that broader use case remains less supported by this source. Practical interpretation for clinicians The reported results are encouraging, but they should be interpreted as prediction-model evidence, not clinical guidance. A model with an AUC around 0.86 for F≥3 fibrosis is potentially useful, especially if it reduces the number of indeterminate results. The abstract’s report of small grey zones is clinically interesting because grey-zone results are a common limitation of non-invasive fibrosis algorithms. If reproduced in full manuscripts and prospective care pathways, reducing uncertainty could help clinicians make clearer decisions about further testing or referral. However, clinicians should not conclude that this model is ready to replace biopsy in every ambiguous case. The authors conclude that machine-learning models integrating routine clinical variables perform better than commonly used non-invasive tests and may provide an efficient, widely applicable alternative to biopsy in routine clinical care. That is the study’s conclusion, but from a clinical implementation standpoint, several questions remain unanswered in the accessible source. The abstract does not report whether use of the model improves patient outcomes. It does not show whether the model changes referral patterns, reduces unnecessary biopsy, prevents missed advanced fibrosis, or improves cost-effectiveness. It also does not provide workflow details: where the model would sit in the MASLD care pathway, which threshold should trigger elastography or hepatology referral, and how clinicians should manage discordant results between the model, FIB-4, liver stiffness measurement, and Agile 3+. Therefore, the best interpretation is balanced: this study supports machine learning as a promising adjunct for fibrosis staging in MASLD, particularly when liver stiffness measurement is available, but it does not establish a new standard of care. Association, prediction, and causation This study should be understood as a prediction study. It does not test a causal intervention. The machine-learning models identify patterns associated with fibrosis stage as defined against biopsy and fibrosis thresholds. They do not prove that any individual variable causes fibrosis progression, nor do they show that using the model will improve liver-related outcomes. That distinction is especially important in AI-based clinical research. A model may predict fibrosis accurately because it identifies statistical patterns in the dataset. Those patterns can be clinically useful, but they are not the same as biological causation. For example, if a variable contributes strongly to prediction, that does not mean modifying that variable will necessarily change fibrosis stage. Prediction can guide risk stratification; causation guides intervention. For GastroAGI readers, the central message is that the model may help identify who is more likely to have advanced fibrosis, but it should not be used to infer mechanistic drivers of fibrosis unless supported by separate biological or interventional evidence. Strengths of the evidence Several strengths make this abstract worth attention. First, the study uses biopsy-proven MASLD patients for model development. Histology-based staging provides a clinically meaningful reference point for fibrosis classification, even though biopsy itself has known practical limitations. Second, the investigators evaluated multiple clinically relevant fibrosis thresholds: F≥2, F≥3, F4, and multiclass F0–F4 staging. This is more informative than focusing only on cirrhosis or only on advanced fibrosis. Third, the models were internally validated with a held-out 25% test split across five MICE-imputed datasets, suggesting attention to missing data and internal performance assessment. Fourth, external validation was performed in three independent cohorts from different centers. External validation is essential for AI models because performance in a development cohort often overestimates real-world utility. Finally, the comparison with established non-invasive tests provides clinical context. A model’s performance is meaningful only if compared with what clinicians already use. The abstract reports comparisons with FIB-4, liver stiffness measurement, and Agile 3+. Limitations and unanswered questions The most important limitation is that the available source is a conference abstract, not a full peer-reviewed manuscript with complete methods and supplementary details. It provides promising performance metrics but not the level of information needed for full clinical appraisal. Several questions remain open. What was the sample size in the training and external validation cohorts? What were the inclusion criteria? Were patients recruited from tertiary hepatology clinics, metabolic clinics, or broader clinical settings? What was the prevalence of F≥2, F≥3, and F4 disease? How did performance