Automated Large Vessel Occlusion Detection Software and Thrombectomy Treatment Times

Abstract

ImportanceThe benefit of endovascular stroke therapy (EVT) in large vessel occlusion (LVO) ischemic stroke is highly time dependent. Process improvements to accelerate in-hospital workflows are critical.ObjectiveTo determine whether automated computed tomography (CT) angiogram interpretation coupled with secure group messaging can improve in-hospital EVT workflows.Design, Setting, and ParticipantsThis cluster randomized stepped-wedge clinical trial took place from January 1, 2021, through February 27, 2022, at 4 comprehensive stroke centers (CSCs) in the greater Houston, Texas, area. All 443 participants with LVO stroke who presented through the emergency department were treated with EVT at the 4 CSCs. Exclusion criteria included patients presenting as transfers from an outside hospital (n = 158), in-hospital stroke (n = 39), and patients treated with EVT through randomization in a large core clinical trial (n = 3).InterventionArtificial intelligence (AI)–enabled automated LVO detection from CT angiogram coupled with secure messaging was activated at the 4 CSCs in a random-stepped fashion. Once activated, clinicians and radiologists received real-time alerts to their mobile phones notifying them of possible LVO within minutes of CT imaging completion.Main Outcomes and MeasuresPrimary outcome was the effect of AI-enabled LVO detection on door-to-groin (DTG) time and was measured using a mixed-effects linear regression model, which included a random effect for cluster (CSC) and a fixed effect for exposure status (pre-AI vs post-AI). Secondary outcomes included time from hospital arrival to intravenous tissue plasminogen activator (IV tPA) bolus in eligible patients, time from initiation of CT scan to start of EVT, and hospital length of stay. In exploratory analysis, the study team evaluated the impact of AI implementation on 90-day modified Rankin Scale disability outcomes.ResultsAmong 243 patients who met inclusion criteria, 140 were treated during the unexposed period and 103 during the exposed period. Median age for the complete cohort was 70 (IQR, 58-79) years and 122 were female (50%). Median National Institutes of Health Stroke Scale score at presentation was 17 (IQR, 11-22) and the median DTG preexposure was 100 (IQR, 81-116) minutes. In mixed-effects linear regression, implementation of the AI algorithm was associated with a reduction in DTG time by 11.2 minutes (95% CI, −18.22 to −4.2). Time from CT scan initiation to EVT start fell by 9.8 minutes (95% CI, −16.9 to −2.6). There were no differences in IV tPA treatment times nor hospital length of stay. In multivariable logistic regression adjusted for age, National Institutes of Health Stroke scale score, and the Alberta Stroke Program Early CT Score, there was no difference in likelihood of functional independence (modified Rankin Scale score, 0-2; odds ratio, 1.3; 95% CI, 0.42-4.0).Conclusions and RelevanceAutomated LVO detection coupled with secure mobile phone application-based communication improved in-hospital acute ischemic stroke workflows. Software implementation was associated with clinically meaningful reductions in EVT treatment times.Trial RegistrationClinicalTrials.gov Identifier: NCT05838456

The Role of Artificial Intelligence-Powered Imaging in Cerebrovascular Accident Detection

“…Clinically relevant reductions in endovascular stroke therapy (EVT) in LVOs were recorded in a cluster randomized stepped-wedge clinical trial in Houston, Texas that assessed the primary outcomes of patients needing acute stroke triaging [ 51 ]. This study was conducted in four comprehensive stroke centers with 243 participants via enabling AI-automated CT angiogram coupled with secure messaging.…”

“…This study was conducted in four comprehensive stroke centers with 243 participants via enabling AI-automated CT angiogram coupled with secure messaging. This system resulted in a 9.8-minute reduction (96% CI, -16.9 to -2.6) in time from CT scan and initiation of EVT, further reducing the length of hospital stay and ultimately causing an 18% decrease in mortality [ 51 ]. With those previous numbers in mind, a 9.8-minute reduction is equivalent to 18.62 million neurons, 135.24 billion synapses, and 117.6 miles of axonal fibers potentially spared, per patient, due to the integration of AI and closed-loop messaging systems.…”

Einsatz von Künstlicher Intelligenz in der Notaufnahme

“…10 Minuten. Allerdings sind echte NeuroradiologInnen derzeit noch etwas genauerund leider langsamer [9,10]. Auch wenn diese konkreten KI-Anwendungen bereits Marktreife erlangt haben und zugelassen sind, gibt es zahlreiche weitere Proof-of-Concept-Studien zur Diagnosevorhersage und Entscheidungsunterstützung, deren prospektive Validierung und Nutzenbewertung noch aussteht [11,12,13].…”

The clinician-AI interface: intended use and explainability in FDA-cleared AI devices for medical image interpretation

“…As providers consider AI adoption, it is especially instructive to be aware of the different CAD types and their advantages and limitations across different diseases. In the case of CADt, several studies have indeed shown the potential for faster turnaround times and improved outcomes for prioritized exams 15 , 16 . However, the FDA has also recently released a letter “reminding health care providers about the intended use of radiological computer-aided triage and notification (CADt) devices for intracranial large vessel occlusion (LVO)”, including statements that such devices are not diagnostic and cannot rule out the presence of an LVO 17 , 18 .…”