Top 5 Papers
#1
Source: ESMO Open | Authors: Nakamura Y, Esaki T, Nishina T, … Kuwata T, Yoshino T (National Cancer Center Hospital East and Japanese multicentre group) | Published: October 9, 2026
Score: 9/20 — ESMO Open (base 6) + Phase II (+2) + biomarker-guided / precision, ctDNA blood-TMB selection (+1) = 9. No survival-benefit bonus: single-arm, ORR primary endpoint. No colleague-engagement bonus: the X sweep surfaced no qualifying post for this paper.
This amended Phase II basket study enrolled patients with gastrointestinal cancers whose blood tumour mutational burden (bTMB) was high on Guardant360 CDx (cohorts 1 and 2) or ultrahigh, the top 2% within each GI tumour type (Ultrahigh TMB GI cohort); after that cohort was discontinued early, a bTMB-High cohort enrolled any solid tumour with a FoundationOne Liquid CDx bTMB score of at least 14. All received nivolumab 360 mg every 3 weeks. Seventy patients were treated (28, 4, 19 and 19 per cohort), with ORRs of 14.3%, 0%, 21.1% and 21.1%; the bTMB-High cohort met its prespecified primary endpoint, the lower bound of the 90% CI exceeding 5%. Grade ≥3 treatment-related adverse events occurred in 8.6%. The more useful signal was kinetic: early ctDNA decline at week 3 was associated with longer PFS (HR 0.51, 95% CI 0.29–0.87), while baseline bTMB did not clearly separate responders from non-responders and higher baseline ctDNA concentration was associated with a lower ORR.
Post angle: A blood TMB cut-off picked a population with a 21% response rate to nivolumab, but it did not tell you who would respond. The week-3 ctDNA drop did (PFS HR 0.51). #GIOnc #ctDNA #LiquidBiopsy #Immunotherapy #PrecisionMedicine
#2
Source: Journal of the National Cancer Institute | Authors: Ruze R, Zayier M, Aboduhelili A, … Shao Y, Aji T (First Affiliated Hospital of Xinjiang Medical University) | Published: October 9, 2026
Score: 7/20 — JNCI (base 5; not on the tier table, Other = 5, and no prior edition has scored it) + meta-analysis (+2) = 7. No survival-benefit bonus: the pooled OS comparison is null (HR 1.03) and GRADE certainty is very low for every outcome.
This PROSPERO-registered systematic review and meta-analysis (CRD420261346604) pooled 48 observational studies from 15 countries comparing early-onset pancreatic cancer (EOPC) with average-onset disease, covering 1,659,432 patients (a non-overlapping estimate of 695,936). Patients with EOPC had less diabetes (OR 0.39, 95% CI 0.21–0.72), more stage IV disease at presentation (OR 1.18, 95% CI 1.02–1.36) and higher chemotherapy use (OR 2.04, 95% CI 1.69–2.45; P < .001). Overall survival was comparable (HR 1.03, 95% CI 0.86–1.23; P = .76), with a wide prediction interval of 0.50–2.12. An exploratory dose-response meta-regression suggested a U-shaped association between the age threshold used and survival (P = .008). GRADE certainty was very low for all outcomes, and the authors call for stage-specific survival data to confirm the pattern.
Post angle: Younger patients with pancreatic cancer arrive sicker, get treated harder, and end up with the same survival (HR 1.03). Youth is not protecting them. #GIOnc #PDAC #EarlyOnsetCancer #PancreaticCancer
#3
Source: Cancer | Authors: Zhu D, Fang Z, Jin G, … Wei Q (Zhejiang Cancer Hospital and collaborators) | Published: October 2026 (online October 8, 2026)
Score: 7/20 — Cancer (base 5; not on the tier table, Other = 5) + retrospective (+1) + biomarker-guided / precision, CLDN18.2 (+1) = 7.
Investigators retrospectively scored claudin 18.2 by immunohistochemistry on paired tumour samples taken before and after systemic therapy in 155 patients with advanced gastric or gastro-oesophageal junction adenocarcinoma, using the zolbetuximab-era definition of positivity: at least 75% of viable tumour cells with moderate-to-strong (2+/3+) membranous staining. Positivity rose from 54 patients (34.84%) before treatment to 61 (39.35%) after (p < .001). Among baseline-negative patients, 56.44% showed increased expression and 22 (21.78%) converted to positive; among baseline-positive patients, 64.81% showed decreased expression and 15 (27.78%) converted to negative. Conversion rates did not differ significantly between chemotherapy alone (n = 25) and chemotherapy plus immunotherapy (n = 130; p = .45). Overall, 131 patients (84.52%) showed some change in CLDN18.2 expression, which led the authors to recommend reassessing status before starting CLDN18.2-targeted therapy.
Post angle: A CLDN18.2 result has a shelf life. 22% of negative gastric cancers became positive after treatment, and 28% of positives became negative. Re-test before a later-line CLDN18.2 drug. #GIOnc #GastricCancer #CLDN182 #Biomarkers
#4
Source: Journal of Gastric Cancer | Authors: Minashi K, Inamoto R, Kawakami H, … Kodera Y, Muro K (Japanese multicentre EN-DEAVOR investigators) | Published: October 2026
Score: 7/20 — Journal of Gastric Cancer (base 5; not on the tier table, Other = 5) + real-world (+1) + biomarker-guided / precision, HER2-selected (+1) = 7. No survival-benefit bonus: a descriptive subgroup analysis with no comparator arm.
