TU-D-BRB-08: A CBCT Projection Matrix Method for Radiation and Imaging Isocenter QA

Abstract

Purpose: To test the use of a cone beam computed tomography(CBCT) projection matrix method for determining the imaging isocenter diameter as a replacement for the traditional gantry star shot for radiation isocenter testing.Method and Materials: The Siemens MVision megavoltage CBCT is calibrated by imaging a geometric reconstruction phantom that contains BBs of various sizes at well characterized positions. This generates several projection matrices, Pq, that define where a point in the reconstruction volume is projected onto the flat panel detector at gantry angle q. The standard reconstruction angles are −90° to 110°. A new protocol, with angles from −30° to 170°, provides information about radiation isocenter from posterior angles. Flat panel positions projected to a plane containing the isocenter are [Uq,Vq] = [0.276(Pq(1,4) − P0(1,4)), 0.276(Pq(2,4) − P0(2,4))]. The room coordinates [xq, yq, zq] = [Uqcos(‐q), Uqsin(‐q), Vq]. The radiation isocenter ellipsoid diameters are (xq max ‐ xq min, yq max ‐ yq min, zq max ‐ zq min), where superscripts max and min refer to maximum and minimum values of the room coordinate, respectively. The maximum diameter is compared to that of a traditional star shot and a Winston‐Lutz type test. Results: Traditional star shots are limited in accuracy due to the subjectivity in the analysis and set‐up error. The maximum radiation isocenter diameters were about 0.8 mm, 0.9 mm and 1.4 mm for the star shot, Winston‐Lutz test and the projection matrix analysis, respectively. The result for the projection matrix includes deviations resulting from flat panel motion and is larger than that of the Winston‐Lutz test, which is unaffected by flat panel motion. Conclusion: The projection matrix method simultaneously checks the stability of the imaging and radiation isocenter, while providing an annual geometric calibration for the CBCT system.