Top 5 Papers
#1
Source: American Journal of Gastroenterology | Authors: Takayama T, Okamoto K, Ishikawa H, … Mutoh M, Sato Y; J-CAP-C study group (25 authors, multicenter Japan) | Published: August 28, 2026
Score: 9/20 — Base 6 (American Journal of Gastroenterology, GI subspecialty flagship — same tier applied to Gastroenterology on 2026-08-20) + large randomized double-blind placebo-controlled multicenter trial (+3) = 9. No clinical-impact bonus: the primary endpoint was not met, so the positive findings are prespecified secondary and exploratory analyses only.
J-CAP-C randomized 577 patients (curcumin 287, placebo 290) after endoscopic resection of colorectal neoplasia to submicron-powdered curcumin (Theracurmin) 180 mg twice daily or identical placebo for two years. The primary endpoint was not met: the adenoma detection rate at two years was essentially unchanged, RR 0.98 (95% CI 0.82–1.16). Prespecified secondary analyses were positive — adenomas ≥6 mm were significantly less frequent, RR 0.56 (95% CI 0.33–0.95), P=0.04, with fewer such adenomas per patient (P=0.03) and a smaller median adenoma size (P=0.003). Exploratory analyses suggested a larger effect in patients with a higher baseline adenoma burden and in men. Registered as jRCTs061180079.
Post angle: The honest read: curcumin did not prevent adenomas, it slowed the ones that recurred. A 577-patient double-blind RCT that misses its primary endpoint and reports the secondary signal plainly is worth more than a positive underpowered study. #CRC #GIOnc #Chemoprevention
#2
Source: Lancet Oncology | Authors: Zhong R, Hong H, Yan M, … Zhong H (31 authors, 39 hospitals in China) | Published: August 25, 2026 (Volume 27, Issue 9, September 2026)
Score: 9/20 — Base 8 (Lancet Oncology, Tier-1 — the base applied on 2026-08-25) + phase 1 dose-escalation and expansion (+1, scaled below the phase 2 bonus) = 9. No clinical-impact bonus: no efficacy figure is disclosed for the oesophageal SCC subset. Included under the Tier-1 safety-net rule as the only Tier-1 paper with GI relevance in the 7-day window.
A phase 1 dose-escalation and expansion trial of the nectin-4-directed antibody–drug conjugate SHR-A2102 enrolled 395 patients across 39 Chinese hospitals between April 2023 and February 2025. The tumour mix was 197 NSCLC, 77 oesophageal squamous-cell carcinoma, 36 TNBC, 32 HR+/HER2− breast, 26 HNSCC and 27 other — making oesophageal SCC the second-largest cohort at 19.5%. The maximum tolerated dose was not reached, with a single dose-limiting toxicity (grade 4 thrombocytopenia) at 10 mg/kg; 6 and 8 mg/kg were carried forward. Grade 3–4 treatment-related adverse events occurred in 212/395 (54%), driven by neutrophil count decrease (30%), white cell decrease (19%) and anaemia (17%). Treatment-related serious adverse events occurred in 25%, and there were two treatment-related deaths (pulmonary embolism, pneumonia). No response rate is reported for the oesophageal SCC subset.
Post angle: The only Tier-1 paper with GI relevance this week, and it is a safety and dose-finding readout, not an efficacy one. The 77-patient oesophageal SCC cohort is the reason to watch the phase 2 — the abstract deliberately does not break out a subset ORR. #EsophagealCancer #ADC #GIOnc
#3
Source: Endoscopy | Authors: Simadibrata DM, … Chandrasekhara V (Mayo Clinic / San Raffaele) | Published: August 29, 2026
Score: 7/20 — Base 5 (Endoscopy, 'Other' tier — consistent with 2026-06-15 and 2026-06-29) + network meta-analysis of randomised trials only (+2) = 7. No clinical-impact bonus: the headline comparison is null, and the actionable findings are secondary safety and patency endpoints.
A frequentist network meta-analysis restricted to randomised controlled trials — deliberately excluding the observational studies that biased earlier syntheses — pooled 10 RCTs and 1659 patients with unresectable extrahepatic malignant biliary obstruction, comparing fully covered (FCSEMS), partially covered (PCSEMS) and uncovered (UCSEMS) self-expandable metal stents. There were no significant differences between stent types in time to recurrent biliary obstruction, incidence of recurrent obstruction, or overall survival. The differences were in failure mode: uncovered stents had the lowest rates of migration, tumour overgrowth and sludge occlusion, while fully covered stents were most effective at preventing tumour ingrowth. Fully covered stents — but not partially covered — were associated with significantly more acute cholecystitis than uncovered stents in patients with an intact gallbladder.
