I-Corps: Algorithm to detect stroke during cardiovascular surgery and reduce the time to effective clinical intervention

The broader impact/commercial potential of this I-Corps project is to detect stroke during surgery and reduce the time to effective clinical interventions. Available approved, lifesaving, mechanical clot removal therapy is not routinely administered for stroke due to delays in detection. This delay has prompted the recommendation of neuromonitoring to detect strokes.

Currently, in high-risk cardiovascular surgery, a trained neurologist visually monitors electroencephalogram (EEG) signals continuously and applies empirical criteria to detect cerebral ischemia and stroke. However, such visual monitoring can be mentally demanding, variable in quality, and limiting in scalability for the neurologist who monitors many surgeries.

Despite the availability of EEG devices, they are not universally available at all medical institutions. There is a clear need for scalable solutions that can support the neurologist to improve stroke detection and the administration of timely lifesaving therapies. Successful development of this project may translate to widespread adoption of neuromonitoring in surgery for not only stroke but spinal cord and peripheral nerve injuries.

This I-Corps project further develops a software system that can display intraoperative electroencephalogram (EEG) signals and use machine learning (ML) to detect stroke and alert the monitoring neurologist in real-time. An artificial intelligence software system with ML capabilities can meaningfully process massive sets of data quickly, accurately, and symbiotically work alongside clinicians.

The system relies on ML models using EEG, and clinical and anesthesia data collected during carotid endarterectomy (CEA). The initial ML models can also be used to detect ischemia, a precursor of stroke.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.