Top 5 Papers
#1
Source: Nature Medicine (PMID 42191879, DOI 10.1038/s41591-026-04376-9) | Authors: Merriam P, Morrow JJ, Demetri G, Bernstein BE, George S (senior) et al. — Dana-Farber + multicenter US consortium (MSK, MD Anderson, City of Hope, U Michigan, WashU, Northwestern, OSU, Miami, U Colorado, Pittsburgh, UMass, NCI) | Published: May 26, 2026
Score: 13/20 — Nature Medicine base 9 + Phase 2 (+2) + epigenetically-guided / biomarker-driven (+1) + colleague engagement on X (+1, @DraMartinezLago) = 13
SDH-deficient GIST accounts for 10–15% of all GIST, lacks KIT/PDGFRA driver mutations, is largely imatinib-refractory, and has had no approved targeted option. This Phase 2 trial (NCT04595747, multicenter US, NCI-coordinated) treated 24 patients with rogaratinib — a pan-FGFR1–4 inhibitor — on the rationale that SDH loss causes genome-wide hypermethylation, which disrupts genomic insulators and drives aberrant expression of FGF3/FGF4 ligands feeding an FGFR1 autocrine loop. Confirmed ORR was 41.7% (10/24 partial responses), median PFS 31.0 months (95% CI 20.2–not reached), and 1-year PFS 77.4% (95% CI 61.7–97.1). Toxicities were manageable (hyperphosphatemia consistent with on-target FGFR engagement, fatigue, diarrhea), and WES of serial biopsies confirmed SDHx alterations as the lone consistent genomic driver. This is the first prospective trial to validate an epigenetics-to-oncogene-activation rationale as a tractable therapeutic strategy.
Post angle: First targeted therapy with real activity in SDH-deficient GIST — ORR 42%, mPFS 31 mo. Mechanism is epigenetic, not somatic mutation. Rethink other SDH-deficient tumors (paraganglioma, RCC). #GIOnc #GIST #PrecisionOncology
#2
Source: AstraZeneca Press Release (April 2, 2026) / ASCO 2026 Annual Meeting Late-Breaking Abstract LBA4000 (presentation Monday June 1, 9:45 AM CT, Hall B1) | Authors: Sangro B, Kudo M, Galle PR, Finn RS et al. — global multicenter steering committee; 760 patients across 171 centers in 22 countries | Published: Topline April 2, 2026 — full data ASCO June 1, 2026
Score: 12/20 — Phase 3 press release base 6 + Phase III RCT (+3) + new SOC potential for intermediate-stage HCC (+3) = 12. Quantitative HR/medians embargoed until ASCO presentation.
EMERALD-3 randomized 760 patients with embolization-eligible unresectable HCC across 22 countries 1:1:1 to STRIDE (single 300 mg priming dose of tremelimumab + durvalumab 1500 mg q4w) + lenvatinib + concurrent TACE, STRIDE + TACE without lenvatinib, or TACE alone. AstraZeneca topline (April 2, 2026): the four-drug arm met its primary endpoint of PFS with a statistically significant and clinically meaningful improvement vs TACE alone, plus a trend toward improved OS at the planned interim analysis (key secondary endpoint). The triplet arm (STRIDE + TACE, no lenvatinib) showed numerical PFS/OS trends but was not formally tested at this readout. Hazard ratios, median PFS, and p-values are embargoed until presentation. EMERALD-3 follows EMERALD-1 (durvalumab ± bevacizumab + TACE, Lancet 2025) and LEAP-012 (pembrolizumab + lenvatinib + TACE) and effectively ends the era of TACE-alone for intermediate-stage HCC. Open question: which IO-plus-TKI-plus-TACE regimen wins on tolerability and which on activity once cross-trial OS matures.
Post angle: Third positive Phase 3 of IO+TKI+TACE in intermediate-stage HCC after EMERALD-1 and LEAP-012. TACE-alone is officially dead. Quadruple regimen — pick your poison. Full LBA4000 data June 1. #GIOnc #HCC #ASCO2026
#3
Source: Nature Communications (PMID 42185305) | Authors: Shao S, Qin J et al. — multicenter, China | Published: May 25, 2026
Score: 11/20 — Nature Communications base 6 + RCT (+3) + reduced surgical morbidity / clinically meaningful safety improvement (+2) = 11
Diverting ileostomy after low anterior resection for rectal cancer is performed in 60–80% of cases to mitigate the consequences of anastomotic leak (AL), but adds a second-stage reversal operation, prolonged morbidity, and substantial QoL burden — and most of those stomas turn out to be retrospectively unnecessary. This multicenter Chinese RCT randomized 872 patients (750 analyzed) with stage I–III rectal cancer undergoing anterior resection to either the RTID machine-learning AL-prediction model or surgeon discretion for the ileostomy decision. RTID cut the overall temporary diverting ileostomy rate from 40.5% to 18.6% (p<0.001) and the unnecessary stoma rate from 41.3% to 17.7% (p<0.001) — a >50% relative reduction. Critically, the AL rate did not increase (2.4% vs 2.7%, p=0.753). Limitations: single-country consortium, primary endpoint was decision quality rather than non-inferiority on AL, and external validation in Western surgical practice is pending. But the effect size is hard to ignore — this is one of the first AI surgical decision-support tools with real RCT-level evidence.
