Introduction: A Year Defined by HER2 and Immunotherapy Combinations

Upper GI cancers remain among the most lethal malignancies worldwide, and the treatment landscape has been evolving rapidly since the approvals of trastuzumab in HER2-positive disease and checkpoint inhibitors across the HER2-negative population. ASCO 2026 brought four abstracts that collectively clarify where the field is heading: a new bispecific antibody combination that may redefine the standard of care for HER2-positive gastroesophageal adenocarcinoma, a negative dual immunotherapy trial, and two trials from Asia that introduce novel mechanisms.

HORIZION-GEA-01: Zanidatamab May be a new standard of care option in HER2-Positive Disease

The current standard of care for HER2-amplified advanced gastroesophageal adenocarcinoma is chemotherapy plus trastuzumab plus pembrolizumab (for patients with PD-L1 CPS of at least one), established by KEYNOTE-811. The median OS with this regimen is approximately twenty months.

HORIZON-GEA-01 evaluated zanidatamab, a HER2-directed bispecific antibody with a mechanistically distinct design from trastuzumab. While trastuzumab binds a single epitope on the HER2 extracellular domain, zanidatamab binds two separate epitopes (ECD2 and ECD4) in a trans configuration, facilitating HER2 clustering on the cell surface, enhanced receptor internalization, and potentially superior antitumor activity. It was combined with tislelizumab, a high-affinity PD-1 inhibitor engineered to minimize Fcγ receptor binding on macrophages, a design intended to reduce tumor-associated immunosuppression.

The trial enrolled patients with metastatic or unresectable HER2-positive gastroesophageal cancer in the first-line setting, randomized to three arms: the control arm (trastuzumab plus chemotherapy), a triplet (zanidatamab plus tislelizumab plus chemotherapy), and a doublet (zanidatamab plus chemotherapy, without IO). The trial used both TAP and CPS for PD-L1 classification, with good concordance between these metrics established in previous studies.

Triplet results: The zanidatamab-tislelizumab-chemotherapy triplet significantly improved PFS over the control arm (HR 0.63; 37% reduction in risk of progression or death). Interim OS also significantly favored the triplet (HR approximately 0.72; 28% reduction in risk of death). The median OS of the triplet was approximately twenty-six months numerically, longer than the twenty-month historical benchmark with trastuzumab-pembrolizumab-chemotherapy from KEYNOTE-811, though cross-trial comparisons carry significant limitations.

Doublet results: The zanidatamab-chemotherapy doublet also significantly improved PFS (HR 0.65), but the interim OS did not reach statistical significance (HR 0.80, just missing the pre-specified alpha). Final OS data are still pending. 

PD-L1 subgroup: A counterintuitive finding. The updated analysis presented at ASCO examined whether PD-L1 expression predicted benefit from this regimen. Interestingly, both TAP-low (less than one) and CPS-low (less than one) subgroups showed hazard ratios numerically more favorable than the PD-L1-high subgroups for both PFS and OS. The TAP less-than-one subgroup had a PFS HR of 0.47 versus 0.65 for TAP at least one; CPS less than one produced a HR of 0.28 versus 0.62 for CPS at least one. Why PD-L1-negative patients appear to benefit more is not fully understood. One hypothesis involves the higher proportion of Asian patients in the PD-L1-negative population in both HORIZON and Attraction-6; differences in patient ethnicity, tumor biology, or testing practices may be contributing. The practical implication is reassuring: even patients with PD-L1-negative disease appear to derive meaningful benefit from the zanidatamab-tislelizumab-chemotherapy combination.

Safety: The triplet had numerically higher grade 3 and 4 treatment-related adverse events and more immune-mediated adverse events compared to control. Diarrhea and immune-mediated events require monitoring; loperamide prophylaxis during cycle one was mandated in the HORIZON protocol. These are manageable toxicities in the context of a regimen that significantly improves OS.

Conclusion: Once approved, zanidatamab plus tislelizumab plus chemotherapy could become the new standard of care for HER2-positive advanced gastroesophageal adenocarcinoma, applicable regardless of PD-L1 status.

ATTRACTION-6: Dual Immunotherapy Does Not Move the Needle in HER2-Negative Disease

The precedent for combining nivolumab and ipilimumab in gastrointestinal cancer came from CHECKMATE-649, which included a nivolumab-ipilimumab arm that was discontinued early due to excess toxicity (ipilimumab at 3 mg/kg every three weeks). ATTRACTION-6 revisited this concept with a lower ipilimumab dose (1 mg/kg every three weeks) added to standard-dose nivolumab and chemotherapy in HER2-negative metastatic gastroesophageal cancer.

The study was conducted exclusively in Japan, Korea, and Taiwan, which introduces important considerations about generalizability. One notable feature of the enrolled population: PD-L1 CPS of at least five was present in only forty-four percent of patients, compared to approximately sixty percent in CHECKMATE-649, a lower-PD-L1 population that may have been less likely to benefit from immunotherapy intensification.

The trial did not meet its primary endpoint. OS and PFS did not significantly improve with the quadruplet versus chemotherapy alone (OS HR 0.90; PFS HR 0.83). The experimental arm also had meaningfully higher rates of adverse events and lower chemotherapy dose intensity; patients in the experimental arm could not tolerate the same intensity of chemotherapy, which may have offset any benefit from the additional immunotherapy.

Even the CPS-at-least-five subgroup analysis did not show statistically significant benefit, though ORR was numerically higher in the quadruplet arm. The likely contributors to the negative result include: greater use of subsequent immunotherapy in the control arm (forty percent received IO after progression, versus three percent in the experimental arm, potentially equalizing OS); the lower-than-expected PD-L1-high population; and the chemotherapy dose attenuation in the experimental arm.

