ESTRO 2024 - Abstract Book

S3192

Physics - Detectors, dose measurement and phantoms

ESTRO 2024

Material/Methods:

Six members of an international Dosimetry Audit Network (DAN) participated in this study. A set of EBT3 Gafchromic films (10Sq: 10 x 10 cm 2 , 15Sq: 15 x 15 cm 2 , calibration film set, and one blank film) were irradiated with a blind dose at a single institution (the Host) and sent to the six DANs. Each DAN was required to produce dose-maps of the films using their standard film dosimetry process: (1) using the Host’s calibration film set (“Host_Cal”) and (2) using their own calibration film set (“DAN_Cal”). A 2 x 2 cm 2 region of interest (ROI) was examined on each dose-map and compared to the blind dose. In order evaluate the variability due to calibration process, scanner, or software influences, three experiments were investigated. Calibration differences were compared between the two calibration film sets used for processing (Host_Cal and DAN_Cal). To address scanner influence, all irradiated films were posted to one DAN and rescanned on a single scanner and processed using their software (“Scanner”). For software variability, the scan file taken by all DANs were re-processed in a single software application (“Software”). For the latter two assessments, the conversion of irradiated films to dose-maps were performed using the Host calibration. A tolerance level of ± 5% difference in the film dose compared to the blind dose was assumed, deviations from this were investigated. The results of this intercomparison with the different assessments for the 10Sq film are shown in Figure 1. Three DANs (DAN1, DAN2, DAN3) measured the film dose to within ±5% of the known dose for the 10Sq film and 15Sq film using both Host_Cal and DAN_Cal. These three DANs use in-house software to process their film dosimetry. Users of Vendor-based software (DAN4, DAN5, DAN6) measured film doses up to 12.2% from the known dose and had larger variations in their results (range of SD: 1.8%-4.8%) compared to in-house developed solutions (range of SD: 0.5%- 1.8%). DAN5 had scanner issues which were identified and corrected, however, when their image scans were processed with no scaling or corrections applied, both of their 10Sq and 15Sq films showed dose values 3.3 Gy higher than the known dose, suggesting that quality checks should be considered to identify scanner variability and film inhomogeneities. DAN4 discovered they were not implementing a multi-channel correction accurately when applying their calibration and re-processed their scans using the green-channel only – their film dose improved to within 1.6% of the known dose for the 10Sq film. When the films were scanned on a single scanner and processed with the same software, the results were all within +5%: -2.8% to 3.2% for 10Sq, and -4.4% to 2.8% for 15Sq. Results: