Deep Learning CAIPIRINHA-VIBE Improves and Accelerates Head and Neck MRI

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Neck MRI is crucial for diagnosing various conditions.
• Conventional MRI sequences can be time-consuming.
• Deep learning reconstruction offers potential for accelerated imaging.

Purpose of the Study:

• To evaluate the image quality of deep learning-reconstructed VIBE sequences (DL4-VIBE and DL6-VIBE) for contrast-enhanced neck MRI.
• To compare these novel sequences against standard VIBE techniques.

Main Methods:

• Prospective study on 68 patients undergoing 3-Tesla neck MRI.
• Acquisition of four CE fat-saturated VIBE sequences: Star-VIBE, VIBE, DL4-VIBE, and DL6-VIBE.
• Radiologist evaluation using a 5-point Likert scale and objective SNR/CNR measurements.

Main Results:

• DL4-VIBE demonstrated superior overall image quality, muscle contour delineation, mucosa conspicuity, and reduced motion artifacts compared to other sequences.
• DL4-VIBE achieved significantly higher SNR and CNR.
• DL6-VIBE showed improvements over standard VIBE in specific aspects despite being inferior to DL4-VIBE.

Conclusions:

• Deep learning-reconstructed DL4-VIBE significantly enhances image quality in contrast-enhanced neck MRI.
• DL4-VIBE achieves this improvement with a fraction of the scan time of conventional sequences.