When Roche touted data for Polivy in previously untreated large B-cell lymphoma, the company billed the drug as the first potential new treatment standard in nearly two decades. Now, though, the FDA appears to be throwing cold water on the company's label expansion application.
In briefing documents (PDF) ahead of an expert meeting, the FDA questioned whether Polivy, used in a five-drug combination, can even be approved in front-line LBCL, let alone change decades of established clinical practice.
Among the topics up for discussion at the Thursday hearing, the agency spotlighted a “modest” tumor progression benefit and a lack of a patient survival showing for the drug.
Based on data from the phase 3 POLARIX trial, the FDA wants members of its Oncologic Drugs Advisory Committee to assess whether the Polivy regimen can be approved in its current form, or whether additional follow-up is needed.
The current standard of care for front-line LBCL is the well-established R-CHOP regimen. Roche is proposing to substitute Polivy for the Oncovin (O) component. Given the lack of competition in the field, Jefferies analysts previously estimated Polivy could reach $2.1 billion in peak sales in front-line LBCL alone.
But upon seeing Roche’s detailed data at American Society of Hematology’s 2021 meeting, the Jefferies team toned down its excitement. Compared with R-CHOP, the Polivy-R-CHP combo reduced the risk of disease progression or death by 27%. When it came to overall survival, the two trial arms saw no significant differences.
After two years, 88.7% of patients in the Polivy arm and 88.6% in the control group remained alive, translating into a mere 6% death risk reduction in favor of the experimental therapy.
In the FDA briefing documents, agency staffers said the progression-free survival benefit was “modest,” and that it was “questionable whether this rate of difference is clinically meaningful.”
The FDA staff also highlighted Polivy’s lack of an overall survival benefit. Roche previously postponed its application—at the FDA’s request—to wait for longer follow-up data. But after a median follow-up of nearly 40 months, the death risk reduction was still low at 6%, according to an update at the American Society of Hematology annual meeting in December.
Even worse for Polivy, the risk of death in the largest patient subgroup—those with diffuse large B-cell lymphoma (DLBCL)—actually favored the R-CHOP regimen. In DLBCL patients, results across several trial endpoints were “either marginal or not indicative of a positive treatment effect,” the FDA staffers said.
The agency noted that about 60% of DLBCL cases can be cured by current standard first-line therapy, hinting at a relatively low need for additional therapies.
For its part, Roche has chosen to look at the 40% of patients who aren’t cured with initial treatment.
“Ultimately for patients, increasing cure rates in the [first-line] setting, thereby sparing more patients from experiencing disease relapse and the burdens of additional therapies, is the most impactful way to address the unmet medical need in DLBCL,” Roche said in its section of the briefing document.
The FDA suspected that about 11% of POLARIX participants who had high-grade B-cell lymphoma drove Polivy’s overall positive data. In this subset, Polivy logged a death risk reduction of 58%. For these patients, current U.S. clinical practice generally prefers more intensive regimens than R-CHOP, the FDA noted.
As for the lack of overall survival benefit, the high cure rate means this endpoint could take about 10 years to complete, and most patients will eventually die of causes unrelated to DLBCL, Roche noted. Therefore, overall survival should carry less weight in assessing Polivy’s benefit, Roche argued.
Polivy, a CD79b-targeted antibody-drug conjugate, in 2019 won an accelerated approval to treat DLBCL after at least two prior therapies. POLARIX is also meant to serve as the confirmatory study for that indication. Before the FDA, the European Commission has already approved the Polivy-R-CHP combo for previously untreated DLBCL.