Page 87 - Imaging of Osteoarthritis and Rheumatoid Arthritis in Hand Joints
P. 87

 High-resolution MRI of cartilage in finger joints
MRI acquisition
Prior to this study, a custom-built multichannel receive coil for high-
resolution finger joint imaging was created in collaboration with Machnet
BV (Roden, The Netherlands). This coil was designed specifically for imaging
of finger joints affected by rheumatic diseases and allowed us to scan two
adjacent metacarpophalangeal (MCP) and the two corresponding proximal interphalangeal (PIP) joints in one session with high resolution on a clinical
3T MRI scanner (Discovery MR 750, GE Healthcare, Milwaukee, WI). Hand
OA patients and healthy volunteers were scanned using this coil in a prone
superman position. Patients were positioned comfortably using torso, head
and arm supports to minimize motion artefacts. The second and third MCP
and PIP joints were scanned. The entire scanning protocol consisted of a
coronal proton density (PD) and sagittal fat-suppressed spoiled gradient echo
(SPGR) images of each joint separately. Additional performed sequences were
not further used in this study. The PD sequence was a Fast recovery Fast Spin
Echo sequence with parameters: repetition time (TR) 1500; echo time (TE) 30;
echo train length (ETL) 4; Field of view (FOV) 8x8 cm; matrix size 320x320; slice 5 thickness (ST) 2mm with a 0.1 mm gap; bandwidth 41; number of excitations
(NEX) 2; with no phase wrap (NPW) and tailored radiofrequency pulse (TRF) options enabled; scan time was 4 min and 28s.The SPGR was a 3D fast SPGR with parameters: minimum TR and TE; flip angle of 30; FOV 3x2.4 cm for the PIP joints and 4x3.2 cm for the MCP joints; ST 0.8mm; matrix 320x224; bandwith 16; NEX 2; with fat suppression and zero filling interpolation processing options on. Scan time per joint was 3m and 28s.
MRI scoring systems
A face-to-face meeting and following online discussions were organized to demonstrate HOAMRIS9 to the MRI readers, modify the definitions of cartilage scoring specifically for our hrMRIs, and test the reliability. Prior to the meeting, JL identified five patients and one healthy control with different amounts of MRI pathology. MSS (radiology resident with training in reading musculoskeletal MRI), GSRM (musculoskeletal radiologist with previous RAMRIS experience) and IKH (co-developer of HOAMRIS) independently read all images. The images were read in two rounds of 3 patients each and after each round, the results were discussed to improve reliability.

   85   86   87   88   89