Gilead’s Trodelvy-Keytruda lung cancer combination misses Phase III endpoints despite higher response rate

Gilead Sciences, Inc. (Nasdaq: GILD) and Merck & Co., Inc. (NYSE: MRK) have received a clearer explanation for the failure of their Phase III EVOKE-03/KEYNOTE-D46 study after full data presented at the World Conference on Lung Cancer showed that Trodelvy plus Keytruda did not significantly improve progression-free or overall survival in first-line PD-L1-high metastatic non-small cell lung cancer. Median progression-free survival was 11.8 months with sacituzumab govitecan plus pembrolizumab compared with 7.7 months for pembrolizumab alone, but the statistical threshold required by the study was not met. Median overall survival at the interim analysis was 21.5 months for the combination and 22.8 months for pembrolizumab alone. The result is commercially important because Gilead continues to grow Trodelvy, which generated $457 million of second-quarter 2026 sales, but the drug now has one less clear route into the large first-line lung cancer market.

The failure itself was not entirely new. Gilead Sciences and Merck & Co. announced in June that the trial would be discontinued following a data-monitoring review showing that the progression-free survival endpoint had not reached statistical significance and that the probability of achieving a significant overall survival result was unlikely. The WCLC presentation adds the numerical detail needed to understand why the programme failed despite several signs of antitumour activity.

How could Trodelvy plus Keytruda improve response rates but still fail the Phase III trial?

The combination produced an objective response rate of 55.6%, compared with 43.7% for pembrolizumab alone, and disease control rates of 83.3% versus 73.1%. Those figures show that adding Trodelvy increased tumour response activity. Median progression-free survival also appeared numerically longer by 4.1 months.

Clinical trials, however, are designed around prespecified statistical thresholds rather than whether one arm produces a larger raw number. The progression-free survival hazard ratio of 0.81 had a reported P value of 0.0252 but failed the threshold required under the study’s statistical design. Overall survival was even less supportive, with a hazard ratio of 1.07 and median survival numerically favouring pembrolizumab monotherapy.

The outcome illustrates why oncology development cannot rely on response rate alone. A drug may shrink more tumours without creating a sufficiently durable improvement in disease progression or survival, especially if added toxicity, treatment discontinuation or underlying disease biology offsets some early benefit. For commercial forecasting, a positive response signal cannot be treated as equivalent to a registrational success.

What does EVOKE-03 mean for Gilead Sciences’ growing Trodelvy franchise?

Trodelvy generated $457 million in second-quarter 2026 sales, up 26% from the same period a year earlier. That makes it a meaningful growth product within Gilead Sciences, although HIV medicines remain far larger, with second-quarter HIV revenue of $5.7 billion.

Expanding Trodelvy into new tumours and earlier treatment lines is important because each successful indication enlarges the addressable population and increases the return from the antibody-drug conjugate platform Gilead obtained through its acquisition of Immunomedics. EVOKE-03 offered access to patients receiving first-line therapy for PD-L1-high metastatic NSCLC, a substantial market already served by Keytruda monotherapy.

The trial failure closes off that particular route unless future biomarker work identifies a subgroup in which the combination has a more favourable benefit-risk profile. It does not erase Trodelvy’s established approvals or revenue, nor does it determine the outcome of unrelated studies. Investors should therefore view EVOKE-03 as an indication-expansion setback rather than evidence that the existing Trodelvy franchise itself has failed.

Why was toxicity especially relevant when pembrolizumab alone already works in these patients?

Treatment-related adverse events occurred in 93.2% of patients receiving the Trodelvy-Keytruda combination compared with 65.4% of those receiving Keytruda alone. Grade 3 or higher treatment-related adverse events occurred in 55.7% of combination patients versus 16.5% with monotherapy. Anaemia, neutropenia, diarrhoea, nausea and alopecia were among the events reported more commonly with the combination.

Higher toxicity can be acceptable when a treatment provides sufficiently large survival gains, particularly in aggressive cancers. EVOKE-03 did not demonstrate that offset. Pembrolizumab alone delivered similar or better interim overall survival with considerably fewer severe treatment-related events, making the combination difficult to justify for the overall population studied.

This is also a reminder of the challenge facing antibody-drug conjugates as companies attempt to move them earlier in treatment. Later-line patients with few alternatives may accept greater toxicity for improved tumour control, while first-line populations often have effective therapies already available. The bar for adding a cytotoxic ADC to immunotherapy is consequently higher when the comparator is an established drug that can be given alone.

Does the failure weaken Merck’s Keytruda position or reinforce it?

For Merck & Co., EVOKE-03 is more supportive of Keytruda monotherapy than damaging to it. The study attempted to improve upon pembrolizumab by adding another active oncology drug, yet failed to produce statistically significant progression-free or overall survival advantages. Keytruda therefore remains a difficult benchmark in PD-L1-high first-line NSCLC.

The commercial context is substantial. Merck reported $8.366 billion in second-quarter 2026 sales from Keytruda and Keytruda Qlex, up 5% year over year. The franchise remains one of the largest medicines globally and a central contributor to Merck’s earnings.

The longer-term competitive challenge is coming from mechanisms that may genuinely improve upon PD-1 monotherapy rather than merely add toxicity. Akeso’s HARMONi-2 data showing an overall survival advantage for the PD-1/VEGF bispecific ivonescimab against pembrolizumab are therefore particularly noteworthy when viewed alongside EVOKE-03. One combination strategy failed to clear Keytruda’s bar, while a different therapeutic architecture has now done so in a China Phase III study.

Key takeaways from the EVOKE-03 Trodelvy-Keytruda Phase III results

• EVOKE-03 enrolled 620 previously untreated patients with PD-L1-high metastatic NSCLC.
• Median PFS was 11.8 months with Trodelvy plus Keytruda versus 7.7 months with Keytruda.
• The PFS result did not meet the study’s prespecified statistical threshold.
• Interim median overall survival was 21.5 months versus 22.8 months.
• Objective response was higher with the combination at 55.6% versus 43.7%.
• Grade 3 or higher treatment-related adverse events were 55.7% versus 16.5%.
• Gilead and Merck had already discontinued the study after the June monitoring review.
• Trodelvy generated $457 million of second-quarter 2026 sales, up 26%.
• The failure limits one expansion opportunity but does not remove Trodelvy’s approved franchise.
• Keytruda remains a formidable first-line benchmark in PD-L1-high NSCLC.