Preview – Tigit lives again
An obscure Roche dataset reignites the Tigit space, and sends up Arcus, Compugen and Iteos.
Jacob Plieth30 May, 2023
Despite showing only the vaguest signals of efficacy Tigit blockade remains a major stock price mover. The latest data to have caught investors’ imagination come from Roche’s Morpheus-liver study, whose results featured among an Asco abstract drop that saw significant gains for many biopharma companies on Friday.
The Roche data – from a trial that until last week had sat firmly under the radar for most – managed to send Iteos stock up 30% even though that Tigit company is not even presenting at Asco; Arcus, which does feature, closed up 26%. However, though Morpheus-liver comes as a positive surprise, those betting on it are perhaps ignoring the earlier lessons of Roche’s confusing mid-stage Cityscape trial.
Cityscape had famously misled the markets into thinking that Tigit plus PD-(L)1 blockade had a significant benefit in front-line lung cancer, but the result was confounded by underperformance of the control arm. Morpheus-liver, a multi-cohort phase 1/2 study in first-line liver cancer, appears to suffer from precisely the same effect.
Despite showing only the vaguest signals of efficacy Tigit blockade remains a major stock price mover.
58% risk reduction
The good news for Tigit fans is that combining Roche’s anti-Tigit MAb tiragolumab with Tecentriq and Avastin yielded median PFS of 11.1 months, with a reduction in risk of progression or death of a remarkable 58% versus Tecentriq and Avastin alone. ORR also impressed, coming in at 43% versus 11%.
Evercore ISI analysts said the data were hard to ignore, and wrote: “We think it’s getting hard to argue there isn’t something there.” They were especially enthused by an apparent benefit in PD-L1-negative subjects, where ORR was 28% against zero remissions among control subjects; this is unexpected, since Tigit blockade has previously tended to enhance responses only in PD-L1-positives.
However, the big caveat is that control patients seem to be performing worse in Morpheus-liver than expected. Imbrave-150, the phase 3 trial backing Tecentriq plus Avastin’s first-line liver cancer use, yielded median PFS of 6.8 months (in Morpheus-liver this was just 4.2 months), and ORR of 28% rather than 11%.
Evercore argues that a cross-trial comparison against Imbrave-150 still yields an impressive result for Morpheus-liver’s active arm. The abstract has a 28 November 2022 data cutoff, and the active cohort comprised just 40 patients; data will likely be updated at Asco.
The other big Tigit dataset to watch at Asco will be an update on Arcus/Gilead’s Arc-7 study combining domvanalimab with zimberelimab in first-line NSCLC.
This yielded positive mPFS and ORR data in December, but that result was flattered by poor performance of zimberelimab, and BTIG analysts expect an Asco data cut including five more months’ follow-up. However,
the abstract has not been updated, so Arcus’s 26% share price surge can be put entirely down to Roche’s Morpheus-liver result.
Also featuring at Asco is Compugen, which will present data from a small study of its Pvrig inhibitor COM701 combined with Bristol Myers Squibb’s anti-Tigit MAb BMS-986207 and Opdivo. This uncontrolled trial, in microsatellite-stable endometrial cancer, has shown two partial remissions among nine patients, and Compugen closed up 11% on Friday.
While all ordinary Asco abstracts are now live, still to be presented are those concerning late-breaking data.
The most industry-relevant Asco late-breakers include Novartis’s Natalee trial of Kisqali, Immunogen’s Mirasol, Astrazeneca’s Duo-O, Servier’s Indigo and J&J/Legend’s Cartitude-4. However, the EHA meeting managed already to leak the Cartitude-4 data, so Natalee looks set to take centre stage.
The most industry-relevant Asco late-breakers include Novartis’s Natalee trial of Kisqali, Immunogen’s Mirasol, Astrazeneca’s Duo-O, Servier’s Indigo and J&J/Legend’s Cartitude-4.
In addition, Jacob Plieth spoke with Brad Loncar on BiotechTV to preview the conference.