FDA scrutiny now hits Parp inhibitors

Glaxosmithkline’s Zejula, already trailing the leading Parp inhibitor, Astrazeneca/Merck & Co’s Lynparza, could soon have its wings clipped further. A US advisory committee will meet on November 22 to discuss whether Zejula’s use in ovarian cancer maintenance should be curtailed.

The basis for this is the Nova trial, whose progression-free survival benefit backed Zejula’s US approval in this setting, but whose overall survival data ostensibly showed Zejula patients doing worse than those on placebo. Though confounding factors are at play, the bottom line is that Lynparza does boast an OS benefit in ovarian cancer maintenance.

Interestingly, the deleterious OS data from Zejula’s Nova trial were first revealed in early 2021, but it is only recently that the FDA has started cracking down on oncology drugs that gain accelerated approval but then fail to show a survival benefit. And the move comes amid tightened scrutiny over Parp inhibitors, whose makers have quietly abandoned late-line ovarian cancer.

What’s the issue?

At issue at the November adcom will be the relevance of Nova’s OS result, a secondary endpoint whose updated numbers GSK says it recently shared with the FDA.

Back at the March 2021 virtual meeting of the Society of Gynecological Oncology it was reported that germline Brca-positive patients in Nova had a 7% reduction in risk of death on Zejula, but in those without gBrca mutations the median OS was 31.1 months for the GSK drug, versus 36.5 months for placebo (hazard ratio of 1.10).

Nova assigned patients to gBrca positive and negative cohorts separately, but the FDA gave Zejula an all-comers label. One option for the adcom, therefore, might be to restrict Zejula’s maintenance approval to gBrca-mutant disease, rather than withdrawing it entirely. Lynparza boasts an all-comers label in this setting, and this is backed by OS benefits in the secondary endpoints of two studies.

However, all is not so simple. Nova spanned a relatively long timeframe, during which Parp inhibitors gained approval in numerous settings, and as a result it suffered from extensive patient withdrawals.

Though the trial did not allow crossover, patients who progressed could leave Nova and receive a rival Parp. Overall, missing data claimed 25% of the Nova population, the SGO meeting was told.

A key question for the panel, therefore, will be what to make of this, and whether any adjustments can be applied to glean the realistic OS benefit such patients might derive on Zejula. A further unknown is whether the OS numbers have deteriorated further, and GSK has not made these data public.

A third Parp inhibitor, Clovis’s Rubraca, is also available for ovarian cancer maintenance in all-comers, based solely on a PFS benefit.

Zejula (and Lynparza) additionally carries a first-line maintenance indication – also on nothing beyond PFS – and this is not under threat from the adcom. However, here Zejula boasts an all-comers label, while Lynparza is limited to Brca-positives, or HRD-positive patients if combined with Avastin.

A separate issue is late-line use in heavily treated ovarian cancer, in which all three Parps had received full approvals on the basis of remission rates in single-arm studies.

This quietly started unravelling in June, when Clovis pulled Rubraca after reporting a numerical 31% survival detriment in the Ariel-4 study. Lynparza fell in an identical battle in August, when the Solo-3 trial yielded a 1.33 hazard ratio, and then last week GSK quietly withdrew Zejula from late-line use citing the “totality of information” from other Parp inhibitors.

It is not known how much weight the FDA had put on the manufacturers to withdraw, and in any case third, fourth or later-line use in a shifting space probably does not represent a huge market. But make no mistake: companies with Parp inhibitors that have not shown an OS benefit are having their feet held to the fire.