vary across age, sex, diabetes status, body mass index, alcohol exposure, ethnicity, and geography? Were calibration plots reported? Were decision-curve analyses performed? How were clinically actionable thresholds selected? These details matter because MASLD populations are heterogeneous. A model performing well in biopsy-enriched cohorts may behave differently in primary care or population screening, where advanced fibrosis prevalence is lower. Similarly, a model requiring liver stiffness measurement may be less applicable where elastography is unavailable or technically unreliable. The source also does not establish prospective clinical utility. Before implementation, a model should ideally be tested in real clinical workflows to determine whether it improves appropriate referrals, reduces unnecessary testing, shortens diagnostic pathways, or identifies advanced fibrosis earlier without increasing harm. What clinicians should and should not conclude Clinicians can reasonably conclude that machine-learning models integrating routine clinical variables and liver stiffness measurement showed promising discrimination for fibrosis staging in biopsy-proven MASLD cohorts. They can also conclude that FT-Transformer and TabNet performed better than FIB-4 and liver stiffness measurement alone for selected fibrosis thresholds in the reported results, but not significantly better than Agile 3+ for F≥3 and F≥2. Clinicians should not conclude that machine learning has replaced established MASLD fibrosis pathways. They should not use these findings to avoid biopsy when biopsy is otherwise clinically indicated. They should not assume the model is validated for every healthcare setting, every ethnicity, every disease prevalence, or every elastography platform. They should also not interpret the model as identifying causal drivers of fibrosis progression. Most importantly, clinicians should resist the common AI trap: equating improved AUC with immediate clinical adoption. AUC is useful, but clinical implementation requires interpretability, calibration, reproducibility, workflow integration, safety monitoring, and clear action thresholds. Future research and implementation priorities The next step should be publication of full methods and prospective validation. Future studies should clarify the exact variables used, model calibration, subgroup performance, decision thresholds, and implementation strategy. Prospective studies should ask whether model-guided care changes decisions. Does it reduce the indeterminate group? Does it improve selection for elastography, hepatology referral, biopsy, or surveillance? Does it perform well in lower-prevalence populations? Can it be embedded into electronic health records without adding clinician burden? Can clinicians understand and trust the output? Another priority is comparison against complete care pathways, not isolated tests. In practice, clinicians do not use FIB-4 in isolation; they combine non-invasive tests with metabolic risk profile, imaging, longitudinal trends, and clinical judgment. Machine learning should be judged against this real-world standard. Clinical Takeaway This Digestive and Liver Disease conference paper suggests that machine-learning models, particularly FT-Transformer and TabNet, may improve fibrosis prediction in biopsy-proven MASLD when routine clinical variables and liver stiffness measurement are integrated. The findings are clinically relevant because fibrosis staging remains central to MASLD management, and current non-invasive tests leave room for uncertainty. However, the evidence remains abstract-level prediction-model research. It is promising, not practice-changing. The study supports further validation and clinical implementation research, but it does not yet justify replacing established non-invasive pathways, biopsy when clinically needed, or clinician judgment. Five key clinical takeaways Verified source: This is a Digestive and Liver Disease conference paper/abstract, not a CGH full article, published in February 2026. Study design: Machine-learning prediction-model development with internal validation and external validation in three independent cohorts. Population: Biopsy-proven MASLD patients; exact sample size was not available in the accessible abstract. Main finding: FT-Transformer and TabNet showed promising AUCs for F≥3, F≥2, and F4 fibrosis, outperforming FIB-4 and liver stiffness measurement in reported comparisons, but not Agile 3+ for F≥3 or F≥2. Clinical interpretation: This is promising AI-based risk stratification evidence, but it remains hypothesis-generating and not yet practice-changing . Source reference and link Petta S, Pennisi G, Celsa C, et al. Machine Learning Prediction of Liver Fibrosis in Patients with Metabolic Dysfunction Associated Steatotic Liver Disease. Digestive and Liver Disease. 2026;58(Suppl 1):S130. DOI: 10.1016/j.dld.2026.01.204.

