Liu A, Williams L E, Raubitschek A A
Department of Radiological Sciences, UCLA Medical Center 90024, USA.
Med Phys. 1996 Nov;23(11):1919-28. doi: 10.1118/1.597755.
A method is described for the determination of radioactivity (microCi or MBq) at an organ site within an object or patient. Using both anatomic image data (CT or MRI scans) and planar gamma camera images, activity at depth is determined using a matrix inversion method based on least squares. The result of the inversion analysis was the unknown set of n linear (uniform) activity densities representative of each organ within the phantom or patient. The problem was overdetermined since the number of unknown activity densities (microCi/cm) was much less than the number of analysis points (N) within the nuclear image. This method, defined as the CT assisted matrix inversion (CAMI) technique, was accurate to within 15% for a three "organ" plastic phantom, wherein the organs were right circular cylinders having activities of 74 to 508 microCi (or 2.74 MBq to 18.8 MBq). This accuracy included image quantitation effects, particularly assumptions concerning attenuation correction. The average absolute percent error of the estimated activity in four distinct radioactive volumes in the phantom was 9.8%. It was found that the background activity within the phantom was estimated to be too high if sampling regions near strong sources were used in the analysis (scatter effect). This was minimized by going at least 2 cm away from such sources. By applying the method to a monoclonal antibody clinical study, activities within the patient's major organs such as liver, spleen, and kidney could be estimated, even in cases where the organ could not be visualized. Here, the CAMI algorithm gave internally consistent results for the patient's left and right lung linear activity concentrations. The CAMI technique resolves the problem of tissue superimposition using depth information from 3-D CT and is applicable in cases where a number of organs overlap in the gamma camera image. Thus, the method should be generally useful to nuclear image quantitation and the estimation of absorbed radiation doses in patients. One particular application is the estimation of radiation doses in radioimmunotherapy (RIT).
描述了一种用于确定物体或患者体内器官部位放射性（微居里或兆贝克勒尔）的方法。利用解剖图像数据（CT或MRI扫描）和平面伽马相机图像，基于最小二乘法的矩阵反演方法确定深度处的放射性。反演分析的结果是代表体模或患者体内每个器官的n个线性（均匀）放射性密度的未知集。由于未知放射性密度（微居里/厘米）的数量远少于核图像内的分析点数（N），该问题是超定的。这种方法被定义为CT辅助矩阵反演（CAMI）技术，对于一个三“器官”塑料体模，其准确度在15%以内，其中器官为右圆柱体，放射性活度为74至508微居里（或2.74兆贝克勒尔至18.8兆贝克勒尔）。该准确度包括图像定量效应，特别是关于衰减校正的假设。体模中四个不同放射性体积内估计放射性活度的平均绝对百分比误差为9.8%。发现如果在分析中使用靠近强源的采样区域，体模内的本底放射性估计过高（散射效应）。通过至少离此类源2厘米远可将此效应最小化。通过将该方法应用于单克隆抗体临床研究，即使在器官无法可视化的情况下，也可估计患者主要器官如肝脏、脾脏和肾脏内的放射性活度。在此，CAMI算法给出了患者左右肺线性放射性活度浓度的内部一致结果。CAMI技术利用来自三维CT的深度信息解决了组织叠加问题，适用于伽马相机图像中多个器官重叠的情况。因此，该方法对于核图像定量和患者吸收辐射剂量的估计通常应是有用的。一个特别的应用是放射免疫治疗（RIT）中辐射剂量的估计。
