HYPER-AI-SCAN Using AI-Assisted Transcranial Ultrasound to Detect Brain Bleeds in Acute Stroke Patients

What this Trial Is About

Researchers are evaluating whether combining transcranial Doppler ultrasound with artificial intelligence AI can help detect intracerebral hemorrhage ICH in adults with acute stroke within 48 hours of symptom onset. The study focuses on the feasibility of performing standardized ultrasound scans in stroke patients and whether AI can accurately distinguish hemorrhagic stroke from non-hemorrhagic stroke by comparing ultrasound predictions with CT scan results, which are the standard diagnostic method. Participants will undergo a standard CT scan to confirm stroke type, followed by a non-invasive transcranial Doppler ultrasound scan using a structured protocol. The ultrasound images will be coded and used to train and test AI models designed to recognize patterns linked to hemorrhagic stroke. No treatments or changes to clinical care are introduced as part of this observational research. During the study, participants will provide clinical and imaging information collected as part of routine care, while researchers will evaluate the feasibility of ultrasound acquisition and the accuracy of AI-based classification. The main outcome is the ability to perform standardized ultrasound scans, with secondary measures including exam acquisition time and AI accuracy. The ultrasound is performed only for research and does not affect participants treatment. The study aims to improve early stroke diagnosis, especially where CT is not immediately available.

AI-assisted Transcranial Duplex Sonography for Early Detection of Intracerebral Haemorrhage: HYPER-AI-SCAN