Vessel Wall Imaging in 1.5 T MRI Using Deep Learning Reconstruction: Prospective Evaluation of Interchangeability...

Area of Science:

• Radiology and Medical Imaging
• Neuroimaging
• Vascular Imaging

Background:

• Vessel wall magnetic resonance imaging (MRI) is crucial for diagnosing intracranial vascular diseases.
• Current clinical implementation often necessitates 3 T scanners, limiting accessibility.
• Established sequences include T1 3D Sampling Perfection with Application optimized Contrast using different flip angle Evolution (SPACE).

Purpose of the Study:

• To evaluate the interchangeability of a 1.5 T MRI sequence with deep learning reconstruction against a 3 T MRI sequence.
• To compare image quality and diagnostic confidence between the two MRI systems.
• To assess the potential for improved clinical availability of vessel wall imaging.

Main Methods:

• Prospective examination of 50 participants using T1-weighted turbo spin echo sequences.
• Comparison of a 1.5 T sequence with deep learning reconstruction and a 3 T sequence (T1 3D SPACE).
• Radiologist assessment of pathologic vessel wall enhancement, image quality, and diagnostic confidence using Likert scales.

Main Results:

• The 1.5 T and 3 T sequences demonstrated interchangeability for detecting pathologic vessel wall enhancement (IEI = 0.008).
• High interreader agreement was observed for both 1.5 T (kappa = 0.72) and 3 T (kappa = 0.78) sequences.
• While image quality and diagnostic confidence were rated superior at 3 T, the 1.5 T sequence provided diagnostic quality in all cases.

Conclusions:

• Deep learning reconstruction enables diagnostic vessel wall imaging at 1.5 T MRI.
• The 1.5 T sequence is interchangeable with established 3 T sequences for detecting pathologic vessel wall enhancement.
• This technological advancement promises to significantly increase the clinical availability of vessel wall imaging.