GastroAGI TeamRead more

5 min readJul 19, 202624reads

Coffee and Gut Microbiota: What the 2024 Metagenomic Data Means for GI Practice

A 42-year-old patient returns from a colon resection and is back on solid food by day three, weeks ahead of the ward's usual timeline. Her surgeon credits the two cups of coffee she insisted on the morning after surgery. She's not wrong to give coffee the credit - but the mechanism she assumes, caffeine kicking the bowel into gear, is likely only half the story. A growing body of metagenomic and clinical data suggests coffee's gut effects run through the microbiota itself. The core clinical problem For decades, coffee's effect on bowel function has been attributed almost entirely to the gastrocolic reflex - caffeine-driven stimulation of colonic motility and transient hormonal signaling. That explanation is tidy, but it doesn't hold up against several inconvenient findings: decaffeinated coffee produces similar motility effects in vitro, caffeine alone fails to replicate coffee's post-surgical recovery benefit, and coffee-associated shifts in gut microbiota composition persist independent of caffeine content. The clinical relevance of coffee and gut microbiota research isn't academic. Patients ask about coffee constantly - after GI surgery, with IBS, with reflux, with IBD - and "it stimulates your colon" is an incomplete answer that misses the polyphenol- and melanoidin-driven mechanisms now documented at scale. Coffee-Responsive Bacteria and the Lawsonibacter Signal The clearest evidence for a direct coffee-microbiota interaction comes from a 2024 Nature Microbiology analysis of metagenomic data from more than 54,000 stool samples paired with dietary records from over 22,000 participants across European and US cohorts. The study identified a consistent, dose-related association between coffee intake and increased abundance of Lawsonibacter asaccharolyticus - a species whose functional role is still being characterized. Critically, this association held across independent cohorts with different dietary backgrounds and was present even among decaffeinated coffee drinkers, which rules out caffeine as the sole driver. In vitro work in the same study showed coffee directly promoting bacterial growth in culture, supporting a biological rather than purely epidemiologic link. A separate 2024 Nutrients review adds texture to this picture: moderate coffee intake (roughly under four cups daily) was associated with higher Bifidobacterium abundance, lower Enterobacteriaceae - a family that signals dysbiosis when it overgrows - and greater overall microbial diversity. For clinicians, that diversity signal matters more than any single taxon; it's one of the more reproducible markers of a resilient gut ecosystem across microbiome literature. Case in point A 55-year-old man with quiescent Crohn's disease reports a new pattern of loose stools and mild abdominal cramping over the past two months. His medications are unchanged, his calprotectin is mildly elevated but not flaring-range, and he mentions, almost as an aside, that he's increased his coffee intake to six or seven cups a day since starting a demanding new job. The workup didn't reveal a flare requiring escalation. Instead, the timeline pointed to intake above the threshold - more than five cups daily - that has been associated with both GERD and Crohn's disease progression in recent population data. Cutting back to two to three cups resolved his symptoms within two weeks, without any medication change. The lesson isn't that coffee is dangerous in IBD; it's that dose matters, and a dietary detail easily dismissed as incidental can be the actual variable driving symptoms. Beyond Caffeine: Mechanism and Dose Thresholds The bioactive compounds most likely responsible for coffee's microbiota effects are chlorogenic acid derivatives and melanoidins. Chlorogenic acid polyphenols reach the colon largely intact, where bacterial metabolism converts them into bioactive metabolites; melanoidins formed during roasting are partially fermentable and appear to act as prebiotic substrates. Together, these compounds plausibly explain why decaffeinated coffee reproduces effects that caffeine alone does not. Dose matters in both directions. Moderate intake correlates with favorable shifts - higher Bifidobacterium , lower Enterobacteriaceae , greater diversity - but the same literature flags more than five cups daily as a threshold associated with increased GERD risk and Crohn's disease progression. A separate 2025 population study linked excessive caffeine intake specifically to chronic constipation, a useful counterpoint for patients who assume more coffee always means more regularity. The practical takeaway for counseling: moderate coffee intake is not something to routinely restrict, but intake above roughly four to five cups daily deserves a specific conversation, especially in patients with reflux or active/quiescent IBD. A frequently overlooked point The instinct in clinic is to treat coffee as a binary - recommend it after GI surgery, discourage it in reflux or functional bowel disease - without asking about quantity or caffeination status. That binary approach misses the dose-response relationship the recent data actually describes, and it ignores that decaffeinated coffee is not a "safe" substitute that removes GI risk; it carries much of the same microbiota-modulating activity as regular coffee, just without the caffeine-driven motility kick. A patient switching to decaf to manage reflux may still be getting a meaningful polyphenol and melanoidin exposure worth discussing. Bottom line for clinical practice Counsel post-GI-surgery patients that moderate coffee intake (caffeinated or decaf) may support earlier bowel recovery through microbiota effects, not caffeine alone. Flag intake above five cups daily as a threshold associated with GERD symptoms and Crohn's disease progression - worth asking about directly in flare or reflux workups. Don't assume decaffeinated coffee is GI-neutral; its polyphenol and melanoidin content still interacts with the microbiota. In patients reporting new or excess caffeine intake with constipation, consider intake reduction before escalating a functional workup. Moderate coffee intake (under ~4 cups/day) is associated with higher microbial diversity and Bifidobacterium abundance - a reasonable talking point when patients ask if coffee is "bad" for their gut. Closing The next time a patient's GI symptoms don't fit the expected pattern, a two-line dietary history on coffee intake and caffeination status might surface the variable the standard workup misses. Walk a case like this through GastroAGI for a guideline-anchored, evidence-checked read in seconds.

GastroAGI TeamRead more

7 min readJul 16, 202621reads

Intrapancreatic Fat Deposition and Mortality: Is the Fatty Pancreas Becoming a Risk Marker?

