The Circle of Willis (CW), visualized via Magnetic Resonance Angiography (MRA), is crucial for assessing cerebral circulation. Accurate artery identification is essential not only for detecting stenosis and pathological changes but also for understanding vascular adaptations in healthy aging. Manual CW assessment is time-consuming, necessitating automated alternatives. This study evaluates intracranial artery diameter estimations from the Express IntraCranial Arteries Breakdown (eICAB) toolbox against manual measurements. eICAB was tested on 631 participants from the Northern Manhattan Study (NOMAS) with 1.5T MRA images (0.293 × 0.293 × 1 mm resolution). We analyzed eICAB's detection and diameter estimation accuracy of the Internal Carotid (ICA), Basilar (BA), Anterior Cerebral (ACA), Middle Cerebral (MCA), Posterior Cerebral (PCA), and Posterior Communicating (PCom). eICAB showed over 95% accuracy in detecting major arteries except for PCA and PCom (∼80%). Diameter discrepancies were generally ≤0.5 mm, with ICA and BA reaching 1 mm. Spearman correlation (p ≪ 0.05) confirmed strong agreement between automated and manual measurements. Resampling at 0.2083 mm improved precision. eICAB accurately identifies CW arteries and estimates diameters, demonstrating strong clinical and research potential.
Keywords: Automated vascular segmentation; Circle of Willis; cerebral artery validation; cerebral diameter estimation; cerebral diameter validation.