Validation and feasibility of fast knee MRI using a deep learning-assisted 3D iterative image enhancement system

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Fast magnetic resonance imaging (MRI) enhances patient comfort and efficiency but can reduce image quality.
• High-resolution knee MRI is crucial for accurate diagnosis.
• Deep learning offers potential solutions for improving fast MRI techniques.

Purpose of the Study:

• To evaluate the feasibility of a deep learning-assisted three-dimensional iterative image enhancement (DL-3DIIE) system for high-resolution fast knee MRI.
• To compare the image quality of DL-3DIIE MRI with conventional and standard fast MRI protocols.

Main Methods:

• A prospective study involving 134 participants undergoing knee MRI.
• Comparison of three protocols: conventional MRI, fast MRI, and DL-3DIIE MRI.
• Objective assessment using signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), peak signal-to-noise ratio (PSNR), and multi-scale structural similarity index (MS-SSIM).
• Subjective assessment of lesion conspicuity, margin delineation, and diagnostic confidence.

Main Results:

• DL-3DIIE MRI demonstrated significantly higher SNRs and CNRs compared to conventional and fast MRI (P<0.05).
• Quantitative metrics (PSNR, MS-SSIM) were significantly improved with DL-3DIIE MRI over fast MRI (P<0.001).
• Subjective assessments showed DL-3DIIE MRI provided higher scores for lesion conspicuity, margin delineation, and diagnostic confidence (P<0.05).

Conclusions:

• DL-3DIIE MRI significantly enhances image quality compared to conventional and fast MRI.
• The system improves objective metrics (SNR, CNR, PSNR, MS-SSIM) and subjective diagnostic quality.
• DL-3DIIE MRI shows promise for accelerating knee MRI scans without compromising diagnostic performance.