Top 5 Papers
#1
Source: New England Journal of Medicine | Authors: Jover R, Bretthauer M, Cubiella J, et al. — EPoS Study Group | Published: September 16, 2026
Score: 16/20 — Base 10 (NEJM) + Phase III RCT (+3) + new standard of care (+3 — directly targets the 3-year surveillance interval that guidelines currently recommend). The interim nature of the analysis is carried as a caveat in the summary rather than as a score deduction, because the prespecified noninferiority margin was formally met at the prespecified interim look.
Across eight European countries, 10,799 patients with high-risk adenomas were randomised to a first surveillance colonoscopy at 5 years (n=5398) or at 3 years (n=5401), with the 3-year group also scoped at 5 years. The 5-year cumulative incidence of colorectal cancer was 0.77% with less-frequent surveillance versus 0.82% with more-frequent surveillance — a difference of −0.05 percentage points, and the upper boundary of the one-sided 99.12% CI was 0.68, inside the prespecified noninferiority margin of 0.7 percentage points. Cancer stage distribution at diagnosis did not appear to differ substantially between groups, and 5 patients died of colorectal cancer in total (3, or 0.06%, in the 5-year group; 2, or 0.04%, in the 3-year group). The critical caveat is that this is a prespecified interim analysis at 5.5 years of a trial whose primary endpoint is 10-year cumulative incidence, with inverse probability weighting used for non-participation at the 5-year scope. Directionally this is the strongest randomised evidence yet for lengthening surveillance intervals after high-risk adenoma removal, but the trial is not finished.
Post angle: 10,799 patients across 8 countries: waiting 5 years instead of 3 for the first surveillance scope after a high-risk adenoma did not increase CRC incidence. Guidelines say 3. This interim says 5 holds. #CRC #Colonoscopy #GIOnc
#2
Source: Nature Cancer | Authors: Marsoni S, Montagut C, Pietrantonio F, Sartore-Bianchi A, et al. | Published: September 14, 2026
Score: 12/20 — Base 8 (Nature Cancer) + Phase II (+2) + biomarker-guided/precision (+1 — ctDNA-directed escalation and de-escalation) + colleague engagement (+1: posted by one GI oncology KOL within 72h). No survival-benefit or new-standard-of-care bonus, because the trial is formally negative on its primary feasibility endpoint.
PEGASUS enrolled 135 patients with resected MSS T4N0 or stage III colon cancer and used post-surgical ctDNA to direct adjuvant therapy, de-escalating ctDNA-negative patients and escalating ctDNA-positive ones. At the post-surgery landmark 26% (n=35) were ctDNA-positive. The prognostic separation was striking: 3-year DFS was 58% in ctDNA-positive versus 83% in ctDNA-negative patients (HR 2.71, 95% CI 1.35–5.45, P=0.0036). The primary endpoint was not met — the observed 2-year relapse-free rate was 88% (90% CI 81–93) against a prespecified threshold of ≥92%, with only 100 of a planned 134 ctDNA-negative cases accrued. Against matched controls from TOSCA, DFS was similar with substantially less neurotoxicity, and relapse tracked with consensus molecular subtype 4, stroma-rich biology. The honest reading is that ctDNA's prognostic power is no longer in question while its ability to safely direct treatment still is.
Post angle: ctDNA tells you who relapses — 3-year DFS 58% vs 83%, HR 2.71. It has not yet told us how to act on it: PEGASUS missed its feasibility endpoint. Prognostic is not the same as actionable, and that gap is where this whole field sits. #CRC #ctDNA #GIOnc
#3
Source: Nature Medicine | Authors: Ai X, Song Z, Wu L, et al. — Lu S (corresponding) | Published: September 15, 2026
Score: 11/20 — Base 9 (Nature Medicine) + Phase 1 (+1, per the established precedent that a first-in-human dose-finding study scores below the +2 given to Phase II) + biomarker-guided/precision (+1 — KRAS G12D-selected enrolment). No survival-benefit bonus: this is a single-arm dose escalation with no comparator arm.
