Frameless stereotaxy for radiosurgical planning and follow-up

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 resides in possible patient movement, CT scan slice thickness and inter-observer error in the identification of natural fiducial markers.