📬 GI Cancer Intelligence Digest — Week 2026-W40

FDA approved lirafugratinib for FGFR2-fusion cholangiocarcinoma, ASCOT's 5-year adjuvant S-1 data held up in resected biliary tract cancer, adjuvant aspirin failed after colorectal liver metastasectomy in ASAC, and EPoS II supported a 5-year surveillance colonoscopy interval.

⚡ TL;DR

📄 Publications

Aspirin after curative-intent treatment of colorectal cancer liver metastases (ASAC): a multicentre, phase 3, randomised, double-blind, placebo-controlled trial. — 466 patients randomised at 14 centres in Norway, Sweden, and Denmark; 428 initiated treatment (217 aspirin, 211 placebo)
Five-Year Follow-Up of the Phase III ASCOT Trial of Adjuvant S-1 After Resection for Biliary Tract Cancer. — 440 patients with resected biliary tract cancer randomised to observation (n=222) or S-1 (n=218) in Japan
First-Line Retlirafusp Alfa Plus Chemotherapy for Human Epidermal Growth Factor Receptor 2-Negative Gastric or Gastroesophageal Junction Adenocarcinoma: A Randomized, Double-Blind, Phase III Study (RELIGHT). — 731 previously untreated patients with unresectable, locally advanced or metastatic gastric or GEJ adenocarcinoma randomly assigned 1:1 to retlirafusp alfa or placebo plus CAPOX

🐦 Twitter / 🦋 Bluesky

EPoS II in NEJM: a 5-year first surveillance colonoscopy after high-risk adenoma removal is non-inferior to 3 years — A randomized trial of 3 versus 5 years showed non-inferiority for the longer interval, with higher frequency detecting more lesions without meaningfully reducing colorectal cancer incidence over 5 years
LAVA-CRLM randomised phase 2: SBRT matches microwave ablation for oligometastatic colorectal liver metastases with better tolerability — LAVA-CRLM was a prospective, randomised, phase II trial of microwave ablation versus stereotactic body radiotherapy for oligometastatic colorectal liver metastases
Cell paper: synthetic IL-21-driven CD4 T cells turn transient KRAS inhibitor responses into durable remissions in pancreatic cancer models — A new paper from Stephanie Dougan with David Baker reports that CD4 T cells convert transient responses to KRAS inhibition into durable remissions in pancreatic cancer
FDA approves lirafugratinib for previously treated FGFR2 fusion/rearrangement cholangiocarcinoma on a 46% response rate — On September 23, 2026, the FDA approved lirafugratinib for adults with previously treated, unresectable, locally advanced or metastatic cholangiocarcinoma harboring FGFR2 fusions or rearrangements
Reduced oxaliplatin exposure meta-analysis: no survival loss, far less neuropathy across GI cancers — Across 24 studies with N=30,395 there was no significant survival difference with reduced oxaliplatin exposure, while grade >=3 neuropathy OR was 0.34 and grade >=3 AE OR was 0.67

Full details below ↓

📄 What's Being Published

Manuscripts:

Aspirin after curative-intent treatment of colorectal cancer liver metastases (ASAC): a multicentre, phase 3, randomised, double-blind, placebo-controlled trial. — Yaqub Sheraz et al., Lancet Gastroenterol Hepatol

Between December 2017 and January 2022, 466 patients were randomised across 14 centres in Norway, Sweden, and Denmark, with 428 initiating treatment (217 aspirin, 211 placebo). Aspirin did not improve disease-free survival (HR 1.08, 95.44% CI 0.85–1.38; p=0.58), and 36-month overall survival was 76% with aspirin versus 85% with placebo (HR for death 1.64, 95% CI 1.01–2.65; p=0.045). Serious adverse events occurred in 17 (8%) of 217 patients in the aspirin group compared with 4 (2%) of 211 patients in the placebo group.

Clinical bottom line: Adjuvant aspirin should not be given after curative-intent treatment of colorectal cancer liver metastases: it did not improve disease-free survival and was associated with worse overall survival and more serious adverse events.

