ASCO 2026: Gastrointestinal Abstracts to Watch

By the time attendees arrive in Chicago, many will have spent weeks combing through abstracts, building tentative session plans, and anticipating which presentations have the potential to change practice. Full abstracts for the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting will be released May 21. At present, the field is in a familiar holding pattern. For community oncologists, five gastrointestinal (GI) abstracts are particularly notable. 

HERIZON-GEA-01: The PD-L1 Question in HER2+ Gastroesophageal Cancer

One standout presentation is the PD-L1 subgroup analysis from HERIZON-GEA-01 (Abstract 4010). The trial evaluated first-line zanidatamab plus chemotherapy with or without tislelizumab against trastuzumab plus chemotherapy in HER2-positive locally advanced or metastatic gastroesophageal adenocarcinoma. The primary readout, presented at the 2026 ASCO Gastrointestinal Cancers Symposium in January, was a meaningful step forward for the zanidatamab backbone in first-line HER2-positive disease. Since KEYNOTE-811, the central question has shifted from whether immunotherapy provides benefit to whether benefit is limited to PD-L1–positive tumors. KEYNOTE-811 sets the precedent for how this presentation should be read. After the trial’s initial positive readout, the FDA narrowed the pembrolizumab indication to tumors with PD-L1 combined positive score (CPS) of 1 or greater on the basis that the benefit in the CPS-low population was not durable enough to justify the toxicity and expense of adding a checkpoint inhibitor. That decision reshaped how to test, sequence, and counsel in first-line HER2-positive gastric and gastroesophageal disease. HERIZON-GEA-01 now arrives with a different HER2 backbone, zanidatamab, a bispecific that engages two non-overlapping HER2 epitopes, and with tislelizumab rather than pembrolizumab as the immuno-oncology (IO) partner. Whether the PD-L1 story echoes KEYNOTE-811 or diverges from it will materially change how this regimen is positioned.

A few considerations deserve attention. Where do the CPS cut points fall, and what is the magnitude of the hazard ratio in the CPS-low population? If the benefit washes out below CPS 1, that argues for the same kind of biomarker-restricted indication that already governs pembrolizumab use. The more practically important question is whether the zanidatamab-plus-chemotherapy doublet without IO holds up well enough in the CPS-low subgroup to become a defensible standalone option, which would meaningfully expand what can be offered to patients in whom checkpoint blockade is undesirable. PD-L1 stratification was not the primary endpoint, so this will be hypothesis-shaping rather than hypothesis-confirming. Still, in a disease where a CPS-restricted IO indication has shaped practice for years, even a directional signal will have an impact in the clinic.

RASolute 302: Pan-RAS Inhibition in Metastatic Pancreatic Cancer

RASolute 302 (LBA5) will report phase 3 primary and final analysis results of daraxonrasib versus chemotherapy in previously treated metastatic pancreatic ductal adenocarcinoma (mPDAC). Pancreatic cancer is roughly 90% KRAS-mutant with KRAS G12D being the most common. Daraxonrasib is a pan-RAS(ON) inhibitor that engages the active GTP-bound state across multiple KRAS variants. A positive phase 3 trial in second-line mPDAC against chemotherapy would be the first time a RAS-targeted strategy has produced a survival win in this disease.

Two things deserve scrutiny: the comparator arm and how the trial defined previously treated. Second-line pancreatic cancer is a setting where the chemotherapy control arm (typically nanoliposomal irinotecan with 5-fluorouracil [5-FU]/leucovorin, or FOLFOX [leucovorin, 5-FU, oxaliplatin] depending on first-line exposure) has a well-characterized but modest benchmark. The magnitude of overall survival (OS) and progression-free survival (PFS) separation matters enormously here, both for regulatory framing and for how a community oncologist would counsel a patient about a new non-chemotherapy option. The toxicity profile also warrants careful attention. Pan-RAS inhibition has potential on-target liabilities (skin, GI, and metabolic effects driven by RAS signaling in normal tissue). Also the question is whether the therapeutic window in a heavily pretreated, often deconditioned pancreatic cancer population is wide enough to be deliverable in practice. A statistically positive trial with a regimen most patients cannot tolerate is a very different outcome than a positive trial with a regimen they can. Regardless, this is a significant step forward for pancreatic cancer.

BREAKWATER Cohort 3 and SWOG S2107 in BRAF 600E mCRC

BRAF V600E metastatic colorectal cancer (mCRC) will get two presentations worth attending in sequence. BREAKWATER cohort 3 (LBA3503) will report PFS and OS for first-line encorafenib plus cetuximab (EC) with FOLFIRI (leucovorin, 5-FU, and irinotecan), the irinotecan-backbone counterpart to the EC plus mFOLFOX6 data that already established this regimen as a new first-line standard in BRAF V600E mCRC. The clinical question is whether FOLFIRI delivers comparable benefits to FOLFOX. That matters in real-world community practice, where oxaliplatin neuropathy, prior adjuvant exposure, or patient preference often dictates the doublet. So far, we see positive PFS and OS have trended favorably but immature at the prior analysis. Thus, updated survival data is what to track. The SWOG S2107 (Abstract 3504) study asks a different question in the same population. Does adding nivolumab to EC produce durable responses in previously treated microsatellite stable (MSS) disease? The biologic hypothesis is reasonable, but MSS colorectal cancer has burned the field on similar combination ideas before. Durability will matter more than headline response rate.

CIRCULATE: ctDNA-Guided Adjuvant Therapy in  Colon Cancer

The CIRCULATE trial (LBA3500) will report disease-free survival and time to recurrence with circulating tumor DNA (ctDNA)-guided adjuvant decisions in stage II colon cancer. This is the setting where the benefit-versus-harm calculus for adjuvant chemotherapy is most closely balanced, and where shared decision-making can be difficult. A positive result would give the community oncologist a defensible way to escalate in ctDNA-positive patients and de-escalate in ctDNA-negative ones, rather than relying on classic T-stage and pathologic risk features. The event rate in the ctDNA-negative arm is the number that matters most. The entire de-escalation argument depends on it being genuinely low. 

Implications for Community Oncology Practice

Across these studies, the common thread is that molecular-defined treatment is no longer a frontier in GI oncology. It is the operating assumption. HER2 and PD-L1 in gastroesophageal disease, KRAS in pancreas, BRAF and mismatch repair/microsatellite instability status, and ctDNA in colorectal: every one of these abstracts is, at its core, a question about which patients are identified and how quickly. For the community oncologist, the most consequential challenges are increasingly downstream. This included turnaround times and access. The everyday infrastructure that determines whether trial results reach patients. That is the part of ASCO that does not fit in an abstract, and the part most worth discussing with colleagues after the meeting concludes.