In their article, Kim et al. concluded that they could use perfusion magnetic resonance imaging (MRI) to identify the extent/location of the femoral head ischemic process at an earlier sclerotic stage of Legg-Calvé-Perthes disease than typically would be possible with conventional radiographs¹. Their MRI scans showed both the location and the extent of head involvement. The authors suggest that, by utilizing perfusion MRI, they could perform their choice of containment osteotomy (proximal femoral osteotomy) earlier in the course of the disease on selected patients (lateral pillar groups B and B/C according to the classification of Herring et al.). They speculated that if surgeons did not have to “wait to classify” (as they seemingly find necessary with conventional radiography), there would be less risk of collapse (while waiting).

Perfusion scans identify epiphyseal blood flow at a single point in time; thus, conceivably perfusion scans of the same epiphysis performed days earlier or later might not show the same blood flow patterns. In other words, different perfusion scans may show different blood flow patterns for the same femoral head if one scan was performed during the sclerosis stage and the other, during fragmentation. Kim et al. have not proven that an MRI perfusion scan at any one point in time is (or is not) a valid determinant of the eventual location/extent of the femoral head necrosis. Approximately 40% of their patients underwent the MRI perfusion scan when the disease was in Waldenström stage IIa. Certainly, by that time, the extent of femoral head involvement would be identifiable on radiographs, obviating the need for sedating a child and performing a perfusion MRI.