Post angle: A null primary comparison is the useful part here. If patency and survival are a wash, the choice is driven by which failure mode you are trying to avoid — and by whether the gallbladder is still in. #Endoscopy #Cholangiocarcinoma #GIOnc
#4
Source: Gastroenterology | Authors: Mo F, Good AL, McDevitt JC, … Engle DD, Posey AD Jr, Stanger BZ | Published: August 27, 2026 (Articles in Press)
Score: 7/20 — Base 6 (Gastroenterology, GI subspecialty flagship — the base applied on 2026-08-20) + biomarker-guided / precision target identification (+1) = 7. No study-type bonus: this is preclinical work with no human data, so the Phase II/III and survival-benefit bonuses do not apply.
CAR T cell therapy has underperformed in pancreatic ductal adenocarcinoma largely because of a shortage of tumour-specific cell-surface targets. This study screened 14 CAR constructs recognising the CA19-9 glycan — an antigen broadly expressed on PDAC and other gastrointestinal tumours — and identified AbLIFT15.28z as the optimal configuration. Cytotoxicity was confined to CA19-9-expressing cells both in vitro and in vivo, with activity demonstrated in human PDAC cell lines, patient-derived organoids and immunocompetent murine PDAC models, against both primary and metastatic tumours. Efficacy extended to CA19-9-expressing tumours of the colon, stomach, oesophagus and biliary tract. The work is entirely preclinical: there are no patients, and no response, progression-free or overall survival data.
Post angle: Solid-tumour CAR T keeps failing on target selection rather than on T cell engineering. A glycan expressed across GI adenocarcinomas — not just PDAC — is a genuinely different swing at the problem. Preclinical only, so read it as a target paper. #PDAC #CART #GIOnc
#5
Source: ESMO Open | Authors: Tan CK, Vogel A, Wong YF, … André F, Muthukkumaran T, Ducreux M (30 authors, 10 Asian oncology societies + ESMO) | Published: Volume 11, Issue 8, August 2026 (indexed in this window)
Score: 7/20 — Base 7 (practice guideline update, per the guideline tier in the scoring table) = 7. No study-type or clinical-impact bonus: a consensus adaptation carries no trial endpoints. Noted for freshness — the guideline itself carries an earlier publication date and entered this window by PubMed entry date.
A Pan-Asian adaptation of the May 2025 global ESMO hepatocellular carcinoma guidelines, produced by consensus across the oncology societies of China, Indonesia, India, Japan, Korea, Malaysia, the Philippines, Singapore, Taiwan and Thailand, coordinated by ESMO and the Malaysian Oncological Society. The document covers epidemiology, diagnosis, staging, management, treatment and follow-up, with applicability of each recommendation and the availability of specific tests and treatments described country by country. Methodologically, the voting on recommendations was held independently of drug access and reimbursement status, with regional approval and reimbursement disparities documented separately rather than folded into the recommendations themselves.
Post angle: The design choice worth noting: recommendations were voted on the evidence, and access and reimbursement gaps were documented separately instead of quietly shaping the guidance. That makes the disparities visible rather than encoding them. #HCC #LiverCancer #GIOnc
Additional Papers of Interest
-
Hepatology — 22,537 genotyped veterans over 352,612 person-years: major adverse liver outcome risk rose with Z-allele burden (Pi*MZ aHR 1.25, Pi*SZ aHR 1.51, Pi*ZZ aHR 1.80 vs Pi*MM), but only Pi*ZZ raised HCC risk — the heterozygous genotypes drove decompensation, transplant and liver-related death without an HCC signal
-
ESMO Gastrointestinal Oncology — 8 phase III trials and 4904 patients, only 14.7% female; chemo-IO improved OS overall (HR 0.68, 95% CI 0.62–0.73) and in both sexes, with a numerically smaller effect in women (HR 0.79, 0.64–0.98) than in men (HR 0.68, 0.63–0.74). The authors conclude the magnitude is less in females, but the interaction was not statistically significant (subgroup p=0.19, meta-regression p=0.20) and sensitivity analysis showed the female estimate rested mainly on two studies
-
ESMO Gastrointestinal Oncology — MSKCC-led 3+3 pilot in 19 patients across three nivolumab schedules relative to deb-TACE: ORR 21% (95% CI 6–44), median PFS 5.9 months, median OS 23 months, and the authors conclude plainly that it did not enhance antitumor activity beyond historical benchmarks for deb-TACE alone
Back to all digests