ESTRO 2022 - Abstract Book

S1438

Abstract book

ESTRO 2022

Conclusion Both qualitative and quantitative results are encouraging and suggest that GANs are a viable approach to deal with missing modalities when applying multi-modal segmentation models.

PO-1642 Validation of photon convolution-superposition algorithm applied to daily in vivo QC in radiotherapy

G.M. Mistretta 1 , F. Itta 1 , P. Ceroni 1 , M.V. Gutierrez 1 , N. Maffei 1 , F. Campanaro 1 , A. Bernabei 1 , F. Marino 1 , F. Lohr 2 , G. Guidi 1

1 University Hospital of Modena, Medical Physics, Modena, Italy; 2 University Hospital of Modena, Radiotherapy Unit– Oncology and Hematology, Modena, Italy Purpose or Objective Implementation of a third-party software for daily independent in vivo quality control of radiotherapy treatments that performs calculation of daily fraction doses with a collapsed cone superposition convolution (CCSC) algorithm by using log- file data and/or EPID doses may further strengthen the QC chain of complex treatment paradigms. The clinical implementation of such a software, requires an appropriate study of the applied algorithm. The aim of this work was the evaluation and validation of the CCSC algorithm of SunNuclear Dose Calculator -DoseCHECK & PerFRACTION- software, for clinical implementation of daily dosimetrical QC of radiotherapy treatments. Materials and Methods The “Raystation” treatment planning system and the “DoseCalculator Algorithm” were commissioned for use with Elekta VersaHD linear accelerators, with beam energies of 6, 10, 15 MV, 6 & 10 MV FFF. All measurements were performed with a water phantom; using Semiflex & PinPoint ionization chambers. Based on international guidelines, several types of treatments were planned to be delivered in order to evaluate agreement between: • calculated and measured point dose in water phantom (IAEA-TECDOC-1540); • calculated and measured dose distribution for different complex beams (TG-119) planned with the techniques IMRT, dMLC and VMAT; Calculated and measured point doses in a Cheese phantom for different complex beams of different dimensions and for different dose prescriptions, with targets On Axis and Off Axis; • calculated and measured point dose in an inhomogeneous phantom for 3DCRT and VMAT techniques for different dose prescriptions. The agreement between calculated and measured dose distributions was performed by using the 2D Array detector MatriXXEVOLUTION (IBA) and myQA software, using Gamma Index (2mm@ 2% Global/Local Dose and 3mm@ 3% Global/Local Dose) as comparison index. Results Tolerance of all obtained values were inside the evaluation criteria ranges suggested by international guidelines. An excellent agreement was found between point dose calculations and measurements for all energies. Mean relative errors comparing measured doses and Raystation calculations with SDC software were (-0.39±0.95) % and (-0.43±1.05) % respectively for On Axis points (-0.29±0.95) % and (-0.68±0.59) % respectively for Off Axis points. Gamma analysis between SDC calculated and measured dose maps, both with individual passing rates higher than 95%, had a mean excursion difference of (0.33±0.12) % with a Gamma Index 3mm@3% Global Dose. Verification in Cheese and inhomogeneous phantoms showed results inside required tolerances.