Room: ePoster Forums
Purpose: To quantify the geometric uncertainties associated with image-guidance and re-positioning for non-coplanar fields for Stereotactic Radiosurgery with BrainLab ExacTrac image-guidance on a Varian TrueBeam linear accelerator.
Methods: A series of Winston-Lutz tests were performed using a phantom consisting of a BB affixed to a carbon rod attached to the treatment table of a linear accelerator. The BB was initially aligned to the isocenter with using x-rays generated by the ExacTrac system, and a pair of orthogonal MV-portal images were acquired at five different treatment table angles. This test was repeated 7 times on different days on two different treatment units. This series of tests was also repeated with x-ray image-alignment performed by the ExacTrac system after each table rotation, applying shifts whenever the deviation was found to be greater than 0.3mm, and repeated with a lower limit of 0.5mm, as well as with errors of known magnitude (range: 0.5-3.5cm) or random errors (with average magnitude of 0.3mm) induced prior to localization after each table rotation. The deviation vector was computed for each image-pair separately and for the image-pairs from all five table angles in a given test, as well as the average magnitude of all tests in each series.
Results: There was no significant difference between the average deviation for the series with no image-guidance after the initial alignment and the series with image-guidance after each rotation with no shifts applied for deviations less than either 0.3mm and 0.5mm, or randomly-induced errors. However, the average deviation for the series with larger, known errors induced was about 0.1mm less. The average deviation between all image-pairs at each table angle was also the similar for each series.
Conclusion: The ExacTrac/TrueBeam system performs well at all table angles and is well-suited to perform intra-fraction image-guidance for Stereotactic Radiosurgery with non-coplanar fields.