To compare magnetic resonance (MR) imaging, computed tomography (CT) and radiography for the detection of arthropathies in patients with haemophilia. Forty-one symptomatic joints in 14 men with haemophilia, ages 11-24 years (mean age 17.5 +/- 3.9 years) were examined with radiography, CT and MR. Images of each joint were acquired on the same day. The precontrast MR examination obtained coronal spin-echo T1 images and gradient echo with rephasing T2* images, as well as sagittal, axial gradient echo with rephasing T2* images using a low-field-strength 0.3-T MR imager. For the postcontrast MR examination, coronal, sagittal and axial images were acquired using the same precontrast T1 sequence. Thirteen joints were also examined on enhanced MR. The severity of damage was classified using conventional radiographical staging. Severely affected haemophilic arthropathy (HA) patients (stage 5) were excluded. Findings of soft tissue swelling, osteoporosis, epiphyseal overgrowth, joint erosion, cysts, joint space narrowing, bone marrow oedema, joint effusion, haemorrhage, synovial hypertrophy and widened intercondylar notches as well as anterior and posterior cruciate ligaments (for the knee) were used in all imaging comparisons. The joints were classified by radiographical criteria into stage 0 (n = 5), stage 1 (n = 7), stage 2 (n = 6), stage 3 (n = 8) and stage 4 (n = 15). Soft tissue swelling or joint effusion was observed in 33 joints on radiographs, in 34 on both CT and MR; joint erosions were observed in 34 joints on MR, 33 on CT and 20 on radiographs. Joint cysts appeared in 21 joints on MR, 18 on CT and 9 on radiographs. Significant differences in detection of erosion and cysts were found between radiography and CT (P < 0.05) and radiography and MR imaging (P < 0.05), not between CT and MR (P > 0.05). MR was better for detecting foci of both erosion and cysts than CT and radiography, and CT was better than radiography. MR imaging, CT and radiography were equally effective in showing the changes of epiphyseal overgrowth in 26 joints, joint space narrowing in 14 joints and widened intercondylar notches in 20 knee joints. However, only MR imaging detected tears in 17 anterior and 13 posterior cruciate ligaments in the 20 knee joints with widened intercondylar notches. Bone marrow oedema in 14 joints, haemorrhage in 34 joints and synovial hypertrophy in 27 joints were seen on MR images, but not on CT or radiography. MR imaging is superior to CT and conventional radiography for detecting abnormal changes and should be considered the first choice among imaging modalities in evaluating HAs.
