Department of Radiological Sciences, University of California, Irvine, CA, 92697, USA.
Int J Cardiovasc Imaging. 2021 Jan;37(1):331-341. doi: 10.1007/s10554-020-01961-y. Epub 2020 Sep 2.
To investigate the feasibility of quantifying the chemical composition of coronary artery plaque in terms of water, lipid, protein, and calcium contents using dual-energy computed tomography (CT) in a simulation study. A CT simulation package was developed based on physical parameters of a clinical CT scanner. A digital thorax phantom was designed to simulate coronary arterial plaques in the range of 2-5 mm in diameter. Both non-calcified and calcified plaques were studied. The non-calcified plaques were simulated as a mixture of water, lipid, and protein, while the calcified plaques also contained calcium. The water, lipid, protein, and calcium compositions of the plaques were selected to be within the expected clinical range. A total of 95 plaques for each lesion size were simulated using the CT simulation package at 80 and 135 kVp. Half-value layer measurements were made to make sure the simulated dose was within the range of clinical dual energy scanning protocols. Dual-energy material decomposition using a previously developed technique was performed to determine the volumetric fraction of water, lipid, protein, and calcium contents in each plaque. For non-calcified plaque, the total volume conservation provides the third constrain for three-material decomposition with dual energy CT. For calcified plaque, a fourth criterion was introduced from a previous report suggesting a linear correlation between water and protein contents in soft tissue. For non-calcified plaque, the root mean-squared error (RMSE) of the image-based decomposition was estimated to be 0.7%, 1.5%, and 0.3% for water, lipid, and protein contents, respectively. As for the calcified plaques, the RMSE of the 5 mm plaques were estimated to be 5.6%, 5.7%, 0.2%, and 3.1%, for water, lipid, calcium, and protein contents, respectively. The RMSE increases as the plaque size reduces. The simulation results indicate that chemical composition of coronary arterial plaques can be quantified using dual-energy CT. By accurately quantifying the content of a coronary plaque lesion, our decomposition method may provide valuable insight for the assessment and stratification of coronary artery disease.
为了在模拟研究中使用双能计算机断层扫描（CT）定量评估冠状动脉斑块的水、脂质、蛋白质和钙含量的化学组成的可行性。基于临床 CT 扫描仪的物理参数开发了一个 CT 模拟包。设计了一个数字胸部体模来模拟直径在 2-5 毫米范围内的冠状动脉斑块。研究了非钙化斑块和钙化斑块。非钙化斑块模拟为水、脂质和蛋白质的混合物，而钙化斑块还含有钙。斑块的水、脂质、蛋白质和钙组成选择在预期的临床范围内。使用 CT 模拟包在 80 和 135 kVp 下为每个病变大小模拟了 95 个斑块。进行半值层测量以确保模拟剂量在临床双能扫描方案的范围内。使用先前开发的技术进行双能物质分解，以确定每个斑块中水、脂质、蛋白质和钙含量的体积分数。对于非钙化斑块，总体积守恒为双能 CT 的三物质分解提供了第三个约束条件。对于钙化斑块，从之前的报告中引入了第四个标准，该标准表明软组织中水和蛋白质含量之间存在线性相关性。对于非钙化斑块，基于图像的分解的均方根误差（RMSE）估计分别为水、脂质和蛋白质含量的 0.7%、1.5%和 0.3%。对于钙化斑块，估计 5 毫米斑块的 RMSE 分别为水、脂质、钙和蛋白质含量的 5.6%、5.7%、0.2%和 3.1%。RMSE 随着斑块尺寸的减小而增加。模拟结果表明，冠状动脉斑块的化学成分可以使用双能 CT 定量。通过准确量化冠状动脉斑块病变的含量，我们的分解方法可能为冠状动脉疾病的评估和分层提供有价值的见解。
