Evaluation of Deep-Learning Reconstructed High-Resolution 3D Lumbar Spine MRI to Improve Image Quality
Simon Sun1, Ek Tsoon Tan1, John A Carrino1, Douglas Nelson Mintz1, Meghan Sahr1, Yoshimi Endo1, Edward Yoon1, Bin Lin1, Robert M Lebel2, Suryanarayan Kaushik2, Yan Wen2, Maggie Fung2, and Darryl B Sneag1
1Radiology, Hospital for Special Surgery, New York, NY, United States, 2GE Healthcare, Chicago, IL, United States
Interobserver agreement for variables of interest among 3D T2w-FSE DLRecon
and SOC reconstructions and SOC 2D T2w-FSE ranged from moderate to very good,
and was similar for all three sequences. Overall image quality was
qualitatively improved on 3D T2w-FSE using DLRecon.
Fig. 2 Top Row, sagittal T2-weighted images of
the lumbar spine categorized by sequence demonstrated markedly improved image
quality using the DLRecon
algorithm for 3D imaging:
A:
2D T2 standard of care (SOC) T2-weighted fast spin echo (T2w-FSE) B: 3D
standard of care (SOC) T2w-FSE C: Deep learning reconstructed (DLREcon) 3D
T2w-FSE
Bottom Row, axial T2-weighted images of
the lumbar spine, categorized by sequence, demonstrate superior image quality
with the DLRecon
algorithm for 3D imaging:
D: 2D T2 standard of care (SOC) E: 3D SOC
T2w-FSE F: DLRecon 3D
T2w-FSE
Fig. 4 Multiplanar reformations of 3D DLRecon T2
lumbar spine in a patient with moderate scoliosis demonstrating the ability to
optimally evaluate each level using orthogonal planes.