13. Assessing long-term stability of Electronic Portal Image Device-based dosimetry

Abstract

Introduction Over the past years, EPID has demonstrated to be an available easy-to-use device revealing a high spatial resolution for dosimetric applications such as both pretreatment controls in open conditions and in vivo dosimetry in transit conditions. Under normal functioning of beam delivery, the relevance of EPID-based dose measurements is inherently related to the quality and reproducibility of the portal image acquisitions which must be periodically controlled. The aim of this study is to investigate the long-term stability of the EPID, in common situation of quality control and maintenance of the device (periodic dark field and flood field calibrations for linearity and differential pixel sensibility respectively). Methods We studied a pool of 44 iViewGT EPID images (Elekta) acquired weekly between May 2016 and April 2017. EPID was irradiated with a 6MV beam (Elekta Synergy), 100 Monitor Units with a large aperture (240 × 240 mm at Source Axis Distance) to cover the panel detector at a Source Imager Distance of 1600 mm. The EPID images were converted to 2D dose-in-water distributions at isocentre level under 50 mm of water using ThinkQA/EpiBeam solution (Dosisoft S.A.). This system corrects well-known EPID image biases, especially using an ideal 2D beam map restoration and an explicit dose calibration factor. All dose images were compared to the first one, taken as the reference, and to the ThinkQA-EpiBeam 2D dose prediction (based on the planned dicom RT-Plan). That way, it was possible to quantify the dose image fluctuations in terms of relative and absolute dose pixel and of image uniformity. Results The EPID dose changes are within the normal tolerances of beam functioning (about 2% asymmetry and output variations). Nonetheless, the whole dose comparisons showed two major discrepancies with the initial reference, requiring special interpretations. The first dose deviation corresponds to a drift of the monitor unit chamber response of the treatment unit which has been corrected. The second one corresponds to a maintenance act concerning the EPID, requiring a new Flood Field and Dark Field calibration. That consequently induced a change in grey levels and a major variation of sensitivity of the detector (−40% for the central pixel values). Conclusions The overall long term stability of the EPID has been confirmed, making possible with confidence its use for dosimetric applications, providing quality controls are periodically performed. Manufacturer maintenance must be taken into account through the introduction of a specific control of the dose calibration factor.