perfDSA: Automatic Perfusion Imaging in Cerebral Digital Subtraction Angiography.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Cerebrovascular Imaging

Background:

• Cerebral digital subtraction angiography (DSA) is crucial for visualizing cerebral blood flow in image-guided interventions.
• Current visual assessment of DSA perfusion is subjective, time-consuming, and prone to errors.
• There is a need for automated, quantitative methods to assess cerebral perfusion.

Purpose of the Study:

• To develop and validate a fully automatic and quantitative framework for perfusion DSA.
• To enable fast and reproducible assessment of cerebral perfusion characteristics.
• To improve therapeutic decision-making in cerebrovascular interventions.

Main Methods:

• The perfDSA framework was developed for automatic deconvolution-based perfusion parametric imaging from cerebral DSA.
• It automatically extracts the arterial input function (AIF) from the supraclinoid internal carotid artery (ICA).
• Perfusion parameters including cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and Tmax are computed.

Main Results:

• The perfDSA framework achieved a Dice score of 0.73 (±0.21) in segmenting the supraclinoid ICA on a dataset of 1006 patients.
• The accuracy of extracted arterial input function (AIF) curves was comparable to manual extraction.
• Extracted perfusion images showed statistically significant associations with favorable functional outcomes in stroke patients (P=2.62e-5).

Conclusions:

• The perfDSA framework offers a promising tool for aiding therapeutic decision-making in cerebrovascular interventions.
• It facilitates the discovery of novel quantitative biomarkers in clinical practice.
• The developed framework supports automated and quantitative analysis of cerebral perfusion from DSA.