Conclusion: Nivolumab plus low-dose ipilimumab plus chemotherapy did not improve outcomes for HER2-negative advanced gastroesophageal cancer and is unlikely to become standard of care.

Two Signals From Asia: PANKU and ONO-4578

Izalontamab Brengitecan (PANKU): A New Standard in China, Not Yet in the US

Izalontamab brengitecan is an EGFR-HER3 bispecific antibody ADC with high-affinity EGFR binding, low-affinity HER3 binding, and a topoisomerase I inhibitor payload (Ed-04). The PANKU trial was a multicenter, randomized, open-label phase 3 study conducted in China, comparing izalontamab brengitecan to standard-of-care chemotherapy as second-line therapy for advanced esophageal squamous cell carcinoma. The current standard produces an ORR of less than ten percent and median OS under six months.

PANKU met its primary endpoints. PFS, OS, and ORR all significantly favored izalontamab brengitecan; ORR was 35.3% versus 13.1% for chemotherapy, and median OS was approximately ten months versus under six months with historical control. This has become the standard of care for previously treated esophageal SCC in China.

The caveat is direct: this trial was conducted exclusively in China and has not been studied in US or European populations. Until a global or US-based trial demonstrates efficacy in a more diverse population, izalontamab brengitecan will not become available to American patients in this setting. It is, however, a biologically compelling agent worth watching as the program expands.

ONO-4578 (EP4 Inhibitor): Prostaglandin Pathway Inhibition as an IO Enhancer

The fourth abstract addressed an entirely novel mechanism. ONO-4578 is a selective inhibitor of EP4, a receptor for prostaglandin E2 (PGE2). The PGE2 pathway drives tumor-associated immunosuppression; inhibiting EP4 may restore immune responsiveness in the tumor microenvironment and thereby enhance checkpoint inhibitor activity.

The double-blind phase 2 randomized trial was conducted in Japan, Korea, and Taiwan in patients with previously untreated HER2-negative advanced gastroesophageal adenocarcinoma, randomized to chemotherapy plus nivolumab plus ONO-4578 versus chemotherapy plus nivolumab plus placebo. The trial met statistical significance for both PFS and OS. The median PFS was nine months in the ONO-4578 arm versus approximately seven months in the CHECKMATE-649 historical benchmark. The median OS was not reached in the ONO-4578 arm. ORR was 62.0% versus 48.7% for placebo (not statistically significant).

This is a phase 2 study conducted in an Asian population, and larger phase 3 studies enrolling more diverse populations will be needed before this approach can become a global standard. But prostaglandin pathway inhibition as a mechanism to enhance immunotherapy is a genuinely novel concept with biologically compelling rationale.

For Patients

If you have been diagnosed with HER2-positive gastroesophageal cancer, a new combination called zanidatamab plus tislelizumab plus chemotherapy is showing impressive results in clinical trials and is expected to seek FDA approval based on HORIZON-GEA-01. If approved, it may be a new standard of care option regardless of whether your tumor expresses PD-L1. If you have esophageal squamous cell carcinoma that has progressed after prior treatment, clinical trial opportunities using new ADC drugs with novel mechanisms are expanding; ask your oncologist about what may be available. For all patients with gastroesophageal cancers, molecular testing for HER2 status and PD-L1 expression at diagnosis remains essential for matching you to the most effective treatment approach.

Key Takeaways

  • HORIZON-GEA-01: Zanidatamab plus tislelizumab plus chemotherapy improves PFS (HR 0.63) and interim OS (HR ~0.72) vs trastuzumab plus chemotherapy in HER2+ gastroesophageal cancer; benefit seen irrespective of PD-L1 status, with stronger hazard ratios in PD-L1-negative subgroups; pending FDA approval.

  • Cross-trial comparison: Median OS ~26 months with zanidatamab triplet vs ~20 months with KEYNOTE-811 standard; subject to all limitations of cross-trial comparison.

  • ATTRACTION-6: Nivolumab plus low-dose ipilimumab plus chemotherapy did not significantly improve OS or PFS in HER2-negative gastroesophageal cancer (Asian-only population); higher toxicity, lower chemo dose intensity, and post-progression IO use in control arm likely contributed to negative result.

  • PANKU: Izalontamab brengitecan (EGFR-HER3 bispecific ADC) is new SOC for 2nd-line esophageal SCC in China; ORR 35.3% vs 13.1%; not yet studied in US populations.

  • ONO-4578 (EP4/PGE2 inhibitor): Phase 2 trial; adding EP4 inhibition to chemo-IO significantly improved PFS and OS in HER2-negative gastroesophageal cancer in Asian populations; phase 3 needed for global application.

References

  1. Xu RH, et al. HORIZON-GEA-01: Zanidatamab plus tislelizumab plus chemotherapy in HER2-positive gastroesophageal adenocarcinoma. ASCO 2026.

  2. Shitara K, et al. ATTRACTION-6: Nivolumab plus ipilimumab plus chemotherapy in HER2-negative gastroesophageal cancer. ASCO 2026.

  3. Janjigian YY, et al. KEYNOTE-811: Pembrolizumab plus trastuzumab plus chemotherapy in HER2-positive gastroesophageal cancer. New England Journal of Medicine, 2024.

  4. Shen L, et al. PANKU: Izalontamab brengitecan in 2L esophageal SCC. ASCO 2026.

  5. Yoshida K, et al. ONO-4578 plus nivolumab plus chemotherapy in HER2-negative gastroesophageal adenocarcinoma. ASCO 2026.

  6. Hsieh RW. "Updates on Upper GI Cancers." Presented at the 2026 Best of Oncology, Seattle, WA, June 2026.