MR Perfusion Index as a Quantitative Method of Evaluating Epiphyseal Perfusion in Legg-Calve-Perthes Disease and Correlation With Short-term Radiographic Outcome

BACKGROUND: Current radiographic prognosticators of the outcome of Perthes disease can only be applied after femoral head deformity has occurred. Quantification of femoral head perfusion using the gadolinium-enhanced subtraction magnetic resonance imaging (MRI) technique may serve as an early prognosticator of outcome. The purposes of this study were 2-fold: (1) to develop a reliable method to quantify femoral head perfusion using this MRI technique; and (2) to determine whether the perfusion at early stages of Perthes disease correlates with radiographic deformity after a 2-year follow-up. METHODS: A total of 20 patients meeting the following inclusion criteria were studied: radiographs and MRI obtained of femoral heads predeformity, age between 5 and 13 years, and unilateral disease. MR perfusion index, a measure of perfusion in the epiphysis, was obtained using digital image analysis of subtraction gadolinium-enhanced MRI. Intraobserver and interobserver agreement of this index was assessed by 2 independent observers. MR perfusion index was correlated with a radiographic deformity index (a measure of femoral head deformity) obtained after a minimum of 2 years. RESULTS: The intraobserver agreement assessed by the intraclass correlation coefficient was 0.96 for observer 1 and 0.97 for observer 2. The interobserver agreement of the MR perfusion index was 0.90 for trials 1 and 2. MR perfusion index in the early stages of Perthes disease was highly variable, ranging from 0 to 0.70. After a minimum of 2 years following MRI acquisition, radiographs were obtained and evaluated using the deformity index, a continuous measure of femoral head deformity, by 2 blinded observers. Deformity index at 2-year follow-up showed moderate correlation with predeformity MR perfusion index (r=-0.56, P=0.01, R=0.31). In those patients who were treated nonoperatively, the correlation was stronger (r=-0.79, P=0.006, R=0.63). CONCLUSIONS: MR perfusion index obtained from gadolinium-enhanced subtraction MR images showed a high interobserver agreement. MR perfusion index is highly variable at early stages of Perthes disease, and a lower MR perfusion index correlated with greater radiographic deformity at the 2-year follow-up. This pilot study shows the promise of predeformity MR perfusion index as a possible early prognosticator of outcome in Perthes disease. LEVELS OF EVIDENCE: Prognostic level II.