Artificial Intelligence-Assisted Software Significantly Decreases All Workflow Metrics for Large Vessel Occlusion Transfer Patients, within a Large Spoke and Hub System

Stavros Matsoukas, Laura K. Stein, Johanna Fifi

Research output: Contribution to journalArticlepeer-review

6 Scopus citations

Abstract

Introduction: Artificial intelligence (AI) software is increasingly applied in stroke diagnostics. Viz LVO (large vessel occlusion) is an AI-based software that is FDA-approved for LVO detection in CT angiography (CTA) scans. We sought to investigate differences in transfer times (from peripheral [spoke] to central [hub] hospitals) for LVO patients between spoke hospitals that utilize Viz LVO and those that do not. Methods: In this retrospective cohort study, we used our institutional database to identify all suspected/confirmed LVO-transferred patients from spokes (peripheral hospitals) within and outside of our healthcare system, from January 2020 to December 2021. The "Viz-transfers"group includes all LVO transfers from spokes within our system where Viz LVO is readily available, while the "Non-Viz-transfers"group (control group) is comprised of all LVO transfers from spokes outside our system, without Viz LVO. Primary outcome included all available time metrics from peripheral CTA commencement. Results: In total, 78 patients required a transfer. Despite comparable peripheral hospital door to peripheral hospital CTA times (20.5 [24.3] vs. 32 [45] min, p = 0.28) and transfer (spoke to hub) time (23 [18] vs. 26 [13.5], p = 0.763), all workflow metrics were statistically significantly shorter in the Viz-transfers group. Peripheral CTA to interventional neuroradiology team notification was 12 (16.8) versus 58 (59.5), p < 0.001, and peripheral CTA to peripheral departure was 91.5 (37) versus 122.5 (68.5), p < 0.001. Peripheral arrival to peripheral departure was 116.5 (75.5) versus 169 (126.8), p = 0.002, and peripheral arrival to central arrival was 145 (62.5) versus 207 (97.8), p < 0.001. In addition, peripheral CTA to angiosuite arrival was 121 (41) versus 207 (92.5), p < 0.001, peripheral CTA to arterial puncture was 146 (53) versus 234 (99.8), p < 0.001, and peripheral CTA to recanalization was 198 (25) versus 253.5 (86), p < 0.001. Conclusion: Within our spoke and hub system, Viz LVO significantly decreased all workflow metrics for patients who were transferred from spokes with versus without Viz.

Original languageEnglish
Pages (from-to)41-46
Number of pages6
JournalCerebrovascular Diseases Extra
Volume13
Issue number1
DOIs
StatePublished - 14 Feb 2023

Keywords

  • Artificial intelligence
  • Artificial intelligence-assisted diagnosis
  • Drip and ship
  • Endovascular thrombectomy
  • Stroke transfers

Fingerprint

Dive into the research topics of 'Artificial Intelligence-Assisted Software Significantly Decreases All Workflow Metrics for Large Vessel Occlusion Transfer Patients, within a Large Spoke and Hub System'. Together they form a unique fingerprint.

Cite this