European neurology · 2026
Artificial Intelligence-Based CTA Software for Real-World Detection of Large-Vessel Occlusion in Acute Ischemic Stroke.
СтатьяНаблюдательное исследованиеНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Morelli Nicola, Rota Eugenia, Spallazzi Marco, Biondi Marina, Immovilli Paolo
Аннотация
INTRODUCTION: Artificial intelligence (AI)-based tools for CT angiography (CTA) have been introduced to support rapid detection of large vessel occlusion (LVO) in acute ischemic stroke. Although regulatory approval has relied mainly on controlled clinical studies, evidence from routine clinical practice remains limited. This study aimed to assess the real-world diagnostic performance and workflow impact of the Brainomix e-CTA software.
METHODS: We conducted a retrospective, single-center observational study including consecutive patients with suspected acute ischemic stroke who underwent multiphase CTA between May 2023 and May 2025. Automated LVO detection using e-CTA was compared with expert neuroradiologist interpretation as a pragmatic clinical reference standard. Diagnostic performance was evaluated using sensitivity, specificity, and accuracy. Sensitivity was additionally analyzed according to occlusion location. Inter-reader agreement was assessed using Cohen's κ, and time to diagnostic conclusion with and without AI assistance was compared across readers with different experience levels.
RESULTS: Among 531 CTA examinations, LVO was identified in 19% of cases. Overall LVO detection showed a sensitivity of 84%, specificity of 95%, and accuracy of 93%. Sensitivity was highest for ICA/proximal M1 occlusions (94%) and progressively decreased for intermediate (87%) and distal M1 occlusions (73%). Agreement between AI output and expert interpretation was excellent for ICA/proximal M1 occlusions (κ = 0.91) and moderate for distal M1 occlusions (κ = 0.48). AI assistance significantly reduced time to diagnostic conclusion across all reader profiles.
CONCLUSION: In routine clinical practice, Brainomix e-CTA showed highest diagnostic performance for ICA/proximal M1 occlusions and reduced time to diagnostic conclusion. However, lower sensitivity for intermediate and distal M1 occlusions and the exclusion of nontarget occlusions from the primary endpoint indicate that e-CTA should be used as an adjunctive workflow tool rather than as a standalone method for treatment decision-making.
METHODS: We conducted a retrospective, single-center observational study including consecutive patients with suspected acute ischemic stroke who underwent multiphase CTA between May 2023 and May 2025. Automated LVO detection using e-CTA was compared with expert neuroradiologist interpretation as a pragmatic clinical reference standard. Diagnostic performance was evaluated using sensitivity, specificity, and accuracy. Sensitivity was additionally analyzed according to occlusion location. Inter-reader agreement was assessed using Cohen's κ, and time to diagnostic conclusion with and without AI assistance was compared across readers with different experience levels.
RESULTS: Among 531 CTA examinations, LVO was identified in 19% of cases. Overall LVO detection showed a sensitivity of 84%, specificity of 95%, and accuracy of 93%. Sensitivity was highest for ICA/proximal M1 occlusions (94%) and progressively decreased for intermediate (87%) and distal M1 occlusions (73%). Agreement between AI output and expert interpretation was excellent for ICA/proximal M1 occlusions (κ = 0.91) and moderate for distal M1 occlusions (κ = 0.48). AI assistance significantly reduced time to diagnostic conclusion across all reader profiles.
CONCLUSION: In routine clinical practice, Brainomix e-CTA showed highest diagnostic performance for ICA/proximal M1 occlusions and reduced time to diagnostic conclusion. However, lower sensitivity for intermediate and distal M1 occlusions and the exclusion of nontarget occlusions from the primary endpoint indicate that e-CTA should be used as an adjunctive workflow tool rather than as a standalone method for treatment decision-making.