The days of broad approvals for PD-1 inhibitors in stomach and esophageal cancers appear to be numbered.
A group of external advisers to the FDA voted 10-2 that the use of PD-1 inhibitors in first-line, HER2-negative gastric cancer is not favorable in patients who have PD-L1-negative tumors, as measured by a combined positive score (CPS) or tumor area positivity (TAP) score below 1. One additional expert abstained.
In first-line advanced esophageal squamous cell carcinoma (ESCC), the vote was 11-1, with the experts again suggesting PD-1 inhibitors should not be allowed in PD-L1-negative tumors. There was one abstention from the same person.
The votes came Thursday during the FDA’s Oncologic Drugs Advisory Committee (ODAC) meeting. Given the similar stance taken by the FDA in its own review, the agency appears poised to narrow existing approvals for Bristol Myers Squibb’s Opdivo and Merck & Co.’s Keytruda and to grant BeiGene’s Tevimbra similar limited nods.
Before Thursday’s vote, an FDA exploratory analysis of three phase 3 clinical trials for each of the above immuno-oncology meds found that evidence of a life-extension benefit, while in favor of the PD-1 inhibitors, was not convincing in patients with PD-L1-negative stomach cancer.
The numerical reductions in the risk of death in PD-L1-negative tumors that were microsatellite stable or mismatch repair proficient were 8% for Opdivo in the CheckMate-649 trial, 8% for Keytruda in KEYNOTE-859 and 2% for Tevimbra in RATIONALE-305, the FDA found.
As an FDA official noted Thursday, there was “not even a hint of a possible plateau” in the overall survival curves of the PD-1 drugs in PD-L1-negative stomach cancer patients. These curves visually record patient deaths for treatment arms in trials, and a flattening curve indicates a long-term survival benefit.
The situation was similar in esophageal cancer. The numerical reductions in the risk of death in PD-L1-negative tumors in ESCC were 0% for Keytruda in KEYNOTE-590 and 7% for Opdivo in CheckMate-648, and an increased 34% risk for Tevimbra in RATIONALE-306.
In stomach cancer, the three drug developers have argued that their phase 3 trials met the statistical significance bar in the overall trial populations and that the PD-L1-negative subgroup analyses were not statistically powered to adequately assess the meds. The drugs didn't show any detriment to survival in PD-L1-negative patients in the analyses, the companies have said.
Several patients also presented moving stories of how they benefited from PD-1 drugs during the public comment portion of the meeting.
“We recognize that the magnitude of benefit may be less in CPS less than 1,” Catherine Pietanza, M.D., Merck's vice president of late-stage oncology global clinical development, said during Thursday’s meeting. However, the overall survival trend was not unfavorable for Keytruda in those patients, and “PD-L1 is a continuum,” she said.
“Merck really wants to keep the label as it is, because it gives patients options,” Pietanza said, adding that clinical guidelines can help guide physicians.
Most members on the ODAC were not convinced, pointing to how the overall survival hazard ratios are close to 1 in the analyses, meaning there are marginal risk reductions in the PD-L1-negative group.
“I’m just not sure we want to let their doctors make this decision when these hazard ratios are almost 1, and there’s financial and toxicity impacts for these patients,” Daniel Spratt, M.D., from Case Western Reserve University, said after casting his “no” vote on the stomach cancer indication during Thursday’s meeting.
“When you look at the tails of where [the curves] converge, there’s less than a 1% absolute difference in this less-than-one subgroup,” Spratt added. “ ... That means you’re treating hundreds of these patients to benefit one.”
Hanna Sanoff, M.D., from the University of North Carolina at Chapel Hill, echoed that comment.
“Even though I cannot even tell you how moving it is to hear everyone come up and speak, the folks we don’t have at the microphone are the folks who have passed away from getting PD-1 inhibitors,” Sanoff said. “These are not just grade 3 and 4 toxicities, these are also grade 5 fatal toxicities, and that is very moving to me when you look at these curves that do not show long-term survivors from these drugs in a PD-L1 negative population.”
During Thursday’s gathering, experts invited by the companies also laid out real-world challenges regarding diagnostics if the FDA were to limit the drugs based on PD-L1 levels in stomach cancer. The discussion among the advisory committee members was heavily focused on that issue.
Robert Anders, M.D., Ph.D., a pathologist at the Johns Hopkins University and a paid consultant for BMS, noted that PD-L1 expression scores may change over time, which could make “the difference between a patient receiving first-line I-O therapy or not.” Besides, pointing to two research papers in stomach cancer, Anders argued that CPS is a “highly subjective” test that can yield different results from the same sample.
Based on these challenges, BMS and the other companies argued that limiting the PD-1 drugs based on patients' PD-L1 levels would risk excluding some patients from the meds’ labels who could benefit.
Thursday’s vote wasn’t catastrophic for the drug developers, given that clinical guidelines already didn’t place strong recommendations for Keytruda or Opdivo in stomach cancer patients with very low expressions of PD-L1.
“I think it’s telling that multiple guidelines—NCCN, ASCO and others—have all actually chosen a cutoff despite a broad indication right now,” said Jeff Meyerhardt, M.D., from Harvard Medical School.
As for the esophageal cancer discussion, the ODAC members spent some time on how difficult it was to interpret the PD-L1-negative analyses because they only included a very small group of patients.
But as Northwestern University’s Bill Gradishar, M.D., summarized, “despite the concerns about small numbers, there was no evidence of benefit, period.”
Editor's Note: The story was updated at 4:40 pm Thursday to include additional information on the esophageal cancer discussion and vote.