A panel of external experts for the FDA voted 11 to 1 in support of a narrow approval for AstraZeneca and Merck’s Lynparza, used in tandem with Johnson & Johnson’s Zytiga and a corticosteroid, in metastatic castration-resistant prostate cancer (mCRPC).
In doing so, the experts endorsed the FDA’s assessment from earlier this week that the drug has yet to prove itself in a broad set of patients with mCRPC. In briefing documents ahead of Friday’s meeting, FDA staffers said the drug has only shown a favorable benefit-risk profile in mCRPC patients with BRCA mutations. And a comment from FDA’s oncology chief Rick Pazdur, M.D., during Friday’s meeting likely played a crucial part in the vote outcome.
If the agency ultimately sticks to its own analysis and limits the approval to just BRCA-mutant patients, Lynparza would only be able to target 10% to 15% of the entire mCRPC popualtion.
For this application, AstraZeneca and the FDA crossed swords on Lynparza’s performance in patients without BRCA mutations in the phase 3 PROpel trial.
In its review, the FDA focused on a subgroup of patients who were confirmed to not have BRCA mutations through two tests measuring either plasma circulating tumor DNA or tissue samples. In this subgroup of patients, who made up 54% of the PROpel trial, the Lynparza regimen reduced the risk of progression or death by 15% but was linked to a 6% increase in death risk, according to a post-hoc analysis.
Pointing to what it viewed as a “modest” tumor progression benefit but a potential detriment to patient survival, the FDA argued that an approval for the Lynparza-Zytiga combo should be restricted to BRCA-mutant patients.
But during Friday’s advisory committee discussion, AZ and its experts argued that the FDA’s double-negative standard doesn’t reflect the entire non-BRCA-mutant population in the real world, where physicians—especially those in the community setting—often rely on just one test to determine patients’ BRCA status. In some cases, tumor tissue samples are simply impossible to obtain, according to AZ’s team.
Instead, AZ focused on an aggregate non-BRCA group that included patients with an undetermined BRCA status and those with just one negative test. In this group, AZ found Lynparza led to a 24% reduction in the risk of progression or death by investigator assessment, or 28% by central review. The drug improved median progression-free survival by 5.1 months and 11 months, respectively, on those two assessments.
In addition, the drug led to a 9% reduction in the risk of death in this aggregate population, suggesting no detriment in overall survival, AZ argued. That result didn’t reach the level of statistical significance.
Countering, the FDA argued questions remained for this undetermined, middle-ground population because some BRCA-positive patients could be misclassified as negative and might have driven the benefit. AZ argued the chance of a misdiagnosis is low.
Meanwhile, deaths from COVID-19 appeared to have hurt the Lynparza arm in the trial. Removing the impact from COVID-related deaths, AZ found that even by the FDA’s stringent double-negative standard, the two trial arms showed basically the same death risk in confirmed BRCA-negative patients.
In voting for a restricted approval, members of the advisory committee agreed with the FDA that AZ hasn’t proven Lynparza works in the non-BRCA population, that uncertainty remains for the drug in these patients.
In a perhaps defining moment during Friday’s meeting, the FDA’s oncology chief Richard Pazdur, M.D., pointed out that the reality of testing and consideration of treatment for the undetermined population don’t fall in the FDA’s purview.
“That’s the practice of medicine, and we do not regulate practice of medicine,” Pazdur said.
“We’re saying that there is a potential for a detrimental effect on overall survival. We’re not saying that this has been statistically proven. We don’t have to show that—it is the responsibility of AstraZeneca to show that their drug is safe and effective,” Pazdur said.
Because the potential gap in outcomes seen between different trial subpopulations, the PROpel trial should have planned separate analysis in its statistical plan, Pazdur said. “This was not done here. So we’re guessing on what this heterogeneous middle population is.”
“When we’re doing drug regulation, we should be approving a drug in a population that, you know, it works, and not that you’re guessing it works,” he added.
Previously, the FDA granted AZ’s application priority review but then in late 2022 extended the target decision date by three months. The FDA appeared to have already missed that target.
Also on Friday, Johnson & Johnson said that its application for a combination of niraparib—sold by GSK as Zejula outside prostate cancer—and Zytiga won a priority review for mCRPC. J&J is only seeking approval for BRCA-mutant patients. The J&J combo was recently cleared for BRCA-positive cases in Europe under the brand name Akeega, whereas the Lynparza-Zytiga cocktail gained an all-comers label regardless of BRCA status.