Implementation of AI program improved stroke treatment response times
Click Here to Manage Email Alerts
Key takeaways:
- RapidAI is a program that analyzes images for vascular occlusions and salvageable tissue.
- Door to intervention time for mechanical thrombectomy was reduced by an average of 30 minutes after implementation.
Implementation of an artificial intelligence platform led to improved workflow and significantly improved response time for those with ischemic stroke, according to a poster from the International Stroke Conference.
“RapidAI is a collection of tools that allows us to screen these patients and allows multiple healthcare providers to review the imaging simultaneously in real time,” Timothy W. Malisch, MD, neurointerventional surgeon at Ascension Alexian Brothers Hospital, told Healio in an email. “There has been limited real-world data to assess whether this technology increases the recognition of stroke victims likely to benefit from these therapies, to assess whether these therapies might be provided more quickly and whether this might translate into improved outcomes for these patients.”
Malisch and colleagues aimed to assess the effectiveness of the RapidAI platform for improved stroke outcomes, including the Rapid mobile app.
Their study culled data from the Get with the Guidelines Stroke database of adults admitted to Alexian Brothers Medical Center in Illinois with acute ischemic stroke between 2019 and 2022. The final analysis included 1,732 individuals, 1,193 prior to installation of the RapidAI platform and 539 after installation of Rapid CTP, CTA, ASPECTS, LVO and RMA.
Primary outcomes for the study were the number of eligible participants with AIS given thrombolysis (IV tPA or IV TNK) and/or mechanical thrombectomy (MT), along with time to treatment, while door to MT puncture times for transfer patients were analyzed separately. Modified Rankin score (mRS) at time of discharge was included as a secondary clinical outcome.
According to results, the median door to femoral puncture time for MT cases decreased from 2 hours, 28 minutes, to 1 hour, 58 minutes, post-Rapid implementation for non-transfer patients; however, door to femoral puncture times for transfer patients were unchanged.
The researchers found that MT procedure volume increased from 9.6% to 14.5%. IV tPA/TNK volume and door-to-needle times were unchanged; 15% treated with tPA/TNK pre-Rapid and 14% post-Rapid, with door-to-intervention time decreased from 48 minutes pre-implementation vs. 47 minutes post implementation.
Data additionally showed improvement in the distribution of discharge mRS scores in post-Rapid patients.
“Our study demonstrates that the use of RapidAI by multiple layers of healthcare providers throughout our healthcare system was associated with an increase in recognition of patients who could benefit from these interventions,” Malisch told Healio. “We found shorter times to mechanical thrombectomy and we also saw improved neurological outcomes in these stroke patients after the implementation.”