Read full article →

Five-Year Follow-Up of the Phase III ASCOT Trial of Adjuvant S-1 After Resection for Biliary Tract Cancer. — Ikeda Masafumi et al., J Clin Oncol

In this multicentre, open-label, randomised trial conducted in Japan, 440 patients with resected biliary tract cancer were assigned to observation (n=222) or S-1 (n=218). At a median follow-up of 61.5 months, 5-year OS was 64% in the S-1 arm versus 52% in the observation arm (HR 0.72, 95% CI 0.55–0.95; one-sided P=.01), with median OS of 8.2 years with S-1 versus 5.9 years with observation. Median relapse-free survival was 6.1 versus 3.5 years (S-1 vs observation); the stratified HR was 0.76 (95% CI 0.59–0.98) and the unstratified HR was 0.80 (95% CI 0.62–1.03).

Clinical bottom line: Five-year follow-up confirms adjuvant S-1 as a standard of care after biliary tract cancer resection in Japanese practice, though the all-Japanese population and pre-checkpoint-inhibitor era limit how far this can be generalised.

Read full article →

🔗 Also discussed on social media →

Chemotherapy and Surgical Outcomes for Ovarian Metastases From Colorectal Cancer: A Narrative Review. — Weil James B E et al., JAMA Oncol

Twenty-four retrospective studies comprising 2705 patients with ovarian metastases from colorectal cancer were evaluated. Ovarian metastases demonstrated lower objective chemotherapy response rates than extraovarian metastases, had substantially lower mean shrinkage, and frequently exhibited discordant growth during treatment. The most favourable outcomes were observed in patients who had complete surgical resection of ovarian metastases and microscopically negative margins, and the literature supports the ovary functioning as a relative sanctuary site such that systemic therapy alone should not be regarded as adequate.

Clinical bottom line: For women with colorectal cancer metastatic to the ovary, plan for metastasectomy with negative margins as part of comprehensive disease control rather than relying on chemotherapy response.

Read full article →

Evorpacept plus trastuzumab, ramucirumab and paclitaxel in HER2-positive gastric cancer: a randomized phase 2 trial. — Shitara Kohei et al., Nat Med

127 patients with pretreated HER2-overexpressing advanced gastric/GEJ cancer were randomised to evorpacept plus trastuzumab, ramucirumab and paclitaxel (evo+TRP; n=63) or TRP alone (n=64). Investigator-assessed ORRs were 40.3% (evo+TRP) versus 26.6% (TRP) in the ITT population and 54.8% versus 23.1% in the fresh biopsy HER2-positive subgroup; however, compared with the historical benchmark of 30%, ORRs in the ITT population (40.3%; P=0.0949, one-sided) and fresh biopsy HER2-positive subgroup (54.8%; P=0.030, one-sided) did not meet the prespecified statistical criterion (one-sided α=0.025). Hematologic toxicities were more common with evo+TRP, but overall safety was similar.

Clinical bottom line: Adding CD47 blockade to trastuzumab/ramucirumab/paclitaxel raised response rates — most in patients whose HER2 positivity was reconfirmed on fresh biopsy — but one of two primary objectives was missed, so this remains a phase 3 question rather than practice.

Read full article →

First-Line Retlirafusp Alfa Plus Chemotherapy for Human Epidermal Growth Factor Receptor 2-Negative Gastric or Gastroesophageal Junction Adenocarcinoma: A Randomized, Double-Blind, Phase III Study (RELIGHT). — Peng Zhi et al., J Clin Oncol

In RELIGHT, 731 previously untreated patients with unresectable, locally advanced or metastatic gastric or GEJ adenocarcinoma were randomly assigned 1:1 to retlirafusp alfa or placebo plus CAPOX. Retlirafusp alfa plus CAPOX significantly prolonged OS in patients with PD-L1 CPS ≥5 (median 16.8 vs 10.4 months; HR 0.53, 95% CI 0.40–0.68; one-sided P<.0001) and in the ITT set (median 15.8 vs 11.2 months; HR 0.66, 95% CI 0.53–0.81; one-sided P<.0001), with PFS benefit in CPS ≥5 (median 7.6 vs 5.5 months; HR 0.52) and the ITT set (median 7.0 vs 5.5 months; HR 0.57). Grade ≥3 treatment-related adverse events were comparable: 62.6% (228 of 364) with retlirafusp alfa plus CAPOX versus 59.0% (216 of 366) with placebo plus CAPOX.

