Evaluation of torque-induced spatial error in the stereotactic definition of the Leksell Gamma Knife ICON system
Carter, Zachary Gray
MetadataShow full item record
The Leksell Gamma Knife ICON is the most recent production model of the Gamma Knife stereotactic radiosurgery system from Elekta AB. This system has the ability to provide both conventional frame-based Gamma Knife treatments and frameless treatments with motion tracking, and provides an additional integrated Cone Beam CT imaging suite. This study addresses recent questions about the spatial accuracy of the stereotactic definition of radiosurgery treatments using the Gamma Knife frame with CT and CBCT when the frame pins are subject to various levels of torque or asymmetrically applied torque. This torque on the frame pins can translate to a further torque about the frame’s base, warping the stereotactic coordinate system away from the orthogonal coordinate system that is assumed by the treatment planning system. Additionally, it can cause further torque on the CT fiducial marker box, which causes error in the planned target dose location versus the actual target location. The project provides a quantitative evaluation of this spatial error as a function of both the torque magnitude and the torque asymmetry from the manufacturer-recommended torque level of 4 inch-lbs up to 14 inch-lbs for both evenly applied torque and asymmetrically applied torque with a “3-post” distribution, such as in the case of a patient who received a craniotomy in the area where the pin would be placed. Results indicate that basing the stereotactic definition on a cone beam CT image using the integrated CBCT arm provide consistent sub-mm accuracy in planned vs. delivered dose with an evenly applied torque distribution, whereas using a CT image for stereotactic definition does not. The practice of imaging with 4 frame posts and delivering with only 3 is also confirmed to cause spatial error well above 1 mm, and it is therefore recommended to always re-image after removing a frame post.