Brainomix says its CT reading picked out which severe stroke patients responded in a Phase 3 trial

The hardest question in a failed trial is whether the drug worked for somebody. Brainomix says its automated reading of plain CT scans pointed to the severe stroke patients who looked more likely to benefit, and it took that analysis to a stroke congress in October.

By Atlas ... AI Editorial Collaborator

What Brainomix Announced About Its CHARM Trial Imaging Analysis

Brainomix said on 28 October 2026 that its AI imaging analysis of the Phase 3 CHARM trial identified what it calls promising treatment signals in severe stroke patients. The company says its reading of non-contrast CT scans may help identify patients more likely to benefit from treatment with CIRARA, and that the data were presented at the 18th World Stroke Conference in Seoul between 21 and 23 October 2026. The page on the Brainomix site is a short summary that links to a newswire release this desk could not open, so the figures, the number of patients, the comparison group and the authors are not available here. Brainomix states no regulatory status for this use of the software.

Why the Brainomix CHARM Trial Analysis Matters to Stroke and Research Leaders

Everything in this one is the company talking about its own analysis, and that is how it should be read. A conference presentation is not a published paper, and a signal is not a result. What makes it worth noting anyway is the job the software is being given. Reading a plain CT scan to sort patients into likely responders is an attempt to rescue information from a trial, and that is a far more valuable use of imaging AI than flagging a scan for a radiologist, because it could change which patients a future trial enrols. It is also the use with the most room for a wrong answer, since a model that sorts patients after the fact can find a pattern that does not repeat. Promising treatment signals carries no number, and until the full analysis is public nobody outside can judge it.

Where the Brainomix Update Comes From

Brainomix is the original record behind this update. It tells us what the company published. This brief adds the market context and the method we would use to test the development against other evidence.

Brainomix original source.

How We Would Research the Brainomix CHARM Trial Analysis

The source gives us the starting point. This is how we would build the next layer of research around it.

We would look for the conference abstract and then the published paper, because a summary on a company site cannot be checked.
Then we would ask Brainomix how many patients the analysis covered and whether the subgroup was chosen before or after the results were known.
We would ask whether the same CT reading has been tested on a separate group of patients, since a pattern found once inside one trial is not yet a finding.