Rusheen Aaron E, Goyal Abhinav, Owen Robert L, Berning Elise M, Bothun Dane T, Giblon Rachel E, Blaha Charles D, Welker Kirk M, Huston John, Bennet Kevin E, Oh Yoonbae, Fagan Andrew J, Lee Kendall H
1Department of Neurologic Surgery, Mayo Clinic, Rochester.
2Medical Scientist Training Program, Mayo Clinic, Rochester.
J Neurosurg. 2022 Mar 25;137(5):1265-1277. doi: 10.3171/2021.11.JNS211078. Print 2022 Nov 1.
Magnetic resonance imaging at 7T offers improved image spatial and contrast resolution for visualization of small brain nuclei targeted in neuromodulation. However, greater image geometric distortion and a lack of compatible instrumentation preclude implementation. In this report, the authors detail the development of a stereotactic image localizer and accompanying imaging sequences designed to mitigate geometric distortion, enabling accurate image registration and surgical planning of basal ganglia nuclei.
Magnetization-prepared rapid acquisition with gradient echo (MPRAGE), fast gray matter acquisition T1 inversion recovery (FGATIR), T2-weighted, and T2*-weighted sequences were optimized for 7T in 9 human subjects to visualize basal ganglia nuclei, minimize image distortion, and maximize target contrast-to-noise and signal-to-noise ratios. Extracranial spatial distortions were mapped to develop a skull-contoured image localizer embedded with spherical silicone fiducials for improved MR image registration and target guidance. Surgical plan accuracy testing was initially performed in a custom-developed MRI phantom (n = 5 phantom studies) and finally in a human trial.
MPRAGE and T2*-weighted sequences had the best measures among global measures of image quality (3.8/4, p < 0.0001; and 3.7/4, p = 0.0002, respectively). Among basal ganglia nuclei, FGATIR outperformed MPRAGE for globus pallidus externus (GPe) visualization (2.67/4 vs 1.78/4, p = 0.008), and FGATIR, T2-weighted imaging, and T2*-weighted imaging outperformed MPRAGE for substantia nigra visualization (1.44/4 vs 2.56/4, p = 0.04; vs 2.56/4, p = 0.04; vs 2.67/4, p = 0.003). Extracranial distortion was lower in the head's midregion compared with the base and apex ( 1.17-1.33 mm; MPRAGE and FGATIR, p < 0.0001; T2-weighted imaging, p > 0.05; and T2*-weighted imaging, p = 0.013). Fiducial placement on the localizer in low distortion areas improved image registration (fiducial registration error, 0.79-1.19 mm; p < 0.0001) and targeting accuracy (target registration error, 0.60-1.09 mm; p = 0.04). Custom surgical software and the refined image localizer enabled successful surgical planning in a human trial (fiducial registration error = 1.0 mm).
A skull-contoured image localizer that accounts for image distortion is necessary to enable high-accuracy 7T imaging-guided targeting for surgical neuromodulation. These results may enable improved clinical efficacy for the treatment of neurological disease.
7T磁共振成像可提高图像的空间和对比度分辨率，有助于观察神经调节中所针对的小脑核。然而，较大的图像几何失真以及缺乏兼容的仪器设备阻碍了其应用。在本报告中，作者详细介绍了一种立体定向图像定位器及配套成像序列的开发，旨在减轻几何失真，实现基底神经节核的精确图像配准和手术规划。
针对9名人类受试者，对磁化准备快速梯度回波（MPRAGE）、快速灰质采集T1反转恢复（FGATIR）、T2加权和T2*加权序列进行了7T优化，以观察基底神经节核，使图像失真最小化，并使目标对比度噪声比和信噪比最大化。绘制颅外空间失真情况，以开发一种嵌入球形硅胶基准标记的颅骨轮廓图像定位器，用于改进磁共振图像配准和目标引导。手术计划准确性测试最初在定制的磁共振成像体模中进行（n = 5个体模研究），最终在人体试验中进行。
在图像质量的整体测量中，MPRAGE和T2加权序列表现最佳（分别为3.8/4，p < 0.0001；和3.7/4，p = 0.0002）。在基底神经节核中，FGATIR在苍白球外侧部（GPe）可视化方面优于MPRAGE（2.67/4对1.78/4，p = 0.008），在黑质可视化方面，FGATIR、T2加权成像和T2加权成像优于MPRAGE（1.44/4对2.56/4），p = 0.04；对2.56/4，p = 0.04；对2.67/4，p = 0.003）。与头部底部和顶部相比，颅外失真在头部中部较低（1.17 - 1.33毫米；MPRAGE和FGATIR，p < 0.0001；T2加权成像，p > 0.05；T2*加权成像，p = 0.013）。在低失真区域将基准标记放置在定位器上可改善图像配准（基准配准误差，0.79 - 1.19毫米；p < 0.0001）和靶向准确性（目标配准误差，0.60 - 1.09毫米；p = 0.04）。定制手术软件和改进的图像定位器在人体试验中实现了成功的手术规划（基准配准误差 = 1.0毫米）。
一种考虑图像失真的颅骨轮廓图像定位器对于实现高精度7T成像引导的手术神经调节靶向是必要的。这些结果可能会提高神经系统疾病治疗的临床疗效。
