Automated MRI-Based Segmentation of the Cranio-Cervical Venous System Using a Dedicated Deep Learning Pipeline

A dedicated nnU-Net v2 deep-learning pipeline segments the cranio-cervical venous system on contrast-enhanced MRI with a mean Dice similarity coefficient of 0.94.

• Why it matters: Automated CCVS segmentation is needed because no dedicated tool exists for this complex venous network, limiting large-scale anatomical studies.
• What they did: In a retrospective single-center cohort of 194 contrast-enhanced cranio-cervical MRI exams (2019–2024), 50 manual 3D Slicer references trained an iterative nnU-Net v2 model, evaluated by Dice similarity coefficient.
• The result: Five-fold cross-validation on 106 reference-labeled examinations yielded mean DSC 0.94 (95% CI, 0.93–0.95), cutting manual work from ~10 hours to <10 minutes and enabling automated 3D reconstructions in VenousNetworkAnalyzer.

AI summary of the abstract · may contain errors · not medical advice · read the paper.