GI Oncology Daily Digest

A landmark Lancet Oncology trio of paired papers from Ward and colleagues uses the Global Cancer Workforce microsimulation model to project incidence, stage, survival, mortality, and workforce-attributable mortality for 17 cancers across 200 countries from 1990 to 2050 — of which six are GI sites (oesophagus, stomach, colon, rectum, anus, liver, pancreas). Diagnosed cancer incidence is projected to rise from 13.58M (95% UI 12.17–14.51) cases in 2025 to 19.32M (17.65–21.33) in 2050. Globally, ~31.5% (95% UI 28.1–36.7) of incident cancers die undiagnosed, ranging from 0.9% in western Europe to 67.4% in western Africa. Pooled 5-year net survival is projected to remain flat globally (47.6%→47.7%), with persistent gaps: 34.4% in Africa vs 70.4% in Oceania by 2050. Workforce-scale-up modeling shows scaling surgeons single-handedly cuts global cancer mortality 3.64% (95% UI 2.68–4.66), while expanding diagnostic and imaging cadres cuts global cancer mortality 7.61% (5.23–9.88). Comprehensive scale-up of all cadres reduces cancer mortality >50% in 55 countries. The accompanying EORTC-ESTRO OligoCare endpoints consensus already framed how to measure MDT trials in oligometastatic disease. Bottom line: ~1 in 3 of the world’s cancer patients still die undiagnosed in 2026 — fixing diagnostics and imaging is a higher-yield lever than any single drug class.

The May 2026 cover article in Cancer Discovery, from the Dave Tuveson lab at CSHL, dissects bi-directional crosstalk between sympathetic nerves and cancer-associated myofibroblasts in the precancer ecosystem of pancreatic ductal adenocarcinoma. The team shows myofibroblasts actively induce sympathetic neuroplasticity — nerve sprouting and remodeling — around pre-invasive PDAC lesions, and that this stroma-driven neuroplasticity is a driver of tumor progression rather than a passive bystander. The accompanying editorial frames neuroplasticity as a previously underappreciated stromal-neural axis with therapeutic implications: targeting sympathetic outgrowth or the myofibroblast–nerve interface could intercept PDAC at its earliest stages, complementing the wave of KRAS G12X and pan-RAS inhibitors now in clinic. Allan’s read: precancer interception is the next frontier in PDAC — first KRAS, now stroma+nerves — we may finally be moving upstream of the diagnosis.

Multicenter single-arm Phase II from the Hokkaido GI Cancer Study Group of ramucirumab (8 mg/kg D1, D15) + docetaxel (60 mg/m2 D1) every 28 days as 2L therapy in 35 patients with advanced gastric cancer refractory or intolerant to first-line. Primary endpoint ORR was 25.7% (90% CI 14.1–40.6), DCR 74.3%, mPFS 3.1 mo (95% CI 2.1–4.2), mOS 11.5 mo (95% CI 9.2–13.9). Notable safety profile: peripheral sensory neuropathy 65.7% but ZERO grade ≥3 — a meaningful improvement vs paclitaxel-based ramucirumab combos that drive the standard 2L. Febrile neutropenia drove a protocol amendment to mandatory primary G-CSF prophylaxis, after which no FN was observed. Bottom line: ramucirumab + docetaxel may be a reasonable taxane partner for patients who can’t tolerate paclitaxel-induced neuropathy, with G-CSF mandatory.

Trial-design publication for COLOSOTO (ENGIC 01 / PRODIGE 107 / FFCD 2306), a multicenter, multinational (France, Italy, Germany, Spain), single-arm Phase II evaluating sotorasib + panitumumab + 5-FU in 1L unresectable KRAS G12C-mutated mCRC patients deemed unfit for doublet/triplet chemotherapy. 37 frail/elderly patients will be enrolled; 2-week cycles until progression or intolerance. Primary endpoint 8-month PFS (target 70%, null 50%); secondaries include mPFS, ORR, OS, DOR, safety, QoL, and geriatric assessment. Rationale: sotorasib + panitumumab is an established 3L+ standard for KRAS G12C mCRC after CodeBreaK 300; this trial uniquely tests it upfront in patients who can’t safely receive doublet/triplet chemo — a frail/older population usually excluded from Phase III trials. Important real-world geriatric oncology question.

Single-cell RNA-seq of paired primary CRC, adjacent tissue, and liver mets from CRC liver metastasis (CRLM) patients vs non-metastatic CRC identifies ENO2+ cancer cells as enriched in CRLM. Mechanistically, the ENO2 protein binds MIF, blocking CHIP-mediated ubiquitination/degradation of MIF and thereby activating MIF signaling that drives M2 macrophage polarization in the metastatic niche. Spatial transcriptomics confirms colocalization of ENO2+ cancer cells and M2 macrophages in human CRLM. ENO2 knockout suppressed tumor growth and liver metastasis in mice; pyrithioxin (an off-the-shelf compound that disrupts the ENO2–MIF interaction) significantly reduced metastatic burden. Bottom line: tractable druggable interaction nominated for CRC liver mets, and an off-the-shelf chemical probe to test the hypothesis quickly.