Evaluation of a Novel Point-of-Care Neuromonitoring Device (AlphaStroke) to Detect Large Vessel Occlusion in Suspected Acute Stroke Patients

Study Objectives: Several prehospital stroke scales have been developed to provide quick and accurate triage to facilitate timely treatment. This study evaluated a portable, experimental electroencephalogram (EEG) device (AlphaStroke, Forest Devices, Pittsburgh, Penn.) using artificial intelligence (AI) as a potential tool for detection of acute stroke and large vessel occlusion (LVO) among patients with neurological deficits. Both device performance and feasibility in the emergent setting were assessed. 

Methods: This observational study enrolled a convenience sample of emergency department (ED) patients evaluated for suspected stroke within 24 hours of symptom onset. LVO status was determined by local neuroradiologists blinded to AlphaStroke’s output. LVO was defined as an acute occlusion of the any of the following arteries: ICA/MCA (M1 or M2)/vertebral/ basilar. Controls were neurologically normal subjects (NIHSS=0). 

Results: From May 2018 to June 2019, eight urban US stroke centers enrolled 100 subjects being evaluated for stroke. The study also enrolled 113 controls. In subjects with acute neurologic deficits, 26 had LVOs (26%). Device performance for detecting stroke and LVO is shown in Table 1. There were no severe adverse events related to use of the device. 

Table 1: AlphaStroke Performance for Identification of LVO

Predicted:

• True positives, 52; false positives, 17
• False negatives, 6; true negatives, 113    
• Sensitivity=92%, positive predictive value=46%, positive likelihood ratio, 6.17
• Specificity=85%, negative predictive value=99%, negative likelihood ratio, 0.09

Conclusion: The AlphaStroke device performed well in identifying LVO in patients presenting with suspected stroke. The performance of the AlphaStroke device in the acute setting indicates it may be able to support prehospital decision-making when triaging suspected stroke subjects. Additional studies with larger sample sizes are needed to validate this study’s findings.