Jemperli aims for ovarian white space

GSK's slowly recovering oncology franchise is in need of further boosts, but whether today's apparent success of the phase 3 First study, testing Jemperli plus Zejula in front-line ovarian cancer, is such a win will be open to debate.

It won't go unnoticed that First has been toplined just two weeks after Merck & Co disappointed in Keylynk-001, a first-line ovarian cancer trial of Keytruda plus Lynparza. The two studies share one unwelcome feature: both were said to have hit on PFS, a primary endpoint in Keylynk-001 and First alike, but both have also now failed to show a benefit on OS, a key secondary.

Given the poor precedent of Merck's other Keylynk trials, that OS failure likely makes Keylynk-001 a negative study. The same might also apply to First, though of course in both cases the decision will be down to regulatory agencies, with whom each company has said it intends to discuss the data.

The setting is important because no anti-PD-(L)1 drugs are approved for ovarian cancer. However, because of the broad market that first-line use would represent, a positive OS result might be necessary for approval; moreover, Keylynk-001 and First have demonstrably failed on OS, not just yielded a result that's too immature to interpret.

Still, there are notable differences between the two trials. Perhaps the most important is that First recruited all-comer patients, while Keylynk-001 enrolled only those who had BRCA non-mutated disease.

Zejula and Lynparza are both approved for ovarian cancer maintenance (patients who are in response to first-line platinum chemo), but while the former can be used in all-comers the latter is approved only in those with a BRCA mutation. Another major difference was that the Merck trial had Keytruda as active comparator, whereas GSK used Zejula or placebo.

Accordingly, what Merck seems to want to do is expand Lynparza, through the addition of Keytruda, into patients without BRCA mutations, while GSK is simply shooting for its PD-1 combo to beat Zejula in all-comers. Given the failure to hit OS, the extent to which either company has succeeded is questionable.

First overhaul

And there are other reasons to be cautious about the First result, including the fact that GSK has extensively tinkered with the trial's design, something that might have been driven by emerging data.

First initially had a sole PFS endpoint in PD-L1-positive patients, but in 2020 this was split to PFS in all-comers and PFS in PD-L1-expressers; later the primary switched to PFS in all-comers. First was also upsized twice, from 912 to 1,228 and then to 1,402 patients, likely to increase its chances of success, delaying readout by years.

An anomaly is that First had two comparator cohorts – Zejula or placebo alone; the latter was closed early owing to the approval of PARP inhibitors in first-line maintenance, which made it unethical to give patients placebo alone. However, if it's found that comparison versus placebo alone contributed significantly to the PFS benefit that could undermine the result further. 

Another red flag is that Jemperli plus Zejula was to have been tested in a separate phase 3 ovarian cancer study, a second-line investigator-sponsored trial called NSGO-Avatar. However, that was withdrawn without any patients enrolled, with the clinicaltrials.gov entry citing "lack of financial support" as the reason.

There is much to play for, given that Jemperli has so far been limited to a somewhat niche use in endometrial cancer, so GSK will want to build on the recent resurgence of its other oncology hope, Blenrep. Whether it succeeds will depend on the stance taken by regulators.

Comparison of two pivotal PD-1 + PARP inhibitor trials in ovarian cancer

Source: OncologyPipeline.