The Bispecific Antibody Landscape | ESMO 2026

Bispecifics are increasingly evolving from an emerging platform into a competitive therapeutic class, but Programmed Death-ligand 1 (PD-L1) × Vascular Endothelial Growth Factor (VEGF) illustrates how quickly platform validation can turn into a crowded race. Ivonescimab has established an important proof of concept through multiple randomized Phase III signals across Non-small Cell Lung Cancer (NSCLC) and Biliary Tract Cancer (BTC), including Overall Survival benefit versus pembrolizumab (KEYTRUDA) in HARMONi-2 and versus tislelizumab plus chemotherapy in HARMONi-6, while HARMONi-GI1 has extended the clinical story beyond lung cancer. However, the HARMONi program also illustrates the remaining uncertainty: the initial OS analysis in EGFR-mutant NSCLC did not reach statistical significance, although the subsequent 2026 analysis showed an updated OS HR of 0.76.

The strategic question is therefore changing. The issue is no longer whether dual-target biology can work, but whether individual molecules can reproduce that benefit across geographies, tumor types, comparators, and treatment settings. As competing PD-(L)1×VEGF programs advance globally, the bar is shifting from demonstrating biological activity to establishing clinically meaningful and reproducible differentiation.

Why is the platform case now credible?

A true platform should do more than deliver a single successful trial; it should demonstrate repeatable clinical benefit across treatment settings, active comparators, and diverse tumor biology. Ivonescimab is approaching that threshold, with four positive Phase III readouts in lung cancer and one in BTC, providing a broader clinical validation of PD-(L) 1×VEGF bispecificity beyond a single tumor setting.

 

Together, these studies support the clinical plausibility of PD-(L) 1×VEGF bispecificity across multiple treatment architectures, with and without chemotherapy and against different PD-(L) 1 backbones. However, the claim that ivonescimab creates a synergistic effect beyond simply combining PD-(L)1 and VEGF inhibition remains a mechanistic hypothesis rather than a directly proven clinical conclusion.

Why it still looks like a crowded race?

The emergence of multiple late-stage programs means that platform validation does not automatically translate into asset-level leadership.

Pumitamig is particularly important because its development is already global and broad, with pivotal studies spanning NSCLC, SCLC, TNBC, CRC, gastric cancer, and other settings. Global Phase II data in first-line NSCLC have shown encouraging activity across PD-L1 expression levels and histologies, while the Phase III portion of ROSETTA Lung-02 is evaluating pumitamig plus chemotherapy against pembrolizumab plus chemotherapy. Tivunatamig adds another credible challenger, with Pfizer highlighting ongoing pivotal development in NSCLC, metastatic CRC and endometrial cancer.

The race is therefore crowded, but not yet mature. Ivonescimab has the most extensive randomized efficacy dataset today; challengers are still establishing whether their clinical profiles can match or exceed that benchmark.

Can efficacy move beyond China?

Ivonescimab has generated some of the strongest Phase III signals in the class, but much of its evidence base comes from China. The critical next step is global reproducibility, whether treatment effects hold across Western populations, treatment patterns, subsequent therapies, and regulatory settings. The differentiator will increasingly be consistency of benefit, not statistical significance in a single population.

Does the benefit hold outside lung?

This is the critical test of whether PD-(L)1×VEGF can develop into a broader multi-tumor strategy rather than remain primarily an NSCLC-led opportunity. HARMONi-GI1 compares ivonescimab plus gemcitabine/cisplatin with durvalumab plus gemcitabine/cisplatin in first-line advanced BTC. The reported interim analysis met its OS, PFS and ORR endpoints, but the detailed OS HR and median OS are essential for understanding the magnitude and durability of benefit.

The BTC result could also increase interest in other angiogenesis-driven tumors, particularly HCC, where immune checkpoint and VEGF-pathway combinations already have clinical precedent. However, activity in BTC should not be interpreted as proof that the same architecture will outperform established combinations across other tumor types.

Can bispecifics coexist with or outperform the ADC revolution?

ADCs are moving rapidly into earlier lines and across tumor types, raising the bar for bispecifics to demonstrate value beyond incremental checkpoint inhibition. The more compelling opportunity may be bispecific and ADC combinations, where immune activation and targeted cytotoxic delivery could be complementary. Early data with pumitamig + BNT325/DB-1305 support feasibility, but larger randomized datasets must establish whether this translates into deeper, more durable responses without compromising safety. The key question is whether bispecifics become competitors to ADCs, or the immune backbone that enables the next generation of ADC regimens.

Will the next wave be bispecific or multispecific?

The platform is already moving toward trispecific and more complex multispecific constructs designed to address antigen escape, tumor heterogeneity, immune exhaustion, and inadequate immune activation.

However, greater molecular complexity does not automatically create greater clinical value. Each additional target increases the challenges around engineering, pharmacology, safety, biomarker selection, manufacturing, and trial design. ESMO 2026 should therefore help determine whether multispecificity represents a meaningful next step in therapeutic design, or whether simpler bispecific architectures can deliver a more attractive balance of efficacy, safety, and manufacturability.

ESMO 2026 Watchlist

Post-ESMO 2026 Catalyst Calendar

 

Investment and business development implication

For pharmaceutical companies, the strategic question is shifting from “Should we enter bispecifics?” to “Which bispecific architecture is defensible?”

Platform ownership alone is unlikely to be enough as competition intensifies. The most valuable assets will increasingly combine validated biology, differentiated design, biomarker strategy, combination flexibility, and compelling global clinical evidence. For investors and BD teams, the key distinction will be between molecules that demonstrate activity and those that establish a durable, differentiated role within the treatment paradigm.

In conclusion, bispecifics have moved beyond technological promise, but the winners are not yet established. PD-(L) 1×VEGF is the clearest example: the mechanism is increasingly validated, while the molecules are entering a race for reproducibility, differentiation, and commercial relevance. ESMO 2026 is therefore less about proving that bispecifics can work and more about identifying which architectures can turn that proof of concept into durable clinical value.