Schwartz M L, Ramani R, O'Brien P F, Young C S, Davey P, Hudoba P
Sunnybrook Health Science Centre, Toronto, Ontario, Canada.
Can J Neurol Sci. 1994 Nov;21(4):319-24. doi: 10.1017/s0317167100040890.
In our centre, 111 patients have been treated with linear accelerator stereotactic radiosurgery. Angiographic, CT and MRI images are generated and the target coordinates calculated in 3 dimensions. For CT scanning, cross sections of perpendicular and oblique fiducial markers are seen. For follow-up CT scans done without the frame, a virtual frame is generated by means of a computer program that places fiducial markers on each CT scan cut, as if the patient had been wearing the OBT frame and the scan produced with the gantry parallel to the base of the frame. The position of the oblique marker may be calculated by knowing the thickness and position of each CT cut. Various natural fiducial markers (bony landmarks) are identified by coordinates in the scan with the patient wearing the real frame and in the scan with the virtual frame applied. A transformation matrix is utilized to establish the equivalence between the original CT scan with the real frame applied and subsequent scans without the real frame but with the virtual frame applied. In effect, the virtual frame is re-applied in exactly the same position as the real frame. Lesion measurements may then be duplicated and growth or regression accurately established. The uncertainty in this system of re-application residues in possible patient movement, CT scan slice thickness and inter-observer error in the identification of natural fiducial markers.
在我们中心，111例患者接受了直线加速器立体定向放射外科治疗。生成血管造影、CT和MRI图像，并在三维空间中计算靶点坐标。对于CT扫描，可以看到垂直和倾斜基准标记的横截面。对于在未使用框架的情况下进行的随访CT扫描，通过计算机程序生成虚拟框架，该程序在每个CT扫描切片上放置基准标记，就好像患者戴着OBT框架并且扫描时机架与框架底部平行一样。通过知道每个CT切片的厚度和位置，可以计算倾斜标记的位置。在患者佩戴真实框架的扫描和应用虚拟框架的扫描中，通过坐标识别各种自然基准标记（骨性标志）。利用变换矩阵来建立应用真实框架的原始CT扫描与随后未应用真实框架但应用虚拟框架的扫描之间的等效性。实际上，虚拟框架被重新应用于与真实框架完全相同的位置。然后可以重复病变测量，并准确确定其生长或消退情况。这种重新应用系统中的不确定性源于患者可能的移动、CT扫描切片厚度以及在识别自然基准标记时观察者之间的误差。
