Top 5 Papers
#1
Source: Lancet | Authors: Shen L, Ji JF et al. (57 hospitals, China) | Published: June 1, 2026
Score: 14/20 — Lancet base (9) + Phase 3 RCT (+3) + survival/EFS benefit (+2) = 14. PD-L1-selected perioperative IO in resectable gastric/GEJ.
ASTRUM-006 randomized 588 patients with PD-L1 CPS ≥5 resectable gastric or GEJ adenocarcinoma to perioperative serplulimab + SOX vs placebo + SOX (3 cycles neoadjuvant, surgery, 3 cycles adjuvant chemo, then up to 15 cycles serplulimab maintenance). In the CPS ≥10 dual primary population (n=312, median FU 42.7 mo), event-free survival was not reached vs 42.0 mo (HR 0.65, 95% CI 0.47–0.90, p=0.0082). In the ITT CPS≥5 population, EFS was NR vs 35.9 mo (HR 0.73, 95% CI 0.56–0.94, p=0.015). Grade ≥3 TRAEs were paradoxically LOWER with serplulimab (47% vs 59%). OS data immature. This is the second perioperative IO Phase 3 to read out positive in gastric after MATTERHORN, and the first to formally enrich for PD-L1.
Post angle: Perioperative IO is now the standard direction of travel in PD-L1+ resectable gastric/GEJ. ASTRUM-006 + MATTERHORN both positive on EFS. #GastricCancer #GIOnc
#2
Source: Journal of Clinical Oncology | Authors: Bekaii-Saab TS, Vogel A et al. (multicenter, global) | Published: June 1, 2026
Score: 12/20 — JCO base (8) + Phase 3 RCT (+3) + biomarker-guided/precision (+1) = 12.
FIGHT-302 randomized 167 patients with previously untreated unresectable or metastatic FGFR2-rearranged cholangiocarcinoma 1:1 to pemigatinib monotherapy vs gemcitabine + cisplatin. Median PFS 8.3 vs 6.8 mo (HR 0.58, 95% CI 0.39–0.87, nominal p=0.0078). ORR 47% vs 15%, duration of response 14.2 vs 6.3 mo. Median OS was numerically similar (24.4 vs 25.0 mo), driven by crossover — 81% of gem/cis patients received an FGFR inhibitor next-line. The trial was closed early after gem/cis/durvalumab became the new SoC. Even so, this is the first randomized 1L data for a targeted monotherapy in FGFR2-rearranged CCA — and the PFS/ORR/DoR magnitude says targeted monotherapy is at least non-inferior to chemo in this molecular subset.
Post angle: 1L pemigatinib monotherapy in FGFR2-rearranged CCA: PFS HR 0.58, ORR 47%, DoR doubled. OS null because of high crossover. The chemo-free 1L option is now formally on the table. #CCA #FGFR2 #PrecisionMedicine
#3
Source: Nature Medicine | Authors: Peng Z, Shen L et al. (multicenter, China) | Published: June 1, 2026
Score: 12/20 — Nature Medicine base (9) + Phase 2 (+2) + biomarker-guided/precision (+1) = 12.
Single-arm Phase 2 of oral MET inhibitor savolitinib in 110 patients with MET-amplified (gene copy number ≥10 by FISH) gastric or GEJ adenocarcinoma after ≥2 prior lines. In the 65-patient pivotal cohort, ORR was 32.3% (95% CI 21.2–45.1%), meeting the predefined ≥15% threshold. mDoR was a respectable 6.3 mo. Grade ≥3 TRAEs 34.5%; one treatment-related death (0.9%). MET-amplified gastric is a 4–6% subset with historically dismal outcomes and no approved targeted option — these data formally support a confirmatory RCT. This is the first dedicated MET-amp G/GEJ pivotal cohort in Nature Medicine.
Post angle: MET-amplified G/GEJ — a 4–6% molecular orphan — finally has a credible targeted option. Savolitinib ORR 32% in pretreated GCN≥10 G/GEJ. RCT next. #GastricCancer #MET #PrecisionMedicine
#4
Source: Lancet Oncology | Authors: Baudin E, Foulon S et al. (10 academic centres, France) | Published: June 1, 2026
Score: 11/20 — Lancet Oncology base (7) + Phase 2 RCT (+2) + survival/PFS benefit (+2) = 11.
