Robertson James, Martin Eleanor, Cox Ben, Treeby Bradley E
Department Medical Physics and Biomedical Engineering, University College London, Gower Street, WC1E 6BT, London.
Phys Med Biol. 2017 Apr 7;62(7):2559-2580. doi: 10.1088/1361-6560/aa5e98. Epub 2017 Feb 6.
High intensity transcranial focused ultrasound is an FDA approved treatment for essential tremor, while low-intensity applications such as neurostimulation and opening the blood brain barrier are under active research. Simulations of transcranial ultrasound propagation are used both for focusing through the skull, and predicting intracranial fields. Maps of the skull acoustic properties are necessary for accurate simulations, and can be derived from medical images using a variety of methods. The skull maps range from segmented, homogeneous models, to fully heterogeneous models derived from medical image intensity. In the present work, the impact of uncertainties in the skull properties is examined using a model of transcranial propagation from a single element focused transducer. The impact of changes in bone layer geometry and the sound speed, density, and acoustic absorption values is quantified through a numerical sensitivity analysis. Sound speed is shown to be the most influential acoustic property, and must be defined with less than 4% error to obtain acceptable accuracy in simulated focus pressure, position, and volume. Changes in the skull thickness of as little as 0.1 mm can cause an error in peak intracranial pressure of greater than 5%, while smoothing with a 1 [Formula: see text] kernel to imitate the effect of obtaining skull maps from low resolution images causes an increase of over 50% in peak pressure. The numerical results are confirmed experimentally through comparison with sonications made through 3D printed and resin cast skull bone phantoms.
高强度经颅聚焦超声是一种经美国食品药品监督管理局（FDA）批准用于治疗特发性震颤的方法，而诸如神经刺激和打开血脑屏障等低强度应用正在积极研究中。经颅超声传播的模拟既用于通过颅骨聚焦，也用于预测颅内场。颅骨声学特性图对于精确模拟是必要的，并且可以使用多种方法从医学图像中得出。颅骨图的范围从分段的、均匀的模型到从医学图像强度得出的完全非均匀模型。在本工作中，使用单元素聚焦换能器的经颅传播模型研究了颅骨特性不确定性的影响。通过数值灵敏度分析量化了骨层几何形状以及声速、密度和声学吸收值变化的影响。结果表明，声速是最具影响力的声学特性，必须以小于4%的误差来定义，才能在模拟的聚焦压力、位置和体积方面获得可接受的精度。颅骨厚度仅0.1毫米的变化就可能导致颅内峰值压力出现大于5%的误差，而用1[公式：见正文]核进行平滑处理以模拟从低分辨率图像获取颅骨图的效果会使峰值压力增加超过50%。通过与通过3D打印和树脂铸造的颅骨骨模型进行超声处理的实验结果进行比较，证实了数值结果。
