Evaluation of vessel tracking and 3D roadmapping software for endovascular treatment of brain arteriovenous malformations

Endovascular treatment of brain arteriovenous malformations (AVMs) is complex due to the intricate angioarchitecture, requiring accurate identification of feeding arteries, draining veins, and the nidus. While 3D digital subtraction angiography (DSA) aids visualization, it lacks vessel-specific segmentation. Embo ASSIST (GE HealthCare, USA) is a software designed to enhance procedural planning and real-time guidance through automatic segmentation and 3D roadmap visualization. This study evaluates the feasibility and clinical utility of Embo ASSIST in brain AVM embolization, focusing on its impact on procedural planning, navigation, radiation exposure, and contrast media use. A retrospective analysis of 19 procedures in 14 patients was conducted between February and August 2021. Images from biplane DSA and cone-beam computed tomography (CBCT) were processed with Embo ASSIST for vascular segmentation. A structured survey, rated by two senior neuroradiologists, assessed the software's effectiveness. Results showed Embo ASSIST improved procedural planning, with a better understanding of angioarchitecture (3.8/5), pedicle definition (3.9/5), and embolization pathway selection (3.9/5). It allows a reduction of contrast media use (3.7/5), radiation exposure (3.8/5), and procedure time (3.8/5) according to the operators. The software was highly rated for ease of use (3.8/5), CBCT processing speed (4.2/5), and integration with fluoroscopic navigation (4.3/5). However, vein segmentation was limited (3.0/5), requiring manual corrections. In conclusion, Embo ASSIST significantly improved AVM embolization procedures, enhancing visualization, navigation, and efficiency. Future advancements in venous segmentation and AI-based dynamic vessel analysis could optimize treatment further.