MRI-based semiquantitative assessment of subchondral bone marrow lesions in osteoarthritis research

Abstract

We would like to comment on magnetic resonance imaging (MRI) assessment of bone marrow lesions (BMLs) in osteoarthritis (OA) research from a radiologic perspective. Selection of pulse sequences that are able to depict the maximum extent of BMLs is of paramount importance for subsequent correct assessment of the size of these lesions. A recently published article investigating a canine OA model by MRI suggested that to accurately assess semiquantitative size and extent of BMLs a combination of different pulse sequences should be used, and that this information could also apply to human clinical studies 1 . In that study the authors used a protocol of T1-weighted threedimensional (3D) fast gradient recalled echo (GRE), sagittal fat-suppressed (FS) 3D spoiled gradient echo at a steady state (SPGR), and sagittal T2-weighted fast spin echo (FSE) with fat saturation sequences and assessed BMLs using a semiquantitative scoring approach. Another article by the same research group published in ‘‘Annals of the Rheumatic Diseases’’ recently investigated subchondral bone lesions with a quantitative MRI approach 2 . Acquisition of BMLs in that publication was performed on a 1.5 T system using a single 3D fast imaging with steady state precession (FISP) e sequence, a GRE-type sequence. While it has been shown that GRE-type sequences such as SPGR, fast low angle shot (FLASH), 3-point Dixon, double echo steady state (DESS) and others are very sensitive in delineating subchondral cysts, ‘‘even with robust FS or WE [water excitation], these sequences are notoriously