Clinical bottom line: The PD-L1/TGF-beta bispecific plus CAPOX produced a large first-line survival gain in HER2-negative gastric/GEJ adenocarcinoma, but with a chemotherapy-only control arm it cannot yet displace chemoimmunotherapy.

Read full article →

Trastuzumab plus nivolumab with gemcitabine and cisplatin as first-line therapy for HER2-positive unresectable biliary tract cancer: a phase 1b/2 trial. — Lee Choong-Kun et al., Nat Med

The multicentre, open-label, phase 1b/2 HERBOT study evaluated first-line trastuzumab, nivolumab, gemcitabine and cisplatin in patients with HER2-positive advanced biliary tract cancer. The primary endpoint was met: objective response rate was 55% (95% CI 38.5–70.7; one complete response and 21 partial responses among 40 participants) and disease control rate was 95% (95% CI 83.5–99.4), with a median duration of response of 12.6 months (95% CI 5.7–not reached). With a median follow-up of 17.0 months, median progression-free survival was 10.6 months (95% CI 7.8–17.4) and median overall survival was not reached; common grade ≥3 treatment-related adverse events included neutropenia (57.5%), anemia (30.0%), and thrombocytopenia (22.5%).

Clinical bottom line: Upfront HER2 targeting added to chemoimmunotherapy gives a 55% response rate in HER2-positive biliary tract cancer, which supports the ongoing phase 3 programs but does not yet change first-line practice.

Read full article →

Personalized Therapy in Advanced Cholangiocarcinoma: Contemporary Treatment Approaches and Future Directions. — Saj Fen et al., JCO Oncol Pract

The addition of durvalumab and pembrolizumab to gemcitabine-cisplatin in the TOPAZ-1 and KEYNOTE-966 trials, respectively, improved overall survival, establishing chemoimmunotherapy as the new first-line standard. FGFR2 fusions, IDH1 mutations, and HER2 amplification are among the most relevant actionable alterations, with targeted agents now available in the second-line setting, and comprehensive molecular profiling has become indispensable for second-line treatment selection and front-line clinical trial eligibility. The immunosuppressive tumor microenvironment and the absence of reliable predictive biomarkers for immunotherapy response remain key challenges.

Clinical bottom line: This review reinforces that every patient with advanced cholangiocarcinoma needs comprehensive molecular profiling at diagnosis, because FGFR2, IDH1, and HER2 status determines second-line options and trial access.

Read full article →

🗂 Every paper considered this week, including the ones that didn’t make it →

🗣️ What People Were Discussing This Week

How these are chosen: filtered for clinical relevance, grouped by the new development discussed, then ordered by engagement (bookmarks and reposts).

EPoS II in NEJM: a 5-year first surveillance colonoscopy after high-risk adenoma removal is non-inferior to 3 years

The EPoS package published in NEJM includes the EPoS II trial, an editorial on surveillance after high-risk adenomas, and correspondence on serrated polyps from EPoS III. In 10,799 patients, colorectal cancer at 5 years was 0.77% with the 5-year strategy versus 0.82% with 3 years — a risk difference of −0.05% — meeting noninferiority criteria; higher-frequency surveillance detected more lesions without meaningfully reducing colorectal cancer incidence over 5 years. The central debate is whether societies should move first post-polypectomy surveillance to 5 years now, or wait for the planned 10-year colorectal cancer outcomes, given that these remain interim results of a planned 10-year trial.

Discussed by: @EladSharonMD, @OncoAlert, @jgong15, @drsarahsam

→ View representative tweet

Sources cited: NEJM · When More Is Not Better — Colonoscopy Surveillance after High-Risk Adenomas (New England Journal of Medicine, 2026) · Colorectal Cancer Risk among Patients with Serrated Polyps (New England Journal of Medicine, 2026)

LAVA-CRLM randomised phase 2: SBRT matches microwave ablation for oligometastatic colorectal liver metastases with better tolerability

LAVA-CRLM was a prospective, randomised, phase II trial comparing microwave ablation versus stereotactic body radiotherapy for oligometastatic colorectal liver metastases. The head-to-head study enrolled patients with colorectal cancer liver metastases under 4 cm, using SBRT at 45 Gy in 3 fractions, and showed similar disease control but better tolerability for SBRT; clinicians summarised the result as SABR being better tolerated than thermal ablation without significant differences in efficacy. The open question for tumor boards is whether a sub-4 cm colorectal liver metastasis should now default to SBRT on tolerability grounds alone, or whether lesion location continues to determine the choice.

