Abstract 018: Preprocedural Prediction Of Impossible Transfemoral Access In Stroke Mechanical Thrombectomy Using Graph Neural Networks

Transfemoral access (TFA) for endovascular thrombectomy is unsuccessful due to complex arterial anatomies that prevent catheter access to the occlusion site in approximately 3‐5% of cases. Radial access can serve as an effective alternative, but delays during the procedure may negatively impact patient outcomes. Recent advancements in image processing algorithms enable detailed characterization of vascular pathways from baseline neuroimaging, offering the potential to predict TFA challenges before arterial puncture.