Schematic illustration of the experimental setup for EPID and gantry sag measurements. The ball bearings (a, b, and c) are 4.8 mm in diameter.
A sample image acquired for the determination of EPID and gantry sag. The shadows of three markers are labeled as (a), (b), and (c). The distance between (a) and (b) is specified as (d).
Schematic illustration of the procedure followed in the detection algorithm on one section of a sample EPID image for the measurement of EPID and gantry sag during rotation.
A sample profile (P u1) for the BB (a) in cross-plane direction used for the determination of the BB center (U a) with subpixel accuracy. The center is identified by the 50% level on each side of the valley.
Results of the combined effect of gantry and EPID sag as a function of gantry angle at SDD = 150 cm for the same linac over a period of 15 months in (a) cross-plane and (b) in-plane directions. The fitted sag values in the in-plane and cross-plane directions are compared in (c). Most error bars are smaller than the symbols used for plotting.
Results of EPID sag measurements in its plane as a function of gantry angle using the algorithm for the same linac over the period of 15 months in (a) cross-plane direction before correction, (b) cross-plane direction after correction for BB misalignment at the isocenter, and (c) in-plane direction. The fitted sag values in the in-plane and cross-plane directions are compared in (d).
The algorithm results for changes in SDD as a function of gantry angle for the same linac. The results are reproducible for each SDD setting.
Comparison of the EPID and gantry sag for different linacs used at two different radiotherapy centers: combined EPID and gantry sag in (a) cross-plane and (b) in-plane directions; EPID sag in (c) cross-plane and (d) in-plane directions.
Left column: The edge pixels of the sum of a sequence of EPID images acquired from a circular collimator during a whole gantry rotation: (a-1) before sag corrections, (a-2) after correction for the EPID sag, and (a-3) after correction for the combined effect of EPID and gantry sag. Images are cropped for better visibility. Right column: Gamma maps comparing the cumulated cine-images acquired during a whole arc delivery with the reference image taken at zero gantry angle: (b-1) before sag corrections, (b-2) after correction for the EPID sag, and (b-3) after correction for the combined effect of EPID and gantry sag.
Comparison of a series of cine EPID images with reference using the Gamma evaluation (3%, 3 mm criteria). The percentage of points with Gamma index <1 are shown before and after corrections for the EPID sag and the combined EPID and gantry sag (±1 SD). Control points at 45° intervals are also compared with the data points in the sequence at the same angles.
The intentional shifts to the BB at the isocenter as detected by the algorithm. Results are listed for the average of three series of measurements (±1 SD).
The Gamma evaluation results (3%, 3 mm criteria) comparing the cumulated cine-images acquired during a whole gantry rotation with the reference image at zero gantry angle. The average of three sets of measurements (±1 SD) have been used for comparison.
Fourier coefficients for the fitted curves through EPID sag and EPID + gantry sag in the cross-plane and in-plane directions for one of the linacs tested in this study
Article metrics loading...
Full text loading...