ESTRO 2022 - Abstract Book

S24

Abstract book

ESTRO 2022

an advantage for some applications. Regardless, it will always be necessary to correct for the LET dependency of both dosimeter compositions.

MO-0050 Optically stimulated luminescence dosimetry as alternative for radiochromic film in UHDR e-beams?

V. Vanreusel 1,2,3 , F. Galante 4 , A. Gasparini 2,3 , P. Leblans 5 , G. Mariani 4 , M. Pacitti 4 , D. Vandenbroucke 5 , G. Felici 4 , L. de Freitas Nascimento 1 , D. Verellen 2,3 1 SCK CEN, Research in Dosimetric Applications, Mol, Belgium; 2 University of Antwerp, Faculty of Medicine and Health Sciences, Antwerp, Belgium; 3 Iridium Netwerk, Medical Radiation Physics, Wilrijk, Belgium; 4 Sordina IORT Technologies S.p.A., Research and Development, Aprilia, Italy; 5 Agfa NV, Corporate Innovation Office, Mortsel, Belgium Purpose or Objective There are strong indications that ultra-high dose rates (UHDR) used for FLASH-radiation therapy (RT) strongly reduce the adverse effects of conventional RT. The radiobiological mechanisms and exact conditions for obtaining the FLASH-effect are, however, under investigation. Radiochromic film is generally used for dosimetry in this field, but is subject to drawbacks such as non-linear calibration and a cumbersome read-out procedure. Optically stimulated luminescence (OSL) dosimetry is therefore investigated as alternative. Materials and Methods Measurements were performed using the ElectronFlash linac Error! Hyperlink reference not valid. ) which generates pulsed electron beams and allows large systematic variation of the important parameters for FLASH-RT research. The linear response to dose per pulse (DPP) and total dose, and the stability with pulse repetition frequency (PRF) were investigated for a phosphor coated OSL film optimized for radiation therapy ( Error! Hyperlink reference not valid. ) and compared with EBT-XD GafChromic film Error! Hyperlink reference not valid. . The DPP ranged from 0.0-9.6 Gy by variation of pulse length and/or applicator diameter. Total dose ranged from 0-20 Gy by variation of the number of pulses, both at conventional- and UHDR. The PRF was varied from 1-245 Hz. The OSL film was irradiated in a dark environment and read out by a computed radiography (CR) digitizer, adapted for conventional RT. The radiochromic films were scanned 48 hours after irradiation using an EPSON 10000XL Error! Hyperlink reference not valid. transmission scanner and their calibration curve was obtained against an Advanced Markus ( Error! Hyperlink reference not valid. ) ionization chamber in a conventional electron beam. Results Radiochromic film and the OSL systems showed good linearity with increasing DPP up to 12 Gy/pulse (R ² > 0.995) and with dose (R ² >0.999). For the OSL system linearity was maintained up to 20 Gy. Radiochromic film has a nominal dynamic range up to 60 Gy. No difference between conventional- and UHDR was observed. PRF stability was examined as the percentage difference from the 1 Hz signal. The OSL system showed maximal deviation of -2.2 % at 245 Hz. Radiochromic film showed maximal deviation of 2.5 % at 5 Hz. Conclusion With the good linearity with DPP and total dose, and good PRF stability, the OSL system has the potential to be a valid alternative to radiochromic film as reference dosimeter in UHDR electron beams. However, further investigation of the important characteristics is needed and currently ongoing. It has strong benefits with the straight forward read-out procedure, linear calibration and reusability of the films. Modification of the coating and digitizer optimization are being investigated to further improve the OSL characteristics in UHDR electron beams. The need for irradiation in dark conditions is currently a drawback of the system. Also a dark fading curve is required to correct for time differences between irradiation and read-out. This work is part of the 18HLT04 UHDpulse project which received funding from the EMPIR programme. The first author is funded via the young potential PhD grant from SCK CEN. 1 University of São Paulo, Department of Physics - FFCLRP, Ribeirão Preto - SP, Brazil; 2 University of São Paulo, Department of Physics, Ribeirão Preto - SP, Brazil Purpose or Objective Three-dimensional dosimetry using gel has an increasingly application in radiotherapy, being the polymeric gels with dose distribution reading using magnetic resonance images (MRI), on class of these dosimeters. An important step in the dosimetry process, which is recommended for chemical dosimeters in general, is the need to calibrate their batches. This process is laborious and requires significant time for its achievement. To reduce the time involved in this process, we developed an algorithm that predicts the calibration curve slope or sensitivity, based on machine learning and on the texture characteristics extracted from the MRIs of the non-irradiated gel samples through radiomics. Materials and Methods 50 transversal MRIs acquired of calibration vials irradiated with different doses between 0 and 10 Gy from 4 different MAGIC-f gel dosimeters batches were used. All the images were acquired with the same MRI protocol and each image contained all the different doses applied for the calibration of that specific batch of gel. Matlab software was used to access all the calibration curves, one per MRI, and determine the coefficients of calibration curve the linear fitting. A set of decision tree (forest) was used to build the regression model that was fed with radiomics data of the zero gray or non- MO-0051 Machine learning prediction of calibration curves in gel dosimetry J. Pavoni 1 , K. Tozzi 1 , J. Lizar 2

Made with FlippingBook Digital Publishing Software