Steggerda Marcel, Schneider Christoph, van Herk Marcel, Zijp Lambert, Moonen Luc, van der Poel Henk
Department of Radiation Oncology, The Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands.
Department of Urology, The Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands.
Med Phys. 2005 Jul;32(7Part1):2262-2270. doi: 10.1118/1.1940147.
To study dose-effect relations of prostate implants with I-125 seeds, accurate knowledge of the dose distribution in the prostate is essential. Commonly, a post-implant computed tomography (CT) scan is used to determine the geometry of the implant and to delineate the contours of the prostate. However, the delineation of the prostate on CT slices is very cumbersome due to poor contrast between the prostate capsule and surrounding tissues. Transrectal Ultrasound (TRUS) on the other hand offers good visualization of the prostate but poor visualization of the implanted seeds. The purpose of this study was to investigate the applicability of combining CT with 3D TRUS by means of image fusion. The advantage of fused TRUS-CT imaging is that both prostate contours and implanted seeds will be well visible. In our clinic, post-implant imaging was realized by simultaneously acquiring a TRUS scan and a CT scan. The TRUS transducer was inserted while the patient was on the CT couch and the CT scan was made directly after the TRUS scan, with the probe still in situ. With the TRUS transducer being visible on both TRUS and CT images, the geometrical relationship between both image sets could be defined by registration on the transducer. Having proven the applicability of simultaneous imaging, the accuracy of this registration method was investigated by additional registration on visible seeds, after preregistration on the transducer. In 4 out of 23 investigated cases an automatic grey value registration on seeds failed for each of the investigated cost functions, and in 2 cases for both cost functions, due to poor visibility of the seeds on the TRUS scan. The average deviations of the seed registration with respect to the transducer registration were negligible. However, in a few individual cases the deviations were significant and probably due to movement of the patient between TRUS and CT scan. In case of a registration on the transducer it is important to avoid patient movement in-between the TRUS and CT scan and to keep the time in-between the scans as short as possible. It can be concluded that fusion of a CT scan and a simultaneously made TRUS scan by means of a three-dimensional (3D) transducer is feasible and accurate when performing a registration on the transducer, if necessary, fine-tuned by a registration on seeds. These fused images are likely to be of great value for post-implant dose distribution evaluations.
为研究I - 125粒子前列腺植入的剂量效应关系，准确了解前列腺内的剂量分布至关重要。通常，植入后计算机断层扫描（CT）用于确定植入物的几何形状并勾勒前列腺轮廓。然而，由于前列腺包膜与周围组织之间对比度差，在CT切片上勾勒前列腺非常繁琐。另一方面，经直肠超声（TRUS）能很好地显示前列腺，但对植入粒子的显示不佳。本研究的目的是通过图像融合研究CT与三维TRUS相结合的适用性。融合的TRUS - CT成像的优点是前列腺轮廓和植入粒子都能清晰可见。在我们诊所中，植入后成像通过同时采集TRUS扫描和CT扫描来实现。当患者躺在CT检查床上时插入TRUS探头，在TRUS扫描后立即进行CT扫描，探头仍在原位。由于TRUS探头在TRUS和CT图像上均可见，通过在探头上进行配准可定义两组图像之间的几何关系。在证明了同时成像的适用性之后，在对探头进行预配准后，通过对可见粒子进行额外配准来研究这种配准方法的准确性。在23例研究病例中，有4例针对每个研究的代价函数，自动灰度值配准在粒子上失败，2例中两个代价函数均失败，原因是TRUS扫描上粒子的可见性差。相对于探头配准，粒子配准的平均偏差可忽略不计。然而，在少数个别病例中偏差很大，可能是由于患者在TRUS和CT扫描之间移动所致。在对探头进行配准时，重要的是避免患者在TRUS和CT扫描之间移动，并使扫描之间的时间尽可能短。可以得出结论，在对探头进行配准（必要时通过对粒子进行配准进行微调）时，通过三维（3D）探头对CT扫描和同时进行的TRUS扫描进行融合是可行且准确的。这些融合图像可能对植入后剂量分布评估具有重要价值。
