With six-month data from the Zuma-1 study KTE-C19 today moved a step closer to becoming the world’s first CAR-T project. Whether it does so depends on the FDA, but Kite now has a dataset that looks strong enough to be filed, and which it can boast shows “durability”.
It was durability that was the main unknown at three months’ follow-up, so the fact that at six months there is no significant fall-off in remission is a major finding (see table below). That said, given the safety fears swirling around all CAR-T projects, perhaps the most important point is that no further suspicious deaths were seen at six months.
At the Ash meeting Kite spelled out three Zuma-1 deaths that were not due to disease progression, two of which were deemed a result of KTE-C19. This is still the case, and today Kite was able to boast that serious cytokine release syndrome dropped from 18% at three months to 13% at six, with neurological events falling from 34% to 28%.
This is likely thanks to doctors and nurses becoming more used to handling the side effects of CAR-T therapy and implementing recently discussed mitigation strategies.
There were no cases of cerebral oedema, Kite said, clearly referring to Juno’s competing asset JCAR015, whose pivotal Rocket study is on hold owing to deaths from this adverse event. Kite intends to complete its rolling US filing for KTE-C19 in March, and says the full Zuma-1 data will be presented at April’s AACR meeting.
Analysis of Zuma-1, a trial in various types of lymphoma, is not helped by its complex design; it had two separate phases, the second of which had interim readout at three months, and was further complicated by a decision to increase enrolment by 50%.
As such the three-month data concerned 62 patients, 51 of whom were in the diffuse large B-cell lymphoma (DLBCL) cohort (Kite flies, but durability remains the key unknown, September 27, 2016). Today’s six-month data, the point for determining its primary endpoint of overall remission rate (ORR), concerns 101 patients, 77 with DLBCL.
With durability holding up at six months the question of whether ORR of just under 40%, and a complete remission (CR) rate of just over 30%, are good enough will be down to the regulator. There are few like-for-like comparators: the Scholar-1 relapsed lymphoma meta-analysis that Kite sets as its benchmark showed 26% ORR and 8% CR, but related to long-term outcomes.
Relapses do continue to be a focus, best illustrated by comparing absolute against best response, a separate measure used by Kite. Across all 101 patients 82% went into partial or complete remission at any point, but at six months absolute ORR was 41%.
|Durability in Kite's Zuma-1 trial|
|Phase I (n=7)||Phase II (n=51-77)*||Phase II (n=62-101)**|
|Note: *DLBCL cohort, comprising 51 pts at 3mth, and 77 pts at 6mth; **combined lymphoma cohort, comprising 62 pts at 3mth, and 101 pts at 6mth.|
In Kite’s favour is that, while some patients relapse, others might take a while to go into remission: one of four partially responding patients at six months improved to a complete remission three months later, Kite said.
Four CR patients did not have a six-month tumour assessment and are thus classified as non-responders, but clearly might be counted as CRs later, improving the data. The AACR presentation might also include 10 patients who have received a second treatment with KTE-C19.
While the safety and efficacy results look sufficient for filing, cost is a separate consideration – Kite said it had already had talks with payers.
Of course, the FDA could raise additional questions about data analysis or about the still relatively unreliable method of manufacturing CAR-T cells. And Kite must separately contend with ongoing legal action against Juno concerning a claim that KTE-C19 infringes a JCAR015 patent; an inter-parties review on this has already favoured Juno.
That said, Kite investors today breathed a sigh of relief, and its stock opened up 16%. Now it is over to the FDA.