Abstract WP299: Clinical and Cost Impact of AI-Based Imaging Software on Reperfusion Treatment Rates in Acute Ischemic Stroke

Abstract

Background: Timely and accurate identification of patients eligible for reperfusion therapies is critical in acute ischemic stroke management. We modelled the potential impact of an AI-powered imaging decision-support platform, Brainomix 360 Stroke (B360S), on the number of Intravenous Thrombolysis (IVT) and mechanical thrombectomy (MT) procedures performed across NHS hospitals in England, alongside associated long-term health outcomes. Methods: We developed a decision-analytic model to simulate an annual cohort of stroke patients managed with or without AI-based support. Treatment decisions influenced by AI imaging outputs were modelled for a cohort across one year (2023-2024), impacting reperfusion rates in those eligible. The downstream impact on lifetime quality of life and costs was projected using a Markov model incorporating mortality and health utility differences across modelled scenarios. Input parameters were based on published literature, including from a recent real-world evidence study. An open-access R model was built with GPT-4 Turbo and GitHub Co-Pilot assistance. Scenario and probabilistic sensitivity analyses were conducted. Results: Use of AI imaging decision support was associated with increased reperfusion procedure volumes: Intravenous thrombolysis (IVT) increased by 9% and mechanical thrombectomy (MT) by 44% compared to standard care, at the national level (Table) and reduced lifetime costs. This translated into improved long-term health outcomes, with a quality-adjusted life year (QALY) gain of 0.57 per patient receiving IVT and 1.01 per MT performed. In total, the intervention arm produced over 1,000 more QALYs and saved over £10 million for the NHS for the modelled year of 2023. Therefore, the AI decision support tool was found to be dominant compared to standard care. These results remain relatively consistent over the different scenarios modelled. Conclusions: Implementation of AI-based imaging support significantly increases IVT and MT utilization in eligible patients, contributing to meaningful improvements in functional outcomes and long-term health quality. From a cost-effectiveness perspective, our findings support the broader clinical benefit of AI tools in enhancing stroke care delivery.

Ähnliche Arbeiten