Deep Learning Enabled Endovascular Stroke Therapy Screening in Community Hospitals

Deep Learning Enabled Endovascular Stroke Therapy Screening in Community Hospitals

After onset of Acute Ischemic Stroke (AIS), every minute of delay to treatment reduces the likelihood of a good clinical outcome. A key delay occurs in the time between completion of computed tomography (CT) angiography of the head and neck and interpretation in the setting of AIS care.

The purpose of this study is to assess the effect of incorporating Viz.AI software, which via via a machine-learning algorithm performs artificial intelligence-based automated detection of large vessel occlusions (LVO) on CT angiography (CTA) images and alerts the AIS care team (diagnosis and treatment decisions will be based on the clinical evaluation and review of the images by the treating physician, per routine standard of care). The hypothesis is that integration of the software into the AIS care pathway will reduce delays in treatment. A cluster-randomized stepped-wedge trial will be performed across 4 hospitals in the greater Houston area.

Inclusion Criteria:

• Male or Female
• 18 years of age or older.
• Patients who present to the emergency department with signs and/or symptoms concerning for acute ischemic stroke.
• Patients who undergo CT angiography imaging
• Patients determined to have a large vessel occlusion acute ischemic stroke. This determination will be made based on official radiology report for the CT angiography imaging.

Exclusion Criteria:

• Patients with incomplete data on the electronic medical record.