It’s the moment of truth for Akeso’s ivonescimab, and to many, a high-stakes test for China’s booming biotech industry. And the first-in-class PD-1xVEGF bispecific didn’t disappoint—at least on its overall survival headline message.

With a statistically significant 34% improvement on overall survival (OS), ivonescimab plus chemotherapy became the first regimen to beat the established standard of a PD-1 inhibitor plus chemo in patients with previously untreated, advanced squamous non-small cell lung cancer.

The latest overall survival data was generated from an interim analysis of the phase 3 Harmoni-6 trial conducted in China, and it marked the first time that a China-alone data set was selected for presentation at the prestigious plenary session at an annual meeting of the American Society of Clinical Oncology (ASCO).

The 34% number exceeded the expectations of many industry watchers, who typically view a 30% improvement as a big win. Before the revelation, one group within the oncology community was skeptical that Harmoni-6 would meet statistical significance. The p-value landed at 0.0017, meeting the prespecified statistical significance boundary of 0.0049.

David Spigel, M.D., president and chief medical officer at Sarah Cannon Research Institute and an ASCO expert in lung cancer, called the results “certainly very encouraging” during a press conference. In an ASCO release, he said the study “provides a vital new path forward for patients with these difficult-to-treat cancers who have limited treatment options.”

The good news

Ever since ivoenscimab made a splash in 2024 by beating Merck & Co.’s Keytruda on progression-free survival (PFS) in a Chinese head-to-head study in first-line, PD-L1-positive NSCLC, questions have swirled around whether the bispecific would be able to extend overall survival when pitted against PD-1 and chemo in a clinically more relevant approach.

Previously, the original Harmoni-2 trial showing a 22.3% death-risk reduction for ivonescimab alone against single-agent Keytruda failed to impress, albeit from an unplanned interim analysis that had a very high statistical significance bar.

What’s more, standalone VEGF inhibitors have historically proven difficult to confer an OS benefit in first-line NSCLC despite being associated with promising PFS results.

In the latest Harmoni-6 trial, ivonescimab and chemotherapy improved median overall survival time by 4.2 months compared with BeOne Medicines’ Tevimbra and chemo, reaching 27.9 months, after a median follow-up of 21.4 months.

The drug’s benefit appeared consistent between PD-L1-negative and PD-L1-positive patients, with OS improvements of 36% and 32%, respectively.

The control arm’s 23.7-month median OS matched the Tevimbra-chemo regimen’s historical performance of 22.8 months, as shown in the phase 3 Rationale-307 trial conducted in China. Besides, as Spigel noted, a combination of Regeneron’s PD-1 drug Libtayo and chemo delivered a 21.9-month median OS in the squamous subgroup of the Empower-Lung3 trial.

As for Keytruda, the median OS for the Keytruda-chemo arm in the China extension portion of the phase 3 Keynote-407 trial reached 30.1 months, which was widely considered an outlier, as it markedly exceeded the 17.1 months observed in the global study.

All considered, past experiences with PD-1 plus chemo suggest ivonescimab’s survival benefit was not driven by an underperforming comparator.

The ASCO-invited discussant of Hamoni-6, Julie Brahmer, M.D., director of the thoracic oncology program at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, also said she didn’t notice any less robust follow-on treatment distorting OS outcomes.

“There was a fair amount of subsequent treatment, so I’m not as worried about that,” she said in an interview with Fierce. “There was a fair amount of continued IO.”

Before ivonescimab, freestanding VEGF inhibitors like bevacizumab have been strictly prohibited in squamous NSCLC due to potentially life-threatening risks of bleeding.

Now, the PD-1xVEGF bispecific approach has made it possible for squamous NSCLC patients to benefit from an anti-VEGF mechanism.

“That’s great for us to have another option, because PD-1-and-chemo doesn’t perform great in the patient population,” Brahmer said.

But that’s about as far as Brahmer’s praises for Harmoni-6 would go.

