ESTRO 2025 - Abstract Book

S3680

Physics - Quality assurance and auditing

ESTRO 2025

Material/Methods: A daily QA study is carried out during six months. A GUI is created comprising: Canny and Hough transforms, segmentations of the images, edges and center contour detectors and threshold and gradient controls for imaging features. The QA test consists in taking kV and MV images of CIRSMobius and Skull phantoms set on a TrueBeam’s (TB) table in seven couch rotations. The QA plan is designed in ECLIPSE. The RESIS-ITS uses the phantom’s image and calculates the centre and discrepancies in x and y axis of Radiation-Field, Portal-Imager and phantoms and its angular deviations for each table position regarding references (couch at 0º) with a tolerance of 1 cm and 0.5º (see Figure 1). All outcomes are verified and compared by using external image analysis software (Pylinac and Mobius-Dose-Lab). Results: The Table 1 collects a portion of the data showed in Figure 1. It presents the displacements in each axis measured by Pylinac, RESIS-ITS, and Dose-Lab, alongside the uncertainties between RESIS-ITS and Dose-Lab. Overall, Pylinac reports higher displacements, especially in x-axis (see Table 1). At 60°, Pylinac shows 1.48mm, while RESIS-ITS and Dose-Lab report 1.36mm and 1.07mm, respectively. Similarly, in y-axis, Pylinac tends to give larger values, such as 1.08mm at 0°, compared to 0.72mm from RESIS-ITS and 0.66mm from Dose-Lab. Differences in each axis tend to be smaller for RESIS-ITS than those reported by Pylinac, especially at 30°, 60°, and 90°, with discrepancies over 0.5mm (1.5 pixels). RESIS-ITS and Dose-Lab show very close values in each axis, with differences generally below 0.2mm. The largest discrepancy occurs at 90° in the y-axis. RESIS-ITS and Dose-Lab uncertainties are compared. The discrepancies are within 0.5mm across all positions and couch rotations. Minor uncertainties appear at 90° and 270° 0.07mm and 0.04mm, indicating strong agreement. The highest uncertainty is at 60° (0.29mm), where RESIS-ITS slightly diverges. This overall consistency highlights RESIS ITS's reliability in QA processes.