The FDA’s Oncologic Drugs Advisory Committee on Thursday voted 10-2 with one abstention that Merck’s Keytruda and Bristol Myers Squibb’s Opdivo are not beneficial as a first-line treatment for stomach cancer patients who test negative for the biomarker known as PD-L1.
The vote followed consensus from the committee members, FDA reviewers and sponsors that there was a higher magnitude of benefit with immune checkpoint inhibitors in this indication of HER2-negative microsatellite stable gastric/gastroesophageal junction cancer when there’s higher PD-L1 expression.
Panelist Jeffrey Meyerhardt, chief clinical research officer at Dana-Farber Cancer Institute, voted against the use of these immunotherapies in patients with low or no PD-L1 expression. He said it’s important to note that national guidelines already advise doctors not to use these therapies on patients with very low or no PD-L1 expression.
The morning vote on the risk-benefit ratio may trigger a label change in gastric cancer from the FDA for Keytruda and Opdivo for this low PD-L1 population. It might also lead to a more limited label for BeiGene’s tislelizumab, if it’s approved in this setting. Currently, Opdivo and Keytruda are approved as first-line stomach cancer treatments, regardless of PD-L1 status.
Nearly all of those speaking, including the sponsors, said that PD-L1 expression is a predictive biomarker for patient selection in this patient population. But part of the problem is that all of the companies used different PD-L1 tests that were difficult to compare across trials. In the real world, not everyone is tested for PD-L1 expression before treatment.
BMS VP Ian Waxman explained to the panel that physicians understand the importance of testing, and less than 5% of those who received Opdivo in this indication were PD-L1 negative. Merck VP Catherine Pietanza similarly acknowledged that PD-L1 testing in the real world varies, but that data from post-hoc subgroup analyses on PD-L1 expression should not supersede the findings of its positive Phase 3 trial.
The panel generally agreed, however, that the data are adequate to revise the blockbusters’ labels.
The FDA’s oncology chief Rick Pazdur weighed in at the end of this morning’s meeting, calling on the companies to do a better job collaborating on PD-L1 tests and make them uniform.
Merck, Bristol Myers and BeiGene offered their support, as Pazdur added, “This has not been a great experience having all these different tests here.”
Esophageal Cancer
In the afternoon, ODAC voted 11 to 1 with one abstention against whether the risk-benefit assessment is favorable for the use of anti-PD-1 antibodies in first line unresectable or metastatic esophageal squamous cell carcinoma with low or no PD-L1 expression.
While the panelists were not voting on label changes in either vote, many expressed their understanding that their “no” votes would likely lead to label changes, and were made based on the best available data.
“This dataset is quite small but it’s a minority of patients who will be treated,” Christopher Lieu, professor of medicine at the University of Colorado, said in explaining his “no” vote on ESCC.
While ESCC is similar to stomach cancer in terms of higher PD-L1 expression correlating with more overall survival benefit, several ODAC panelists and both Merck and Bristol Myers cautioned that in ESCC many of these patients are extremely sick and treatment has to proceed quickly.
In an earlier discussion prior to the vote, panelists also raised concerns about the sample size of data in this indication, and several panelists debated whether a “no” vote could mean that the approximately 10% of patients with ESCC who don’t express PD-L1 or express small amounts may be holding back treatment with immune checkpoint inhibitors for the nearly 90% of patients with ESCC who do express higher levels of PD-L1 because more testing would have to occur.
Editor’s note: Article updated with information on the afternoon ODAC vote.