Introduction The “fatty pancreas” has traditionally received far less clinical attention than fatty liver. In routine practice, pancreatic fat is often noticed incidentally on cross-sectional imaging, mentioned briefly, and rarely acted upon. That may be changing. A new prospective UK Biobank study published in The American Journal of Gastroenterology examined whether intrapancreatic fat deposition , or IPFD, is associated with mortality. The study, titled “Association of Intrapancreatic Fat Deposition with Mortality: A Prospective Cohort Study with Genetic Risk Profiling,” used MRI-based pancreatic fat quantification, survival analysis, genome-wide association analysis, and polygenic risk scoring. The headline finding is clinically intriguing: excessive baseline IPFD was associated with a higher risk of all-cause mortality and mortality from vascular diseases. The authors also reported consistent genetic-risk findings using an IPFD-linked polygenic risk score. This does not mean that pancreatic fat measurement should become a routine screening test tomorrow. But it does suggest that pancreatic fat may be more than an incidental imaging descriptor. Why this update matters Pancreatic fat sits at the intersection of pancreatology, metabolic disease, diabetes risk, vascular risk, imaging biomarkers, and preventive gastroenterology. For years, clinicians have understood ectopic fat mainly through the lens of metabolic dysfunction-associated steatotic liver disease. The pancreas, however, is also metabolically active and closely linked to both endocrine and exocrine disease. The same research group previously reported, using UK Biobank data, that elevated IPFD was associated with higher risks of acute pancreatitis, pancreatic cancer, and diabetes mellitus. In that earlier prospective cohort, IPFD was quantified using MRI and a deep learning-based framework, and fatty change of the pancreas was reported in 17.86% of 42,599 participants. That prior study made a pancreatology-focused argument: pancreatic fat may matter for pancreatic outcomes. The newer mortality study expands the question: could pancreatic fat also carry broader prognostic information about survival, particularly vascular mortality? For gastroenterologists, the relevance is not that we now have a new treatment target. Rather, the update encourages us to think more carefully about incidental pancreatic fat, metabolic risk clustering, and the possibility that pancreatic steatosis may reflect systemic cardiometabolic vulnerability. What the study did The investigators analyzed participants from the UK Biobank and divided them into two cohorts based on whether MRI-quantified IPFD was available. They used Kaplan-Meier survival analysis and multivariable Cox proportional hazards modeling to examine mortality outcomes. They also performed a genome-wide association study for IPFD and used Mendelian randomization methods. An IPFD-linked polygenic risk score was then applied in an MRI-naïve cohort. The study included 55,058 participants in the MRI-based analysis. During a median follow-up of 4.9 years, 695 participants died, representing 1.26% of the analyzed cohort. Excessive baseline IPFD was significantly associated with increased all-cause mortality, with a reported hazard ratio of 1.081, and with mortality from vascular diseases, with a reported hazard ratio of 1.247. The genetic component is important. The authors identified 38 significant IPFD-associated single nucleotide polymorphisms. A polygenic risk score derived from these variants showed significant associations with all-cause mortality and vascular mortality. In a larger MRI-naïve cohort of 354,761 participants, the authors reported consistent results when the polygenic score was used as a genetic proxy for IPFD. In plain clinical language, the study asks two related questions. First, do people with more pancreatic fat on MRI have higher mortality risk? Second, does genetic predisposition to higher pancreatic fat point in the same direction? What the study found The main finding was that excessive IPFD was associated with future mortality, especially vascular mortality. The reported effect size for all-cause mortality was modest, while the association with vascular disease mortality appeared stronger. This distinction matters. A modest association with all-cause mortality should not be overinterpreted as a direct causal pathway. All-cause mortality is influenced by many factors, including age, cardiometabolic health, smoking, alcohol use, socioeconomic context, malignancy, and comorbidity burden. The vascular mortality signal is more biologically plausible in the context of ectopic fat and metabolic risk. Pancreatic fat may coexist with visceral adiposity, dyslipidemia, insulin resistance, hypertension, hepatic steatosis, and systemic inflammatory changes. The study’s genetic-risk approach strengthens the signal, but it does not convert IPFD into a ready-to-use clinical decision tool. The best interpretation is cautious: IPFD may be a measurable imaging biomarker that captures a component of metabolic and vascular risk not fully represented by traditional risk factors. Clinical interpretation For the practicing gastroenterologist, the key question is not “Should we treat pancreatic fat?” The better question is: What should