Discussed by: @jryckman3, @Sushilberiwal, @NiuSanford, @JoeChangMD

→ View representative tweet

Sources cited: The Lancet

Cell paper: synthetic IL-21-driven CD4 T cells turn transient KRAS inhibitor responses into durable remissions in pancreatic cancer models

A paper from Stephanie Dougan with David Baker reports that CD4 T cells convert transient responses to KRAS inhibition into durable remissions in pancreatic cancer, with the mechanism being an IL-21 mimic that harnesses CD4 T cells to sustain KRAS inhibitor responses in mice. A separate preprint using digenic CRISPR screens identifies attenuating FGFR signaling as the most effective pairing with a KRAS inhibitor in pancreatic cancer. With KRAS inhibitor resistance now the central problem in PDAC, the debate is whether the next combination trials should prioritize immune partners such as IL-21 mimetics or vertical pathway partners such as FGFR inhibition.

Discussed by: @Aiims1742, @Aiims1742, @EladSharonMD, @break_cancer

→ View representative tweet

Sources cited: Cell Press · bioRxiv · Selective Inhibition of KRASG13C Reveals an Increased Dependence on Wild-type RAS Isoform… (Cancer Discovery, 2026)

FDA approves lirafugratinib for previously treated FGFR2 fusion/rearrangement cholangiocarcinoma on a 46% response rate

On September 23, 2026, the FDA approved lirafugratinib for adults with previously treated, unresectable, locally advanced or metastatic cholangiocarcinoma harboring FGFR2 fusions or rearrangements. Efficacy was evaluated in the REFOCUS trial (NCT04526106), a multicenter, open-label, single-arm trial in 116 patients who had received prior chemotherapy or chemoimmunotherapy and were naïve to FGFR inhibitor treatment; ORR was 46% (95% CI 36–55) and median duration of response was 11.8 months (95% CI 7.5–13.0) by independent review. The recommended dose is 70 mg orally once daily until disease progression or unacceptable toxicity, and key toxicities include ocular toxicity and hyperphosphatemia; with a third FGFR inhibitor now approved alongside pemigatinib and futibatinib, the debate is whether selectivity justifies preferring lirafugratinib in FGFR2-fusion cholangiocarcinoma.

Discussed by: @VivekSubbiah, @ArndtVogel, @ilyassahinMD, @OncBrothers, @hongjaechon, @ozdogan_md

→ View representative tweet

Sources cited: RLY-4008, the First Highly Selective FGFR2 Inhibitor with Activity across FGFR2 Alteratio… (Cancer Discovery, 2023) · FDA press release · OncoDaily

Reduced oxaliplatin exposure meta-analysis: no survival loss, far less neuropathy across GI cancers

A meta-analysis across GI cancers and settings spanning 24 studies (N=30,395) found no statistically significant survival difference with reduced oxaliplatin exposure, while grade ≥3 neuropathy OR was 0.34 and grade ≥3 adverse event OR was 0.67. The authors' position is not to continue oxaliplatin beyond 3–4 months, with these findings supporting limiting exposure to improve the therapeutic index of oxaliplatin-based chemotherapy across GI settings. The debate is whether any remaining GI setting — adjuvant colon, gastric, or metastatic — justifies pushing oxaliplatin beyond that threshold.

Discussed by: @DraMartinezLago, @UGrewalMD

→ View representative tweet

Sources cited: Elsevier · ScienceDirect

Why 5-FU underperforms in dMMR colorectal cancer: failure of damage sensing, not true resistance

Without mismatch repair, 5-FU-induced DNA damage is not recognized, producing less apoptosis — a failure of damage sensing rather than true resistance. A separate study reports that arachidonic acid-driven lipophagy sustains 5-fluorouracil-resistant colorectal cancer cells, suggesting a strategy to restore chemotherapy sensitivity. Multimodal testing also improves MSI/MMR assessment in metastatic colorectal cancer, resolving most discordant results and helping to identify patients most likely to benefit from immunotherapy; the question is whether the damage-sensing explanation for 5-FU failure in dMMR tumors justifies skipping fluoropyrimidine entirely in stage II dMMR colon cancer.

