Faculty of Computer Science & Research Campus STIMULATE, University of Magdeburg, Germany.
Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, China.
Comput Methods Programs Biomed. 2021 Jul;206:106117. doi: 10.1016/j.cmpb.2021.106117. Epub 2021 Apr 25.
Liver tumor ablation is often guided by ultrasound (US). Due to poor image quality, intraoperative US is fused with preoperative computed tomography or magnetic tomography (CT/MR) images to provide visual guidance. As of today, the underlying 2D US to 3D CT/MR registration problem remains a very challenging task.
We propose a novel pipeline to address this registration problem. Contrary to previous work, we do not formulate the problem as a regression task, which - for the given registration problem - achieves a low performance regarding accuracy and robustness due to the limited US soft-tissue contrast and the inter-patient variability on liver vessels. Instead, we first estimate the US probe angle roughly by using a classification network. Given this coarse initialization, we then improve the registration by formulation of the problem as a segmentation task, estimating the US plane in the 3D CT/MR through segmentation.
We benchmark our approach on 1035 clinical images from 52 patients, yielding average registration errors of 11.6 and 4.7 mm, which outperforms the state of the art SVR method[1].
Our results show the efficiency of the proposed registration pipeline, which has potential to improve the robustness and accuracy of intraoperative patient registration.
肝脏肿瘤消融术通常由超声（US）引导。由于图像质量较差，术中 US 与术前计算机断层扫描或磁共振成像（CT/MR）融合，以提供可视化指导。截至目前，二维 US 到三维 CT/MR 的配准问题仍然是一项极具挑战性的任务。
我们提出了一种新的流水线来解决这个配准问题。与以往的工作不同，我们没有将该问题表述为回归任务，由于 US 软组织对比度有限以及肝脏血管的个体间变异性，对于给定的配准问题，回归任务在准确性和鲁棒性方面的性能较低。相反，我们首先使用分类网络粗略估计 US 探头角度。有了这个粗略的初始化，我们通过将问题表述为分割任务来改进配准，通过分割来估计 3D CT/MR 中的 US 平面。
我们在 52 名患者的 1035 组临床图像上对我们的方法进行了基准测试，平均配准误差为 11.6 和 4.7 毫米，优于 SVR 方法[1]的最新水平。
我们的结果表明，所提出的配准流水线具有效率，有望提高术中患者配准的鲁棒性和准确性。
