Stroke: AI software saves about 11 minutes before the clot-removal procedure

Why does every minute count in a stroke?

An ischaemic stroke happens when an artery in the brain is blocked. When it is a large artery, it is called a large vessel occlusion. The standard treatment is thrombectomy: a doctor threads a thin tube through an artery in the groin to remove the clot. Its benefit depends heavily on time: the faster, the better.

Hospitals therefore try to speed up every step: reading the scan, calling the teams, moving the patient to the treatment room.

What did the researchers do?

Between January 2021 and February 2022, four specialist stroke centres in the Houston area of Texas switched on an AI software one after another, in a randomly drawn order. This is called a cluster randomised trial with a stepped rollout.

The software automatically analyses the CT angiogram (a scan that shows the blood vessels of the brain) and, within minutes, sends an alert to the phones of clinicians and radiologists when an occlusion is possible. The main measure was the door-to-groin time: the time between arrival at the emergency department and the puncture for thrombectomy.

Of 243 patients included, 140 were treated before the software was switched on and 103 after.

What do the results show?

Before the software, the median door-to-groin time was 100 minutes. With it, that time was reduced by 11.2 minutes. The time from the start of the scan to the start of thrombectomy fell by 9.8 minutes.

On the other hand, the time to the clot-dissolving drug given through a vein, and the length of hospital stay, did not change. In an exploratory analysis, the likelihood of regaining functional independence at 90 days was not different, but the study was not large enough to settle that question.

What does this change for patients?

The authors conclude that automated detection combined with secure phone messaging improves acute stroke workflows and cuts treatment times by a clinically meaningful amount. For a patient, this means the clot can be removed sooner. Those minutes matter because the benefit of thrombectomy is highly time dependent, as the authors point out.

The effect on patients' long-term health is not shown by this study. So it is not a promise of better recovery.

What are the limits?

The sample is modest (243 patients) and the study took place in a single region. Patients transferred from another hospital, those whose stroke happened in hospital and a few other cases were excluded.

The result mostly concerns the speed of organisation, which also depends on teams and their habits. Larger studies are needed to know whether these minutes translate into better recovery.