Correlation of Magnetic Resonance Imaging to Arthroscopic Findings of Stability in Juvenile Osteochondritis Dissecans

Purpose: To determine the ability of magnetic resonance imaging (MRI) to characterize the stability of osteochondritis dissecans (OCD) fragments in juveniles. Methods: Twenty-eight consecutive patients underwent surgery for OCD between 2004 and 2008. Of these, 23 patients had adequate preoperative imaging. There were 14 boys and 9 girls with a mean age of 12.9 years. Of the 23 lesions, 21 were located in the knee and 2 were located in the talus. On the basis of MRI, a single radiologist (1) indicated the presence or absence of 4 established magnetic resonance signs of instability, (2) classified each lesion according to a staging system for OCD stability, and (3) described the lesion as stable or unstable. These findings were compared with the arthroscopic findings. Arthroscopy was considered the gold standard for diagnosing fragment stability. Results: Of the OCD lesions, 13 were found to be stable and 10 were found to be unstable. The final MRI impression was unstable in 21 patients and stable in 2 patients. This yielded a sensitivity of 100% and a specificity of 15% for diagnosing fragment instability. When 2 or more criteria were present, the specificity of MRI to classify lesion instability improved to 92%. The sensitivity, however, dropped to 50%. Concordance between arthroscopic stage and MRI stage was 30% (7 of 23). Conclusions: MRI predicted 21 of 23 lesions to be unstable, whereas arthroscopy found only 10 of these 23 lesions to be unstable. The most common pattern of false-positive findings involved lesions with an area of high signal intensity at the bone-fragment interface. MRI should not be used in isolation to determine lesion instability in young patients with juvenile OCD. Level of Evidence: Level IV, therapeutic case series.