Deep learning-enhanced zero echo time MRI for glenohumeral assessment in shoulder instability: a comparative study...

Area of Science:

• Radiology
• Medical Imaging
• Orthopedics

Background:

• Shoulder instability often involves bone loss requiring accurate assessment.
• Conventional zero echo time (ZTE) MRI has limitations in image quality and lesion conspicuity.
• Computed tomography (CT) is the gold standard for bone loss measurement but involves radiation exposure.

Purpose of the Study:

• To compare the image quality and lesion conspicuity of deep learning (DL) reconstructed ZTE MRI against conventional ZTE MRI.
• To evaluate the performance of DL-reconstructed ZTE MRI for bone loss measurements in shoulder instability compared to CT.

Main Methods:

• Forty-four patients with anterior glenohumeral instability underwent same-day ZTE MRI and CT.
• ZTE MRI datasets were reconstructed using conventional and DL algorithms.
• Musculoskeletal radiologists assessed image quality and lesion conspicuity using Likert scales.
• Glenoid track and Hill-Sachs interval measurements were performed on DL ZTE MRI and CT datasets.

Main Results:

• DL-reconstructed ZTE MRI demonstrated significantly improved perceived resolution and glenoid lesion conspicuity compared to conventional ZTE MRI.
• Inter-reader agreement for image quality assessments was substantial.
• DL-reconstructed ZTE MRI showed near-perfect agreement with CT for bone loss measurements (Glenoid track and Hill-Sachs interval).

Conclusions:

• Deep learning-based reconstruction enhances ZTE MRI image quality for glenohumeral instability assessment.
• DL-reconstructed ZTE MRI provides osseous evaluation and quantification equivalent to CT.
• This technique may simplify preoperative workflows and reduce CT radiation exposure.