Enhertu Nearly Doubles Progression-Free Survival in HER2-Mutant Lung Cancer Trial, But Survival Data Trends the Wrong Way

A large international trial found that starting HER2-mutant lung cancer patients on trastuzumab deruxtecan, sold as Enhertu, delayed disease progression far longer than the current standard combo of pembrolizumab and chemotherapy. But a preliminary survival readout complicated the celebration, and the researchers who ran the study are openly debating what it means.

The phase 3 DESTINY-Lung04 trial results were presented at the International Association for the Study of Lung Cancer's 2026 World Conference on Lung Cancer, held September 12-15, 2026, in Seoul, South Korea, by Julia Rotow, MD, clinical director of the Lowe Center for Thoracic Oncology at Dana-Farber Cancer Institute, according to Dana-Farber's own release on the trial, which notes her presentation took place on Sept. 14, 2026.

The numbers that matter

The trial enrolled 454 treatment-naive patients with unresectable, locally advanced, or metastatic nonsquamous non-small cell lung cancer carrying HER2 exon 19 or exon 20 mutations, according to Oncozine. Patients were randomly split, 227 to an arm, between single-agent T-DXd every three weeks and pembrolizumab plus platinum chemotherapy and pemetrexed, the current standard of care.

Median progression-free survival, assessed by blinded independent central review, was 14.3 months with T-DXd versus 8.3 months with the standard combo, a hazard ratio of 0.63 with a p-value under .0001, according to Medscape and multiple other outlets covering the presentation. That's a 37% reduction in the risk of progression or death.

The objective response rate was 70.0% with T-DXd versus 44.5% with standard treatment, an odds ratio of 2.93, according to cancertherapyadvisor. Median duration of response ran 13.4 months versus 9.7 months.

Rotow called the benefit consistent across nearly all prespecified subgroups, including patients with brain or liver metastases, those over 65, and smokers, per Medscape's reporting.

The overall survival problem

At a data cutoff of June 9, 2026, with survival data only 46.9% mature, median overall survival ran 29.3 months with T-DXd versus 33.1 months with the standard regimen, a hazard ratio of 1.15 with a 95% confidence interval of 0.88 to 1.52, according to AllSci and cancertherapyadvisor. That interval crosses 1.0, meaning the trend isn't statistically significant, but it points the wrong direction for the drug that won on every other measure.

Rotow flagged a likely explanation: more than twice as many patients in the pembrolizumab arm went on to receive HER2-directed therapies after their disease progressed, 48.0% compared with 23.3% in the T-DXd arm, according to cancertherapyadvisor. Her argument is that patients on the standard regimen who progressed were able to access effective HER2-targeted drugs afterward, potentially propping up their survival numbers, while T-DXd patients who progressed had fewer options left.

Whether this fully explains a survival gap this size is not something the current data can prove. AstraZeneca and Daiichi Sankyo, the drug's co-developers, say formal statistical testing of overall survival is planned for later analyses once more data matures, according to AllSci.

A pointed rebuttal from the trial's own discussant

James Chih-Hsin Yang, MD, of National Taiwan University Hospital, served as the invited discussant on the trial and called the 1.15 survival hazard ratio "disturbing," according to MedPage Today.

Yang compared the pembrolizumab-chemotherapy control arm's performance against KEYNOTE-189, an earlier trial of the same regimen in broader lung cancer populations. The control arm here hit 83% survival at one year versus 69.2% in KEYNOTE-189, and a median overall survival of 33.1 months versus roughly 19 to 20 months previously, per MedPage Today's account of Yang's remarks. He acknowledged the comparison isn't perfectly apples-to-apples, since KEYNOTE-189 enrolled all comers rather than only HER2-mutant patients, but said the gap was striking enough to ask what changed.

Rotow's own explanation leaned on the same idea from a different angle: broader access to effective HER2-directed drugs across the field over the past several years has pushed survival up for everyone with this mutation, control arm included, according to MedPage Today.

Safety and where this leaves patients

Grade 3 or higher treatment-related adverse events were similar between arms, 34.1% with T-DXd versus 33.6% with standard therapy, according to AllSci. But interstitial lung disease or pneumonitis showed up in 20.8% of T-DXd patients versus a much lower rate on chemoimmunotherapy, and four of those cases were fatal, a 1.8% rate of deadly lung inflammation tied directly to the drug, per AllSci and cancertherapyadvisor.

T-DXd, jointly developed by AstraZeneca and Daiichi Sankyo, is already FDA-approved in the U.S. for previously treated HER2-mutant NSCLC under accelerated approval. Whether DESTINY-Lung04 supports expanding that label to newly diagnosed patients is now a question for regulators, not conference attendees.

The field isn't short on competition, either. Boehringer Ingelheim's zongertinib, sold as Hernexeos, picked up FDA accelerated approval in February 2026 for treatment-naive HER2-mutant NSCLC patients and is already recommended by the National Comprehensive Cancer Network as initial therapy, according to AllSci. Sevabertinib is running its own first-line trial, SOHO-02, after gaining accelerated approval for previously treated patients in November 2025.

The unresolved question is straightforward: does T-DXd actually help patients live longer, or does it just delay progression while a shortage of follow-up options works against it? AstraZeneca and Daiichi Sankyo say the next survival analysis, once the data matures further, should start answering that.