Diagnostic Accuracy of Ultrasound for Assessment of Hemophilic Arthropathy: MRI Correlation

Abstract

OBJECTIVE. The purpose of this article is to assess the reliability of interpretation of ultrasound findings according to data blinding in maturing hemophilic joints and to determine the diagnostic accuracy of ultrasound compared with MRI for assessing joint components. SUBJECTS AND METHODS. Ankles (n = 34) or knees (n = 25) of boys with hemophilia or von Willebrand disease (median age, 13 years; range, 5-17 years) were imaged by ultrasound, MRI, and radiography in two centers (Toronto, Canada, and Vellore, India). Ultrasound scans were performed by two operators (one blinded and one unblinded to MRI data) and were reviewed by four reviewers who were unblinded to corresponding MRI findings according to a proposed 0- to 14-item scale that matches 14 of 17 items of the corresponding MRI scale. MRI examinations were independently reviewed by two readers. RESULTS. When data were acquired by radiologists, ultrasound was highly reliable for assessing soft-tissue changes (intraclass correlation coefficient [ICC], 0.98 for ankles and 0.97 for knees) and substantially to highly reliable for assessing osteochondral changes (ICC, 0.61 for ankles and 0.89 for knees). Ultrasound was highly sensitive (> 92%) for assessing synovial hypertrophy and hemosiderin in both ankles and knees but had borderline sensitivity for detecting small amounts of fluid in ankles (70%) in contrast to knees (93%) and variable sensitivity for evaluating osteochondral abnormalities (sensitivity range, 86-100% for ankles and 12-100% for knees). CONCLUSION. If it is performed by experienced radiologists using a standardized protocol, ultrasound is highly reliable for assessing soft-tissue abnormalities of ankles and knees and substantially to highly reliable for assessing osteochondral changes in these joints.