Development of Automated Vessel Analysis Platform for Middle Meningeal Artery Based on Magnetic Resonance Angiography

Motivation: Accurate assessment of middle meningeal artery (MMA) is crucial for investigating cerebrovascular disorders; however, automated analysis of smaller but anatomically significant arteries is challenging. Goal(s): This study aimed to develop automated vessel analysis platform for segmenting, labeling, and quantifying MMA using MRA. Approach: The automated procedure included segmentation, labeling, and quantification of vessel morphology and diameter. The platform was trained on 70 cases and tested on 30 cases, employing 3D U-Net and PointNet++ for segmentation and labeling. Results: The platform achieved high accuracy and efficiency in MMA quantitative assessment (segmentation accuracy = 0.92, labeling accuracy = 0.84), supporting clinical diagnosis of MMA-related diseases. Impact: This study developed an automated vessel analysis platform capable of segmenting, labeling, and quantifying middle meningeal arteries (MMA) and offered insights into future applications in the diagnosis and treatment of MMA-related diseases.