we do when pancreatic fat is identified? At present, there is no guideline recommending routine MRI quantification of pancreatic fat for mortality prediction. There is also no validated intervention that specifically targets pancreatic fat independent of broader metabolic risk management. However, the finding may still be clinically useful. When pancreatic fat is reported on imaging, it should not be dismissed automatically. It may be a prompt to look for cardiometabolic risk: diabetes or prediabetes, obesity, dyslipidemia, hypertension, alcohol exposure, metabolic syndrome, hepatic steatosis, and history of pancreatitis. This is similar to how incidental hepatic steatosis has evolved. A finding that was once treated as benign background noise is increasingly understood as a signal for metabolic disease, fibrosis risk, and cardiovascular outcomes. Pancreatic fat is not at that level of clinical maturity, but it may be moving in that direction. The study also fits with a growing body of data suggesting that the pancreas may be affected by ectopic fat in clinically meaningful ways. The earlier UK Biobank analysis found that elevated IPFD was associated with incident acute pancreatitis, pancreatic cancer, and diabetes mellitus, even after adjustment for multiple cardiometabolic variables. Together, these studies support the idea that pancreatic fat deserves more research attention, particularly in metabolic GI clinics, pancreatitis cohorts, diabetes-risk populations, and imaging biomarker studies. Practical implications for gastroenterologists The immediate implication is awareness, not protocol change. If a radiology report mentions fatty pancreas or pancreatic steatosis, gastroenterologists can use it as a cue to assess the broader metabolic context. That may include checking glycemic status, lipid profile, body weight trajectory, alcohol history, liver steatosis, and cardiovascular risk factors. For patients with pancreatitis risk, pancreatic fat may eventually become part of a more refined risk model. The prior UK Biobank study linked fatty change of the pancreas with higher risks of acute pancreatitis and pancreatic cancer, although these associations still require validation before routine clinical use. For researchers, this study is more provocative. It suggests the need for standardized definitions of pancreatic steatosis, reproducible imaging thresholds, ethnic and geographic validation, longitudinal follow-up, and mechanistic studies. We also need to know whether reducing ectopic fat through weight loss, metabolic therapy, alcohol reduction, or diabetes prevention changes pancreatic fat and whether that change modifies outcomes. For radiologists and endosonographers, the study raises an important reporting question. Should pancreatic fat be described more consistently? At present, many reports mention pancreatic atrophy, fatty replacement, or lipomatosis without standardized quantification. MRI-based research tools may not be directly transferable to routine CT, MRI, or EUS workflows. Limitations and caution This study should not be presented as proof that pancreatic fat causes death. It is an observational prospective cohort study with genetic-risk profiling. Even with multivariable modeling and genetic analyses, residual confounding and selection effects remain possible. UK Biobank participants are not a perfect representation of the general population. The median follow-up of 4.9 years is useful but not long enough to fully capture lifetime pancreatic, metabolic, cancer, and vascular outcomes. The all-cause mortality association was statistically significant but modest, which means clinical interpretation should remain measured. Another limitation is clinical actionability. We do not yet have accepted IPFD thresholds for routine care, nor do we know whether reporting pancreatic fat improves outcomes. There is also no pancreas-specific therapy analogous to a guideline-based drug intervention. The genetic-risk findings are interesting, but polygenic risk scores are not ready for day-to-day gastroenterology use in this context. They are valuable for research, risk stratification, and biological inference, but not yet for deciding who should undergo surveillance or intervention. GastroAGI takeaway Intrapancreatic fat deposition is emerging as more than a radiologic curiosity. This new UK Biobank study suggests that excessive IPFD is associated with higher future mortality, particularly vascular mortality, and that genetic predisposition to higher IPFD may show a similar direction of risk. For now, the clinical message is cautious but important: fatty pancreas is not yet a stand-alone diagnosis that changes management, but it may be a marker of broader metabolic and vascular risk. Gastroenterologists should not overdiagnose or alarm patients based on incidental pancreatic fat. But we should also avoid ignoring it completely. The sensible approach is to interpret pancreatic fat in context: metabolic health, diabetes risk, liver fat, cardiovascular risk, alcohol exposure, and pancreatitis history. The fatty pancreas may be entering the same conversation that fatty liver entered years ago. The evidence is earlier, the pathways are less defined, and the clinical tools are not mature. But the signal is becoming harder to dismiss.

