- April 26, 2019
- Category: Scientific Publications
Delivery error sensitivity of an EPID based pre-treatment control for FFF dynamic arc therapy
Achraf Ouakkad1, Maxime Goubert1, Laure Vieillevigne1, François Husson2, Laure Parent1
1 IUCT Oncopole, 1 avenue Irène Joliot-Curie,31059 TOULOUSE Cedex 9
2 Dosisoft, 45 Avenue Carnot, 94230 Cachan
Introduction: EPIbeam is a new algorithm based on a superposition/convolution algorithm and developped for pretreatment quality control with electronic portal imaging device (EPID). It was tested in this study for dynamic conformal arc therapy with flattening filter free (FFF) photon beams in the context of stereotactic radiotherapy. Its sensitivity to delivery errors was assessed and compared to 3D phantom measurements.
Material and methods: A Varian TrueBeam STx linear accelerator equipped with HD120 MLC was used for the measurements. EPID images were acquired with Varian aSi 1000 detector and analysed with Dosisoft EPIbeam software. 3D phantom measurements were performed with PTW 1000 SRS array inserted in PTW Octavius 4D phantom. Analysis was performed in PTW Verisoft software. Varian Eclipse treatment planning system (version 13.7 AAA algorithm) was used to calculate the reference dose distribution. EPID and phantom pretreatment controls were first compared for ten 6 MV FFF lung plans (6.0 to 59.0 cm3 PTV size) and ten 10 MV FFF liver plans (9.8 to 327.5 cm3 PTV size). Delivery error sensitivity was then tested by modifying the initial plans to introduce errors on dose (+1%, 2% and 3%), leaf bank shifts (1mm and 2mm), 10 mm central leaf shift, central leaf blockage, gantry rotation (+5° and +15°) as well as collimator rotation (+5° and 15°). For each energy, these errors were introduced for the largest and smallest PTV. Gamma agreement indices (GAI) were calculated with 2% local dose difference, 2 mm distance-to agreement and 10% threshold.
Results: EPIbeam gave gamma index passing rates similar to those with 3D phantom : for 6 MV FFF, the GAI were (98,79±0,61)% for EPIbeam and (99,86±0,26)% for 3D phantom and for 10 MV FFF, the GAI were (98,55±0,47)% and (99,55±0,86)% respectively. Delivery error sensitivity varied with PTV size but not with energy. For small lesions (6-59 cm3), EPIbeam is more sensitive to dose errors compared with 3D phantom, spotting errors from 1% difference whereas for the largest lesion (327 cm3), a 3% difference was necessary. Leaf bank errors had to be at least 2 mm to fail the test with EPIBEAM whereas the 3D phantom test spotted a 1 mm error for small lesions. Central leaf 10 mm shift was spotted for the small lesions but not for the large lesion with both techniques. Leaf blockage was identified as error with both detectors. As expected, EPIbeam was completely insensitive to gantry rotation errors, unlike 3D phantom. EPIbeam is also less sensitive to collimator errors, compared to 3D phantom.
Conclusion: Once the treatment planning system has been validated with 3D phantom measurements, EPID based pre-treatment quality insurance can be achieved with EPIbeam for fluence verification, provided that independent QA of collimator and gantry rotations is performed on a regular basis on the machine.