FDA Panel Rejects Broad Olaparib Indication in Metastatic CRPC

In a near-unanimous vote on Friday, outside advisors to the FDA recommended a narrow indication should the agency approve olaparib (Lynparza) plus abiraterone (Zytiga) as initial treatment for metastatic castration-resistant prostate cancer (mCRPC).

In an 11-1 vote (with one abstention), the Oncologic Drugs Advisory Committee (ODAC) recommended that any approval be restricted to patients with tumors that have a BRCA mutation.

AstraZeneca, developer of the PARP inhibitor olaparib, was seeking a broad indication for the combination treatment that would include BRCA-negative patients and those without homologous recombination repair (HRR) mutations. However, in explaining their votes, panel members said they had issues with the trial design of the study — PROpel — supporting the company’s application, and that the phase III study’s results did not explicitly demonstrate a benefit for patients without a BRCA mutation.

“Does PROpel prove that patients with non-BRCA-mutated prostate cancer benefit from olaparib?” asked Christopher Lieu, MD, of the University of Colorado Cancer Center in Aurora. “And I believe the answer is: We don’t know. If the applicant, and the GU oncology community are convinced that there is evidence of meaningful benefit in the biomarker-negative cohort, I think this study would be feasible and could be completed.”

“BRCA mutations predicting PARP-inhibitor sensitivity is a hallmark cancer biomarker and underpins the whole synthetic lethality paradigm in cancer,” commented Neil Vasan, MD, PhD, of the Columbia University Medical Center in New York City, in explaining his vote. “If testing if the lack of the biomarker still predicts for drug efficacy — which is counter to this fundamental paradigm — it requires a high level of scientific rigor, and I don’t feel this high level of rigor was met in this trial, given that it did not prospectively perform BRCA testing or power the study around BRCA status.”

PROpel met its primary endpoint, showing a significant radiographic progression-free survival improvement in the study’s intention-to-treat (ITT) population with olaparib plus abiraterone, at 24.8 months versus 16.6 months with placebo plus abiraterone (HR 0.66, 95% CI 0.54-0.81, P<0.0001).

An overall survival (OS) analysis further demonstrated a trend favoring the olaparib-treated group (HR 0.81, 95% CI 0.67-1.00).

However, in presenting the FDA’s position on the question, Chana Weinstock, MD, of the FDA’s Office of Oncologic Diseases, noted that if PROpel were conducted today, the agency would consider it inappropriate.

“BRCA status should have been prospectively evaluated with efficacy results analyzed separately for biomarker-selected subgroups,” she said, calling this “a significant design flaw.”

Post hoc analyses performed by the agency showed that the BRCA-positive group, which made up just 11% of the ITT population, drove the bulk of the survival benefit in PROpel:

• BRCA-positive: HR 0.30 (95% CI 0.15-0.60)
• BRCA-uncertain: HR 0.73 (95% CI 0.52-1.03)
• BRCA-negative: HR 1.06 (95% CI 0.81-1.39)

Thus, Weinstock said, the FDA is “concerned about a potential OS detriment in patients with non-BRCA mutated tumors, based on the HR of 1.06.

Jorge Nieva, MD, of the University of Southern California’s Keck School of Medicine in Los Angeles, was the lone panel member voting that the combination should not be limited to the BRCA-positive population.

“This was a positive clinical trial designed in conjunction with FDA guidance on the endpoints,” he said. “The FDA has proposed that there should be a restriction to 11% of the patient population, and I don’t think this level of restriction is justified.”

“Patients with homologous recombination deficient [HRD] cancers gain significant benefits from PARP therapy, and this has been seen in multiple clinical trials,” he added. “A more reasonable restriction might have been to reduce the indication to HRD tumors.”

Disagreement Over the Definition of Non-BRCA

It became clear during the meeting that a major bone of contention between the FDA and AstraZeneca was the way in which BRCA-negative patients were defined in the post hoc analysis.

The FDA defined these patients as those who were BRCA-negative on tissue and blood-based circulating tumor DNA (ctDNA) testing.

However, said AstraZeneca’s Laurence Toms, MD, this meant that the FDA characterized about 35% of patients as “undetermined,” even though the vast majority of those individuals had one negative test result.

Given the hazard ratio for OS in this group (0.73), “there is a minimal risk that the effect in this group is driven by misclassified BRCA-mutant patients,” Toms said, adding that this suggests there is a clinically meaningful OS benefit derived in patients beyond those with a BRCA mutation.

“In reality, this undetermined BRCA subset, particularly those with one negative blood test, represents the vast majority of the mCRPC patients that I currently treat in my practice,” said Daniel George, MD, of the Duke Cancer Institute in Durham, North Carolina, who spoke during AstraZeneca’s presentation.

“If you restrict the PROpel approval to the BRCA-mutant subgroup, this large proportion of non-BRCA patients in the real world will be denied the option to receive this combination, and any hope of additional clinical benefit,” George added.

When asked why the FDA called those patients “undetermined,” Weinstock said that for the majority of those patients it was due to tissue testing.

“There is a lot of uncertainly and that was borne out when we looked at … patients with two negative BRCA tests based on ctDNA and tumor-tissue testing then the results, to us, looked very different,” she said. “So, there was something else going on there, and a lot of it was based on tumor tissue results that were indeterminate or missing.”

“We should be approving a drug in a population we know it works in,” emphasized Richard Pazdur, MD, of FDA’s Oncology Center of Excellence. “The whole purpose of a clinical trial is really to define a homogenous group of patients, and when you have a situation where you may have differential outcomes based on BRCA status, or other biomarkers, you would want to stratify, and also do separate analysis planned in a statistical plan. This was not done here. So, we’re guessing what this heterogeneous, middle population is.”

While the FDA is not required to follow the advice of its advisory committees, it typically does.