Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, New York, USA.
Magn Reson Med. 2024 Jul;92(1):269-288. doi: 10.1002/mrm.30022. Epub 2024 Mar 23.
To determine whether the spatial scale and magnetic susceptibility of microstructure can be evaluated robustly from the decay of gradient-echo and spin-echo signals.
Gradient-echo and spin-echo images were acquired from suspensions of spherical polystyrene microbeads of 10, 20, and 40 μm nominal diameter. The sizes of the beads and their magnetic susceptibility relative to the medium were estimated from the signal decay curves, using a lookup table generated from Monte Carlo simulations and an analytic model based on the Gaussian phase approximation.
Fitting Monte Carlo predictions to spin-echo data yielded acceptable estimates of microstructural parameters for the 20 and 40 μm microbeads. Using gradient-echo data, the Monte Carlo lookup table provided satisfactory parameter estimates for the 20 μm beads but unstable results for the diameter of the largest beads. Neither spin-echo nor gradient-echo data allowed accurate parameter estimation for the smallest beads. The analytic model performed poorly over all bead sizes.
Microstructural sources of magnetic susceptibility produce distinctive non-exponential signatures in the decay of gradient-echo and spin-echo signals. However, inverting the problem to extract microstructural parameters from the signals is nontrivial and, in certain regimes, ill-conditioned. For microstructure with small characteristic length scales, parameter estimation is hampered by the difficulty of acquiring accurate data at very short echo times. For microstructure with large characteristic lengths, the gradient-echo signal approaches the static-dephasing regime, where it becomes insensitive to size. Applicability of the analytic model was further limited by failure of the Gaussian phase approximation for all but the smallest beads.
确定从梯度回波和自旋回波信号的衰减中是否可以稳健地评估微观结构的空间尺度和磁化率。
从标称直径为 10、20 和 40μm 的球形聚苯乙烯微球悬浮液中获取梯度回波和自旋回波图像。使用基于蒙特卡罗模拟生成的查找表和基于高斯相位逼近的分析模型，从信号衰减曲线中估计微球的大小及其相对于介质的磁化率。
将蒙特卡罗预测拟合到自旋回波数据，可对 20 和 40μm 微球的微结构参数进行可接受的估计。使用梯度回波数据，蒙特卡罗查找表为 20μm 微球提供了令人满意的参数估计，但对于最大微球的直径则得到不稳定的结果。自旋回波和梯度回波数据均无法准确估计最小微球的参数。分析模型在所有微球尺寸上的表现都不佳。
磁化率的微观结构源在梯度回波和自旋回波信号的衰减中产生独特的非指数特征。然而，从信号中提取微观结构参数的反问题并不简单，并且在某些情况下条件不佳。对于具有小特征长度尺度的微观结构，由于在非常短的回波时间很难获取准确的数据，因此参数估计受到阻碍。对于具有大特征长度的微观结构，梯度回波信号接近静态去相位区域，在该区域中，其对尺寸不敏感。由于除最小微球外，高斯相位逼近均失效，因此分析模型的适用性进一步受到限制。