GastroAGI TeamRead more

5 min readJul 16, 202623reads

13 Expert Tips on Electrosurgery in Therapeutic Endoscopy: What Every Endoscopist Should Revisit

Introduction Electrosurgery is used every day in therapeutic endoscopy, but it is often learned informally: by observing seniors, memorizing generator settings, or copying unit-specific presets. The AGA’s recent expert update is a useful reminder that electrosurgery is not simply about pressing the blue or yellow pedal. It is about understanding how electrical energy interacts with tissue, how generator settings alter clinical effect, and how technique influences bleeding, perforation, and delayed adverse events. The AGA Clinical Practice Update emphasizes that endoscopists should understand fundamental electrosurgical principles and know the specific electrosurgical unit and settings they are using. This matters because polypectomy, EMR, ESD, APC, sphincterotomy, hemostasis, and third-space endoscopy all depend on controlled energy delivery. Why this update matters Therapeutic endoscopy has become increasingly complex. Endoscopists are now expected to remove large lesions, manage intraprocedural bleeding, ablate residual margins, perform ERCP sphincterotomy, and in advanced units, perform third-space procedures such as POEM and ESD. In each of these settings, electrosurgery can be both therapeutic and harmful. The same current that cuts or coagulates tissue can also cause deep injury, delayed perforation, post-polypectomy syndrome, pancreatitis risk during sphincterotomy, or unintended tissue damage. The AGA update is therefore especially relevant for fellows, early-career consultants, therapeutic endoscopists, and endoscopy units trying to standardize safety. What the AGA expert update highlights The first message is foundational: endoscopists should understand basic electrosurgical principles. This includes tissue resistance, current density, waveform, voltage, duty cycle, and how contact pressure, device position, tissue hydration, and activation time influence the final tissue effect. The second message is practical: the endoscopist should know the electrosurgical unit being used and the related settings. This is important because different generators may use different terminology, modes, and algorithms, even when the clinical procedure appears similar. The update also addresses clinical applications. It notes that either cut- or coagulation-predominant current may be considered in some endoscopic applications, depending on the desired tissue effect and procedural context. In other words, there is no universal “safe setting.” Safety comes from matching energy mode, device, tissue target, and technique. A particularly important caution is soft coagulation. The update notes that soft coagulation can be effective for hemostasis and ablation, but endoscopists should recognize its potential for deep tissue injury and delayed perforation. This is highly relevant after EMR, during visible vessel treatment, and when applying thermal therapy to thin-walled areas such as the right colon or duodenum. The AGA update also discusses adjunctive techniques, including hot forceps avulsion with cutting current for focal non-lifting fibrotic areas, argon plasma coagulation as a non-contact method of delivering electrosurgical energy, ERCP sphincterotomy, and third-space endoscopy. Clinical interpretation The main clinical message is simple: electrosurgery should be treated as a procedural skill, not just an equipment setting. For polypectomy and EMR, the decision is not only cold versus hot. When electrosurgery is used, the endoscopist must understand whether the goal is cutting, coagulation, blended effect, vessel sealing, margin ablation, or rescue hemostasis. Each goal requires different thinking. For APC, the update reinforces that gas flow and power parameters influence the tissue effect. This is clinically important because APC is sometimes perceived as “gentle” because it is non-contact, but it can still create clinically significant thermal injury. For ERCP, the update emphasizes optimizing electrosurgical delivery during sphincterotomy while minimizing immediate or delayed bleeding and pancreatitis risk. This is a useful teaching point: sphincterotomy technique is not only about cannulation and anatomy; energy delivery matters. For third-space endoscopy, the AGA advises that endoscopists performing these procedures should have an in-depth understanding of how to optimize desired tissue effects while minimizing off-target tissue damage. This is particularly relevant for POEM, G-POEM, Z-POEM, and ESD, where dissection planes are narrow and complications can occur quickly. Practical implications for gastroenterologists For general gastroenterologists, this update is a reminder to avoid passive use of generator presets. Before using electrosurgery, the operator should know the indication, device, current type, expected tissue effect, and safety concern. For fellows, this is a useful curriculum topic. Electrosurgery teaching should include more than “use forced coag here” or “use Endocut there.” Trainees should learn why a setting is chosen, what tissue effect is expected, and what complication it is trying to prevent. For endoscopy units, the update supports standardization. Teams should know which generator is used, where settings are documented, how foot pedals are assigned, how return electrodes are placed when required, and how communication occurs during activation. For advanced endoscopists, the update reinforces the need for procedure-specific energy planning. EMR, ESD, ERCP sphincterotomy, APC, hemostasis, and third-space procedures should not be approached with the same mental model. Limitations and caution This AGA Clinical Practice Update is expert guidance, not a randomized trial. AGA notes that Clinical Practice Updates include evidence-based information and, where evidence is limited, best consensus opinion. Therefore, the update should be used as a practical safety and education framework rather than as a rigid protocol. Local equipment, generator model, accessory device, lesion location, tissue thickness, patient factors, and operator experience all influence the safest approach. Endoscopists should follow institutional policies, manufacturer guidance, and local training standards. GastroAGI takeaway Electrosurgery is one of the most important but under-discussed skills in therapeutic endoscopy. The AGA’s 13 expert tips are valuable because they move the conversation beyond memorized settings and toward principled, intentional energy use. For GastroAGI readers, the key takeaway is this: safe electrosurgery requires understanding the generator, the waveform, the tissue target, the accessory device, and the complication you are trying to avoid. In therapeutic endoscopy, energy is not just delivered. It is prescribed.

