Malison R T, Miller E G, Greene R, McCarthy G, Charney D S, Innis R B
Department of Psychiatry, Yale University School of Medicine, New Haven, CT.
J Comput Assist Tomogr. 1993 Nov-Dec;17(6):952-60. doi: 10.1097/00004728-199311000-00018.
We have developed and validated in a phantom a method of computer-assisted coregistration using multislice SPECT and MR images. Reusable fiducial markers were fabricated from nylon-based plastic and consist of two parts: a base that remains fixed to the skin with adhesive between scans and a removable, spherical cavity insert that can be filled with contrast agents appropriate for multiple imaging modalities. Markers external and internal to a three-dimensional brain phantom provided a means of quantifying the method's accuracy. A computer algorithm was used to derive transformation matrices for image sets by minimizing the root mean squared deviations obtained for multiple permutations (n = 10) of increasing numbers (range 3-11) of external SPECT/MRI point pairs. As defined by the average +/- SEM mean residual deviations of noncoregistered internal fiducials, the minimal accuracy was 2.4 +/- 0.3 mm (no marker > 3.6 mm) for three coregistration points and did not improve beyond seven to eight fiducials (2.2 +/- 0.1 mm; no marker > 3.2 mm). The method's true accuracy is likely to be better than estimates of minimal accuracy, however, since such measures reflect surmountable random errors in fiducial location. With use of identical MRI (or SPECT) data sets, measures of intraoperator (0.5 +/- 0.1 and 0.5 +/- 0.1 mm, respectively) and interoperator (0.5 +/- 0.3 and 0.8 +/- 0.1 mm) reliability were also obtained, establishing the method as highly reproducible and objective. Preliminary results in a human subject suggest its feasibility for clinical studies.
我们已经在体模中开发并验证了一种使用多层单光子发射计算机断层扫描(SPECT)和磁共振成像(MR)图像的计算机辅助配准方法。可重复使用的基准标记由尼龙基塑料制成,由两部分组成:一个在扫描之间用粘合剂固定在皮肤上的基座,以及一个可移除的球形腔插入物,该插入物可以填充适用于多种成像模态的造影剂。三维脑体模内外的标记提供了一种量化该方法准确性的手段。通过最小化从增加数量(范围为3 - 11)的外部SPECT/MRI点对的多个排列(n = 10)获得的均方根偏差,使用计算机算法来推导图像集的变换矩阵。根据未配准的内部基准标记的平均±标准误平均残余偏差定义,三个配准点的最小精度为2.4±0.3毫米(无标记> 3.6毫米),并且在七个到八个基准标记之后精度没有提高(2.2±0.1毫米;无标记> 3.2毫米)。然而,该方法的真实精度可能优于最小精度估计,因为这些测量反映了基准标记位置中可克服的随机误差。使用相同的MRI(或SPECT)数据集,还获得了操作员内(分别为0.5±0.1和0.5±0.1毫米)和操作员间(0.5±0.3和0.8±0.1毫米)可靠性的测量结果,确立了该方法具有高度可重复性和客观性。在一名人类受试者中的初步结果表明了其在临床研究中的可行性。