Magnetic resonance imaging of the knee in children and adolescents. Its role in clinical decision-making

BACKGROUND: Recent studies have questioned the utility of magnetic resonance imaging in the diagnosis of pediatric knee disorders because of the morphologic changes during growth and the low accuracy of the formal interpretation of the magnetic resonance imaging scan by a radiologist. The purpose of this study was twofold: (1) to report the accuracy of formal interpretations of magnetic resonance imaging scans of the knee in children and adolescent patients by a radiologist, and (2) to determine the benefit, if any, of a personal review of the magnetic resonance imaging scan of the knee by the orthopaedic surgeon, as a routine part of the diagnostic evaluation. METHODS: A three-year prospective study of all patients who underwent knee arthroscopy performed by a single surgeon, at two children's hospitals, was completed. The analysis focused on the six most common diagnoses: anterior cruciate ligament tear, lateral meniscal tear, medial meniscal tear, osteochondritis dissecans, discoid lateral meniscus, and osteochondral fracture. The preoperative diagnosis of the surgeon was determined by integrating the history and the findings on the clinical examination, plain radiographs, and magnetic resonance imaging scans (including the radiologist's interpretation). RESULTS: Ninety-six patients with ninety-six abnormal knees were included. The mean age was 14.6 years at the time of surgery. Relative to operative findings, kappa values for the formal interpretations of the magnetic resonance imaging scans by a radiologist were 0.78 for an anterior cruciate ligament tear, 0.76 for a medial meniscal tear, 0.71 for a lateral meniscal tear, 0.70 for osteochondritis dissecans, 0.46 for discoid lateral meniscus, and 0.65 for osteochondral fracture. Relative to operative findings, kappa values for the preoperative diagnoses by the surgeon were 1.00 for an anterior cruciate ligament tear, 0.90 for a medial meniscal tear, 0.92 for a lateral meniscal tear, 0.93 for osteochondritis dissecans, 1.00 for discoid lateral meniscus, and 0.90 for osteochondral fracture. The preoperative diagnosis by the surgeon was better (p < 0.05) than the formal interpretation of the magnetic resonance imaging scans by the radiologist with respect to an anterior cruciate ligament tear, lateral meniscal tear, osteochondritis dissecans, and discoid lateral meniscus. CONCLUSIONS: Integration of patient information with an orthopaedic surgeon's review of the magnetic resonance imaging scan of the knee in children and adolescent patients improves the identification of pathological disorders in four of the six categories evaluated. This study questions the necessity for and appropriateness of a routine interpretation of a magnetic resonance imaging scan of the knee in children and adolescents by a radiologist.