OCLURANDOM is the first randomized head-to-head of PRRT vs an oral TKI in pancreatic neuroendocrine tumors. 84 patients with progressive somatostatin-receptor-positive metastatic pNET after ≥1 prior line were randomized 1:1 to [¹⁷⁷Lu]Lu-DOTATATE (4 cycles q8w) vs sunitinib 37.5 mg/day. At 12 months, PFS was 80.5% (Lu-DOTATATE) vs 41.9% (sunitinib). QoL favored PRRT (Global Health Status 10.3 points higher, p=0.018). Grade 3–4 AEs 44% vs 72%. One treatment-related acute leukemia in the Lu-DOTATATE arm (median FU 72.5 mo) is the long-tail signal to watch. With NETTER-2 already positive in midgut and now OCLURANDOM positive in pancreas, PRRT is now Phase 2/3 evidence-based across both major NET sites.
Post angle: OCLURANDOM — first randomized PRRT vs sunitinib in progressive metastatic pNET. 12-mo PFS 80.5% vs 41.9%, less G3–4 toxicity, better QoL. NET field keeps consolidating around Lu-DOTATATE. #NET #pNET #GIOnc
#5
Source: Science Translational Medicine | Authors: Nehme Z, Baumert TF, Llovet JM, Goyal L, Heikenwälder M et al. (Strasbourg + Mount Sinai + DKFZ + Stanford + Harvard) | Published: June 3, 2026
Score: 7/20 — Sci Transl Med base (6) + biomarker-guided/precision (+1) = 7.
Multi-institution translational study identifies Claudin-1 (CLDN1) as a previously undiscovered oncogenic driver in cholangiocarcinoma. scRNA-seq + spatial transcriptomics of patient CCA tissue showed CLDN1 is upregulated in cancer cells and associates with stemness. Genetic gain-of-function in orthotopic mouse models accelerated tumor growth and worsened survival. Anti-CLDN1 monoclonal antibodies (Alentis Therapeutics) inhibited tumor growth, migration, and extrahepatic metastasis across CDX, PDX, and patient-derived CCA organoids — including CLDN1-medium/low tumors. Mechanistically, surface CLDN1 blockade inhibited Notch1 and TROP2/STAT3 signaling, reducing stemness and EMT. Notably the authors include the Llovet, Goyal, and Heikenwälder labs — a credible bench-to-clinic translational coalition. Anti-CLDN1 mAbs are already in early clinical development for liver fibrosis; CCA is the logical next indication.
Post angle: CCA has its next surface target candidate: Claudin-1. Llovet + Goyal + Heikenwälder labs validate anti-CLDN1 mAbs across PDX + organoids in CCA. Notch1 + TROP2/STAT3 downstream. Bench-to-clinic translation already in motion via Alentis. #CCA #Cholangiocarcinoma #TranslationalOncology
Additional Papers of Interest
-
Gastroenterology — 8 Best Practice Advice statements: keep US + AFP as baseline, multicancer blood panels NOT for HCC surveillance, GALAD/radiomic biomarkers not yet ready for routine use, PAGE-B/REAL-B for HBV without cirrhosis
-
Gastroenterology — 24-hospital ITT RCT: AI vs standard EGD, primary endpoint (post-review detection rate) negative (RR 1.13, p=0.25); AI did reduce blind spots 2.52→1.07 (p<0.001) and benefit was suggested in less-experienced endoscopists and during fatigue periods
-
Journal of Hepatology — Charles Rice (Rockefeller) + Scott Lowe (MSKCC) close a long-standing translational gap with the first immunocompetent infectious model that recapitulates the HCV → cirrhosis → HCC continuum on a C57BL/6 background, opening the door to mechanistic + therapeutic studies of post-SVR HCC risk
-
JCO — the chemo-free 1L conversation in CCA is now formally open: pemigatinib delivers triple the ORR and more than double the DoR vs gem/cis even though OS reads null because 81% of chemo patients crossed over to an FGFR inhibitor next-line
Back to all digests