用于前列腺近距离放射治疗的MRI/TRUS数据融合。初步结果
MRI/TRUS data fusion for prostate brachytherapy. Preliminary results
摘要：
前列腺近距离放射治疗涉及将放射性种子（例如I125）永久植入腺体中以治疗局部前列腺癌，例如，cT 1c-T2 a N 0 M0具有良好的预后因素。治疗计划和粒子植入通常基于经直肠超声（TRUS）成像的密集使用。这是不容易的，因为前列腺可视化是困难的，在这种成像模式，特别是关于腺体的顶点，并从内部和观察者之间的变异性的立场。在一些中心，放射性粒子被植入开放式介入MR机器内。由于MRI对前列腺成像具有敏感性和特异性，而开放式MR对大多数中心来说是禁止的，并且使外科手术非常复杂，因此这项工作建议将MR与MRI/TRUS数据融合虚拟地带入手术室。这涉及为医生提供双模态图像（TRUS + MRI），旨在从数据配准阶段改进治疗计划。本文介绍了在PROCUR系统中开发和实现的方法。报告了体模和第一系列患者的结果。幻影实验有助于表征该过程的准确性。患者实验已经表明，使用与TRUS数据相关联的MRI数据改进了TRUS图像分割，特别是关于前列腺的顶端和底部。这可能会显著修改前列腺体积定义，并对治疗计划产生影响。
英文摘要：
Prostate brachytherapy involves implanting radioactive seeds (I125 for instance) permanently in the gland for the treatment of localized prostate cancers, e.g., cT1c-T2a N0 M0 with good prognostic factors. Treatment planning and seed implanting are most often based on the intensive use of transrectal ultrasound (TRUS) imaging. This is not easy because prostate visualization is difficult in this imaging modality particularly as regards the apex of the gland and from an intra- and interobserver variability standpoint. Radioactive seeds are implanted inside open interventional MR machines in some centers. Since MRI was shown to be sensitive and specific for prostate imaging whilst open MR is prohibitive for most centers and makes surgical procedures very complex, this work suggests bringing the MR virtually in the operating room with MRI/TRUS data fusion. This involves providing the physician with bi-modality images (TRUS plus MRI) intended to improve treatment planning from the data registration stage. The paper describes the method developed and implemented in the PROCUR system. Results are reported for a phantom and first series of patients. Phantom experiments helped characterize the accuracy of the process. Patient experiments have shown that using MRI data linked with TRUS data improves TRUS image segmentation especially regarding the apex and base of the prostate. This may significantly modify prostate volume definition and have an impact on treatment planning.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
