Abstract

OBJECTIVES:

To compare the ability of two different dedicated extremity MRI (E-MRI) units and conventional radiography (CR) for identifying bone erosions in rheumatoid arthritis (RA) metacarpophalangeal (MCP) and wrist joints.

METHODS:

CR and two MRI examinations (using 0.2 T Esaote Artoscan and 0.2 T portable MagneVu MV1000 units) of 418 bones in the dominant wrist and second to fifth MCP joints of 15 patients with RA and 4 healthy controls were performed and evaluated blindly for bones being visible and for erosions.

RESULTS:

In MCP joints, MagneVu visualised 18.5% of bones entirely and 71.1% were 67-99% visualised. In wrists, MagneVu visualised 1.5% of bones entirely, 39.8% were 67-99% visualised and 19% were not visualised at all. Artoscan and CR visualised all bones entirely. Artoscan, MagneVu and CR found 22, 19 and 15 bones with erosions in MCP joints and 66, 40 and 13 bones with erosions in wrist joints, respectively. With the previously validated Artoscan unit as standard reference, MagneVu and CR had sensitivities of 0.82 and 0.55, respectively, in MCP joint bones and 0.41 and 0.14 in wrist bones. Specificities of CR and MagneVu were comparable (0.82-0.99). The MagneVu unit was particularly more sensitive than CR for metacarpal heads and carpal bones. MagneVu MRI and CR detected 100% and 89%, respectively, of large erosions (Outcome Measures in Rheumatoid Arthritis Clinical Trials-Rheumatoid Arthritis MRI Scoring System (OMERACT-RAMRIS) score >1 on Artoscan) in MCP joints and 69% and 15.8% of large erosions in wrists.

CONCLUSIONS:

Both E-MRI units detected more erosions than CR, in particular due to a higher sensitivity in metacarpal heads and carpal bones. The MagneVu unit detected fewer erosions than the Artoscan unit due to a lower average image quality and a smaller proportion of bones being visualised.