Ultra-high temporal resolution 4D angiography using arterial spin labeling with subspace reconstruction

Area of Science:

• Medical Imaging
• Biophysics
• Radiology

Background:

• Current 4D dynamic angiograms often compromise spatiotemporal fidelity due to temporal binning of k-space data.
• Arterial spin labeling (ASL) enables non-contrast 4D angiography but faces limitations in temporal resolution.
• Achieving ultra-high temporal resolution is crucial for detailed analysis of physiological parameters and blood flow dynamics.

Purpose of the Study:

• To develop and validate an ultra-high temporal resolution non-contrast 4D angiography technique.
• To improve spatiotemporal fidelity in dynamic angiograms reconstructed from ASL data.
• To enable more accurate physiological parameter estimation from 4D angiographic data.

Main Methods:

• Utilized continuous 3D golden-angle sampling with ASL preparation for flexible 4D dynamic angiogram reconstruction.
• Developed a subspace compression method using an angiographic kinetic model to represent voxelwise signal timecourses without temporal binning.
• Estimated physiological parameters via Bayesian fitting and validated in vivo results against numerical simulations.

Main Results:

• Achieved 4D time-resolved angiography with significantly higher temporal resolution (14.7 ms) compared to previous methods (approx. 50 ms).
• Maintained high spatial resolution (1.1 mm isotropic) while improving depiction of blood flow dynamics and thin vessel visibility.
• Demonstrated more realistic spatial patterns in estimated physiological parameters.

Conclusions:

• Incorporating a subspace-compressed kinetic model into 4D ASL angiogram reconstruction substantially enhances temporal resolution and spatiotemporal fidelity.
• The improved image quality facilitates more accurate physiological modeling.
• This advancement offers a promising approach for detailed non-contrast 4D angiography.