HERIZON-GEA-01 PD-L1 Subgroup Data Support Zanidatamab Combo as a New Frontline SOC in HER2+ GEA

Zanidatamab-hrii (Ziihera) plus tislelizumab (Tevimbra) and chemotherapy provided progression-free survival (PFS) and overall survival (OS) benefits over trastuzumab (Herceptin) plus chemotherapy in the frontline treatment of patients with HER2-positive locally advanced or metastatic gastroesophageal adenocarcinoma (GEA), regardless of PD-L1 status, according to findings from a preplanned subgroup analysis of the phase 3 HERIZON-GEA-01 trial (NCT05152147).¹

Previously published findings showed that in the intention-to-treat (ITT) population, the median PFS was 12.4 months (95% CI, 9.8-18.5) with zanidatamab plus tislelizumab and chemotherapy (n = 302) vs 8.1 months (95% CI, 7.0-8.9) with trastuzumab plus chemotherapy (n = 308; HR, 0.63; 95% CI, 0.51-0.78; P < .0001), a greater than 4-month improvement.² The median OS was 26.4 months (95% CI, 21.5-30.3) vs 19.2 months (95% CI, 16.8-21.8), respectively (HR, 0.72; 95% CI, 0.57-0.90; P = .0043), a greater than 7-month prolongation.

The subgroup data, which were shared during the 2026 ASCO Annual Meeting, indicated that PFS improvements with the zanidatamab regimen vs trastuzumab plus chemotherapy were consistent across PD-L1 subgroups regardless of whether the disease was assessed by tumor area positivity (TAP) score and combined positive score (CPS).¹

In the TAP score analysis, the median PFS was 11.3 months (95% CI, 9.6-18.5) with the zanidatamab combination vs 8.3 months (95% CI, 6.9-9.7) with trastuzumab plus chemotherapy in patients with a TAP of 1% or higher (HR, 0.65), and 18.5 months (95% CI, 9.7-25.2) vs 7.9 months (95% CI, 5.8-9.6) in those with a TAP of lower than 1% (HR, 0.47). In the CPS analysis, the median PFS was 12.3 months (95% CI, 9.7-18.5) vs 8.2 months (95%CI, 6.9-9.1) in those with CPS of 1 or higher (HR, 0.62), and 18.5 months (95% CI, 9.7-25.2) vs 8.1 months (95% CI, 5.8-9.8) in those with CPS below 1 (HR, 0.48).

OS results were similarly consistent. In the TAP score subgroup analysis, the median OS with the zanidatamab combination was 26.4 months (95% CI, 18.7-35.9) in patients with a TAP score of 1% or higher vs 21.2 months (95% CI, 17.7-25.2) with trastuzumab plus chemotherapy (HR, 0.82); in those with a TAP score below 1%, the median OS in the respective arms was 29.7 months (95% CI, 24.7-not evaluable [NE]) vs 15.8 months (95% CI, 12.6-21.4; HR, 0.49). In the CPS analysis, the median OS was 26.4 months (95% CI, 18.7-34.6) with the zanidatamab regimen vs 20.8 months (95% CI, 17.3-23.9) with trastuzumab plus chemotherapy in those with a CPS of 1 or higher (HR, 0.82); in those with a CPS below 1, the median OS in the respective arms was 30.3 months (95% CI, 25.7-NE) and 15.7 months (95% CI, 12.6-21.4; HR, 0.43).

Presenting author Sun Young Rha, MD, PhD, of Yonsei Cancer Center at Yonsei University College of Medicine, in Seoul, South Korea, noted that OS in the control arm may be impacted by increased use of subsequent therapies. Among those with a TAP score of 1% or higher, 55.9% in the control arm received any subsequent therapy vs 32.6% in the zanidatamab triplet arm; in the TAP score below 1% group, these rates were 58.2% vs 30.0%, respectively. Moreover, in those with a TAP of 1% or higher, 11.7% vs 2.7% of patients had subsequent immune checkpoint inhibitor use, and 29.3% vs 13.4% of patients received subsequent HER2-targeted therapy.

“Zanidatamab plus tislelizumab plus chemotherapy demonstrated a clinically meaningful and statistically significant prolongation of PFS and OS compared to trastuzumab and chemotherapy, and these benefits were consistent in patients with PD-L1–positive and PD-L1–negative disease determined by TAP score and CPS,” Rha said.

Rha added that PD-L1–independent benefit may be explained by the immune activation induced by zanidatamab, which could potentiate the efficacy of tislelizumab.

Why do PD-L1 subgroup data matter in first-line HER2-positive mGEA?

In HER2-positive locally advanced or mGEA, the addition of anti–PD-1 immunotherapy to standard first-line treatment has historically conferred survival benefit primarily in patients with PD-L1–positive tumors. For example, final OS results from the phase 3 KEYNOTE-811 trial (NCT03615326) showed that adding pembrolizumab (Keytruda) to trastuzumab and chemotherapy reduced the risk of death by 20% in the ITT population (HR, 0.80; 95% CI, 0.67-0.94; P = .004);³ however, the benefit was enriched in those with PD-L1 CPS of 1 or greater, raising clinical questions about which patients are best served by the addition of immunotherapy.