GastroAGI TeamRead more

6 min readJul 15, 202631reads

Maddrey's Discriminant Function: Deciding Steroid Candidacy in Alcohol-Associated Hepatitis

A 47-year-old man with a 20-year history of heavy drinking is admitted with new jaundice, a total bilirubin of 9.8 mg/dL, and a prothrombin time nine seconds above control. He has no ascites, no encephalopathy, and no signs of active bleeding. The team's first question isn't whether he has alcoholic hepatitis - the history and labs already answer that. The question is whether he needs prednisolone, and Maddrey's discriminant function is the number that answers it. Maddrey's discriminant function was never built to predict who lives and who dies from alcohol-associated hepatitis in general. It was built for one narrower, more useful purpose: identifying which patients had enough to gain from corticosteroid therapy that the risk was worth taking. That distinction gets lost constantly on the wards, where a DF gets calculated, filed as "severe," and used loosely as a catch-all severity marker instead of what it actually is - a steroid-eligibility gate with a single, specific cutoff. Clinicians who treat it as a general prognostic score end up either under-treating patients who clear the threshold or, more often, anchoring on a DF above 32 as sufficient justification for steroids without checking the contraindications that make that decision reversible. The formula itself is simple; using it correctly is not. Calculating Maddrey's Discriminant Function Correctly The formula is DF = 4.6 × (patient prothrombin time − control prothrombin time, in seconds) + total serum bilirubin (mg/dL). Both inputs sound straightforward, and both are where errors creep in. The control prothrombin time is not standardized across labs, and a higher control value systematically lowers the calculated DF - meaning the same patient can cross or miss the treatment threshold depending on which reference range a hospital lab uses. This is not a theoretical concern; retrospective data have shown that the choice of control PT measurably changes how often clinicians offer steroids, without changing the underlying disease. Confirm your institution's current control PT before trusting a borderline result, and recalculate if a patient is transferred between systems using different reference values. Total bilirubin should be the most recent value, not an admission value if several days have passed - AH can progress quickly, and a DF calculated on stale labs can understate current severity. A DF of 32 or higher is the conventional threshold established in the original McCullough and O'Connor criteria and defines severe alcoholic hepatitis. Below 32, corticosteroids are not indicated regardless of how sick the patient otherwise appears; the mortality benefit steroids offer has only been demonstrated in the severe subgroup. Have you checked out the Free Maddreys DF Calculator ? Case in point A 52-year-old woman presents with a DF of 38, no infection on screening cultures, and no gastrointestinal bleeding. Per guideline-directed care, she starts prednisolone 40 mg daily. Her bilirubin is rechecked on day 7, and her Lille score comes back at 0.31 - a responder, meaning bilirubin trended down enough to justify completing the full 28-day course. Contrast that with a 61-year-old man, also with a DF of 38, who is found to have a low-grade urinary tract infection on admission labs. The infection is treated first, and only once source control is confirmed does the team revisit steroid candidacy - DF alone does not override an active, untreated infection as a contraindication. This is the operational reality of Maddrey's DF: it opens the door to a treatment decision, but it does not close the loop on patient selection by itself. Sepsis, active variceal bleeding, and renal failure all modify whether a DF-eligible patient should actually receive steroids that day. Where Maddrey's DF Falls Short - and What to Check Next The STOPAH trial, the largest randomized study of alcoholic hepatitis treatment to date, found that a DF of 32 or higher did not reliably separate patients who benefited from prednisolone from those who didn't. Patients with a Glasgow Alcoholic Hepatitis Score (GAHS) above 8, in the absence of sepsis or GI bleeding, showed a more consistent treatment response than DF-based selection alone predicted. This doesn't mean DF is obsolete - it remains the standard trial-inclusion criterion and the number most guidelines still reference - but it means a DF just above 32 shouldn't be read as a confident, isolated treatment signal. Recalculating GAHS alongside DF gives a second, complementary check before committing a borderline patient to a month of corticosteroids with real risk of infection, hyperglycemia, and GI bleeding. The other half of correct DF use happens after treatment starts, not before. A baseline DF only tells you whether to start steroids. Whether to continue them is a separate question, answered by the Lille score at day 4 or day 7, which incorporates the bilirubin trend on treatment rather than a single admission snapshot. A Lille score above 0.45 identifies non-responders, in whom continuing prednisolone offers no survival benefit and only accumulates infection risk. Treating DF as a one-time gate and Lille as the ongoing checkpoint keeps the decision-making sequence in the right order. A frequently overlooked point The most common error isn't miscalculating DF - it's stopping the workup once DF clears 32. A discriminant function above the threshold confirms eligibility for steroids; it says nothing about whether the patient has an occult infection that makes steroids dangerous, whether renal function is deteriorating in a pattern consistent with hepatorenal syndrome, or whether the diagnosis is even correct. Alcoholic hepatitis is a clinical diagnosis, and a DF calculated on a patient whose jaundice actually stems from choledocholithiasis or drug-induced liver injury is a number computed on the wrong condition entirely. Before treating a DF as decision-grade, confirm the diagnostic gestalt still fits - recent heavy intake, AST: ALT ratio typically above 1.5 to 2, AST rarely above 400, and imaging that doesn't point to an alternative biliary or infiltrative process. Bottom line for clinical practice Use DF ≥ 32 as the threshold for considering corticosteroids in alcoholic hepatitis, not as a general severity or mortality score. Confirm your institution's control prothrombin time before trusting a DF near the cutoff - a different reference range can move a patient across the treatment threshold. Rule out sepsis, active GI bleeding, and uncontrolled renal failure before starting steroids, even in a clearly DF-eligible patient. Recalculate GAHS in borderline cases; STOPAH data suggest it selects steroid responders more precisely than DF alone. Reassess with the Lille score at day 4–7 on treatment - a poor Lille response means stopping steroids regardless of the starting DF. Next time a jaundiced patient with a heavy drinking history lands on your service, run the labs through GastroAGI - it will calculate Maddrey's DF alongside GAHS and Lille, and walk through the guideline-anchored reasoning for steroid candidacy in seconds.