GFH375 is an oral small molecule that binds both the GTP-bound ('ON') and GDP-bound ('OFF') states of KRAS G12D, a broader mechanism than the state-selective inhibitors that preceded it. Across 74 heavily pretreated patients (PDAC 43, NSCLC 16, CRC 10) the maximum tolerated dose was not reached, no dose-limiting toxicities occurred in the 22-patient escalation, and the recommended phase 2 dose was set at 600 mg once daily. In pancreatic cancer the confirmed ORR was 35.1% (13/37) with mDoR 5.6 months, mPFS 5.5 months and mOS 9.9 months; NSCLC responded comparably (confirmed ORR 35.7%, mPFS 5.5 months, mOS 13.4 months). Safety is the number to watch: grade ≥3 treatment-related adverse events occurred in 36.5% (27/74) with one grade 5 event (septic shock), and toxicity was more prevalent in patients with prior checkpoint-inhibitor exposure. Phase 2 expansion is ongoing under NCT06500676. The benchmark this paper does not name — because it was submitted before it existed — is daraxonrasib, approved on August 26, 2026 for metastatic pancreatic adenocarcinoma after at least one prior systemic therapy and, importantly, without any biomarker restriction. In the randomised RASolute 302 trial (n=500) daraxonrasib delivered ORR 30% versus 11% and median OS 13.2 versus 6.7 months (HR 0.40). GFH375's 35.1% single-arm ORR is numerically higher; its 9.9-month median OS is not. A single-arm phase 1 cannot be scored against a randomised phase 3, and the honest position is that G12D-selectivity and oral dosing are the differentiators to test, not an established advantage.
Post angle: An oral G12D-selective inhibitor returned 35.1% ORR in pretreated pancreatic cancer. The comparator that matters is daraxonrasib, approved in August for this exact population with no biomarker restriction: ORR 30% but median OS 13.2 months in a randomised phase 3, against 9.9 months here. Single-arm ORR is the easy number. OS is the hard one. #PDAC #KRAS #PrecisionMedicine #GIOnc
#4
Source: ESMO Open | Authors: Rha SY, Wyrwicz LS, Yañez P, et al. | Published: September 16, 2026
Score: 11/20 — Base 6 (ESMO Open) + Phase III RCT (+3 — this is the trial's own extended follow-up of its primary OS endpoint, not an ancillary correlative analysis) + survival benefit (+2: OS HR 0.78 in the ITT population and 0.64 at CPS≥10).
With median study follow-up now 54.8 months (Q1–Q3 46.8–62.1), KEYNOTE-859's 1579 randomised patients show a first-line benefit that has held. Median OS was 12.9 versus 11.5 months in the ITT population (HR 0.78, 95% CI 0.70–0.86), 13.0 versus 11.4 months at CPS ≥1 (HR 0.74, 0.66–0.84), and 15.8 versus 11.8 months at CPS ≥10 (HR 0.64, 0.53–0.77). The absolute medians are modest, but the tail is where this trial earns its place: 48-month OS was 14% versus 8% in the ITT population and 26% versus 10% at CPS ≥10, with the Kaplan–Meier curves separating through 48 months and a modest plateau emerging in the CPS ≥10 group. Grade 3 or 4 treatment-related adverse events occurred in 58.3% versus 49.3%, and grade 5 treatment-related events were lower with pembrolizumab (1.0% versus 2.0%). The authors' framing is worth adopting: PD-L1 expression enriches for benefit rather than selecting who may receive it.
Post angle: Four and a half years out, KEYNOTE-859 still separates. The median moves about 1.4 months — a number that has always undersold this trial. The tail is the story: 48-month OS 26% vs 10% at CPS≥10. #GastricCancer #Immunotherapy #GIOnc
#5
Source: Cancer Cell | Authors: Shi S, Wei M, Du N, et al. — Xu J, Yu X (corresponding) | Published: September 16, 2026
Score: 10/20 — Base 7 (Cancer Cell — Cell-family flagship oncology journal, matching the base used for this title on 2026-08-11: above the Clinical Cancer Research / Nature Communications tier at 6 and below the Nature / Nature Medicine tier at 9) + Phase 1/2 with an RP2D expansion cohort reporting confirmed response (+2) + biomarker-guided/precision (+1, KRAS G12D-selected). Tied on score with the NEJM esophageal review and the Gut snare meta-analysis; placed in Top 5 ahead of both because Top 5 favours primary trial data over a review, and because it is the second independent KRAS G12D readout of the week.
