A large double-blind, placebo-controlled study found a clinically meaningful benefit from combining the anti-programmed cell death protein 1 (PD-1) immune checkpoint inhibitor pembrolizumab with standard platinum-containing chemotherapy as initial therapy for patients with advanced HER2-negative gastric or gastroesophageal junction cancer.
Data from the Phase III KEYNOTE-859 study reported at an ESMO Virtual Plenary session found treatment with pembrolizumab plus a fluoropyrimidine/platinum chemotherapy regimen brought modest, statistically significant improvements in overall survival (OS), progression-free survival (PFS), and objective response rates (ORR) among patients with locally advanced or metastatic, HER2-negative gastric or gastroesophageal junction adenocarcinoma.
In the overall population, the study found around a month and a half extension of life and there also was a similar extension of PFS. Preliminary analysis noted that treatment efficacy improved in proportion to tumor expression levels of programmed death-ligand 1 (PD-L1), the molecular marker associated with PD-1 targeted by pembrolizumab. Around a quarter of patients whose tumor PD-L1 combined positive scores (CPS) were less than 1 had only marginal benefit from the addition of pembrolizumab.
Study leader Sun Young Rha, MD, PhD, Professor of Medical Oncology and Director of Songdang Institute for Cancer Research in the Yonsei University College of Medicine at Yonsei University Health System in Seoul, Korea, told Oncology Times the benefit of adding this immunotherapy had been clinically meaningful, with no new safety signals. In addition, the data supported pembrolizumab plus chemotherapy as a new initial treatment option for this population.
Study Details
Out of 1,579 patients in total, half were randomized to pembrolizumab plus chemotherapy and the other half to placebo with chemotherapy. After a median study follow-up of 31 months, OS was 12.9 months in those treated with the combination of immunotherapy plus chemotherapy compared and 11.5 months among patients receiving placebo with their chemotherapy (HR: 0.78, p<0.0001).
The median PFS was 6.9 months with immunochemotherapy compared with 5.6 months for the control group (HR: 0.76). There was also superiority in median duration of response to treatment, which increased from 5.7 months in the control group to 8.0 months in patients who had both chemotherapy and immunotherapy. ORR was 51.3 percent in patients receiving pembrolizumab compared with 42.0 percent for placebo.
There were more Grade 3-5 treatment-related adverse events (AEs) among patients receiving immunotherapy: 59.4 percent compared with 51.1 percent in the control group. But only 1 percent of patients in the investigational arm died from treatment-related AEs compared with 2 percent of patients in the control arm.
Rha noted that PD-1 checkpoint inhibition with nivolumab, sintilimab, and pembrolizumab had been investigated in Phase III trials among patients with human epidermal growth factor receptor 2 (HER2)-positive and HER2-negative tumors. Regulatory approvals existed for the addition of nivolumab with chemotherapy for patients with HER2-negative tumors, and also for chemotherapy with pembrolizumab plus the anti-HER2 agent trastuzumab for tumors testing positive for HER2. The new study was conducted because 5-year survival rates among patients with advanced gastric cancer remained below 10 percent despite current preferred chemotherapy protocols with fluoropyrimidine/platinum regimens.
The main KEYNOTE-859 study looked at benefits for all subgroups as a whole, irrespective of tumor PD-L1 CPS, Rha noted. But she said more detailed analysis of the relationship between outcomes with PD-L1 expression was yet to be completed. She mentioned that the small number of patients with high microsatellite instability scores benefited greatly from anti-PD-1 immunotherapy (HR: 0.34).
When she was asked whether an HR of 0.78 in the overall population was enough to inform clinical practice for the majority of patients with advanced gastric tumors, Rha acknowledged there was still work to be done. "We have subset analysis with different CPS levels, and we can see how we can define better subgroups [for] improving survival," she noted.
Examining the Findings
In a comment about the KEYNOTE-859 study ESMO Plenary panelist Elizabeth Smyth, MD, a consultant in gastrointestinal oncology at Cambridge University Hospitals NHS Foundation Trust in the U.K., explained that higher PD-L1 CPS levels were associated with better outcomes from anti-PD-1 immunotherapy.
"We know that the association between PD-L1 and benefit is kind of linear-almost. The higher the better. From my perspective, PD-L1-negative patients do not appear to have any consistent benefit. So I would be quite happy not to treat those patients with an immune checkpoint inhibitor."
She noted that patients classified as "PDL-high" (having a CPS of more than 10) should definitely be treated with an anti-PD-1. But for patients who had intermediate CPS-around four or five, for example, she was less sure. PD-L1 testing had limitations caused by a range of factors, including tumor heterogeneity that could bring inconsistencies.
The ESMO Virtual Plenary Chair, Florian Lordick, MD, PhD, Director of the Leipzig University Medical Center in Germany, said KEYNOTE-859 was a positive study with outcomes that differed depending on tumor PD-L1 expression levels.
"However, the groups were stratified according to PD-L1 expression," he stated. "And we can see that there is a differential effectiveness of pembrolizumab according to the PD-L1 status. Those patients presenting with PD-L1 CPS of one or greater-78 percent of the population-had a benefit with a hazard ratio of 0.72. And those with a CPS of 10 or greater (35% of the population) had a hazard ratio of 0.64. So this means [that] the higher the PD-L1 expression was, the better: the more the benefit of pembrolizumab."
Lordick regarded the hazard ratio of 0.78 for the overall study population as statistically significant, but not amazing." He said they were all seeking the subgroup of patients that benefited more than that. And he mentioned the subpopulation of patients in the study with no PD-L1 expression who did not derive a major benefit. "The hazard ratio was 0.92, which is close to ineffectiveness," he stated.
When asked about the future of pembrolizumab combination therapy in gastric and GE-junction cancer, he told Oncology Times he thought it would be approved, but he recommended using the immunotherapy according to PD-L1 expression data from each patient.
"My personal clinical opinion is that for a PD-L1 with a CPS of 10 or greater there will be a clear indication to add an immune checkpoint inhibitor-either pembrolizumab or nivolumab," he noted. "For negative PD-L1 expression, I would not use immune checkpoint inhibition."
Lordick described combining pembrolizumab plus chemotherapy for patients with intermediate tumor PD-L1 expression levels (CPS from 1-9-approximately 40% of patients), as a "grey zone where we really need to weigh the potential benefits against the potential toxicities [and] potential costs."
Peter M. Goodwin is a contributing writer.