ESTRO 2024 - Abstract Book

S3338

Physics - Detectors, dose measurement and phantoms

ESTRO 2024

In radiotherapy, the risk of radiation pneumonitis following total body irradiation (TBI) is reduced by the use of lung shield blocks during each treatment session. Currently, MegaVolt (MV) imaging with dedicated silver halide based films serves as the prevailing technique for lung delineation on the simulation image and position assessment of the shielding blocks during each treatment fraction. However, the availability of these films has recently become a problem. This study investigates the clinical performance of a computed radiography (CR) solution for TBI in comparison to the radiographic films, as well as exploring the potential enhancement of image quality through filtering and post-processing methods.

Material/Methods:

To conduct this study, a standing position setup was employed at a source-to-surface distance of 5 meters. We captured images for both simulation (6 MV) and treatment (15 MV) verification purposes using a CR 15-X Digitizer and imaging plates (Agfa NV) and traditional film (Carestream). Our evaluation was divided into three stages. Initially, we examined an aluminum block and penetrameter to assess contrast and sharpness parameters, utilising ImageJ for image analysis to derive the edge spread function (ESF). Secondly, we employed the PTW EPID QC phantom to objectively evaluate image quality through the corresponding epidSoft software, examining aspects such as signal linearity, signal-to-noise ratio (SNR), contrast resolution, and spatial resolution. In the third phase, we captured images of an anthropomorphic chest phantom and subjected them to a clinical performance study with two independent observers who conducted a visual grading analysis (VGA). In the case of CR, the influence on the image quality was explored by employing different filter materials (copper, lead, Ba-based coating), positioning filters in front or behind the active phosphor layer, and inside or outside of the cassette. The impact of post-processing was examined with dedicated Musica (Agfa NV) processing on the chest phantom images. Additionally, a comparative analysis between the workflows involving film and CR was conducted. The CR images exhibited appropriate linearity and high SNR. However, spatial resolution was inferior to the traditional film and intrinsic contrast resolution was also found to be poor (Figure 1). Nonetheless, application of filtering methods yielded enhancements in contrast resolution as depicted in Figure 1, and the SNR demonstrated a significant increase of approximately 15% when employing filters positioned in front of the cassette. Furthermore, positioning filters inside the cassette had a positive effect on the 10-90% rise distance of the ESF, with values of 1.32 mm for 'in' versus 4.72 mm for 'out'. The scoring of the performance study, as depicted in Figure 2, demonstrated that the intrinsic CR image scored worse than the traditional film. However, the application of image processing had a substantial impact on the VGA scores, particularly on the contrast and sharpness criteria. Moreover, when combined with filtering, the VGA scores exhibited further improvements due to higher scoring on spine visibility. Finally, the TBI simulation workflow could be shortened by approximately 10 minutes, representing a reduction of 25% in time, using CR compared to traditional film. Additionally, a substantial dose reduction on the order of 0.1 Gy, corresponding to a 95% decrease, was achieved for the simulation due to the lower number of monitor units required for adequate image quality. Results: