Enhanced Visualization of Intracranial Cortical Arteries Using Deep Learning Reconstruction in Vessel Wall MR Imaging

Area of Science:

• Neuroimaging
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Cerebral arterial system visualization is crucial for diagnosing neurological conditions.
• Current vessel wall imaging (VWI) techniques face challenges in resolving small vessels like cortical arteries.
• Deep learning reconstruction (DLR) offers potential for improving image quality in medical imaging.

Purpose of the Study:

• To evaluate the effectiveness of deep learning reconstruction (DLR) in enhancing vessel wall imaging (VWI) of the entire cerebral arterial system.
• To assess DLR's utility in visualizing intracranial cortical arteries with high resolution.
• To compare image quality with and without DLR using both qualitative and quantitative metrics.

Main Methods:

• Seventeen patients underwent post-contrast 3D T1WI-CUBE VWI at 0.5 mm isotropic resolution.
• Images were acquired both with and without deep learning reconstruction (DLR).
• Qualitative assessment rated 29 arterial segments; quantitative analysis measured SNR and CNR of the vertebral artery wall.

Main Results:

• Deep learning reconstruction (DLR) significantly improved overall image quality compared to standard VWI.
• Cortical arteries were rated as optimal in all cases utilizing DLR (P < 0.001).
• Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were significantly higher with DLR (P = 0.004).

Conclusions:

• Deep learning reconstruction (DLR) substantially enhances the visualization of the intracranial cerebral arterial system, including cortical arteries.
• DLR enables high-resolution vessel wall imaging (VWI) within clinically feasible scan times.
• This advanced imaging technique holds promise for improved diagnosis and management of cerebrovascular diseases.