Roche could not afford to fall behind its immuno-oncology peers, most of whom have been frantically amending study designs in a race for the moving target of first-line lung cancer. Sure enough, the Swiss firm yesterday upped the ante, revealing that all trials of Tecentriq in this setting were now looking at overall as well as progression-free survival as co-primary endpoints.
This development highlights the importance of study powering, and it has become apparent that all of Tecentriq’s key lung cancer trials had originally been designed to enrol far more patients than necessary – a fortuitous overpowering that could put more pressure on Astrazeneca.
Astra’s latest move was to amend the way it analyses its all-important Mystic trial of durvalumab and tremelimumab, seeking to interrogate the dataset as monotherapy as well as in combination, across different levels of PD-L1 expression, and looking at both PFS and OS (Astra’s Mystic mystery and other surprises in store, January 17, 2017).
Of course, for this trial to be sufficiently powered to be cut in so many ways it has had to be upsized, something the UK group had apparently started doing a while ago. Astra’s chief medical officer, Sean Bohen, today said he remained confident of Mystic’s powering for all cuts of the data.
However, Roche yesterday played up its own statisticians’ excessive early caution, which by pure luck has now given it flexibility.
The ability to add OS as co-primary endpoint to the Impower-150 trial – with neither a statistical penalty or a delay to allow recruitment of more patients – means that, if Mystic and Impower-150 both read out positive, Roche could file in the US at roughly the same time as Astra.
It is anybody’s guess what the FDA would do in this case, though Astra insists that two randomised phase III trials will stand on their own. Then there is the separate question of how it handles Merck & Co’s recent Keytruda plus chemo filing.
Astra’s timeline is for Mystic to yield PFS results around mid-2017, with OS next year, though if the study fails to show a PFS benefit it will likely proceed straight to OS readout without an announcement. There also remains the possibility of undisclosed interim readouts yielding an OS benefit earlier.
Meanwhile, Impower-150 is still driven primarily by PFS events, expected to yield data in the third quarter. And there is a good possibility of the OS benefit being sufficiently mature to show a statistically significant benefit already at this point, Roche’s head of pharma, Daniel O’Day, told yesterday’s 2016 financial conference.
Before that, of course, the FDA will give its verdict on Merck’s surprise filing of Keytruda plus chemo on the strength of the Keynote-021 trial. Though this is itself risky, Bristol recently said it now assumed that it would result in approval; in a breakout session yesterday Mr O’Day said the Keynote-021 data were indeed very encouraging.
|Selected first-line NSCLC readouts, including latest study amendments|
|Study||Company||Intervention||Enrolment||Population||Primary endpt||Data||Trial ID|
|Mystic||Astrazeneca||Durvalumab +/- tremelimumab||1,092||All-comers||PFS, OS||PFS mid-2017, OS 2018||NCT02453282|
|Impower-150||Roche||Tecentriq + chemo + Avastin||1,200||Non-squamous||PFS, OS||Q3/Q4 2017||NCT02366143|
|Checkmate-227||Bristol-Myers Squibb||Opdivo +/- Yervoy||2,220||All-comers||OS, PFS||Jan 2018||NCT02477826|
|Neptune||Astrazeneca||Durvalumab + tremelimumab||800||All-comers||OS||Oct 2018||NCT02542293|
|Impower-130||Roche||Tecentriq + Abraxane + chemo||650||Non-squamous||PFS, OS||2018||NCT02367781|
|Impower-132||Roche||Tecentriq + chemo + pemetrexed||568||Non-squamous||PFS, OS||2018||NCT02657434|
|Impower-131||Roche||Tecentriq + chemo +/- Abraxane||1,025||Squamous||PFS, OS||2018||NCT02367794|
|Impower-110||Roche||Tecentriq monotherapy||570||PDL1 +ve||PFS, OS||Mar 2019||NCT02409342|
|Pearl||Astrazeneca||Durvalumab monotherapy||440||PD-L1 >25%||PFS, OS||Jul 2020||NCT03003962|
This raises the separate question of whether an anti-PD-1/PD-L1 agent should be combined with chemotherapy or with an anti-CTLA4 MAb like Yervoy or tremelimumab. “We don’t find the scientific hypothesis around [a chemo combo] particularly compelling,” Astra’s Mr Bohen told EP Vantage at the company's earnings press conference today.
On the other hand Roche, whose first-line combo strategy revolves around chemo regimens, had “studied CTLA4 preclinically, but early on decided that there were more attractive targets to go after”, Mr O’Day revealed yesterday. The issue is relevant also in light of the serious stumble of Bristol’s Opdivo plus Yervoy combo.
It is clear that in such a fast-shifting field companies no longer have the luxury of reacting to developments by starting new trials. Bets were placed early on, and now the wait is on to see who made the right decisions.