Top 5 Papers
#1
Source: Gut | Authors: Boer LS, Han M, Alkhalaf A, et al. — Barrett Expert Center Registry + Netherlands Cancer Registry | Published: September 21, 2026
Score: 8/20 — Base 7 (Gut — BMJ's flagship GI journal, the peer of Gastroenterology) + retrospective nationwide registry cohort (+1) = 8. No clinical-impact bonus applied: the study quantifies a care-delivery gap rather than demonstrating a survival benefit or establishing a new standard of care. Placed first over the equally-scored SBAGA because it is the only candidate this week combining an in-window entry date, a true national denominator, and an immediately actionable message. Cited to its PubMed record rather than its DOI because the BMJ DOI route is Cloudflare-walled.
Among 8,914 Dutch patients treated with curative intent for Barrett-related dysplasia or oesophageal adenocarcinoma between 2008 and 2018, 90% (95% CI 89–91; 8,025/8,914) presented de novo — with no prior endoscopy of any kind. Stage at diagnosis diverged sharply: 70% of the de novo group presented at stage II or beyond, against 11% of those already under surveillance (p<0.001). Even among patients whose disease was still amenable to endoscopic therapy, 65% (95% CI 62–67) had never been scoped, and 97% (95% CI 97–98) of those who ultimately required non-endoscopic treatment were de novo presentations. Critically, surveillance itself performed well where it was applied: only 11% (96/889) of patients under surveillance had advanced neoplasia missed. The bottleneck is not the surveillance interval — it is that the overwhelming majority of patients never enter a surveillance pathway at all.
Post angle: The Barrett debate has been about intervals and modalities for the 10% already in the system. This nationwide cohort says the 90% who were never scoped are where the stage shift actually happens. #BarrettsEsophagus #EsophagealCancer #GIOnc
#2
Source: Nature Communications | Authors: Cheng M, Zhang P, Hu Q, et al. — West China Hospital, Sichuan University | Published: September 2026 (PubMed entry September 14, 2026)
Score: 8/20 — Base 6 (Nature Communications) + phase Ib/II trial meeting its pre-specified primary endpoint (+2) = 8. No clinical-impact bonus: the design is single-arm and non-randomised, which the authors name as their principal limitation, so no comparator-based survival-benefit claim is available. PubMed entry date (09-14) falls outside this run's adaptive window of 09-19 to 09-22; included because the paper is genuinely unpublished in this digest and was missed by both the 09-17 and 09-20 runs.
This phase Ib/II trial (NCT05640609) added the anti-PD-1 sintilimab and the bevacizumab biosimilar IBI305 to CapeOX as first-line therapy in HER2-negative advanced gastric or gastroesophageal junction adenocarcinoma. The phase Ib portion recorded no dose-limiting toxicities (N=9). In phase II, the primary endpoint of objective response rate reached 87.0% (95% CI 75.1–94.6) in 54 efficacy-evaluable patients, meeting the pre-specified target, with a disease control rate of 98.1% (95% CI 90.1–100.0). At a median follow-up of 20.3 months, median PFS was 12.2 months (95% CI 9.2–18.1) and median duration of response 12.5 months (95% CI 9.4–NE); median OS was not reached. Clinical benefit was seen irrespective of PD-L1 CPS, Claudin-18.2 status, or peritoneal metastases. Grade ≥3 treatment-related adverse events occurred in 66% of patients, most commonly neutropenia (14%), fatigue (13%) and thrombocytopenia (11%).
Post angle: An 87% ORR in HER2-negative gastric cancer is the kind of number that stops you — and the kind that most needs a control arm. Anti-VEGF plus anti-PD-1 has a real mechanistic rationale here; it now needs a randomised test. #GastricCancer #GIOnc
#3
Source: Scandinavian Journal of Gastroenterology | Authors: Cao J, Jiang T, Wu Q, et al. | Published: September 21, 2026
Score: 7/20 — Base 5 (Scandinavian Journal of Gastroenterology — no row in the scoring table and no precedent anywhere in the corpus, so it defaults to Other) + systematic review and meta-analysis (+2) = 7. No clinical-impact bonus was applied despite a pooled OS hazard ratio of 0.71, precisely because the benefit is not reproduced in the pooled randomised subgroup; awarding a survival-benefit bonus here would score the headline rather than the evidence. Cited to PubMed because the Taylor & Francis route could not be confirmed to render.
