GI Oncology Daily Digest

CAPP2 established that 600 mg of daily aspirin reduces colorectal and overall cancer incidence in people with Lynch syndrome, but that dose carries real bleeding risk. CaPP3 asked whether lower doses hold up. It randomized 1,879 Lynch syndrome carriers across the UK, Australia, Finland, Israel, and Spain to 100 mg, 300 mg, or 600 mg daily (3:3:4), double-blinded for 2 years then open-label on the same dose, with a non-inferiority margin of 1.5 for new mismatch-repair-deficient (Lynch) cancers. Over a median 66.4 months, 176 participants developed 216 Lynch cancers. The 100 mg dose was non-inferior to 600 mg in the intention-to-treat analysis (HR for time to first Lynch cancer 0.97, 95% CI 0.67-1.42; incidence-rate ratio for cancer burden 0.94, 95% CI 0.65-1.38), and non-inferior for cancer burden in the per-protocol analysis, though per-protocol time-to-first-cancer just missed the margin. The 300 mg dose did not meet non-inferiority on either measure. Crucially, harm scaled with dose: serious bleeding events occurred in 0 patients on 100 mg, 3 on 300 mg, and 11 on 600 mg (p=0.004). The authors are careful to note that formal non-inferiority could not be concluded for either lower dose, but the totality — comparable efficacy signal at 100 mg with markedly less bleeding — is a strong argument that the lowest dose is the pragmatic choice for chemoprevention in this high-risk population.

HBsAg seroclearance — a functional cure of chronic hepatitis B — lowers but does not eliminate hepatocellular carcinoma (HCC) risk, and guidelines have been vague about who still needs lifelong surveillance. This Hong Kong territory-wide study followed all 13,379 adults with chronic hepatitis B who cleared HBsAg between 2000 and 2022 (mean age 59.7, 59.8% male, 14.6% with cirrhosis). Over follow-up, 274 (2.0%) developed HCC at a median of 5.6 years. Risk stratified sharply by cirrhosis and age at seroclearance: patients with cirrhosis had an annual HCC incidence of roughly 0.43-0.62%, well above the ~0.2% threshold at which surveillance is cost-effective. But non-cirrhotic patients who cleared HBsAg before age 50 had an annual incidence of just 0.03% — indistinguishable from the general population — and among 1,317 non-cirrhotic women who cleared before age 50, not a single HCC occurred over 15 years of follow-up. The practical message is a de-escalation one: men who clear before 40 and women who clear before 50, without cirrhosis or other risk factors, have a very low HCC risk and may reasonably be spared routine surveillance, while cirrhosis and older age at clearance still warrant ongoing screening.

Iron-metabolism genes have long been suspected in gastrointestinal cancer, and this study lands on a genuinely counterintuitive finding in pancreatic ductal adenocarcinoma (PDAC). Across three independent cohorts (n=795 total), the common germline HFE H63D polymorphism was substantially enriched in patients with resectable PDAC (37.7%, 37.4%, and 34%) compared with roughly 22-24% in patients with unresectable disease and in the general population. One might expect a variant associated with more operable disease to be favorable — but the opposite was true: H63D independently predicted shorter disease-free survival after resection (p=0.002). To explain the paradox, the authors turned to a genetically engineered mouse model and spatial transcriptomics, which showed that H63D drives a TWIST1-dependent epithelial-mesenchymal transition (EMT) program — a route to early invasion and recurrence even in tumors that present as resectable. The clinical appeal is that H63D is a cheap, binary germline test already familiar from hemochromatosis workups; if validated prospectively, it could help flag resected PDAC patients at higher recurrence risk who might warrant intensified adjuvant strategies or closer surveillance.

First-line atezolizumab plus bevacizumab transformed advanced hepatocellular carcinoma (HCC), but many tumors still fail to respond, and the immunosuppressive machinery behind that resistance is incompletely mapped. This study from the Llovet group identifies a metabolic vulnerability in regulatory T cells (Tregs) infiltrating HCC. In the lactate-rich, hostile tumor microenvironment, these Tregs activate the transcription factor Nrf2 (NFE2L2) as a metabolic adaptation that supports their survival and immunosuppressive function. Genetically deleting Nfe2l2 specifically in Tregs — or pharmacologically inhibiting Nrf2 — suppressed tumor growth in preclinical models, implicating the pathway as causal rather than a bystander. Clinically, the translational hook is the most interesting part: patients whose HCCs were enriched for Tregs with high Nrf2 activation had shorter progression-free survival on atezolizumab/bevacizumab, suggesting Nrf2-Treg activation could serve as both a resistance biomarker and a therapeutic target to combine with existing immunotherapy. This is preclinical/correlative work, not a trial, but it is a clean mechanism-to-target story in a tumor where better patient selection and combination strategies are badly needed.

Cholangiocarcinoma has quietly become the most targetable-alteration-rich cancer in GI oncology, and this Gastroenterology review — from a group including Chiara Braconi and Lewis Roberts — argues that the field's surgical taxonomy (intrahepatic, perihilar, distal) no longer captures the biology that should drive treatment. The authors show that intrahepatic cholangiocarcinoma is itself heterogeneous, spanning small-duct cholangiolar carcinomas (which can arise from dedifferentiating hepatocytes via cellular plasticity) and intermediate/large-duct ductular or mucinous tumors that carry distinct genomic profiles. That biology matters clinically: intrahepatic tumors are relatively enriched for FGFR2 fusions and IDH1/2 mutations — both now druggable — alongside other actionable aberrations. The review ties molecular subgroups to mutational signatures reflecting oxidative stress, aging, and exogenous exposures, and covers the roles of germline predisposition, epigenetics, and DNA-methylation heterogeneity. The take-home for practice is a familiar refrain made sharper: comprehensive molecular profiling belongs upfront in every advanced biliary cancer, because anatomic classification alone will miss the alterations that determine which targeted therapy a patient should receive.