Unprecedented PFS in HER2-Mutant Lung Cancer, but With a Touch of Controversy

Progression-free survival (PFS) improved in advanced HER2-mutant non-small cell lung cancer (NSCLC) when treatment started with trastuzumab deruxtecan (T-DXd; Enhertu) instead of pembrolizumab (Keytruda) and chemotherapy, a large randomized trial showed.

Median PFS improved from 8.3 months with pembrolizumab and chemotherapy to 14.3 months with T-DXd, representing a 37% reduction in the hazard for progression or death. However, an unfavorable survival trend (HR 1.15) in a preliminary analysis raised questions about the eventual role of T-DXd as first-line therapy for HER2-mutant NSCLC with exon 19/20 mutations.

More than twice as many patients in the pembrolizumab arm subsequently received HER2-directed therapies at progression, which might have confounded interpretation of the overall survival (OS) data, said Julia Rotow, MD, of the Dana-Farber Cancer Institute in Boston, at the World Conference on Lung Cancer in Seoul, South Korea.

“There was overlap of the survival curves until approximately 24 months, at which point there was late divergence,” said Rotow. “Changes in HER2-directed standard of care occurred during the course of the study and may have impacted overall survival findings. I will highlight that the median overall survival of 29 to 33 months [in the two treatment arms] is strikingly better than that historically reported for this patient population, reflecting advances in the overall treatment options available for patients with HER2-mutant lung cancer and overall progress in the field.”

“These data establish DESTINY-Lung04 as the first phase III trial to show benefit with a HER2-directed therapy versus standard-of-care treatment, establishing T-DXd as a new first-line treatment option for patients with HER2-mutant non-small cell lung cancer,” she added.

Yes, but …

The 14.3-month median PFS is unprecedented, and the 70% response rate with T-DXd is also impressive, said invited discussant James Chih-Hsin Yang, MD, of the National Taiwan University Hospital in Taipei. The results clearly establish T-DXd as a first-line choice.

Even so, “I think this overall survival hazard ratio of 1.15 is disturbing,” noted Yang.

The KEYNOTE-189 trial evaluating pembrolizumab plus chemotherapy versus chemotherapy and placebo offers the best reference for the DESTINY results, he added. Looking at landmark OS analyses, the pembrolizumab-chemotherapy control arm of DESTINY-Lung04 far outperformed the same combination (then the experimental arm) in KEYNOTE-189 at 1 year (83% vs 69.2%).

“If you look at 2, 3, 4, 5 years and even the median survival of 33.1 months, the control arm is much better than the previous 19 or 20 months median overall survival in KEYNOTE-189,” said Yang, who acknowledged that DESTINY-Lung04 was limited to patients with HER2 exon 19 or 20 mutations, whereas KEYNOTE-189 enrolled all comers. “Something must have happened during that 5 years. One simple answer is that we have a lot of effective drugs that can be provided to the patient.”

Whether the OS difference should be a real concern awaits further follow-up, as the data remain immature. Additionally, the impact of crossover remains to be determined.

“There was a lot of crossover, but the question is whether the crossover is balanced between the two arms,” said Yang. “If that crossover was balanced, then we can say that the overall survival hazard of 1.15 can be trusted, but it is still too early to know. Second, there probably is a huge imbalance.”

Survival data for DESTINY-Lung04 and other ongoing studies are “in transition, meaning that we probably don’t need to remember the point-estimate numbers. These point estimates will change in the future,” he added.

Background, Key Findings

HER2 mutations occur in 2-4% of nonsquamous NSCLC, leading to aggressive disease associated with poor prognosis, Rotow said in her introduction to the study. Immunotherapy plus platinum-based chemotherapy remains the worldwide first-line standard, but response rates remain low and disease progression frequent, supporting the need for HER2-directed treatment options.

T-DXd already has a second-line indication for HER2-mutant NSCLC. DESTINY-Lung04 investigated the performance of T-DXd in the first line. The trial included patients with unresectable or metastatic nonsquamous NSCLC associated with HER2 exon 19 or 20 mutations, and no prior treatment for advanced disease.

Investigators in North America, Europe, and Asia randomized 454 patients to single-agent T-DXd or to pembrolizumab plus platinum-pemetrexed chemotherapy. The primary endpoint was PFS by blinded independent committee review. Secondary endpoints included OS, objective response rate (ORR), duration of response (DOR), and investigator-assessed PFS2.

After a median follow-up of about 21 months, an interim analysis showed the trial had met the primary endpoint, showing that T-DXd reduced the PFS hazard by 37% versus the control arm (95% CI 0.50-0.79, P<0.0001). The PFS difference reached 6 months after about 15 months of follow-up, and the survival curves began to diverge thereafter. A subgroup analysis showed a consistent benefit with T-DXd, with no prespecified subgroup favoring the control arm.

The ORR was 70% with T-DXd and 44.5% for the control group (OR 2.93, 95% CI 2.00-4.34). Median DOR also favored T-DXd (13.4 vs 9.7 months), as did PFS2 (22.7 vs 17.3 months).

A virtually identical proportion of patients in both arms received subsequent therapy (71.5% in the T-DXd arm vs 72.4% in the control arm), but substantially more patients in the control group received HER2-directed therapy at progression (48% vs 23.3%). Substantially more patients in the T-DXd arm received immunotherapy at progression (28.6% vs 5.3%).

Grade ≥3 adverse events (AEs) occurred in 34% of patients in each arm. The AE-related discontinuation rate was 15.9% in each group. Four treatment-related deaths occurred in the T-DXd group versus one in the control arm.

Putting the results into perspective, Yang noted that HER2-mutant NSCLC accounts for a “small, tiny” patient population. The number of HER2 alterations is far greater than originally believed.

“We previously thought most of these patients had tyrosine kinase domain mutations,” he said. “When we applied NGS [next-generation sequencing], we found there are more mutations in transmembrane domains. We don’t know whether they are activating or non-activating. We don’t know their efficacy toward small molecules. We don’t know anything about their efficacies after treatment with ADCs [antibody-drug conjugates]. There is still a lot of room to go into for our non-small cell lung cancer patients.”