GastroAGI TeamRead more

We are pioneers in clinical intelligence, dedicated to helping gastroenterologists harness the power of artificial intelligence to drive precision, efficiency, and patient growth.

Gastroenterology Insights — Clinical Updates, Guides & Resources — Page 4

CKM Syndrome Guidelines 2026: What They Mean for MASLD Screening in Gastroenterology

CYP2C19 Genotype and PPI Response in Pediatric EoE: Does Dose Matter?

Risk-Based Pathology Reporting After ESD for Early GI Cancer: Why Standardisation Matters

Cryptogenic Steatotic Liver Disease in Lean Patients: A Hidden Risk Signal for Hepatology Practice

GVAX Pancreatic Cancer Vaccine: What the 2026 Trial Data Means for PDAC Care

Machine Learning for Liver Fibrosis Prediction in MASLD: Promising Tool or Premature Clinical Leap?

Coffee and Gut Microbiota: What the 2024 Metagenomic Data Means for GI Practice

Intrapancreatic Fat Deposition and Mortality: Is the Fatty Pancreas Becoming a Risk Marker?

13 Expert Tips on Electrosurgery in Therapeutic Endoscopy: What Every Endoscopist Should Revisit

Maddrey's Discriminant Function: Deciding Steroid Candidacy in Alcohol-Associated Hepatitis

For You

Core Tools

Explore

Gastroenterology Insights — Clinical Updates, Guides & Resources — Page 4

CKM Syndrome Guidelines 2026: What They Mean for MASLD Screening in Gastroenterology

CYP2C19 Genotype and PPI Response in Pediatric EoE: Does Dose Matter?

Risk-Based Pathology Reporting After ESD for Early GI Cancer: Why Standardisation Matters

Cryptogenic Steatotic Liver Disease in Lean Patients: A Hidden Risk Signal for Hepatology Practice

GVAX Pancreatic Cancer Vaccine: What the 2026 Trial Data Means for PDAC Care

Machine Learning for Liver Fibrosis Prediction in MASLD: Promising Tool or Premature Clinical Leap?

Coffee and Gut Microbiota: What the 2024 Metagenomic Data Means for GI Practice

Intrapancreatic Fat Deposition and Mortality: Is the Fatty Pancreas Becoming a Risk Marker?

13 Expert Tips on Electrosurgery in Therapeutic Endoscopy: What Every Endoscopist Should Revisit

Maddrey's Discriminant Function: Deciding Steroid Candidacy in Alcohol-Associated Hepatitis

For You

Core Tools

Explore