ASCO 2026: A Bispecific Antibody Just Cracked One of Lung Cancer’s Hardest Problems — OS Data Are In, But Does the Chinese Data Translate to American Patients

Squamous non-small cell lung cancer (NSCLC) has always been the harder problem. Targeted therapies rarely apply, PD-L1 expression is an imperfect guide, and the gains from adding checkpoint inhibitors to chemotherapy, while real, have plateaued. For community oncologists managing these patients day-to-day, the first-line toolkit has felt essentially frozen since KEYNOTE-407. New data on ivonescimab from the HARMONi-6 trial appears to challenge that status quo.

At the ASCO 2026 Annual Meeting plenary session on May 31, Dr. Shun Lu, Chief of the Shanghai Lung Cancer Center at Shanghai Chest Hospital in China, presented the overall survival results of HARMONi-6, a phase 3 randomized trial testing ivonescimab plus chemotherapy against tislelizumab plus chemotherapy in previously untreated advanced squamous NSCLC.

The data, simultaneously published in The Lancet, showed a statistically significant OS benefit: median overall survival of 27.9 months with ivonescimab versus 23.7 months with tislelizumab (HR 0.66; 95% CI, 0.50–0.87; p=0.0017).

Dr. Pasi Jänne, Director of the Lowe Center for Thoracic Oncology at Dana-Farber Cancer Institute, tells SurvivorNet Connect that the findings are “a game-changer,” and “a significant finding for the field,” but oncologists will need to weigh limitations alongside these headline numbers.

This new data follow already-impressive progression-free survival data presented at ESMO 2025 and published in The Lancet. The ASCO readout answered the question that actually matters: did patients live longer?

They did, and the survival curves separated at six months and stayed separated throughout follow-up.

Ivonescimab: Why It’s Different

Most immunotherapy drugs, like pembrolizumab or nivolumab, work by blocking a single protein called PD-1. Cancer cells use PD-1 to hide from the immune system. Block it, and the immune system can attack the tumor again. That’s the mechanism behind the checkpoint inhibitor revolution of the past decade.

Ivonescimab hits two targets at once. It’s a bispecific antibody, engineered to simultaneously block both PD-1 and VEGF (vascular endothelial growth factor). The idea is that cutting off both the immune escape route and the tumor’s blood supply at the same time produces a stronger, more durable effect than blocking either alone.

“I think it’s the cooperative binding that happens when the two are together, where you don’t get that at each individual level. And certainly we’re starting to see that in the results from the clinical trials, so I think that speaks volumes about this new approach,” Dr. Jänne says.

Until now, anti-VEGF therapy has been essentially off-limits in squamous NSCLC. Bevacizumab (brand name Avastin), the most widely used anti-VEGF drug, carries a serious bleeding risk in squamous tumors, many of which sit close to major blood vessels. Ivonescimab appears to deliver anti-VEGF activity through a different molecular mechanism, with a more acceptable hemorrhagic profile. The bleeding risk is not zero, but it was low enough to run the trial and low enough to be manageable with appropriate patient selection.

HARMONi-6 Yields Promising Results

HARMONi-6 enrolled 532 patients across 50 centers in China. All had previously untreated advanced squamous NSCLC (stage IIIB–IV), no prior systemic therapy, ECOG PS 0-1, and no EGFR or ALK alterations. Importantly, 63% had centrally located tumors, the very patients who have historically been excluded from anti-VEGF trials.

Patients were randomized 1:1 to ivonescimab plus platinum-based chemotherapy or tislelizumab plus chemotherapy for four cycles, followed by maintenance with the immunotherapy agent alone.

• PFS results (already established at ESMO 2025 and published in The Lancet): Median PFS was 11.1 months with ivonescimab versus 6.9 months with tislelizumab — a 40% reduction in the risk of progression or death (HR 0.60; 95% CI, 0.46–0.78; p<0.0001). That benefit held regardless of PD-L1 expression level, which is clinically useful: it means you don’t need a high PD-L1 score to expect benefit.
• Overall survival (ASCO 2026, The Lancet May 31, 2026): At a median follow-up of 21.4 months, median OS was 27.9 months with ivonescimab versus 23.7 months with tislelizumab (HR 0.66; 95% CI, 0.50–0.87; p=0.0017), meeting the prespecified interim boundary (p<0.0049).
• • Two-year OS rates were 64.7% versus 48.6%
  • Twelve-month OS: 78.9% versus 72.2%
  • The OS benefit was consistent across prespecified subgroups, including PD-L1 TPS <1% (HR 0.64; 95% CI, 0.43–0.96), stage IV disease (HR 0.64; 95% CI, 0.48–0.86), and patients with three or more metastatic sites (HR 0.47; 95% CI, 0.26–0.85)
  • Notably, a larger benefit was observed in patients under 65 (HR 0.43; 95% CI, 0.28–0.67) compared to those aged 65-75 (HR 0.93; 95% CI, 0.64–1.36) — a finding worth watching in future analyses, though the subgroup data are descriptive and not formally powered

One important caveat on the OS readout is that the median OS in the ivonescimab arm was reached following a single late event, the last death in the ivonescimab group led the Kaplan-Meier curve to drop sharply below the median point. Had that event not occurred, median OS in the ivonescimab group would not yet have been reached, suggesting the true long-term OS benefit may be even larger than these interim data reflect.

Treatment-related adverse events of grade 3 or higher occurred in 69% of patients in the ivonescimab group versus 59% in the tislelizumab group. Critically, rates of treatment discontinuation due to adverse events were similar: 5% in each arm, and immune-related adverse events of grade 3 or higher were 14% in both groups.

