Abstract Quantitative susceptibility mapping and phase imaging are used to identify multiple sclerosis lesions with paramagnetic rims that slowly expand over time, and are associated with earlier progression to disability, decreased brain volume, and increased frequency of clinical relapse. However, the presence of iron-laden microglia/macrophages at the lesion rim and demyelination within the lesion both contribute to phase and quantitative susceptibility mapping images. Therefore, simultaneous pathological validation is needed to assess accuracies in identifying iron-positive lesions. MRI was performed on fifteen multiple sclerosis brain slabs; thirty-two lesions of interest were processed for myelin, iron, and microglial markers. Three experienced readers classified lesions as rim positive or negative on quantitative susceptibility mapping and phase; these classifications were compared with Perls’ stain as the gold standard. All ten of the quantitative susceptibility mapping-positive lesions had iron-positive rims on histology. Of the sixteen phase-positive lesions, only ten had iron-positive rims on histology. Using Perls’ stain as the ground truth, the positive predictive value was 100% for quantitative susceptibility mapping and 63% for phase; the negative predictive value was 95% for quantitative susceptibility mapping and 94% for phase. Post-mortem imaging results demonstrate that quantitative susceptibility mapping is a more reliable indicator of an iron-positive rim compared to phase imaging.
Read full abstract