Post angle: AI surgical decision-support finally has RCT data, not just slides. RTID halved unnecessary rectal stomas (41% → 18%) in 750 pts — no increase in leaks. Practical, not flashy. #GIOnc #CRC #RectalCancer #SurgOnc
#4
Source: ASCO 2026 Annual Meeting Oral Abstract Session (Sunday May 31, 9:12–9:24 AM CT) / SWOG Cooperative Group / NCTN — abstract data released May 21, 2026 | Authors: Kasi PM, Lieu CH et al. — SWOG-led NCTN multicenter US/Canada | Published: May 21, 2026 (abstract release) — oral presentation May 31, 2026
Score: 9/20 — ASCO oral abstract base 6 + randomized Phase 2 (+2) + biomarker-guided / precision (+1) = 9. A definitive negative read in a worst-prognosis CRC subset has high clinical value — scoring does not penalize negative results.
MSS BRAFV600E mCRC is one of the worst-prognosis subsets in mCRC, and a single-institution lead-in (n=26, ORR 50%, mPFS 7.2 mo) had raised the hypothesis that BRAFi+EGFRi depletes MAPK-driven immune suppression and primes MSS tumors for anti-PD-1 response despite their low TMB. SWOG S2107 was designed to test that signal in a randomized NCTN trial. Among 85 evaluable patients (57 E+C+N, 28 E+C) with previously treated MSS BRAFV600E mCRC randomized to encorafenib + cetuximab + nivolumab vs encorafenib + cetuximab, median PFS was 5.8 mo (95% CI 4.0–7.8) with the triplet vs 6.3 mo (5.0–11.4) with the doublet (HR 1.10, 95% CI 0.65–1.86, p=0.64) — numerically worse with nivolumab. ORR was 35% vs 32%, mOS 13.5 mo (11.0–18.9) vs 11.6 mo (9.2–15.6) (HR 0.85, 95% CI 0.47–1.52, p=0.29) — non-significant. Grade 3–4 TRAEs occurred in 54% with E+C+N vs 36% with E+C — meaningfully more toxicity with no efficacy gain. Bottom line: clinicians should not add PD-1 inhibition to BRAFi+EGFRi in MSS BRAFV600E mCRC outside of clinical trials. SWOG S2107 joins MEDETREME and a growing list of negative IO-in-MSS-mCRC trials, and refutes the single-arm lead-in that motivated the design.
Post angle: SWOG S2107 NEGATIVE — adding nivo to enco+cetux in MSS BRAFV600E mCRC doesn't help PFS (HR 1.10), OS (HR 0.85), or ORR — and adds tox (54% vs 36% G3-4). Don't do it off-trial. Another IO-in-MSS-mCRC bust. #GIOnc #CRC #BRAF #ASCO2026
#5
Source: Clinical Cancer Research (PMID 42189892) | Authors: Topham JT, Renouf DJ et al. — BC Cancer / Canadian Cancer Trials Group, multicenter | Published: May 26, 2026
Score: 8/20 — Clinical Cancer Research base 6 + retrospective / correlative analysis (+1) + biomarker (+1) = 8
Retrospective correlative analysis of plasma cell-free DNA tumor fraction (CTF) in 166 patients with metastatic PDAC enrolled on the (negative) Phase 3 PA.7 trial of atezolizumab + durvalumab + gemcitabine/nab-paclitaxel (NCT02879318). High CTF (>30%, 21% of patients) was independently associated with shorter OS (HR 1.87, 95% CI 1.27–2.75, p=0.0014), more frequent liver metastases, and upregulation of CDK4 and metallothionein-2A (MT2A). High-CTF tumors were enriched for the glycolytic and basal-like molecular subtypes — the most aggressive PDAC transcriptomic states. CTF added prognostic information independent of CA19-9. Practical implication: ctDNA quantification at diagnosis may help risk-stratify mPDAC for trial enrollment and surface a biological rationale for CDK4-axis combinations in the basal-like subset. CTF is gaining traction across solid tumors as a cheap, single-tube "global tumor burden" biomarker; PA.7 is one of the first Phase 3-anchored validations in PDAC.
Post angle: ctDNA tumor fraction >30% in mPDAC → OS HR 1.87, links to basal-like subtype + CDK4/MT2A upregulation. Cheapest "biological burden" biomarker we have. PA.7 correlative. #GIOnc #PDAC #LiquidBiopsy
Additional Papers of Interest
-
Annals of Surgical Oncology (PMID 42189439) — Single-center retrospective cohort (Hokkaido, 1990–2025): AP factor ≥100,000 (AFP × PIVKA-II) identified HCC patients with 0% 5-year OS after pulmonary metastasectomy vs 51.1% for AP <100,000 (p<0.0001); independent OS predictor on multivariate analysis. Useful selection biomarker for the increasingly rare patient still being considered for surgical metastasectomy after recurrence.
-
Gastroenterology (PMID 42191049) — Editorial/analysis on post-colonoscopy CRC incidence in patients aged ≥75, addressing the long-standing tension between USPSTF screening stopping age (75) and the older surveillance population, where competing comorbidities, procedural risk, and life expectancy increasingly dominate the decision.
-
Journal of Hepatology (PMID 42191458) — EASL multinational position paper categorizing and harmonizing in vivo, in vitro, and ex vivo SLD models (MASLD, MetALD, ALD, monogenic, drug-induced) for use across the MASLD → MASH → cirrhosis → HCC axis. Practical model-selection criteria for translational HCC research now that MASLD is the leading global cause of chronic liver disease.
Back to all digests