Abstract TP106: Diagnostic Capability Of An Emerging Portable Stroke Detection Device In Clinical Practice

Abstract

Introduction: Portable stroke detection devices may serve as important tools in stroke triage. With the capacity to differentiate stroke type and severity, these devices may allow for reductions in time-to-treatment, leading to improved patient outcomes in time sensitive neurologic diseases. One such device is the Cerebrotech Visor, an FDA-approved technology that aims to identify severe stroke by detecting bioimpedance asymmetry in brain tissue across a wide range of electromagnetic wavelengths. This study aims to evaluate the ability of this device to detect stroke types in a pilot set of patients presenting as acute stroke codes. Methods: Trained operators performed scans using the Visor during stroke codes from November 2020 to July 2021 on eligible patients. Clinical and radiologic characteristics as well as the Visor output, an algorithmically derived asymmetry score, were prospectively entered into a quality assurance database. CT findings and subsequent physician notes were used to classify patient stroke type and severity. T-tests were performed to assess the ability of the Visor score to detect various types of CT-confirmed stroke. Results: Seventy-eight patients were included and scanned. Using CT results, patients were classified into categories: ischemic stroke (IS) (16), intracerebral hemorrhage (ICH) (12), subdural hematoma (2), brain tumor (0), prior neurosurgical procedure (0), intraventricular hemorrhage (2), hydrocephalus (1), and no findings (49), with 3 patients experiencing a combination of these neurological findings. Visor scores for IS and ICH were individually compared to the cohort without IS and ICH, respectively. A t-test comparing patients with IS to the rest of the cohort showed significantly higher Visor scores (9.66 vs 7.13, p = 0.014). Although patients with ICH also had higher Visor scores, the difference was not statistically significant (8.16 vs 7.55, p = 0.30). Conclusion: The Cerebrotech Visor was able to differentiate between IS and non-IS with the current bioimpedance asymmetry-based device algorithm. Future studies aim to increase patient cohort size and use this patient-centered data to develop a machine learning algorithm to better differentiate other stroke types with this bioimpedance technology.