Imaging The Canine Spine With Space, A 3d T2-Weighted Spin Echo Sequence With Variable Flip Angle Refocusing .

Shaikh LS, Holmes SP, Selberg KT.

in Scientific Proceedings (Abstract). American College of Veterinary Radiology 2014.

Introduction/Purpose: Three-dimensional volume imaging in MRI has the advantage of multiplanar reconstructions with a single acquisition , which makes shortened scan times a reality. Rapid volume imaging has been limited to gradient-recalled echo (GRE) sequences. Fast-spin echo (FSE) sequences are preferred for neuroimaging due to superior image contrast and fewer artifacts. New FSE sequences using variable flip angle refocusing and very long echo train lengths allow 3D volumetric scanning. The goal of this study was to optimize the 3D Sampling Perfection with Application optimized Contrasts using different flip angle Evolution (SPACE) sequence in canine spine MR and compare it to conventional T2-weighted (T2w) FSE sequences.

Methods: Seven young adult, healthy and neurologically normal dogs were examined post-mortem using a 1.5T Siemens Symphony with TIM® MRI unit. Conventional T2w FSE sequences in the sagittal and transverse planes and a sagittal plane T2w SPACE sequence were acquired centered on the thoracolumbar junction. The imaging protocol was set to have similar scan times for the sagittal plane series and so the resultant transverse plane pixel resolution was similar. Transverse image reconstructions were made from the sagittal 3D SPACE images to match the location of the FSE transverse images , with the following slice thickness/gap: 3mm/3mm , 2mm/2mm , 1mm/1mm, 3mm/1mm, 1mm/0.5mm. Signal (SNR – spinal cord) and contrast (CNR – spinal cord:muscle) to noise ratios were calculated and compared for all sequences using repeated measures. Transverse 3D SPACE and 20 FSE images were individually scored by 3 reviewers for overall image quality and visibility of specific neuromuscular structures using a 5-point scale. Each reviewer’s preferred 3D SPACE transverse reconstruction slice thickness/gap was also evaluated with repeated measures on 3 days temporally separated by at least one week.

Results: 3D SPACE scan time was 6min. 40sec. The total scan time for sagittal and transverse 20 FSE scans was 10min. 46sec. SNR and CNR was higher for sagittal 20 FSE compared to 3D SPACE (p<0.0001). SNR and CNR for all 3D SPACE transverse reconstructions were higher than transverse 20 FSE (p<0.01). The SNR and CNR values of the 3mm and 2mm transverse 3D SPACE reconstructions were very similar, both being higher than the 1mm slices, but still were significantly different. 20 FSE transverse sequences were consistently less highly rated compared to 3D SPACE reconstructions (odds ratio 95% CI 0.060-0.859) for overall diagnostic quality and visibility of most anatomic structures. The 20 FSE was preferred for epidural space and grey-white matter distinction. No statistically significant preference was identified within or amongst reviewers for the SPACE transverse reconstructions.

Discussion/Conclusion: The 3D SPACE pulse sequence with multiplanar and thin- slice reconstructions is a viable diagnostic MR tool for the canine spine. With fast acquisition times, clinical application of SPACE will rival computed tomography (CT) exams, while providing superior soft tissue contrast that is ideal for spine imaging.