Characterization of EPID software for VMAT transit dosimetry.
Marco EspositoAndrea BruschiPaolo BastianiAlessandro GhirelliSilvia PiniSerenella RussoGiovanna ZatelliPublished in: Australasian physical & engineering sciences in medicine (2018)
Dosimetry check (DC) is a commercial software that allows reconstruction of 3D dose distributions using transit electronic portal imaging device (EPID) images. In this work, we evaluated the suitability of DC software for volumetric modulated arc therapy (VMAT) transit dosimetry. The volumetric gamma agreement index 3%/3 mm between twenty VMAT dose distributions reconstructed by DC and calculated with treatment planning system (TPS) were compared to those obtained using PTW OCTAVIUS®4D to assess DC accuracy in VMAT quality assurance (QA). The sensitivity of DC in detecting VMAT delivery and set-up errors and anatomical variations has been investigated by measuring the variation of the gamma agreement index before and after the introduction of specific errors in four VMAT plans related to different anatomical sites. The influence of dose computation algorithm in presence of density inhomogeneity was also assessed. The assessment of VMAT QA shows agreements with TPS maps comparable to OCTAVIUS® 4D (OCT) in homogeneous phantom (p < 0.001). DC mean gamma agreement index was 94.2% ± 3.4, versus 95.6% ± 2.5 of OCT, lower dose threshold was set to 10%. Introduction of deliberate errors resulted in lower gamma agreement index and in 38/56 cases the gamma agreement index was over the detection threshold. The dose computation algorithm of DC is accurate in all anatomical sites except lung. However in lung cases, the aqua vivo approach used in this work reduced the algorithm dependence of DC results. DC accurately reproduced VMAT 3D dose distributions in phantom and is sensitive to detect errors caused by delivery inaccuracy and anatomical variations of patients.
Keyphrases
- dendritic cells
- deep learning
- machine learning
- patient safety
- optical coherence tomography
- adverse drug
- end stage renal disease
- high resolution
- immune response
- chronic kidney disease
- emergency department
- mesenchymal stem cells
- data analysis
- quantum dots
- magnetic resonance imaging
- magnetic resonance
- convolutional neural network
- neural network
- single molecule