This systematic review and meta-analysis (PROSPERO CRD420261332944) pooled nine studies — four randomised trials and five cohort studies — examining hyperthermic intraperitoneal chemotherapy in gastric cancer with peritoneal metastasis. Overall survival favoured the addition of HIPEC, with a pooled HR of 0.71 (95% CI 0.56–0.91; p=0.008), strengthening when combined with cytoreductive surgery (HR 0.64, 95% CI 0.50–0.80; p<0.001) and in Western studies (HR 0.67, 95% CI 0.53–0.85; p<0.001). The cohort-study subgroup showed the largest effect (HR 0.63, 95% CI 0.55–0.72; p<0.001). But no significant overall survival benefit was found in the randomised-trial subgroup, nor in Eastern studies, nor without cytoreductive surgery, nor in any drug-regimen subgroup. The pooled estimate is therefore carried by observational data, and the authors say so themselves.
Post angle: A pooled HR of 0.71 that evaporates when you restrict to randomised trials is not a positive result — it is a description of confounding by selection. This lines up with PERISCOPE II, which this digest covered in July and which was flatly negative. #GastricCancer #HIPEC #GIOnc
#4
Source: Journal of Gastroenterology and Hepatology | Authors: Chon HK, Jang DK, Paik WH, et al. | Published: September 20, 2026
Score: 7/20 — Base 5 (Journal of Gastroenterology and Hepatology — no row in the scoring table, but the corpus already answered this on 2026-04-27, where it was scored at base 5 as Other; adopting that precedent rather than setting a second, conflicting value) + multicentre randomised crossover trial (+2, the randomised-comparative tier rather than the registrational phase III tier) = 7. No clinical-impact bonus: this is a null result by design and by outcome.
The heparin wet suction technique has been proposed to improve tissue quality during EUS-guided fine-needle biopsy by preventing clot formation inside the needle. In this multicentre, single-blinded randomised crossover trial, 174 patients with solid pancreatic masses ≥1 cm were randomised to heparin-first or conventional-first, with 170 receiving one pass by each technique. Diagnostic yield was 82.9% (95% CI 76.6–87.9) with heparin wet suction versus 83.5% (95% CI 77.2–88.4) with conventional suction — a paired difference of −0.6 percentage points (95% CI −6.7 to 5.5; exact McNemar p=1.000). Technical success was 100% in both arms and adequacy for immunohistochemistry was comparable (80.6% vs 78.8%), with a single self-limiting bleeding event in the conventional arm. Combined yield after sequential sampling with both techniques reached 91.2% (95% CI 86.0–94.6), a secondary observation the authors explicitly flag as requiring further validation.
Post angle: A well-run randomised crossover that says the trick does not help. Modern 22G FNB needles appear good enough that suction technique has stopped being the limiting variable — a useful negative for anyone still optimising it. #PDAC #EUS #GIOnc
Additional Papers of Interest
-
Cell Reports Medicine — 41,627 single-cell lipidomic profiles from 92 samples; label-free metabolite-based mass cytometry plus a graph neural network distinguishes free gastric-cancer cells from benign peritoneal cells and proposes a HER2 subtype-specific metabolic classification. Discovery-stage: the report contains no sensitivity, specificity or AUC figures.
-
ESMO Gastrointestinal Oncology — Trial design paper, not a results readout. ZELDA (jRCTs071250038, West Japan Oncology Group / Kyushu Study Group) will randomise a planned 210 patients 1:1 to second-line nab-paclitaxel + ramucirumab with or without continued zolbetuximab after progression on first-line zolbetuximab, with overall survival as the primary endpoint; a six-patient safety lead-in is complete. No efficacy data have been reported.
-
NEJM Correspondence — Dekker, Jover, Ijspeert and Bretthauer, writing for the EPoS Study Group, report that patients with serrated polyps had a low 5-year colorectal cancer risk of 1.0% in this prospective trial. A companion signal to the EPoS II surveillance-interval trial covered in the September 17 edition.
-
NEJM Clinical Decisions — Teaching feature, not a study. A case vignette of a 46-year-old man is paired with two opposing essays, one arguing for screening colonoscopy and the other for a stool-based test. Useful as a framing of where the first-line screening argument currently sits.
-
JCO — Commentary, not primary data. Aujla and Ahmed examine what it would take for circulating tumour DNA surveillance after curative colorectal resection to convert molecular residual disease detection into an actual curative opportunity rather than earlier bad news.
-
Annals of Oncology — Editorial accompanying the DeepComp multimodal deep-learning model this digest covered on July 17. Rabasco Meneghetti frames the central problem: up to a third of gastrectomy patients suffer postsurgical complications, Clavien-Dindo grade II or higher delays adjuvant therapy and worsens survival, and prediction alone changes nothing unless it is wired to an action.
Back to all digests