Comparison of Radiographic and Pathologic Diagnosis of Osteonecrosis of the Femoral Head

OBJECTIVE. The purpose of this study was to assess the utility of radiography in diagnosing osteonecrosis of the femoral head with pathologic examination as the reference standard. MATERIALS AND METHODS. Radiography and pathology reports of 253 consecutive femoral head resections were reviewed. A subset of 128 cases in which the diagnosis of osteonecrosis was made or suggested radiographically or pathologically were reviewed to evaluate for factors that might influence correlation. A total of 23 patients in this subset had also undergone MRI, and those reports and images were reviewed. RESULTS. There was 93.9% agreement between radiography and pathologic examination overall (k = 0.67). When grade 3 osteoarthritis was present, 95.0% agreement was found, but because of the large number of patients with severe osteoarthritis, the kappa value decreased to 0.51. In the subset of cases in which osteonecrosis was diagnosed or suspected, radiologic-pathologic correlation decreased as osteoarthritis grade increased, and the diagnostic uncertainty for both evaluation methods increased. One patient without osteoarthritis had osteonecrosis diagnosed in both hips at radiography and MRI, but osteonecrosis was absent at pathologic examination. CONCLUSION. Radiography depicts osteonecrosis in most patients who have osteonecrosis and subsequently undergo femoral head resection. False-positive and false-negative radiographic findings occur, however. Diagnosis is most difficult in patients with advanced osteoarthritis or subchondral fractures. The number of patients who underwent MRI was not sufficient for evaluation of the accuracy of MRI.

Automated summary

Radiography is useful in depicting osteonecrosis in most patients undergoing femoral head resection, but there are false-positive and false-negative findings. Diagnostic difficulty increases in patients with advanced osteoarthritis or subchondral fractures, leading to increased uncertainty for both radiologic and pathologic evaluations. The accuracy of MRI for osteonecrosis evaluation requires further study due to limited patient numbers.