Article Text

Download PDFPDF

O16/158  Artificial Intelligence based nationwide centralised decision supporting system for improving stroke case efficiency in Hungary
  1. Istvan Szikora1,
  2. Bence Magyar1,
  3. György Pápai2,
  4. Gábor Szudi1,
  5. Máté Kondor1,
  6. Máté Czencz1,
  7. Sándor Nardai1,
  8. Zoltan Chadaide3,
  9. Edvard Benes4,
  10. Csaba Ovary5,
  11. Lorand Eross1
  1. 1National Institute of Mental Health, Neurology and Neurosurgery, Neurointerventions, Budapest, Hungary
  2. 2National Ambulance Service, Budapest, Hungary
  3. 3Brainomix Ltd, Oxford, UK
  4. 4eRAD-BB, Budapest, Hungary
  5. 5Bajcsy-Zsilinszky Hospital, Budapest, Hungary


Introduction Completion and interpretation of imaging studies may delay treatment and negatively impact outcome following Mechanical Thrombectomy.

Aim of Study The purpose of this study is to demonstrate the feasibility of building a nationwide stroke imaging network providing fast automated image analysis for all stroke centers of the country using Artificial Intelligence (AI).

Methods A total of 28 stroke centers including 24 primary and 4 comprehensive ones were incorporated within the same teleradiology network (eRAD, Hungary). CT scans completed under an acute stroke protocol are automatically transmitted through the network to a central server which applies an AI based software (eStroke, Brainomix Ltd.) to analyze native CT-s, CT Angiograms and CT Perfusion studies providing ASPECT scores, Large Vessel Occlusion site (LVO), collateral score and standard CTP parameters. All results become immediately available via PACS and through the cloud by both the sender and the relevant thrombectomy center.

Results The system was installed within 6 months, including all technical, legal and data protection preparations. Since October, 2022 till April, 2023 a total of 18 568 scans on 7 959 cases were processed by the system. For drip&ship patients, the decision time from arriving in the first hospital till starting the second transportation has been reduced by 20 minutes in the first 3 months of operation.

Conclusion Using AI based centralized stroke imaging network fast and reliable decision support can be provided to a large nationwide stroke care system. The system was installed with the support of an EU grant EFOP 5.2.6.

Disclosure of Interest Istvan Szikora has a consulting agreement with Brainomix Ltd.

Zoltan Chadaide is a former employee of Brainomix Ltd.

Edvard Benes is owner and CEO of eRAD-BB

The system was built by the support of an EU grant EFOP 5.2.6.

The other authors have no relevant conflict of interest.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.