Discovery mr750w 3t

MRIs were acquired on a 3.0-T scanner (GE Discovery MR750 W 3 T). The same protocol, which included one T1- and three T2-weighted sequences, was performed on all participants. The T1-weighted sequence had an isotropic voxel size of 1 mm³ with a matrix size of 256 × 256 × 160 (TR/TE/TI = 8.0/3.7/450 ms, NSA = 1, flip angle = 8°). T2 and T2*-weighted sequences, with a voxel size of 1 × 1 × 3 mm, were as follows: fluid attenuation inversion recovery (FLAIR: TR/TE/TI = 11,000/90/2600 ms, flip angle = 160°), fast spin echo (TR/TE = 5000/85 ms, flip angle = 110°), and gradient echo (GRE: TR/TE = 1300/23 ms, flip angle = 15°). All scans were visually assessed for quality and incidental findings by a trained neuroradiologist [2 (link)].

Structural imaging was acquired in 67 participants, either on a research-dedicated 3 T hybrid PET-MR scanner (SIGNA (TM), GE Healthcare, Chicago, IL) at the CUB Hôpital Erasme (Brussels, Belgium) using whole-brain axial 3D T1 sequence, or on a Discovery MR750w 3 T (GE Medical Systems, Milwaukee, WI, U.S.A.) or 3 T Ingenia (Philips Medical Systems, Best, Netherlands) at Universitair Ziekenhuis Brussel (UZ Brussel) using sagittal 3D T1-weighted (T1w) MR sequence and a sagittal 3D fluid attenuated inversion recovery (FLAIR) sequence. Some patients had already undergone a brain MRI for clinical routine use or for another research protocol. In that case, the scanner used was the 1.5 T Achieva dStream (Philips Medical Systems, Best, The Netherlands) or 3 T Skyra (Siemens Medical Solutions, Pennsylvania, U.S.A.) with sagittal 3D T1-weighted (T1w) MR sequence and a sagittal 3D fluid attenuated inversion recovery (FLAIR) sequence. For one participant, MRI was performed in another hospital using a sagittal 3D T1-weighted (T1w) MR sequence. All MRI scans were performed within 1 year after neuropsychological testing, except for one MCI patient who had the MRI 14 months after neuropsychological testing.

T₁-weighted structural brain MRI images were acquired from each participant using 3 T MRI (GE Healthcare Discovery MR750W 3 T, Milwaukee, USA). Images were acquired with a 24-channel head coil using an IR-SPGR sequence (TA = 9 m 59 s, TR = 3 s, TE = 3.1 ms, TI = 400 ms, flip angle = 11°, field of view = 256 mm × 256 mm, image matrix = 256 × 256, 1 mm³ isotropic voxels).

MR acquisitions were all performed on 2 imaging machines (MRI) from the same manufacturer (General Electric^®, Milwaukee, USA): Optima MR450w 1.5T and Discovery MR750w 3T. MRI data included at least: a post-contrast (gadoterate meglumine Dotarem, Guerbet, Villepinte, France), a three-dimensional T1-weighted fast spoiled gradient recalled (FSPGR) acquisition (post-contrast 3DT1), post-contrast 3DT1, and fat-suppressed FLAIR images. Only MR images were used as inputs of the radiomics classifier. To ensure image quality, neuroradiologists analysed all the available imaging sequences. Table 2 details the MRI parameters for both machines.

Orbital MRI was performed without the use of quasi-coronal MRI [17 (link)]. The patient’s head was immobilized, and their eyes were closed. MRI scans were acquired by Discovery MR750w 3T (GE Healthcare, Waukesha, WI, USA) using two-dimensional fast spin echo in a T2 sequence without fat suppression and MAGNETOM Skyra 3T (Siemens Healthineers, Erlangen, Germany) using two-dimensional turbo spin echo in a T2 sequence. The parameters of acquisition for the Discovery MR750w 3T were as follows: repetition time, 3600 ms; echo time, 102 ms; number of excitation, 2, 16 slices; slice thickness, 3 mm; slice gap, 1 mm; field of view, 160 mm; acquisition matrix, 320 × 224; reconstruction matrix, 512 × 512; acquisition time, 2 min 17 s. The parameters of acquisition for the MAGNETOM Skyra 3T were as follows: repetition time, 4600 ms; echo time, 64 ms; number of excitation, 2, 15 slices; slice thickness, 3 mm; slice gap, 1 mm; field of view, 160 mm; acquisition matrix, 320 × 240; reconstruction matrix, 640 × 480; acquisition time, 2 min 47 s.

MRI scans were acquired on a 3.0 T scanner (GE Discovery MR750 W 3 T) using the same protocol for all participants, which included one T1- and three T2- weighted sequences. The 3D-T1w sequence was designed with an isotropic voxel size of 1 mm³ and a matrix size of 256x256x160 (TR/TE/TI = 8.0/3.7/450 ms, NSA = 1, flip angle = 8°). Three 3D-T2w sequences, with a voxel size of 1 mm × 1 mm × 3 mm, were also used: fluid attenuation inversion recovery (FLAIR: TR/TE/TI = 11,000/90/2600 ms, flip angle = 160°), fast spin echo (FSE: TR/TE = 5000/85 ms, flip angle = 110°) and, gradient echo (GRE: TR/TE = 1300/23 ms, flip angle = 15°).
All scans were visually assessed to verify their quality and to detect incidental findings by a trained neuroradiologist and have been reported elsewhere (Brugulat-Serrat et al. 2017 (link)). In this study, ten participants were excluded due to the presence of a meningioma, as well as 37 participants due to susceptibility, motion artefacts or segmentation problems, resulting in a total of 561 images available for subsequent analysis. The medial temporal lobe atrophy was assessed by Medial Temporal Atrophy scale (Scheltens et al. 1992 (link)).