Authors: Dr. Räty Silja
The recent ESO Educational Webinar, held in collaboration with the ESO AI committee, continued the topic of artificial intelligence (AI) applications in stroke management and research. With insightful presentations from Prof. Bruce Campbell and Prof. Willian Whiteley, moderated by Prof. Robin Lemmens and Dr. Smriti Agarwal, the session emphasised how AI can advance acute stroke treatment and health data science.
Prof. Campbell focused on the application of automated imaging to enhance the efficiency of acute stroke treatments. He highlighted the potential of automated processing to facilitate the acute stroke workflow by allowing for quicker and more accurate imaging interpretation and by notifying different members of the stroke team. This can mean fewer missed opportunities and faster care for patients who are candidates for reperfusion therapies. Automated tools have been developed for detecting early ischaemic changes or haemorrhage in non-contrast CT, for perfusion interpretation, and for assessing collaterals or vessel occlusions in CTA, and some of them are already operational and proving beneficial in clinical settings.
Moreover, Prof. Campbell discussed the current automated applications of perfusion imaging in acute stroke care. These can aid in differential diagnostics, risk-benefit assessments for reperfusion therapies, and help in identifying patients who could benefit from intravenous thrombolysis or endovascular thrombectomy beyond the established time windows. He emphasised that while automation is helpful, good image acquisition quality and interpretation skills remain vital components of acute stroke management.
Prof. Whiteley shifted the focus to the broader implications of AI in health data science. He outlined the potential to harness health data on a large scale, enabling researchers to analyse complete population, or even multination cohorts without selection bias. By linkage between different data repositories, large-scale analyses can aid in estimating disease epidemiology, improving quality of care, and follow-up of randomised controlled trials. Pivotal steps to this direction were taken during COVID-19, when demand for fast and accurate epidemiological data was immense. However, Prof. Whiteley stressed the importance of sustaining conditions for successful large-scale analyses also after the pandemic, including governance, data availability, collaboration, and citizen support.
Other advantages of AI in health data science addressed in the talk include its potential to accelerate analysis of image and text data. Natural language processing can be leveraged in reviewing large data sets for epidemiological data or for efficient assembly of research cohorts. AI can also help in developing tools for image preparation and analysis. Prof. Whiteley concluded that for successful implementation of AI, health data scientists and clinicians must ask questions that resonate with citizens and patients, rely on scientific evidence, and foster collaborative efforts. On the other hand, there is a need for streamlined and secure governance, as well as access to affordable, high-power computing resources in secure data environments. Thus, advocating for these principles should be encouraged in the stroke community.
Related links:
BHF Data Science Centre: https://bhfdatasciencecentre.org/
Scottish Medical Archive: https://publichealthscotland.scot/resources-and-tools/health-intelligence-and-data-management/scottish-medical-imaging-smi/overview/purpose/
Missed the webinar? Education webinars are available on demand for ESO members on the ESO eSTEP platform.
Not a member yet? Sign up now.
ESOC is Europe’s leading forum for advances in research and clinical care of patients with cerebrovascular diseases. ESOC 2026 will live up to its expectation, and present to you a packed, high quality scientific programme including major clinical trials, state-of-the-art seminars, educational workshops, scientific communications of the latest research, and debates about current controversies. Learn more.