Discussed by: @ChandrakanthMv, @GenesNDiseases

→ View representative tweet

Sources cited: Elsevier · MDPI

ASCOT 5-year follow-up: adjuvant S-1 sustains a survival benefit in resected biliary tract cancer, but clinicians say it is not enough

In 440 randomised patients (S-1 218, observation 222) with a median follow-up of 61.5 months, 5-year OS was 64% with S-1 versus 52% with observation; median OS was 8.2 versus 5.9 years (HR 0.72, 95% CI 0.55–0.95; one-sided p=.01). RFS favours S-1 less decisively: the stratified HR was 0.76 (95% CI 0.59–0.98) while the unstratified HR was 0.80 (95% CI 0.62–1.03), crossing 1.00, with median RFS 6.1 versus 3.5 years and benefit generally consistent across subgroups. A stated limitation is that all participants were Japanese — so generalizability and S-1 tolerability may differ across populations — and the study predates widespread checkpoint inhibitor use, leaving open whether non-Japanese centres should adopt adjuvant S-1 or hold for capecitabine and immunotherapy-containing adjuvant trials.

Discussed by: @DrAllanPereira, @OncoAlert, @DraMartinezLago

→ View representative tweet

Sources cited: PubMed · Five-Year Follow-Up of the Phase III ASCOT Trial of Adjuvant S-1 After Resection for Bili… (J Clin Oncol)

KEYNOTE-859 4.5-year follow-up confirms durable first-line pembrolizumab benefit in HER2-negative gastric/GEJ adenocarcinoma

The 4.5-year follow-up of the randomised phase III KEYNOTE-859 study of pembrolizumab plus chemotherapy as first-line therapy in advanced HER2-negative gastric or gastroesophageal junction adenocarcinoma reported median OS of 12.9 versus 11.5 months (HR 0.78) in 1579 patients, with HR 0.74 in CPS ≥1 and median OS of 15.8 versus 11.8 months (HR 0.64) in CPS ≥10. Benefit was durable at 4.5 years, with the largest magnitude seen in CPS ≥10. The debate for clinicians is whether a 4.5-year HR of 0.78 overall but 0.64 at CPS ≥10 still makes CPS ≥1 a defensible threshold for adding pembrolizumab in first-line gastric cancer.

Discussed by: @ronanhsieh, @DraMartinezLago

→ View representative tweet

Sources cited: ESMO Open

ASTRO26 GI003: protons show no advantage over photons in hepatocellular carcinoma, so photon radiotherapy stays standard

GI003, comparing proton versus photon radiotherapy in HCC, showed no difference in OS, PFS, or grade 3+ adverse events. With a negative randomised trial and cost and access differences, photon radiotherapy should remain standard of care; protons represent one option in the broader radiotherapy toolbox — alongside MRgRT and adaptive approaches — to optimise the therapeutic ratio, with each case individualised by tumor and liver function. Radiation oncologists in the same discussion also noted that SABR is better tolerated than thermal ablation in oligometastatic colorectal liver metastases without significant differences in efficacy.

Discussed by: @NiuSanford, @JoeChangMD

→ View representative tweet

Sources cited: The Lancet

NHS-Galleri published in NEJM misses its primary endpoint as the FDA convenes its first MCED panel, with colorectal signal hardest to interpret for US practice

The FDA's first-ever multicancer early detection advisory panel met on GRAIL's Galleri as NEJM published the full NHS-Galleri trial overnight. The 142,250-person NHS-Galleri trial missed its primary endpoint of reducing stage III/IV diagnoses; in the same week, PATHFINDER 2 reported a positive predictive value of 60.3% and specificity of 99.64% across 32,007 people aged 50 and over. Clinicians argued the issues are more market than medical — the test does not yet facilitate early detection — and that the biggest signal was in colorectal cancer but the UK, where the study was done, lacks widespread screening colonoscopy as in the US, leaving the meaning for American health unclear.

Discussed by: @drgandara, @ilyassahinMD, @sciqst

→ View representative tweet

Sources cited: The Lancet · Performance and safety of a multi-cancer early detection test: the PATHFINDER 2 study (Nature Medicine, 2026) · Liquid biopsy for early detection of pancreatic ductal adenocarcinoma (Nature Medicine, 2026)

NEST trial staging debate: is CT accurate enough to select MSS colon cancers for neoadjuvant immunotherapy?

