Background In Latin America (LATAM), the prevalence of stroke is alarmingly high, but the implementation of reperfusion therapies is low. The advancement of artificial intelligence (AI) technologies has potential to dramatically improve the care of these patients and be rapidly disseminated. RapidAI is a software solution that utilizes AI to quickly analyze imaging data to identify vascular occlusions and salvageable tissue independent of time. This could expand the access to quality care by increasing the number of patients with acute ischemic stroke (AIS) that can be treated with reperfusion therapies that have limited time windows. Our aim is to analyze the impact of the implementation of AI in our center for the selection of patients with AIS amenable to reperfusion treatment beyond the usual time windows.

Methods This was a retrospective analysis of 1,672 consecutive AIS treated at Clínica Alemana de Santiago between 2014–2023. Primary outcomes were the number of patients treated with thrombolysis (IV tPA), mechanical thrombectomy (MT), or rescue decompressive craniectomies (DC) as well as time to treatment in each period. Secondary outcomes included discharge modified Rankin Score (mRS) and hospital length of stay (LOS).

Results This analysis included 824 pre-RapidAI installation and 848 after installation of RapidAI. The baseline NIHSS and last know well to hospital arrival was balanced between the 2 study cohorts. IV tPA utilization increased from 23% to 30% post-Rapid (p = 0.003). MT procedure utilization increased from 5% pre-Rapid to 10% (p < 0.0001). DC were reduced from 13 to 4 (p=0.05), LOS was unchanged. There was a favorable shift in the distribution of mRS scores at discharge in post-Rapid patients (p = 0.03) and a significant increase in excellent outcomes (mRS= 0–1; 60% vs 55%, p=0.03).

Discussion Following implementation of the AI innovative software platform in a comprehensive stroke center like Clínica Alemana de Santiago, there was a significant increase in patients treated with reperfusion therapies and a decrease in the number of rescue DC. These findings support the implementation of AI imaging technology for improving stroke care in LATAM as it has the potential be of widespread use.