Centre for Medical Imaging, University College London, United Kingdom.
Arthritis Research UK Centre for Adolescent Rheumatology, University College London, United Kingdom.
Magn Reson Med. 2019 May;81(5):3094-3107. doi: 10.1002/mrm.27634. Epub 2019 Jan 7.
To evaluate the relationship between bone mineral density (BMD) and magnetic susceptibility, and between proton density fat fraction and susceptibility, in inflamed trabecular bone.
Two different phantoms modeling the fat fraction (FF) and BMD values of healthy bone marrow and disease states were scanned using a multiecho gradient echo acquisition at 3T. After correction for fat-water chemical shift, susceptibility mapping was performed, and susceptibility measurements were compared with BMD and FF values using linear regression. Patients with spondyloarthritis were scanned using the same protocol, and susceptibility values were calculated in areas of inflamed bone (edema) and fat metaplasia, both before and after accounting for the contribution of fat to the total susceptibility.
Susceptibility values in the phantoms were accurately described by a 2D linear function, with a negative correlation between BMD and susceptibility and a positive correlation between FF and susceptibility (adjusted R = 0.77; P = 3·10 ). In patients, significant differences in susceptibility were observed between fat metaplasia and normal marrow, but these differences were eliminated by removing the fat contribution to the total susceptibility.
BMD and proton density fat fraction both influence the total susceptibility of bone marrow and failure to account for the fat contribution could lead to errors in BMD quantification. We propose a method for removing the fat contribution from the total susceptibility, based on the observed linear relationship between susceptibility and FF. In inflamed bone, the overall increase in susceptibility in areas of fat metaplasia is at least partly due to increased fat content.
评估炎症性小梁骨中骨矿物质密度（BMD）与磁化率、质子密度脂肪分数与磁化率之间的关系。
使用 3T 多回波梯度回波采集对模拟健康骨髓和疾病状态下的脂肪分数（FF）和 BMD 值的两个不同体模进行扫描。在对脂肪-水化学位移进行校正后，进行磁化率测绘，并使用线性回归将磁化率测量值与 BMD 和 FF 值进行比较。使用相同的方案对患有脊椎关节炎的患者进行扫描，并在考虑脂肪对总磁化率的贡献后，分别计算炎症性骨（水肿）和脂肪化生区域的磁化率值。
体模中的磁化率值可以通过二维线性函数准确描述，BMD 与磁化率呈负相关，FF 与磁化率呈正相关（调整后的 R = 0.77；P = 3·10）。在患者中，脂肪化生与正常骨髓之间的磁化率存在显著差异，但通过消除脂肪对总磁化率的贡献，这些差异被消除。
BMD 和质子密度脂肪分数都会影响骨髓的总磁化率，如果不考虑脂肪的贡献，可能会导致 BMD 定量的错误。我们提出了一种从总磁化率中去除脂肪贡献的方法，该方法基于磁化率与 FF 之间观察到的线性关系。在炎症性骨中，脂肪化生区域整体磁化率的增加至少部分是由于脂肪含量的增加。