This subgroup analysis of the Japanese real-world EN-DEAVOR study (UMIN000049032) stratified 312 patients with HER2-positive advanced gastric cancer treated with trastuzumab deruxtecan by CT-defined ascites severity: none (173), low (81), moderate (34) and high (20). Median OS was 12.35, 6.77, 6.97 and 5.13 months, and median real-world PFS 5.32, 3.68, 3.50 and 2.76 months, across those groups. The overall ascites response rate was 23.0%; responders had a median OS of 11.83 months (95% CI 7.82–18.96) and rwPFS of 6.64 months (95% CI 4.17–9.56), against 6.28 months (95% CI 4.90–7.03) and 3.15 months (95% CI 2.17–3.94) in non-responders. Grade ≥3 adverse event rates were comparable across subgroups. Patients with ascites are often excluded from clinical trials; this analysis is observational, and several authors are Daiichi Sankyo employees.
Post angle: Ascites roughly halves survival on T-DXd in HER2+ gastric cancer (12.35 → 5.13 mo), but it is not a reason to withhold the drug: 23% respond, and responders do almost as well as patients with no ascites. #GIOnc #GastricCancer #HER2 #TDXd
#5
Source: JHEP Reports | Authors: Ningarhari M, Leray M, Coilly A, … Dharancy S, Lassailly G (15 French transplant centres) | Published: October 9, 2026
Score: 7/20 — JHEP Reports (base 6) + retrospective / real-world (+1) = 7. No survival-benefit bonus: the mTOR-inhibitor association is observational and the authors themselves call for a trial.
This nationwide retrospective study included every patient diagnosed with HCC recurrence after liver transplantation between 2008 and 2020 at 15 French centres — 388 patients. Median time to recurrence was 19.1 months, and 26.3% of recurrences occurred beyond 3 years. Earlier recurrence was associated with tumour factors (AFP, Milan criteria, AFP kinetics), later recurrence with patient factors (HCV recurrence, alcohol relapse, grade 2 obesity), and time to recurrence shortened after France adopted the AFP model in 2013 (p < 0.001). Surgical or ablative treatment was associated with better survival in patients with single-organ involvement of liver, lung or lymph nodes. Post-recurrence survival did not improve across eras (before 2016, 2016–2018, after 2018; p = 0.09) despite newer TKI options, and mTOR-inhibitor use was associated with longer survival (20.7 vs 13.2 months, p = 0.001), whereas calcineurin inhibitors and mycophenolate were not.
Post angle: After transplant, 1 in 4 HCC recurrences arrives after year 3, beyond the usual surveillance window. Second-generation TKIs did not move post-recurrence survival; local therapy for single-organ disease and an mTOR-inhibitor signal did. #GIOnc #HCC #LiverTransplant
Additional Papers of Interest
-
European Journal of Cancer (published online October 9, 2026, ahead of PubMed indexing) — 1,423 patients with recurrence after PDAC resection from every Dutch pancreatic centre and NYU Langone; after propensity matching (344 vs 344), recurrence-focused treatment was associated with post-recurrence survival of 10 vs 4 months (adjusted HR 0.38, 95% CI 0.30–0.48), in both the worse- and better-prognosis subgroups. Observational, so confounding by who was fit enough to treat remains.
-
European Journal of Cancer (published online October 7, 2026, ahead of PubMed indexing) — 162 resected stage I–III GEP-NECs from the Spanish GETNE registry, 93-month median follow-up; perioperative chemotherapy, modelled as a time-dependent exposure, was associated with longer EFS (HR 0.61, 95% CI 0.40–0.96) and RFS (HR 0.59), but not clearly OS (HR 0.73, 95% CI 0.46–1.15). A Bayesian cure model put the population cure fraction at 22.7% for EFS.
-
European Journal of Cancer (published online October 9, 2026, ahead of PubMed indexing) — 85 patients from the French prospective FFCD-AC cohort; 5FU-based first line was associated with longer PFS (HR 0.53, 95% CI 0.32–0.88) but not OS, and absence of liver metastases was the only independent OS factor (HR 0.58). Histological subtype was not associated with outcome.
-
Alimentary Pharmacology & Therapeutics — 166,863 symptomatic patients with a FIT in Welsh primary care; 1.72% were diagnosed with CRC within a year, 90.2% after FIT ≥10 µg Hb/g. In women under 50 with FIT ≥10, iron-deficiency anaemia raised 1-year CRC risk to 6.45% vs 2.20% without it, and the Welsh baseline risk exceeded English and Scottish cohorts.
-
Investigational New Drugs — 710 patients across zolbetuximab, CLDN18.2 ADCs and CAR-T; pooled ORR 30.3%, with zolbetuximab monotherapy ORR 4.8% vs 63.7% combined with chemotherapy, ADCs 36.7% in pretreated disease and CAR-T 54.2% from a single study. Cross-trial pooling of single-arm data, so read it as a landscape map, and alongside today's paper showing CLDN18.2 status itself shifts with treatment.
Back to all digests