The race to develop the next generation of checkpoint inhibitors that could rival Keytruda has become one of the most watched in oncology. Who wouldn’t want a slice of the drug’s gushing multibillion-dollar revenue stream?
And at this year’s European Society of Medical Oncology annual congress, AstraZeneca has its track shoes laced up. The company presented early phase 1/2 data showing that its bispecific antibody targeting CTLA-4 and PD-L1, MEDI5752, in combination with chemotherapy may rival Merck’s prize possession in the first-line non-small cell lung cancer setting. But toxicities that have long-plagued CTLA-4-targeting medicines persisted with the first dose tested, leading to a lowering.
Among 20 patients treated with 1,500 mg of the antibody, the objective response rate was 50%. Among a subset of nine patients with PD-L1 levels lower than 1%, the ORR was 55.6%. In the Keytruda and chemo arm, ORR was 47.6% for the overall group of 21 patients and 30% for the 10-patient PD-L1-low subgroup. And progression-free survival was slightly more than six months longer for the bispecific arm in the overall patient population and almost 4.5 months more in the PD-L1 subset.
In addition, after a follow-up of almost 23 months and 14.5 months for the MEDI5752 and Keytruda arms, respectively, MEDI5752's durability was double Keytruda's in the overall patient population. In the PD-L1-low population, though, the durability was nearly 14 months for MEDI5752 and was not reached for Keytruda.
“[W]hat I'm encouraged by is that response rate and the durability of response that we're seeing in PD-L1 less than 1% patient population,” AstraZeneca’s executive vice president of oncology R&D, Susan Galbraith, Ph.D., said in an interview. “We're seeing activity in places where you'd expect PD-1 to work as well.”
But there’s a critical catch. The 1,500-mg dose was plagued with high levels of toxicity that have dragged down past CTLA-4 treatments—80% of treated patients experienced either grade 3 or 4 side effects, and 70% discontinued treatment as a result of treatment-related adverse events. As a result, the company tried half that dose, testing the 750 mg version in a single-arm cohort of 50 patients.
With a median follow-up of nearly four months, AstraZeneca found that the ORR at the lower dose was nearly 41%, roughly equal to Keytruda's showing from the randomized, higher-dose part of the trial. But in patients with PD-L1 less than 1%, it appeared to maintain an advantage, with an ORR of 44.4%. The grade 3 or 4 side effects that plagued that higher dose also fell—though remained high—declining to 50% and 10 patients, or 20%, dropping out of the trial due to treatment-related side effects. Notably, there was one death in the 750 mg treatment as a result of the side effects.
But the switch means that the company does not yet have the durability data that made the 1,500-mg dose so head-turning. Galbraith says she’s optimistic, pointing out that T-cell proliferation at the 750 mg dose—and an even smaller 500 mg dose that the company has begun testing—are higher than levels seen with other approved drugs. But ultimately, longer-term data will inform future dose selection.
“So we're confident about that pharmacodynamic effect of those doses, and it's just really about seeing that long-term durability and safety profile evolve in order for us to select the right dose,” she said.
As for future clinical development plans, Galbraith said the company is “accelerating this program” and is also exploring combining MEDI5752 with one of its antibody-drug conjugates. She wouldn’t say whether the company is likely to launch into a phase 3 NSCLC study next year.
Editor's note: This story was updated with new safety and response rate data from AstraZeneca's full presentation.