Brainomix presented findings at the World Stroke Conference in Seoul suggesting that automated analysis of ordinary non-contrast CT can help pick out severe stroke patients more likely to respond to an experimental drug.
The Oxford spinout worked with Remedy Pharmaceuticals on a post-hoc analysis of the Phase 3 CHARM trial, which tested intravenous glyburide, or CIRARA, in large hemispheric infarction, one of the most disabling forms of stroke.
Sizing the infarct from the scan most hospitals have
Earlier looks at CHARM hinted that patients with smaller imaging-defined infarct cores did better. Those measures leaned on CT perfusion and MRI, modalities that many emergency departments cannot run around the clock.
Brainomix instead used its 360 Stroke software and in-house AI Core Lab to measure acute infarct volume from non-contrast CT in 428 patients, then compared that with outcomes.
In a per-protocol group of 362 patients, CIRARA showed a borderline association with better scores on the modified Rankin Scale. Among 191 patients whose core volume sat below 125 mL, the link became significant, including a greater likelihood of reaching mRS 0 to 3 at 90 days.
What it means for trial design
George Harston, Brainomix’s chief medical and innovation officer, said the appeal is access: if selection works on the most widely available stroke scan, trials can run at far more hospitals.
The caveats are substantial. CHARM was terminated early after sponsor Biogen pulled back, and the program has since returned to Remedy. Post-hoc subgroup findings are hypothesis-generating, not proof.
Remedy CEO Sven Jacobson called the signals encouraging and said the company intends to investigate them further.
