GI Oncology Daily Digest

The FDA has granted Fast Track Designation to givastomig — a first-in-class CLDN18.2 × 4-1BB bispecific antibody — in combination with nivolumab and chemotherapy for previously untreated HER2-negative advanced or metastatic gastroesophageal adenocarcinoma whose tumors are both CLDN18.2- and PD-L1-positive. Unlike CLDN18.2 monoclonals such as zolbetuximab, givastomig is engineered to conditionally activate 4-1BB costimulation within the tumor microenvironment, aiming to convert CLDN18.2 targeting into an active T-cell response while limiting systemic immune toxicity. The designation is anchored to a Phase 1b combination cohort that the sponsor reports showed deep, durable responses with favorable tolerability; detailed Phase 1b data are expected at a major medical conference in the second half of 2026, and a registrational Phase 3 trial is slated to begin as early as Q4 2026. Fast Track builds on a prior FDA agreement on an accelerated-approval pathway, and — if it reads out — would add a costimulatory bispecific to a 1L gastric landscape that has so far been defined by zolbetuximab and PD-1 blockade.

This individual-patient-data pooled analysis combined four phase II randomized trials (PanaMa, Valentino, PRODIGE-28/TIME, IMPROVE) to ask a practical de-escalation question: after first-line induction in left-sided, non-MSI-H, RAS/BRAF wild-type mCRC, is anti-EGFR maintenance necessary, or is a treatment-free 'stop-and-go' strategy just as good? Among 378 patients across three strategies — 5-FU/LV + anti-EGFR maintenance (n=166), anti-EGFR monotherapy maintenance (n=109), or stop-and-go (n=103) — overall survival did not differ significantly: stop-and-go 26.3 months vs 29.7 months for 5-FU/LV + anti-EGFR (HR 1.24, p=0.530) and vs 30.1 months for anti-EGFR alone (HR 1.39, p=0.165). Progression-free survival was significantly shorter with stop-and-go, but time to failure of strategy — the more patient-centered endpoint that captures the full induction-pause-reintroduction arc — was comparable across arms. The takeaway: planned treatment de-escalation is a defensible option in optimally selected left-sided MSS mCRC, trading some PFS for time off therapy without a clear OS penalty.

CONTROL NETS is the first randomized trial to test adding peptide receptor radionuclide therapy (Lu-DOTATATE) to CAPTEM chemotherapy in well-differentiated GI neuroendocrine tumours. Seventy-two patients with pancreatic (pNET) or small-bowel (SBNET) NETs were randomized 2:1 to combination PRRT+CAPTEM versus the relevant single modality. The primary 12-month PFS endpoint was not met in pNETs (83.3% combination vs 88.9% CAPTEM; HR 0.41, p=0.08), but with extended follow-up the curves separated meaningfully in favor of the combination — 27-month PFS 61.1% vs 33.3% and objective response 72.2% vs 33.3%. In SBNETs, adding CAPTEM to PRRT added nothing (15-month PFS 90.4% vs 92.3%; 36-month 60.4% vs 61.5%). Two of 63 PRRT-treated patients (3%) developed treatment-related myeloid neoplasms. The signal is histology-specific: combination therapy may deepen and prolong responses in pancreatic NETs, but the benefit must be weighed against a real myeloid risk, and small-bowel NETs derive no added value.

Why are some mismatch-repair-deficient (MSI-H) colorectal cancers immune-rich yet still progress despite checkpoint blockade? Using an engineered MMRd CRC mouse model with human MSI-H validation, this study shows that immune pressure shapes these tumors through two simultaneous escape routes. First, immunoediting: in immunocompetent hosts the most immunogenic clones are selectively eliminated early, lowering tumor incidence but leaving a less visible residuum. Second, local immunosuppression: surviving tumors build a suppressive microenvironment with fewer cytotoxic lymphocytes and upregulated checkpoint proteins. In immunodeficient hosts — where there is no immune pressure — tumors instead retained higher mutational burden and more neoantigen-encoding mutations. Human MSI-H CRCs mirrored the pattern: retained immunogenic mutations alongside elevated PD-1/PD-L1 expression and reduced effector infiltration. The work reframes dMMR immune escape as a dual genetic-plus-microenvironmental process and argues for pairing neoantigen-directed or immune-reactivation strategies with checkpoint blockade rather than relying on checkpoint inhibition alone.

This study makes the diet–cancer link mechanistic. Quantifying total fatty acids by gas chromatography–mass spectrometry in tumor versus normal mucosa across a discovery cohort (n=152) and an independent validation cohort (n=28), the authors found that long-chain polyunsaturated fatty acids — notably the pro-inflammatory precursor arachidonic acid — accumulate in colorectal tumors, particularly right-sided ones, independent of sex, molecular subtype, or early- versus late-onset disease. Using Apc-mutant mouse models with stable-isotope-labeled fatty-acid tracing, they showed these lipids are absorbed directly from the intestinal lumen into the tumors. Strikingly, germ-free Apc mice developed fewer tumors and survived longer, implicating the microbiome in this lipid flux, and inhibiting fatty-acid import or β-oxidation reduced cancer-cell proliferation in 2D and 3D models. The findings position dietary fatty-acid uptake — and its microbiome dependence — as a candidate metabolic and preventive target in CRC, and offer a biological mechanism behind the diet/right-sided-tumor epidemiology.