Top 5 Papers
#1
Source: Annals of Oncology | Authors: Al-Batran S-E, Bencivenga M, Crosby T, et al.; Ducreux M, on behalf of the ESMO Guidelines Committee | Published: October 15, 2026 (published online, ahead of PubMed indexing)
Score: 10/20 — Annals of Oncology (base 7, matching the 2026-08-20 ESMO Clinical Practice Guideline edition) + guideline codifying a new standard of care (+3: durvalumab–FLOT is now the preferred perioperative regimen [I, A], and zanidatamab–tislelizumab–chemotherapy is recommended first line for HER2-positive disease [I, A]) = 10. No study-type bonus: a guideline is not a trial. No colleague-engagement bonus: no GI oncology KOL posted it within the 72h window (X sweep 6/6). Tier-1 safety-net paper.
ESMO's full Clinical Practice Guideline for gastric and oesophageal cancer was published online today. It splits management into oesophageal squamous-cell carcinoma and gastro-oesophageal adenocarcinoma (GOAC). For resectable GOAC (≥cT2 or cN+), perioperative FLOT remains the chemotherapy standard [I, A], but adding durvalumab (D-FLOT, then durvalumab maintenance) is now the preferred option where available [I, A; ESMO-MCBS A]. That rests on MATTERHORN: 2-year EFS 67.4% vs 58.5% (HR 0.71), pCR 19.2% vs 7.2%, and final OS HR 0.78 (P = 0.021). Citing ESOPEC (median OS 66 vs 37 months, HR 0.70), FLOT is preferred over CROSS-type chemoradiotherapy for oesophageal and junctional adenocarcinoma. In first-line HER2-positive advanced GOAC, zanidatamab–tislelizumab–chemotherapy is recommended where available [I, A; ESMO-MCBS 4], noted as FDA-approved but not EMA-approved, on the HERIZON-GEA-01 OS HR of 0.72. HER2, MSI/dMMR, PD-L1 (CPS and TAP) and CLDN18.2 testing are all recommended [I, A]. FOLFIRI is endorsed after progression during or shortly after perioperative (D)-FLOT [I, B]. For oesophageal SCC, CROSS chemoradiotherapy and surgery remain the standard, with adjuvant nivolumab for residual disease.
Post angle: ESMO's new gastric and oesophageal guideline puts D-FLOT first for resectable adenocarcinoma, FLOT ahead of CROSS at the junction, and zanidatamab–tislelizumab–chemo in first-line HER2+ disease. The European guideline is recommending a regimen the EMA has not yet approved, which shows how fast the HER2 field is moving. #GIOnc #GastricCancer #EsophagealCancer #ESMO
#2
Source: Journal of Clinical Oncology | Authors: Shroff RT, Kennedy EB, Alarcon Rozas AE, et al.; Greten TF | Published: October 14, 2026
Score: 9/20 — Journal of Clinical Oncology (base 8) + biomarker-guided / precision (+1: the guideline is organised around tissue NGS and biomarker-selected second-line therapy) = 9, matching the 2026-08-18 ASCO gastroesophageal living guideline. No study-type bonus: a living-guideline synthesis is not a primary trial. No colleague-engagement bonus: no GI oncology KOL posted it within the 72h window (X sweep 6/6). Tier-1 safety-net paper.
ASCO has issued a living guideline for advanced biliary tract cancer, covering liquid biopsy and systemic therapy. The evidence base includes 6 tissue–plasma concordance studies, 9 RCTs for first-line therapy, and 1 Phase III plus 4 Phase II RCTs for second-line chemotherapy. Broad tissue DNA and RNA NGS is recommended at diagnosis at any stage. Plasma liquid biopsy is generally recommended only when tissue is unavailable. In first line, an immune checkpoint inhibitor plus chemotherapy is recommended for pMMR/MSS disease, and pembrolizumab monotherapy is an option for dMMR/MSI-H tumours. Options are listed for patients who cannot receive an ICI. After first-line progression with no actionable biomarker, FOLFOX with active symptom control, or active symptom control alone, is recommended. Biomarker-specific second-line options cover dMMR/MSI-H, BRAF V600E, HER2 overexpression and selected mutations and fusions.
