Composite MRA: statistical approach to generate an MR angiogram from multiple contrasts

Area of Science:

• Medical Imaging
• Radiology
• Image Processing

Background:

• Magnetic Resonance Angiography (MRA) is crucial for visualizing blood vessels.
• Sequence-specific artifacts and limitations in vessel depiction can hinder accurate MRA analysis.
• Combining information from different MRI contrasts offers potential to overcome these limitations.

Purpose of the Study:

• To develop a novel composite MRA method integrating multiple MRI contrasts.
• To reduce sequence-specific artifacts and enhance vessel depiction in MRA.
• To improve the accuracy of vascular segmentation.

Main Methods:

• A Bayesian approach was used to determine vessel probability from Black Blood (BB), CE-MRA, and PC-MRA data.
• Composite angiograms were generated by weighting voxel-wise probabilities with a confidence measure.
• Segmentation accuracy of composite MRA was evaluated against CE-MRA using Dice Similarity Coefficient (DSC) for internal carotid artery (ICA) segments.

Main Results:

• Composite MRA effectively suppressed venous contamination, flow artifacts, and velocity aliasing.
• Significantly improved segmentation accuracy for ICA segments compared to CE-MRA (DSC 0.908 vs. 0.765).
• Demonstrated conservative changes in vessel appearance with threshold variations, though small or weakly visible vessels were less consistently depicted.

Conclusions:

• Composite MRA leverages complementary information from multiple MRI contrasts to reduce artifacts and improve vessel contrast.
• This technique shows promise for semi-automatic segmentation of challenging vessels affected by artifacts.
• Further refinement may be needed for optimal depiction of small or weakly contrasted vessels.