ESTRO 2022 - Abstract Book

S1439

Abstract book

ESTRO 2022

Conclusion Implementation of the “SunCHECK” platform for daily, in-vivo quality control of patient’s radiotherapy treatments is possible, dose calculation results differ only minimally from those of the clinically commissioned reference treatment planning system and correspond thus with dose measurements of plans treated with a regularly functioning Linac.

PO-1643 Feasibility study of Modulated Arc Beam TBI treatment rescheduling between two Linear Accelerators

M.V. Gutierrez 1 , P. Ceroni 1 , E. Cenacchi 1 , G.M. Mistretta 1 , L. Manco 1 , L. Morini 1 , L. Boni 1 , S. Pratissoli 2 , G. De Marco 2 , N. Maffei 1 , F. Itta 1 , F. Campanaro 1 , 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 In our department modulated arc beam Total Body Irradiation (TBI) is performed on two matched beam 6MV Elekta Versa linear accelerators. If necessary, conventional VMAT treatments are rescheduled from Linac1, to the other commissioned system, Linac2. For commissioning of Linac 1, originally, a beam profile, parallel to the direction of gantry rotation and generated by an arc between 290º and 70º using a 40x40cm 2 field size, had been characterized. The aim of the present analysis was to evaluate the variability of VMAT TBI treatments between two Linacs fully matched for conventional VMAT treatments without additional profile beam characterization measurements and replanning for Linac 2. Materials and Methods Beam characterization measurements were performed with a PMMA solid phantom (Acrylic Slab T.2967, 30x30x30cm 3 , ρ =1.19g/cm 3 ) and a 0.6cm 3 TM30006 Farmer type ionization chamber. In order to provide a modulated arc beam, individual contribution and weighting factors of each sub-arc were determined. Profile acquisitions were performed for dose estimation at patient`s skin and mid prescription plane. Acquisition of a PDD curve in TBI setup was performed. In-Vivo dosimetry was implemented with diodes controlled by the dedicated Software PTW- VivoSoft . In order to verify the feasibility of rescheduling treatment between Linacs, on Linac2 only profile acquisition for dose estimation at patient`s skin and mid-plane; homogeneity; and PDD curve, were performed. For every clinical TBI treatment, pre-treatment quality assurance and dose evaluation were performed in order to validate the calculated monitor units, as well as in vivo measurements during first fraction. Results Theoretical curves proved to be accurate in prediction of weighing factors for determination of modulated arcs for both Linac1 and Linac2, with homogeneities of 2,3% & 1.42% in L-R profiles, respectively. Homogeneities derived from measurements in phantom surface were 7,06% & 4.67% with respect to CAX and 7,11% & 3.94% with respect to prescription mid-plane. PDD curves in TBI conditions of both Linacs matched within 0.4%. For 22 treated patients, 18 Linac1 and 4 Linac2, pre-treatment dose measurements differed from prescribed doses within 1.1% ± 1.86% on average. For Linac1 & Linac2, pre-treatment dose measurements were within 4.5% & 3.5% of prescribed dose; while clinical in vivo dosimetry, within 3.0% & 4.5% of prescribed dose, respectively. For clinical treatments, pre- treatment dose evaluation and in vivo dosimetry were within 5% of the prescribed dose regardless of the Linac. Conclusion VMAT-TBI Dosimetry is robust between Linacs that are fully matched for conventional VMAT treatments, adding therefore easy patient transferability to other advantages of the technique, such as Independence of treatment room size; optimal dose homogeneity throughout the body, comfort and reproducibility of patient position and short treatment times. 1 University of Texas Health Science Center at San Antonio, Department of Radiation Oncology, San Antonio, USA; 2 University of Texas at San Antonio, Department of Management Science and Statistics, San Antonio, USA Purpose or Objective The aim of this study is to compare and evaluate the performance of various machine and deep learning models to predict the delivered multileaf collimator (MLC) positions for volumetric-modulated arc therapy (VMAT) plans delivered on an Elekta linear accelerator. Materials and Methods In this study, 100 log files (70 for training and 30 for testing) containing 8,000 control points for VMAT treatment plans delivered on an Elekta linear accelerator were retrospectively obtained from a single institution. From the log files, eight planned parameters were extracted: gantry angle, collimator angle, Y1 and Y2 jaw positions, leaf gap, leaf position, leaf velocity, and leaf acceleration. These parameters were used as inputs to the models and the delivered leaf position from the log files were used as the target. The regression models examined were the linear regression, support vector, random forest, extreme gradient boosting (XGBoost), and artificial neural network (ANN). The models were trained with data from Y1 and Y2 banks and tested on leaves from both banks. Validation of the model performance was done using the mean PO-1644 Comparison of machine and deep learning models in predicting Elekta MLC leaf positions for VMAT S. Sivabhaskar 1 , R. Li 1 , N. Kirby 1 , A. Roy 2 , N. Papanikolaou 1