FAST-MRI

Stroke is a common cause of disability and death worldwide. Fast diagnosis of the cause of stroke is crucial for the patient’s prognosis. Diagnosis requires a brain scan, and magnetic resonance imaging (MRI) is increasingly used as imaging for patients with acute stroke. However, increasing radiological workloads risk causing delayed diagnosis or either missed or erroneous diagnosis. 

Aim

The FAST-MRI project aimed to investigate the accuracy, workload changes and time-to-diagnosis when implementing an artificial intelligence (AI) system for stroke magnetic resonance imaging (MRI) diagnostics in a Danish setting.

The project tested a Danish AI system for MRI that offers scan decision support via a smart protocol system and image reviewing support. The AI system needs to be validated to scientific and clinical standards prior to implementation.

The project included a systematic review of available AI systems for brain-MRI of stroke patients, and a retrospective validation study for the MRI AI system and prospective clinical trial to uncover workflow optimisation. This was, to our knowledge, the first of its kind.

Results

The systematic review assessed all available scientific research, from which it was concluded that the current AI technology could confidently diagnose ischaemia in MRI with a sensitivity and specificity of 93%. Furthermore, we determined that limited evidence existed for diagnosing hemorrhage. Lastly, the study highlighted that only one CE-marked AI solution had a scientific publication that examined its detection ability.

The detection Study investigated a commercially available AI algorithm solution for detecting MRI lesions compatible with stroke in a comprehensive stroke centre treating all stroke types. The study examined the AI's ability to categorise scans based on lesion presence and its ability to detect single ischaemic and hemorrhagic lesions. The study found a significant difference in sensitivity between the categorisation of scans and the detection of single lesions. This study concluded that ischaemic and hemorrhagic lesions can be identified but are not an optimal solution for use in a comprehensive stroke centre.

The combined studies in this thesis conclude that AI can be used to assess stroke-suspected patients, and prospective trials should be conducted to investigate.

Participants

Department of Radiology, OUH