
NEW YORK –An updated analysis from the Phase III KEYNOTE-811 trial has given oncologists further reason to limit a triplet regimen, comprising Merck's checkpoint inhibitor Keytruda (pembrolizumab), Genentech's HER2-targeted drug Herceptin (trastuzumab), and chemotherapy, to only the subset of HER2-positive gastric cancer patients whose tumors express PD-L1.
During a presentation at the European Society for Medical Oncology (ESMO) Congress on Friday, Yelena Janjigian, a medical oncologist at Memorial Sloan Kettering Cancer Center, shared the most recent progression-free survival data from the randomized trial, which was simultaneously published in the Lancet.
In KEYNOTE-811, investigators enrolled 698 patients with advanced, newly diagnosed HER2-positive gastric or gastroesophageal cancer and gave them either Keytruda-Herceptin-chemo or Herceptin-chemo. Patients weren't required to have PD-L1-expressing tumors to enroll in the study, but they had to provide a tumor sample for biomarker analysis. All patients had to have HER2-positive cancer as determined by central review. HER2 positivity was defined in the trial as an immunohistochemistry 3+ score or a IHC 2+ score with positive in situ hybridization or fluorescence in situ hybridization results.
After a median follow-up of 38.5 months, the median progression-free survival in the overall population, regardless of PD-L1 expression, was 10 months in the Keytruda-Herceptin-chemo arm, versus 8.1 months in the Herceptin-chemo arm.
Among patients whose tumors expressed PD-L1 with a combined positive score of at least one, the median progression-free survival was 10.9 months with the triplet versus 7.3 months in the control arm. "The addition of pembrolizumab to first-line trastuzumab and chemotherapy led to meaningful improvement in progression-free survival … particularly in patients with dual over-expression of HER2 and PD-L1 in their tumors," Janjigian said.
Janjigian also shared interim overall survival data but emphasized that the findings aren't mature. As of the data cutoff, the median overall survival in allcomers in the Keytruda-Herceptin-chemo arm was 20 months, versus 16.8 months in the control arm. In the PD-L1-expressing subset, the median overall survival was also 20 months in the Keytruda-containing arm, versus 15.7 months in the control arm.
In terms of toxicities, Janjigian said "there were no new surprises," although the side effects observed in the Keytruda arm were slightly higher because of immune-related toxicities.
Biomarker significance
In a discussion following Jangijian's presentation, Florian Lordick, a medical oncologist at the University Cancer Leipzig in Germany, focused on the important, albeit numerically small group of patients enrolled in the KEYNOTE-811 trial whose cancers did not express PD-L1. Only 15 percent of patients in both treatment arms fell into this PD-L1-negative category, marked by tumors expressing PD-L1 with a combined positive score that was lower than one.
"It's important to note that a biomarker is available that at least gives us a direction of who benefits from this treatment and who does not benefit," he said. "In patients where the PD-L1 [combined positive score] is less than one, we cannot expect any benefit from adding pembrolizumab to trastuzumab and chemotherapy."
In the PD-L1-negative subgroup, there was no progression-free survival benefit from adding Keytruda to Herceptin and chemo. The median progression-free survival after 38.5 months in this subgroup was 9.5 months in the Keytruda-containing and control arms.
"Many people would say that PD-L1 is not an optimal biomarker, and many would agree, but still, it's a clinically meaningful biomarker that we should use in this setting," Lordick said, underscoring to his oncologist colleagues at the meeting: "Don't try to give pembrolizumab to patients who are PD-L1 negative and HER2-positive."
Data from studies across numerous cancer types has shown that the higher patients' tumor PD-L1 expression is, the better they do on immune checkpoint inhibitors. As an audience member at the meeting pointed out following Janjigian's presentation, oncologists can't rule out the possibility that the benefit seen in this trial on Keytruda-Herceptin-chemo isn't largely attributable to the patients with particularly high PD-L1-expressing tumors. Janjigian noted that she and her team will take a closer look at how different levels of PD-L1 expression in KEYNOTE-811 impacted patients' outcomes.
Regulatory discrepancies
In the meantime, European regulators recently approved the Keytruda-Herceptin-chemotherapy triplet for advanced HER2-positive gastric cancer patients whose tumors express PD-L1 with a combined positive score of one or more. As Lordick announced in his discussion of the data, the ESMO clinical practice guidelines have been updated to recommend this biomarker approach, too, for this treatment context.
Across the Atlantic, where regulators in the US are often not aligned with their European colleagues when it comes to requiring biomarker testing, the US Food and Drug Administration in 2021 approved this same regimen for advanced HER2-positive gastric cancer patients regardless of their PD-L1 expression status. The FDA based this accelerated approval on overall response rates at the time, and Merck is required to provide confirmatory data on the Keytruda-containing cocktail's benefit if it wishes to achieve full approval in this setting.
In a rare move this past June, based on the updated KEYNOTE-811 results, Merck said it had asked the FDA to change the approved indication for the triplet so that it is available as first-line option for advanced gastric cancer patients who have both PD-L1-expressing and HER2-positive tumors.
As of October 2023, Keytruda's label still allows for PD-L1-agnostic prescribing, though Merck reiterated in a statement on Friday that it "has discussed these findings [from KEYNOTE-811] with the FDA and is working with them to update the current indication for Keytruda."