Feasibility of MRI-based carotid wall shear stress mapping in a population cohort

A semi-automated CFD pipeline using deep-learning MRI assessed carotid wall shear stress in 100 Rotterdam Study participants. Lumen segmentation Dice 0.89±0.02 and inter-rater ICC 0.89–0.98, demonstrating feasibility for large-scale studies.

• Pipeline integrated deep-learning lumen segmentations from BlackBlood MRI with participant-specific 3D phase-contrast MRA flow data.
• Intra- and inter-rater reliability showed good to excellent reproducibility (ICC range 0.89–0.98).
• Automated segmentations achieved Dice similarity coefficient of 0.89 ± 0.02 compared to manual segmentation.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.