Investigators argue that confusion about colon cancer staging accuracy stems from lumping MSI-High and MSS tumors together: for MSI-High, lymph node enlargement can reflect immune response and overestimate nodal staging, whereas CT underestimates in MSS. In NEST3, real-time central radiology review is performed within 24 hours the same day with no delays; with botensilimab/balstilimab, Treg depletion plus multi-arm immune activation within weeks has been observed in cold MSS colorectal tumors. Commenters note that FOxTROT paved the way and that neoadjuvant immunotherapy could become routine for high-risk stage 2/3 colorectal cancer, though the risk of overtreating early-stage disease through CT-based staging remains a barrier to routine adoption.

Discussed by: @pashtoonkasi, @pashtoonkasi, @FernandoOnco

→ View representative tweet

Sources cited: Neoadjuvant Botensilimab/Balstilimab for localized mismatch repair proficient and deficie… (Clinical Cancer Research, 2026)

🔥 Top Ranked Tweets

Vivek Subbiah, MD @VivekSubbiah

🚨Another big win for #precisionmedicine
🚨👉🏼Today the FDA approved lirafugratinib for FGFR2‑rearranged cholangiocarcinoma.
⭐️I still remember the first patient in the universe we enrolled on the trial during the peak of the pandemic. Three days in - I called the amazing Ben- the then CMO of @Relay_Tx
“Your drug is working, working too good. We need to work on the right dosing & get it right for patients. I think this baby will graduate someday.” That day is today 🙌🏼 Thankful to all patients who volunteered & the amazing teams that got this to finish line 🙏🏽 !!! @OncoAlert @oncodaily @OncBrothers @ElizSMcKenna @matthewherper @curecc @Aiims1742 @MiteshBorad @AliSchram
👉🏼Link to FDA approval https://t.co/8WTjhiTD98
👉🏼Link to @CD_AACR @AACR bench to bedside publication https://t.co/HeME5d8mRZ
[Quoting @VivekSubbiah: 📣Super thrilled to share our paper on RLY-4008, the First Highly Selective FGFR2 Inhibitor with Activity across FGFR2 Alterations & Resistance Mutations, published in @CD_AACR !
👉🏼Huge thanks to team at Cancer Discovery & @AACR for making the front page & the entire paper so visually appealing 👏🏼#PrecisionMedicine 🧬🎯#cancerresearch @OncoAlert @ElizSMcKenna @Relay_Tx
@TaliLev123
👉🏼Free Link: https://t.co/DmMttMep3o]

🔖 4 ❤️ 58 🔁 24 👁 7,005

📬 GI Cancer Intelligence Digest — Week 2026-W40

⚡ TL;DR

📄 What's Being Published

🗣️ What People Were Discussing This Week

EPoS II in NEJM: a 5-year first surveillance colonoscopy after high-risk adenoma removal is non-inferior to 3 years

LAVA-CRLM randomised phase 2: SBRT matches microwave ablation for oligometastatic colorectal liver metastases with better tolerability

Cell paper: synthetic IL-21-driven CD4 T cells turn transient KRAS inhibitor responses into durable remissions in pancreatic cancer models

FDA approves lirafugratinib for previously treated FGFR2 fusion/rearrangement cholangiocarcinoma on a 46% response rate

Reduced oxaliplatin exposure meta-analysis: no survival loss, far less neuropathy across GI cancers

Why 5-FU underperforms in dMMR colorectal cancer: failure of damage sensing, not true resistance

ASCOT 5-year follow-up: adjuvant S-1 sustains a survival benefit in resected biliary tract cancer, but clinicians say it is not enough

KEYNOTE-859 4.5-year follow-up confirms durable first-line pembrolizumab benefit in HER2-negative gastric/GEJ adenocarcinoma

ASTRO26 GI003: protons show no advantage over photons in hepatocellular carcinoma, so photon radiotherapy stays standard

NHS-Galleri published in NEJM misses its primary endpoint as the FDA convenes its first MCED panel, with colorectal signal hardest to interpret for US practice

NEST trial staging debate: is CT accurate enough to select MSS colon cancers for neoadjuvant immunotherapy?

🔥 Top Ranked Tweets

🧪 Newly Recruiting Trials