OBJECTIVE: Our objective was to compare the sensitivity of unenhanced radiography, CT, and MR imaging in revealing subchondral fractures. SUBJECTS AND METHODS: Forty-five subjects with stage I and stage II osteonecrosis of the femoral head were included in the study as part of a multicenter clinical trial to evaluate the effectiveness of recombinant human bone morphogenetic protein as an adjuvant treatment to core decompression. Patients were evaluated with radiography, CT, and MR imaging 6 and 12 months after surgery. RESULTS: At 6 months, 18 fractures were shown on CT scans, but only 12 were detected on radiographs and six, on MR images. At 12 months, 20 subchondral fractures were detected on CT scans, but only 17 were seen on radiographs and 11, on MR images. Compared with CT, MR imaging has a sensitivity and specificity of 38% and 100%, and unenhanced radiography has a sensitivity and specificity of 71% and 97%, respectively. On T2-weighted MR images, the subchondral fractures were visualized as crescentic high-signal-intensity lines, and in all patients, on the corresponding CT scans, the fracture clearly breached the femoral cortex. CONCLUSION: CT reveals more subchondral fractures in osteonecrosis of the femoral head than unenhanced radiography or MR imaging. The high-signal-intensity line seen on T2-weighted MR images appears to represent fluid accumulating in the subchondral fracture, which may indicate a breach in the overlying articular cartilage.
OBJECTIVE: Our objective was to compare the sensitivity of unenhanced radiography, CT, and MR imaging in revealing subchondral fractures. SUBJECTS AND METHODS: Forty-five subjects with stage I and stage II osteonecrosis of the femoral head were included in the study as part of a multicenter clinical trial to evaluate the effectiveness of recombinant humanbone morphogenetic protein as an adjuvant treatment to core decompression. Patients were evaluated with radiography, CT, and MR imaging 6 and 12 months after surgery. RESULTS: At 6 months, 18 fractures were shown on CT scans, but only 12 were detected on radiographs and six, on MR images. At 12 months, 20 subchondral fractures were detected on CT scans, but only 17 were seen on radiographs and 11, on MR images. Compared with CT, MR imaging has a sensitivity and specificity of 38% and 100%, and unenhanced radiography has a sensitivity and specificity of 71% and 97%, respectively. On T2-weighted MR images, the subchondral fractures were visualized as crescentic high-signal-intensity lines, and in all patients, on the corresponding CT scans, the fracture clearly breached the femoral cortex. CONCLUSION: CT reveals more subchondral fractures in osteonecrosis of the femoral head than unenhanced radiography or MR imaging. The high-signal-intensity line seen on T2-weighted MR images appears to represent fluid accumulating in the subchondral fracture, which may indicate a breach in the overlying articular cartilage.