A comparison of dedicated 1.0 T extremity MRI vs large-bore 1.5 T MRI for semiquantitative whole organ assessment of osteoarthritis: the MOST study

Objective: To date semiquantitative whole-organ scoring of knee osteoarthritis (OA) relies on 1.5 Tesla (T) Magnetic resonance imaging (MRI) systems. Less costly 1.0 T extremity systems have been introduced that offer superior patient comfort, but may have limitations concerning field-of-view and image quality. The aim of this study was to compare semi-quantitative (SQ) scoring on a 1.0 T system using 1.5 T MRI as the standard of reference.

Methods: The Multicenter Osteoarthritis Study (MOST) is a longitudinal study of individuals who have or are at high risk for knee OA. A sample of 53 knees was selected in which MRI was performed on a 1.0 T extremity system as well as on a 1.5 T scanner applying a comparable sequence protocol. MRIs were read according to the Whole Organ Magnetic Resonance Imaging Score (WORMS) score. Agreement was determined using weighted kappa statistics. Sensitivity, specificity and accuracy were assessed using the 1.5 T readings as the reference standard. In addition the number of non-readable features was assessed.

Results: Agreement (w-kappa) for seven main WORMS features (cartilage, bone marrow lesions (BMLs), osteophytes, meniscal damage and extrusion, synovitis, effusion) ranged between 0.54 (synovitis) and 0.75 (cartilage). Sensitivity ranged between 68.1% (meniscal damage) and 88.1% (effusion). Specificity ranged between 63.6% (effusion) and 96.4% (BMLs). Although the overall rate of non-readable features was very low, it was higher for the 1.0 T system (1.9% vs 0.2%).

Conclusions: Semiquantitative whole organ scoring can be performed using a 1.0 T peripheral scanner with a moderate to high degree of agreement and accuracy compared to SQ assessment using a 1.5 T whole body scanner. Our results are comparable to the published inter- and intra observer exercises obtained from 1.5 T systems. Sensitivity to change of longitudinal scoring was not evaluated in this cross-sectional design and should be investigated in future validation studies.