Astra’s Mystic mystery and other surprises in store
With Astrazeneca and Bristol-Myers Squibb recently changing their plans in first-line lung cancer, and Merck KGaA and Pfizer about to do so, timelines in this all-important immuno-oncology setting have entered the realms of guesswork.
Still, the latest amendment to the Mystic study does send vital hints about Astra’s thinking; for one thing the group’s confidence in the tremelimumab element of its combo might be fading, though it is likely also trying to maximise chances of success. But it now looks like all companies will compile data packages behind closed doors, meaning that filings could take investors by surprise.
This is precisely the strategy used by Merck & Co, which last week caught the markets off guard by filing its Keytruda plus chemo combination for accelerated approval based on the remission rate seen in its Keynote-021 trial (JP Morgan – Merck muscles into Bristol’s combo space, January 11, 2017).
When Bristol-Myers Squibb’s Opdivo failed in the Checkmate-026 study Keytruda seized the upper hand with approval in PD-L1-high, first-line NSCLC patients. This means that there is everything to play for in all-comers, and combinations – with chemo or an anti-CTLA4 agent – are the way forward.
Early Opdivo filing?
However, Bristol’s own plans for a second bite at the cherry are in doubt; an early filing based on remission rates seen in the expanded Checkmate-568 trial was thought by analysts to be capable of yielding a registrational dataset in the wake of Checkmate-026’s failure.
Checkmate-568 has a primary completion date this month, but Bristol told investors at last week’s JP Morgan conference not to expect anything in 2017. A spokesperson subsequently confirmed: “You shouldn’t assume that if we have data this year we would disclose it this year.”
The group still insists that it has identified a path to earlier submission in first-line NSCLC, so this presumably involves either a surprise filing based on Checkmate-568 or an early readout of Checkmate-227. The latter study is projected to yield progression-free survival data next year.
Astra’s plans, revealed today, muddy the waters further. The UK company has amended the statistical analysis plan for Mystic, the hotly awaited phase III trial of durvalumab plus tremelimumab, in what looks like a move designed to maximise chances of success.
Mystic earlier had overall survival added to PFS as a co-primary endpoint, and it now prespecifies analyses in all-comers as well as PD-L1-high patients. Meanwhile, PD-L1-high patients on durvalumab monotherapy will also be subjected to PFS and OS analyses as a primary rather than the earlier secondary measure.
PFS readout has been slightly delayed to mid-2017, with OS data next year at the latest, plus several undisclosed interim OS analyses. Bernstein analysts said a PFS hit would likely result in data being presented at September’s Esmo meeting, though a miss would likely see Mystic proceed straight to OS readout.
|Selected immuno-oncology trials in 1st-line NSCLC
|Chemo combo; PFS data in 2017
|Merck & Co
|To serve as confirmatory trial for ORR benefit seen in Keynote-021 chemo combo arm; PFS data H2 2017
|Now to look at OS & PFS in PD-L1-high & all-comer patients given combo, and PFS & OS in PD-L1-high patients on durvalumab monotherapy; PFS data H2 2017, OS in 2018 "at the latest"; undisclosed interim OS analyses
|New study with durvalumab monotherapy in Asia, measuring PFS & OS in PD-L1-high patients
|Uncontrolled study. ORR data likely in house soon, but unlikely to be announced in 2017
|PFS data due 2018, but some analysts hint at possible readout in 2017
|Javelin Lung 100
|Merck KGaA & Pfizer
|Design being changed; PFS data no longer expected in H2 2017
The analysts also said Astra was cutting the data in so many ways that it risked having Mystic run out of statistical power. The other vital point is the elevation in importance of durvalumab monotherapy – also backed by the initiation of the monotherapy Pearl trial, to support approval in China.
The risk here is that Astra is getting cold feet about tremelimumab, an asset that has yielded relatively little hard data so far. Tremelimumab yielding a toxicity profile in line with Yervoy, but with less efficacy, could spell Astra’s nightmare scenario.
The UK group is not alone in study design tinkering. Merck KGaA’s head of R&D, Luciano Rossetti, recently revealed that his group’s first-line NSCLC trial of avelumab, Javelin Lung 100, was being modified to boost chances of success after Opdivo’s failure, resulting in a delayed readout.
Then there is Roche’s Impower-150 trial of Tecentriq plus chemo. Since this has a primary completion date this month the timing of its data release is crucial in light of the May 10 date for FDA action on Merck & Co’s Keytruda plus chemo filing.
If this is now all about surprising the competition with stealth attacks, Merck & Co needs to watch its back.