Determining the Signal Source in Magnetic Resonance Imaging of Myelofibrotic Bone Marrow

Magnetic resonance imaging (MRI) is a promising diagnostic method for evaluation of bone marrow myelofibrosis for offering non-invasive, wider-area coverage over conventional histology. Although the most striking feature in histology of myelofibrotic bone marrow is the proliferation of reticulin fibers, the spectrum of findings involves abnormalities in cell number and composition, vessel proliferation, as well as possible pre-fibrotic abnormalities in matrix such as inflammation/edema. Correlation of MRI observations with such histological findings to accurately define the pathogenic phases of myelofibrosis has been challenging. Here, we present the first MRI analysis of myelofibrosis in a mouse model (Gata-1^low mice) in an attempt to correlate MRI signal observations with histological findings. T2-weighted MRI images of femur bone marrow of Gata-1^low animals at different ages (12 to 46 weeks) showed high MR signal intensity, absent in age-matched wild-type controls, that is detected as early as 12 weeks of age, before onset of histologically-defined fibrosis in this mouse model. MRI acquisitions employing fat suppression, a technique that suppresses signals from lipids, together with the absence of chemical shift artifacts in the images, excluded the possibility of lipids as the source of the prominent MR signal. To determine the source of this prominent non-lipid signal in Gata-1^low bone marrow, cellular composition, vascularization and vascular permeability were tested. Our findings demonstrate a multi-source nature of MRI signals in bone marrow, which evolves from young (12 weeks) to old (40 weeks) Gata-1^low bone marrow consistent with the pathologic progression of myelofibrosis. Data support the importance of an integrative approach to further the development of this non-invasive modality for following the dynamics and progression of this multifactorial pathology in humans.