Diagnostic image quality of shoulder magnetic resonance imaging at 0.55T in healthy subjects and patients with rotator cuff injury: Prospective comparison with 1.5T.
Authors
Affiliations (5)
Affiliations (5)
- Department of Radiology, Peking University First Hospital, Beijing, China.
- MR Collaboration, Siemens Healthcare, Ltd., Beijing, China.
- MR Clinical Marketing, Siemens Healthineers, Beijing, China.
- Siemens Shenzhen Magnetic Resonance Ltd., Shenzhen, China.
- Department of Orthopedics, Peking University First Hospital, Beijing, China.
Abstract
Lower‑field 0.55 T MRI scanners with large‑bore design could offer patients more comfortable feeling and flexible position. With the advances in deep‑learning reconstruction and hardware optimization, 0.55T have greatly improved their image quality and cost‑effectiveness for musculoskeletal applications. However, its feasibility for shoulder evaluation remains unstudied. To evaluate the image quality of shoulder using 0.55T MRI in volunteers and patients with rotator cuff injury with comparison of conventional 1.5T. Fifty participants (25 volunteers, 25 patients) were prospectively enrolled. All participants underwent shoulder MRI both at open 0.55T with contour coils and deep learning (DL) and 1.5T MR without DL. The nonparametric method was utilized to compare image quality scores and quantitative metrics (signal-to-noise ratio (SNR) and contrast-to-noise ratio (|CNR|)). Arthroscopy have been conducted in 16 patients. Kappa statistics was used to assess interobserver and interstudy agreement for lesion evaluation. Comfort feeling was also compared. The scan time was 11'25″ for 0.55T MRI and 11'42″ for 1.5T MRI. The interobserver analyses showed high consistency (κ = 0.771-1.000, P: 0.008 ∼ < 0.001) among 3 readers. The scores for coronal fat suppressed T2 weighted imaging (fsT2WI) and coronal T1 weighted imaging (T1WI) were significantly higher for 0.55T MRI. Most sequences at 0.55T demonstrated significantly higher SNRs (P:0.04 ∼ < 0.001). Additionally, |CNR| values for lesions were significantly higher in fs proton weighted imaging (PDWI) and fsT2WI at 0.55T. Perfect interstudy agreement between 0.55T and 1.5T was observed between 2 readers (Kappa = 1, P < 0.001) for lesion assessment. Agreement between MRI findings and arthroscopic results was perfect or good. The 0.55T system received higher scores in comfort ratings. The 0.55T MRI with DL, showed comparable or better image quality in healthy subjects and patients with rotator cuff injury to 1.5T MRI without DL.