Forty-eight pretreated patients with metastatic KRAS G12D-mutant PDAC received HRS-4642 together with the anti-PD-L1 antibody adebrelimab, with no chemotherapy in the regimen. In the RP2D expansion cohort (n=37) the confirmed ORR was 43.2% (95% CI 27.1–60.5) and DCR 81.1% (64.8–92.0), with mDoR 6.9 months (4.8–8.6), mPFS 5.7 months (4.0–7.5) and mOS 11.5 months (9.2–16.9). There were no dose-limiting toxicities and no treatment-related deaths; hypercholesterolemia and anemia were the most common treatment-related adverse events at 50% each. Read this alongside GFH375 above: two independent KRAS G12D readouts in pretreated pancreatic cancer landed in the same week, at 43.2% and 35.1% ORR, in a setting where second-line response rates have historically sat in the single digits. Note that this digest covered a different HRS-4642 trial on July 11 — intravenous liposomal HRS-4642 with gemcitabine/nab-paclitaxel in treatment-naive patients (NCT06520488); today's is a distinct chemo-free combination in previously treated disease under NCT06427239. Both of this week's readouts now have to answer to daraxonrasib, approved on August 26, 2026 for pretreated metastatic pancreatic adenocarcinoma with no biomarker restriction: ORR 30% and median OS 13.2 months versus 6.7 with chemotherapy (HR 0.40) in the randomised RASolute 302 trial. Set side by side, the single-arm ORRs here and in GFH375 (43.2% and 35.1%) exceed daraxonrasib's 30%, while the single-arm median OS figures (11.5 and 9.9 months) fall below its 13.2 — which is the clearest possible illustration of why cross-trial ORR comparisons against a randomised phase 3 should not be made. The chemotherapy-free design is the genuinely distinguishing feature, and it is the one worth testing head to head.
Post angle: Two KRAS G12D readouts in pretreated pancreatic cancer in one week: 43.2% and 35.1% ORR, both chemo-sparing. Both now sit against daraxonrasib, approved in August with ORR 30% but a randomised median OS of 13.2 months. Higher ORR, lower single-arm OS — which is exactly why you cannot read across trials. #PDAC #KRAS #GIOnc
Additional Papers of Interest
-
New England Journal of Medicine — Xie, Smyth and Lagergren review the epidemiology, risk factors, prevention and clinical management of a disease causing roughly 500,000 deaths a year, where population 5-year survival still rarely exceeds 20% and adenocarcinoma and squamous-cell incidence trends are diverging by region
-
The Lancet Oncology — 48 cancer types across all 54 African countries: 1.54M incident cases, 923,000 deaths and 33.9M DALYs in 2023, with liver and colorectal among the leading male cancers and risk-attributable deaths up 216.0% since 1990
-
Gut — 1509 polyps in 1423 patients across 50 centres: cold snare cut major adverse events from 5.9% to 1.7% (OR 0.26) but doubled residual or recurrent adenoma from 12.8% to 26.5% (OR 2.61), a genuine safety-versus-durability trade-off rather than a winner
-
Gut — 777 serial stool metagenomes from 315 patients on ICI-based therapy: a 16-species panel generalised across nine public cohorts (n=1204) and stratified OS (HR 0.49) and PFS (HR 0.44)
-
ESMO Open — indirect comparison of five first-line regimens suggesting zolbetuximab is numerically favourable below CPS 10 and checkpoint inhibitors above it; manufacturer-sponsored, and the authors themselves urge that the results be interpreted cautiously
-
FDA / Revolution Medicines — Breakthrough Therapy Designation for daraxonrasib with gemcitabine and nab-paclitaxel in treatment-naive metastatic PDAC; no efficacy figures were disclosed in the announcement, which cites only encouraging preliminary antitumor activity
Back to all digests