Post angle: ASCO's biliary living guideline is built around testing: tissue DNA + RNA NGS at diagnosis, with plasma only as a fallback. ICI + chemo is the first-line default for pMMR/MSS, and two days after the KEYNOTE-966 three-year update the evidence still supports that. #GIOnc #BTC #Cholangiocarcinoma #PrecisionOncology
#3
Source: Med (Cell Press) | Authors: Yao X, Sun H, Fang C, et al.; Han Y | Published: October 13, 2026
Score: 7/20 — Med (base 5; Other = 5, matching the three 2026-06 editions that scored it) + observational diagnostic study (+1) + biomarker-guided / precision (+1) = 7. No survival or standard-of-care bonus: this is a hospital-based diagnostic-accuracy study, not a screening trial. No colleague-engagement bonus: no GI oncology KOL posted it within the 72h window (X sweep 6/6).
Red blood cells have no nucleus but can retain DNA fragments, and this study tests whether those fragments (rbcDNA) can detect gastric cancer. The authors ran shallow whole-genome sequencing on rbcDNA from 1–2 mL of blood in 865 people: 434 treatment-naive gastric cancer patients and 431 controls with benign gastric conditions. They identified 1,000 reproducible read-depth features and trained a classifier with the cutoff fixed at 90% specificity. AUC was 84% in nested cross-validation and 95% in the held-out test cohort (n=109). In three independent hospital-based multicentre validation cohorts (n=321), overall sensitivity was 83–90%, and stage I sensitivity was 82–89%. The classifier also separated early-stage cancer from precancerous lesions and other tumours, and detected about 80% of cancers that were negative for conventional tumour markers. The main caveat is the setting: these were hospital cohorts, not an asymptomatic screening population, so positive predictive value at real-world prevalence is untested.
Post angle: A gastric cancer blood test that doesn't use plasma: DNA left behind in red blood cells. It reached 82–89% stage I sensitivity at 90% specificity across 3 validation cohorts, from 1–2 mL of blood. The next step is a screening-population study where prevalence is low. #GIOnc #GastricCancer #EarlyDetection #LiquidBiopsy
#4
Source: MedComm | Authors: Cui Y, Bao X, Yu G, et al.; Song T | Published: October 13, 2026
Score: 7/20 — MedComm (base 5; Other = 5, matching the 2026-06-11 edition) + Phase II (+2) = 7. No clinical-impact bonus: single-arm, with no comparator for DFS. No colleague-engagement bonus: no GI oncology KOL posted it within the 72h window (X sweep 6/6).
This single-arm Phase II enrolled 31 patients with resectable hepatocellular carcinoma and high-risk features such as large tumours, multinodular disease or vascular invasion. Patients received four cycles of neoadjuvant camrelizumab (200 mg every 2 weeks) plus apatinib (250 mg daily), then resection, then the same regimen for up to a year. The primary endpoint was major pathological response (≤50% viable tumour). Of 31 patients, 26 went to surgery, and the MPR rate among them was 46.2%. Median DFS after resection was 24.1 months (95% CI 14.4–not estimable), and median PFS across all 31 patients was 18.3 months. Grade ≥3 adverse events in the neoadjuvant phase occurred in 12.9%. Caveats: 5 of 31 patients never reached resection, the trial is small and single-arm, and it reports no comparison against upfront surgery.
