Background Mechanical thrombectomy is considered the gold standard for treatment of large vessel occlusion (LVO) stroke. Early identification leads to faster puncture times as communication with the team is key. The integration of artificial intelligence (AI) platforms into acute stroke evaluation protocols may lead to more rapid and accurate detection of LVOs resulting in earlier notification of the neurointerventional (NIR) team, as was demonstrated in the VALIDATE study. In this study, we will evaluate if these findings can be generalized to other platforms such as RapidAI®. Methods Acute stroke consultations seen in the emergency department in 103 facilities (35 RapidAI® and 68 non‐AI) in 24 states from July 1, 2021 to December 31, 2021 were extracted from the TeleCare by TeleSpecialistsTM database. The encounters were reviewed for LVO, arrival‐to‐NIR notification time (ATN), and whether RapidAI® software was utilized. Patients were classified into two groups based on use of RapidAI® or no AI software. Then subcategorized further based on if they were seen at a thrombectomy capable center vs non‐thrombectomy capable center. The median arrival to NIR notification times were compared. Results A total of 14,159 patients (6,726 AI and 7,433 non‐AI) were included. The median ATN for AI was 57.5 min [40.3, 85.0] and non‐AI was 88.5 min [63.0, 116.3; Δ min = 31.0 min, p < 0.0001]. The median ATN for those at a thrombectomy center (TC) with AI was 49.0 min [36.0, 75.0] and non‐AI was 81.0 min [52.5, 97.8], p = 0.0077. The median ATN for non‐thrombectomy centers (non‐TC) with AI was 73.0 min [56.0, 115.5] and non‐AI was 90.0 min [66.5, 120.5], p = 0.0563. [TS1] . There was no difference in the AI vs non‐AI groups for door to needle times with thrombolytics at thrombectomy or non‐thrombectomy centers. Conclusions This large, multicenter study demonstrated a 31‐minute faster arrival to neurointerventionalist notification time with the aid of RapidAI®. A significant reduction in NIR notification times at both TC and non‐TC sites highlights the impact of AI software on expediting acute stroke care.
Read full abstract