The VEGF-related safety signal was present but manageable:

• Grade 3 or higher hemorrhagic events occurred in 3% of the ivonescimab arm versus 1% with tislelizumab
• Among patients with tumor cavitation at baseline, respiratory tract bleeding events were more frequent with ivonescimab (17% vs. 4%), though all were grade 1 or 2 in that subgroup
• Hypertension grade 3 or higher: 4% versus 2%
• Proteinuria grade 3 or higher: 7% versus 0%

“When you have two drugs in one, you can also get additional side effects from those two drugs. We want to see what is the additional cost in terms of side effects, and are those tolerable and manageable side effects,” Dr. Jänne explains.

The discontinuation parity between arms suggests they are, but pre-treatment screening for hemorrhagic risk factors remains an essential practice.

The Depth The Numbers Don’t Capture

The OS benefit in PD-L1-negative tumors deserves specific attention. The HR of 0.64 in patients with TPS <1% suggests that the VEGF-blocking component is contributing meaningfully to efficacy in patients where PD-1 inhibition alone would be expected to have a limited impact. This is biologically coherent: VEGF-mediated immunosuppression within the tumor microenvironment is not PD-L1 dependent, and disrupting it through simultaneous blockade may rescue immune responses that standard checkpoint inhibition cannot.

This finding also partially addresses one of the open strategic questions in squamous NSCLC: what to do for the roughly 40% of patients with low or absent PD-L1 expression. For that subgroup, which has historically derived less benefit from pembrolizumab-based regimens, ivonescimab’s PD-L1-independent efficacy signal is clinically meaningful, if it replicates in a global population.

The quality-of-life data are also worth noting. Updated analyses showed delayed deterioration in global health status in the ivonescimab arm, with better treatment adherence overall, a finding that matters for patients with a median age of 64 who will be living on this treatment for an extended period.

The Data’s Limits

It is important to reiterate that every patient in HARMONi-6 was enrolled in China. China-only trials have a track record of not always replicating in Western populations, partly because of differences in tumor biology, patient genetics, smoking patterns, and how treatment is delivered. We’ve seen this play out before in lung cancer.

“This is a Chinese-only study. We are seeing more of these studies in the U.S. as well now, and I think it’s important to get data on both our worldwide populations, as there are medicines that can be metabolized differently in different ethnic groups of individuals, which may impact both side effects and efficacy,” Dr. Jänne says.

The trial also enrolled relatively few female patients which is a reflection of the demographics of squamous NSCLC in China, but a constraint on generalizability to a more diverse U.S. population.

Additionally, ivonescimab was compared to tislelizumab, not pembrolizumab. In your clinic, the relevant comparator for first-line squamous NSCLC is pembrolizumab plus chemotherapy, based on KEYNOTE-407. Tislelizumab and pembrolizumab produced similar PFS numbers in their respective trials, about 6.9 versus 8.0 months, which is reassuring for indirect comparison. But it’s still an indirect comparison, and those have real limitations.

Essentially, HARMONi-6 tells us ivonescimab beats another PD-1 inhibitor in a Chinese patient population. It does not yet tell us ivonescimab beats pembrolizumab in American patients.

Moving Forward: Watch HARMONi-3

The trial to watch is HARMONi-3, a global phase 3 study comparing ivonescimab plus chemotherapy directly against pembrolizumab plus chemotherapy. It includes both squamous and non-squamous cohorts, analyzed separately.

Final PFS data from the squamous cohort are expected in the second half of 2026. An FDA decision on ivonescimab in the U.S. is expected in November 2026.

If HARMONi-3 shows that ivonescimab beats or meaningfully improves on pembrolizumab in a global population, the conversation changes quickly and significantly. Until then, HARMONi-6 is best understood as strong, directional evidence, but not yet a reason to change what physicians in the U.S. are prescribing.

“It’s really the first randomized trial to show a benefit of this drug over chemotherapy-immunotherapy alone in squamous cell cancer specifically,” Dr. Jänne says. “I think, I hope, a bit of an inflection point that we’ll continue to see more after this, but it’s a significant finding for the field.”

For now, if a patient with squamous NSCLC asks you about ivonescimab, the honest answer is that the early data look genuinely promising, it’s not yet approved in the U.S., and the trial that will determine whether it belongs in American clinics is expected to report later this year.

Fast Facts: HARMONi-6 OS Analysis

• • • Drug: Ivonescimab (20 mg/kg Q3W) + carboplatin/paclitaxel × 4 cycles → ivonescimab maintenance vs. tislelizumab (200 mg Q3W) + carboplatin/paclitaxel × 4 cycles → tislelizumab maintenance
    • Population: Previously untreated advanced squamous NSCLC; 532 patients; 50 centers; China only; 92% stage IV; 63% central tumor type; 39% PD-L1 TPS <1%
    • Median follow-up: 21.4 months
    • Median OS: 27.9 months (ivonescimab) vs. 23.7 months (tislelizumab); HR 0.66 (95% CI, 0.50–0.87); p=0.0017
    • 2-year OS rate: 64.7% vs. 48.6%
    • Median PFS (ESMO 2025): 11.1 vs. 6.9 months; HR 0.60; p<0.0001; PD-L1 independent
    • Grade ≥3 TRAEs: 69% vs. 59%; discontinuation rate equal at 5% in each arm
    • Grade ≥3 hemorrhage: 3% vs. 1%
    • OS benefit in PD-L1 TPS <1%: HR 0.64 (95% CI, 0.43–0.96)