J Ultrasound Med 2005 Jul;24(7):941-6
Endocrinology and Metabolism Research Center, Tehran University of Medical Sciences, North Kargar Avenue, Tehran 14114, Iran.
Objectives: In this study, we sought the proper cutoff level for quantitative ultrasonography (QUS) of the heel in identifying bone mineral density (BMD) categories as determined by dual-energy x-ray absorptiometry (DXA) in postmenopausal women.
Methods: With the use of DXA, BMD categories of the lumbar spine and different areas of the left femur of 420 healthy women according to World Health Organization definitions were determined. Quantitative ultrasonography of the heel was also performed in each subject. Receiver operating characteristic curves were plotted, and sensitivity and specificity of QUS to diagnose osteoporosis were examined at different points to identify the best cutoff level. The diagnostic agreement between the two techniques in identifying osteoporosis was assessed with kappa scores.
Results: The kappa scores were 0.31 for the lumbar region and 0.5 for the femoral neck region. On the receiver operating characteristic study, a score of -1 was found to be the appropriate cutoff point for QUS studies, in which the sensitivity of QUS to diagnose BMD osteoporosis varied between 78% and 87.5% depending on the site of the DXA study. With the proposed cutoff point (-1), sensitivity and specificity of QUS in detecting osteoporosis at the lumbar spine were 83.9% and 51%, respectively, and at the femoral neck were 84% and 50%, respectively.
Conclusions: Insufficient agreement between QUS and DXA led to uncertainty on expected BMD in people tested by QUS. The proposed cutoff value could achieve higher sensitivity but only by accepting higher rates of false-positive results.