Zanidatamab is a HER2-directed bispecific antibody engineered to bind simultaneously to HER2 extracellular domain (ECD) 2 and ECD4 in a trans configuration, facilitating the formation of distinct HER2 receptor clusters on the tumor cell surface, Rha explained. This architecture enhances HER2 internalization, reduces downstream phosphorylation of EGFR, HER2, and HER3, and amplifies immune-mediated tumor killing through complement-dependent cytotoxicity, antibody-dependent cellular cytotoxicity, and antibody-dependent cellular phagocytosis. Tislelizumab is a high-affinity anti–PD-1 antibody specifically engineered to minimize Fc-gamma receptor binding on macrophages, reducing macrophage-mediated T-cell depletion.

The preplanned PD-L1 subgroup analysis presented at ASCO, which used both the prespecified TAP score and the exploratory CPS with a cutoff of 1, sought to shed light on whether the survival benefit achieved with the zanidatamab regimen is truly independent of PD-L1 expression.

How was the HERIZON-GEA-01 PD-L1 subgroup analysis conducted?

The global, open-label, randomized phase 3 HERIZON-GEA-01 trial enrolled previously untreated patients with locally advanced, unresectable, recurrent or metastatic HER2-positive (IHC 3+ or IHC 2+/ISH+) GEA. A total of 914 patients were randomized 1:1:1 to trastuzumab plus chemotherapy (arm A; n = 308), zanidatamab plus chemotherapy (arm B; n = 304), or zanidatamab plus tislelizumab and chemotherapy (arm C; n = 302). The chemotherapy backbone was investigator's choice of capecitabine plus oxaliplatin (CAPOX) or fluorouracil plus cisplatin. Dual primary end points were PFS assessed by blinded independent central review (BICR) and OS.

The current PD-L1 subgroup analysis compared arms A and C. PD-L1 expression was assessed retrospectively in a preplanned analysis using the VENTANA SP263 assay. Results were stratified by two scoring methods: TAP score (the prespecified primary analysis) and CPS (an exploratory analysis), both with a cutoff of 1%.

Although PD-L1 status was not a stratification factor at randomization, baseline demographics and clinical characteristics were well balanced across PD-L1 subgroups, according to Rha. She added that concordance between PD-L1 status by TAP score of at least 1% vs less than 1% and CPS of at least 1 and below 1 was high, at 93.3%.

What did the safety analysis show?

The safety profile with zanidatamab plus tislelizumab and chemotherapy was manageable, Rha said. The median duration of treatment was 43.1 weeks in the zanidatamab arm (n = 294) vs 30.0 weeks in the trastuzumab plus chemotherapy arm (n = 302). Any-grade treatment-related adverse effects (TRAEs) occurred in 98.3% vs 96.4% of patients, respectively; grade 3 or higher TRAEs occurred in 71.8% vs 59.6%. Serious treatment-emergent adverse effects (TEAEs) were reported in 58.5% vs 42.4% of patients, respectively.

Discontinuation of any treatment component due to TEAEs occurred in 42.5% of patients in the zanidatamab arm vs 29.1% of those in the control arm; discontinuation of zanidatamab specifically occurred in 11.9% of patients, and 2.3% discontinued trastuzumab. Diarrhea was the most common grade 3 or higher toxicity across all groups, Rha said.

Moreover, adverse effects of special interest occurred in 34.7% of patients in the zanidatamab arm vs 18.5% of patients in the trastuzumab/chemotherapy arm. Infusion-related reactions were reported in 25.2% and 13.2% of patients, respectively. Noninfectious pulmonary toxicities occurred in 6.8% of patients in the zanidatamab arm vs 1.0% of those in the trastuzumab/chemotherapy arm, and left ventricular dysfunction was observed in 8.8% of patients vs 4.3% of those in the trastuzumab plus chemotherapy arm.

Disclosures: Rha reported consulting or advisory roles with Amgen, Arcus Biosciences, Astellas Pharma, AstraZeneca, BeOne, Daiichi Sankyo, Eisai, Indivumed, LG Chem, MSD Oncology, Ono Pharmaceutical, and Toray Industries; a speakers’ bureau role with Amgen, Arcus Biosciences, Astellas Pharma, AstraZeneca, BeOne, Bristol-Myers Squibb/Ono Pharmaceutical, Daiichi Sankyo/UCB Japan, Eisai, and MSD Oncology; and institutional research funding from ASLAN Pharmaceuticals, Astellas Pharma, AstraZeneca, Bayer, BeiGene, BeOne, Beringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Eisai, Indivumed, Jazz Pharmaceuticals, Lilly, MSD Oncology, Roche/Genentech, and Sillajen.

References

1. Rha SY, Shitara K, Lin S, et al. Zanidatamab + chemotherapy (CT) ± tislelizumab for first-line HER2-positive locally advanced or metastatic gastroesophageal adenocarcinoma (mGEA): PD-L1 subgroup analysis from HERIZON-GEA-01. J Clin Oncol. 2026;44(suppl 16):4010. doi:10.1200/JCO.2026.44.16_suppl.4010
2. Elimova E, Rha SY, Shitara K, et al. Zanidatamab + chemotherapy (CT) ± tislelizumab for first-line (1L) HER2-positive (HER2+) locally advanced, unresectable, or metastatic gastroesophageal adenocarcinoma (mGEA): Primary analysis from HERIZON-GEA-01. J Clin Oncol. 2026;44(suppl 44):LBA285. doi:10.1200/JCO.2026.44.2_suppl.LBA285
3. Janjigian YY, Kawazoe A, Bai Y, et al. Pembrolizumab in HER2-positive gastric cancer. N Engl J Med. 2024;391(14):1360-1362. doi:10.1056/NEJMc2408121