Abstract TMP67: Automated Measurement Of Vascular Tortuosity: Association With Baseline Clinical Characteristics And Impact On Thrombectomy Procedures

Abstract

Introduction: Vascular tortuosity is considered one of the main barriers for mechanical thrombectomy in stroke. We developed a fully automatic algorithm for vascular tortuosity characterization and explored the association of several anatomical features with baseline characteristics and procedural outcomes. Methods: An automatic software integrating segmentation, centerline extraction and vessel labelling, was used to extract features directly from computed tomography angiograms (CTA) of stroke patients undergoing transfemoral endovascular treatment. Mean diameter and tortuosity index (TI) of the carotid arteries and brachiocephalic trunk were used as relevant tortuosity indicators. Patients were compared according to seven different clinical conditions and four procedural measures. Results: 518 patients (51% female, 74.8±12.9 years) were included in this study. Women presented significantly smaller vessels and increased TI in both carotid arteries ( p <.01). Hypertension and atrial fibrillation were associated with higher TI in both carotid arteries ( p <.01). Age correlated with TI in all vessels (.35< r< .52, p <.01). Patients with longer time to first angiographic run and time to revascularization presented higher TI in all ipsilateral carotid segments ( p <.05). Conclusion: An automated software was successfully used to characterize anatomical vascular features from CTA. Gender, age, hypertension and atrial fibrillation were found to influence vascular tortuosity. The algorithm was able to detect higher TI in ipsilateral carotid arteries of patients with further difficult catheter access or longer time to revascularization. This information could potentially be used to plan best arterial access site.