Mousavi Seyed Reza, Raahemifar Kaamran, Pautler Stephen, Samani Abbas
Department of Electrical and Computer Engineering, Western University, London, ON, Canada,
Int J Comput Assist Radiol Surg. 2013 Nov;8(6):1053-61. doi: 10.1007/s11548-013-0898-3. Epub 2013 May 5.
Accurate Transrectal Ultrasound (TRUS)-guided prostate needle biopsy requires registering preoperative 3D TRUS or MR image, in which tumors and other suspicious areas are visible, to intraoperative 2D TRUS images. Such image registration is time-consuming while its real-time implementation is yet to be developed. To bypass this registration step, robotic needle biopsy systems can be used to place the US probe at the same position relative to the prostate during the 3D and 2D image acquisition to ensure similar prostate deformation. To have such similar deformation, only visual feedback is not sufficient as such feedback can be used to only guarantee that the whole prostate is within the field of view irrespective of the probe's orientation. As such, contact pressure feedback can be utilized to ensure consistent minimum contact between the probe and prostate.
A robotic system is proposed where a TRUS probe with pressure sensor array is used. The contact pressure can be measured during imaging and used to provide feedback in conjunction with an optimization algorithm for consistent probe positioning. The robotic system is driven by the feedback to position the probe such that pressure pattern of the sensors during 2D image acquisition is similar to the pressure pattern during 3D image acquisition. The proposed method takes into account the patient's body movement expected during image acquisition. In this study, an in silico phantom is used where the simulated contact pressure distribution required in the optimization algorithm is obtained using a prostate finite element model.
Starting from an arbitrary position where the probe contacts the phantom, this position was varied systematically until a position corresponding to maximum pressure pattern similarity between contact pressure patterns corresponding to the 2D and 3D imaging was achieved successfully.
Results obtained from the in silico phantom study indicate that the proposed technique is capable of ensuring having only minimal relative prostate deformation between preoperative image acquisition and intraoperative imaging used for guiding needle biopsy, paving the way for faster and more accurate registration.
准确的经直肠超声（TRUS）引导下前列腺穿刺活检需要将术前可见肿瘤及其他可疑区域的三维TRUS或MR图像与术中二维TRUS图像进行配准。这种图像配准耗时，且实时实现尚未开发出来。为绕过这一配准步骤，可使用机器人穿刺活检系统在三维和二维图像采集期间将超声探头放置在相对于前列腺的相同位置，以确保前列腺有相似的变形。要实现这种相似变形，仅视觉反馈是不够的，因为这种反馈只能用于确保整个前列腺在视野范围内，而不考虑探头的方向。因此，可利用接触压力反馈来确保探头与前列腺之间保持一致的最小接触。
提出一种使用带有压力传感器阵列的TRUS探头的机器人系统。在成像过程中可测量接触压力，并结合优化算法用于提供反馈，以实现探头的一致定位。机器人系统由反馈驱动，将探头定位，使二维图像采集期间传感器的压力模式与三维图像采集期间的压力模式相似。所提出的方法考虑了图像采集期间患者身体预期的移动。在本研究中，使用了一个计算机模拟模型，其中优化算法所需的模拟接触压力分布是通过前列腺有限元模型获得的。
从探头接触模拟模型的任意位置开始，系统地改变该位置，直到成功获得与二维和三维成像对应的接触压力模式之间压力模式相似度最大的位置。
计算机模拟模型研究获得的结果表明，所提出的技术能够确保在术前图像采集和用于引导穿刺活检的术中成像之间前列腺仅有最小的相对变形，为更快、更准确的配准铺平了道路。