Despite the positive topline readout and historical overall survival win, Brahmer picked apart the newest ivonescimab data in a critical assessment about patient selection, toxicity and its applicability to a global population.

“Just in the fact that the data was positive, it was in my mind groundbreaking, particularly for squamous,” Brahmer said. “However, after looking at the details, I would say it’s provocative.”

The not-so-good news

First, on a less negative matter, Brahmer said she hopes to see a longer follow-up from Harmoni-6.

Usually in oncology clinical trials, being able to meet statistical significance early is a feat worth celebrating because it shows the power of a drug with merely a few events. However, given experiences with VEGF inhibitors, in which positive early signals in OS often disappear later, a longer follow-up would be more reassuring for doctors considering the PD-1xVEGF approach.

At this interim analysis, the median follow-up was less than two years and shorter than the median OS of both arms.

Leerink Partners analyst Daina Graybosch, Ph.D., raised a similar point in a preview interview with Fierce before seeing the Harmoni-6 data. Assuming a median follow-up of around 19 to 20 months, Graybosch said, “I happen to think Harmoni-6 is very immature, so the tail of the curve, we’re not going to have a lot of data in, so not sure we can conclude anything.”

At least for now, Harmoni-6’s OS curves chronicling deaths for the two treatment arms separated clearly and did not show any sign of closing back up, according to a slide shared by study lead author Shun Lu, M.D., Ph.D., from Shanghai Chest Hospital, Jiao Tong University School of Medicine, during a press briefing. That suggests an improving benefit of the ivonescimab arm versus control over time, and Brahmer said she hopes the curves will continue to widen.

Harmoni6

Brahmer is also not so impressed by the ivonescimab regimen’s VEGF-related safety profile. The bispecific clearly solved the bleeding restriction tied to separate VEGF inhibition, as grade 3 or above hemorrhages only happened in 2.6% of patients in the ivonescimab-chemo group, versus 0.8% in the Tevimbra-chemo arm. But overall, Brahmer said she didn’t see much difference in VEGF-specific adverse events for the ivonescimab regimen versus historical freestanding anti-VEGF drugs.

Brahmer railed against Harmoni-6’s eligibility criteria, too, saying they create uncertainties for practicing oncologists.

Harmoni-6 excluded patients whose tumors significantly invade, surround or encase major blood vessels, as well as those with a history of significant hemoptysis, all of which increase the risk of bleeding, Brahmer noted. But the study allowed some milder cases with those risk factors.

“The good thing is we don’t see a huge safety signal because they included some of those,” she said. “But on the flip side is, how did they decide who to put on versus not? In the clinic, what am I going to do, and how am I going to figure that out, particularly with squamous, since squamous is more likely to be central [part of the lung], more likely to have cavitation, and more likely to have hemoptysis?”

Harmoni-6 was conducted almost exclusively in men, whereas the typical expectation of a squamous trial is to enroll at least 20% women, Brahmer added.

An age-based exclusion became a big issue for her, as well. Even though the median age of Harmoni-6’s patients was similar to global experience, the Chinese study uniquely excluded patients older than 75 years of age.

In the U.S., the median age of lung cancer patients is much higher, approaching 70, compared with Chinese patients, Brahmer noted. In individuals older than 65, ivonescimab plus chemo “failed to benefit” patients, as the regimen’s hazard ratio versus control was nearly 1, she said. Fierce did not receive the full data set as of publication time and was therefore unable to independently verify this datapoint.

Brahmer draws a comparison to the landmark ECOG 4599 trial, which tested adding bevacizumab to chemo for NSCLC and also found no benefit in elderly patients.

In that trial, investigators found that the older patients had higher toxicity rates. That prompted Brahmer to wonder if the same phenomenon happened in Harmoni-6, too, information that’s not available at this time.

All of those led Brahmer to question the applicability of the Harmoni-6 OS findings to a global population.

“Right now, I would say no, it’s not applicable to the global patient population,” she said. “We really do need that next wave of Harmoni studies.”