van de Ven Wendy J M, Hu Yipeng, Barentsz Jelle O, Karssemeijer Nico, Barratt Dean, Huisman Henkjan J
Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen 6525 GA, The Netherlands.
Centre for Medical Image Computing, University College London, London WC1E 6BT, United Kingdom.
Med Phys. 2015 May;42(5):2470-81. doi: 10.1118/1.4917481.
Adding magnetic resonance (MR)-derived information to standard transrectal ultrasound (TRUS) images for guiding prostate biopsy is of substantial clinical interest. A tumor visible on MR images can be projected on ultrasound (US) by using MR-US registration. A common approach is to use surface-based registration. The authors hypothesize that biomechanical modeling will better control deformation inside the prostate than a regular nonrigid surface-based registration method. The authors developed a novel method by extending a nonrigid surface-based registration algorithm with biomechanical finite element (FE) modeling to better predict internal deformations of the prostate.
Data were collected from ten patients and the MR and TRUS images were rigidly registered to anatomically align prostate orientations. The prostate was manually segmented in both images and corresponding surface meshes were generated. Next, a tetrahedral volume mesh was generated from the MR image. Prostate deformations due to the TRUS probe were simulated using the surface displacements as the boundary condition. A three-dimensional thin-plate spline deformation field was calculated by registering the mesh vertices. The target registration errors (TREs) of 35 reference landmarks determined by surface and volume mesh registrations were compared.
The median TRE of a surface-based registration with biomechanical regularization was 2.76 (0.81-7.96) mm. This was significantly different than the median TRE of 3.47 (1.05-7.80) mm for regular surface-based registration without biomechanical regularization.
Biomechanical FE modeling has the potential to improve the accuracy of multimodal prostate registration when comparing it to a regular nonrigid surface-based registration algorithm and can help to improve the effectiveness of MR guided TRUS biopsy procedures.
将磁共振（MR）衍生信息添加到标准经直肠超声（TRUS）图像中以指导前列腺活检具有重大临床意义。通过使用MR-US配准，MR图像上可见的肿瘤可以投影到超声（US）上。一种常见的方法是使用基于表面的配准。作者假设，与常规的基于非刚性表面的配准方法相比，生物力学建模将能更好地控制前列腺内部的变形。作者通过用生物力学有限元（FE）建模扩展基于非刚性表面的配准算法，开发了一种新方法，以更好地预测前列腺的内部变形。
从10名患者收集数据，将MR和TRUS图像进行刚性配准，使前列腺方向在解剖学上对齐。在两张图像中手动分割前列腺并生成相应的表面网格。接下来，从MR图像生成四面体体积网格。以表面位移作为边界条件，模拟TRUS探头引起的前列腺变形。通过配准网格顶点计算三维薄板样条变形场。比较由表面和体积网格配准确定的35个参考标志点的目标配准误差（TRE）。
基于表面的配准并进行生物力学正则化时，TRE的中位数为2.76（0.81 - 7.96）mm。这与未进行生物力学正则化的常规基于表面的配准的TRE中位数3.47（1.05 - 7.80）mm有显著差异。
与常规的基于非刚性表面的配准算法相比，生物力学有限元建模有潜力提高多模态前列腺配准的准确性，并有助于提高MR引导的TRUS活检程序的有效性。