Post angle: Perioperative camrelizumab + apatinib in high-risk resectable HCC: MPR 46.2%, median DFS 24.1 months, and only 12.9% grade ≥3 toxicity before surgery. But 5 of 31 patients never reached the operating room, and that matters more than the MPR. Neoadjuvant HCC needs a randomised trial against upfront resection. #GIOnc #HCC #LiverCancer #Immunotherapy
#5
Source: MedComm | Authors: Peng J, Li Y, Huang L, et al.; Cai S | Published: October 13, 2026
Score: 7/20 — MedComm (base 5; Other = 5, matching the 2026-06-11 edition) + Phase II (+2) = 7. No clinical-impact bonus: single-arm with high grade ≥3 toxicity and no comparator. No colleague-engagement bonus: no GI oncology KOL posted it within the 72h window (X sweep 6/6).
This prospective, multicentre, single-arm Phase II (NCT05004831) treated 50 patients with metastatic colorectal cancer that had progressed on at least two standard therapies. Patients received fruquintinib plus trifluridine–tipiracil (TAS-102). At a median follow-up of 25.1 months, median PFS (the primary endpoint) was 6.07 months (95% CI 4.07–7.87) and median OS was 14.10 months (95% CI 9.84–23.28). ORR was 10.0% and DCR 66.0%. Every patient had a treatment-related adverse event, and 68.0% had grade ≥3 events, though no serious adverse events or deaths were reported. This digest covered a separate Phase II of the same doublet, with or without SBRT, on May 1, 2026 (Cancer Medicine, median PFS 8.6 months). The two trials now agree that the combination is active. They also agree that it needs a randomised test against the current third-line standard, TAS-102 plus bevacizumab.
Post angle: A second Phase II of fruquintinib + TAS-102 in refractory mCRC: mPFS 6.07 months, mOS 14.1 months, ORR 10%. But 68% had grade ≥3 toxicity. A VEGFR TKI swapped in for bevacizumab only matters if it beats TAS-102 + bev head to head. #GIOnc #CRC #ColorectalCancer
Additional Papers of Interest
-
Nature Genetics — single-cell RNA sequencing of 62 primary PDACs (443,451 cells), 11 with matched spatial transcriptomics and 6 with paired pre- and post-chemotherapy samples. Multiple genetically distinct clones coexisted within tumours. Focal 8q24 (MYC) amplification was recurrently associated with classical-to-basal transitions and with depletion of immune-response programmes, and MYC knockdown restored classical programmes in murine and primary human models.
-
Cancer Letters — 126 patients from four prospective trials, ypN0 in 56%, 24-month EFS 76.4% and OS 90.5%. In junctional cancer, distal perigastric stations (No. 4, 5, 6, 12) rarely carried metastases, supporting a prospective trial of proximal gastrectomy. In gastric cancer, nodal involvement was broader, supporting standard D2 lymphadenectomy.
-
JNCCN — the full NCCN guideline text. Tumour NGS is recommended for every patient who is a candidate for anticancer therapy, with RNA sequencing added because it detects fusions better. Neoadjuvant PAXG is listed as a useful-in-certain-circumstances option on the Italian Phase 3 result against mFOLFIRINOX (median EFS 16.0 vs 10.2 months, HR 0.63). After neoadjuvant therapy, resection decisions are now framed by biological response (RECIST, CA 19-9, clinical status) rather than vascular contact alone. The panel declines to make a consensus recommendation on postoperative ctDNA.
-
JNCCN — the HCC panel's Insights article on the most recent updates to the NCCN HCC guideline, which covers screening, evaluation, treatment and surveillance.
-
Gastric Cancer — a modified Delphi consensus with 37 international voting members. It does not recommend population-based endoscopic screening in low- to moderate-prevalence regions. Where screening is done, it prefers EGD over upper-GI series and emphasises endoscopy quality indicators and risk stratification with serum biomarkers.
-
The Oncologist — 137 patients (99 IRE, 38 standard of care). With overlap weighting, IRE was associated with lower mortality, HR 0.28, but the time-dependent model attenuated this to HR 0.59. IRE patients enrolled a median of 8 vs 4 months from diagnosis, so selection and immortal-time bias remain likely.
Back to all digests