A methodology for generating four-dimensional arterial spin labeling MR angiography virtual phantoms

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Biology

Background:

• Four-dimensional arterial spin labeling magnetic resonance angiography (4D ASL MRA) provides non-invasive cerebrovascular analysis but generates large datasets requiring complex processing.
• Evaluating medical image processing methods for 4D ASL MRA necessitates annotated data, which is costly and time-consuming to acquire.
• Virtual simulations offer a faster, more cost-effective alternative for the initial assessment of these image processing techniques.

Purpose of the Study:

• To propose a methodology for generating annotated virtual phantoms for 4D ASL MRA.
• To simulate realistic cerebrovascular geometries and blood flow dynamics for virtual phantom creation.
• To evaluate the utility of these virtual phantoms in assessing the accuracy of blood flow parameter estimation methods.

Main Methods:

• Generation of annotated 4D ASL MRA virtual phantoms using real cerebrovascular geometries from healthy volunteers.
• Simulation of blood flow using a mathematical model tailored for 4D ASL MRA signal characteristics.
• Incorporation of realistic noise using a homomorphic approach to mimic multi-coil acquisition noise.

Main Results:

• Virtual phantoms were generated across various scenarios with different acquisition parameter settings.
• The accuracy of blood flow parameter estimation (relative blood volume, transit time) was evaluated using the generated phantoms.
• Higher accuracy in blood flow parameter estimation was observed with longer magnetic labeling times and increased temporal resolution.

Conclusions:

• The developed methodology provides a valuable tool for creating annotated virtual phantoms for 4D ASL MRA.
• These virtual phantoms serve as an effective alternative for the initial evaluation of medical image processing methods.
• The findings highlight the impact of acquisition parameters on the accuracy of hemodynamic measurements derived from 4D ASL MRA data.