ESTRO 2024 - Abstract Book

S3422

Physics - Dose calculation algorithms

ESTRO 2024

Hubert Szweda 1 , Liu Hong 2 , Matthias Meichtry 1 , Zema Chowdhuri 1 , Michele Togno 1 , Jan Hrbacek 1 , Dario Veghini 1 , Damien Weber 1,3,4 , Arturs Meijers 1 1 Paul Scherrer Institut, Center for Proton Therapy, Villigen, Switzerland. 2 Ion Beam Applications S.A., Dosimetry Particle Therapy, Louvain-la-Neuve, Belgium. 3 University Hospital of Zürich, Department of Radiation Oncology, Zürich, Switzerland. 4 Inselspital, Bern University Hospital, University of Bern, Department of Radiation Oncology, Bern, Switzerland

Purpose/Objective:

Patient-specific quality assurance (PSQA) is a critical component of every radiotherapy department QA program, ensuring the accuracy and safety of treatments. Traditionally, PSQA involves comparing the calculated dose distribution of each treatment field with the measured dose, acquired with a suitable 2D detector array. This process can be time and resource consuming, requiring additional beam time, which is particularly expensive in proton therapy facilities. However, recent advancements have led to the development of more efficient and sensitive techniques. One such technique is the utilization of independent Monte Carlo (MC) dose calculations, combined with treatment delivery log files analysis. This solution not only enhances the level of independence in 3D dose verification but also streamlines the entire PSQA process, making it more efficient and automated, significantly reducing beam time utilization. Treatment logs analysis allows for checking the actual delivered spot map, spot MU and position error for each energy layer. This study focuses on the potential replacement of the conventional PSQA method with myQA iON, the commercially available web-based PSQA platform developed by IBA Dosimetry. The MC dose engine integrated into myQA iON eliminates the need for extra dose measurements, simplifying implementation for clinical practice. Additionally, we developed and validated the integration of IBA Dosimetry software with the Varian ProBeam vers. 4.1 treatment log files. To validate the beam model of the Monte Carlo dose engine, we compared 50 SOBP-type fields calculated in a water phantom. The rectangular targets differed in size and depth, ranging from 1.5cm to 32cm in a phantom. Additionally, we retrospectively investigated 25 clinical treatment plans, encompassing various parameters, such as the size of the irradiated volume, use of a range shifter, number of monitor units and number of fields. Exported plans from Eclipse Treatment Planning System vers. 16.1 were compared with Monte Carlo calculated dose distributions in myQA iON software, using 3%/3mm gamma criteria. Furthermore, we selected one head and neck patient and tracked treatment log files over a span of nine fractions. The MC dose distribution reconstructed with each fraction log file was compared against the nominal TPS dose, for each treatment field. The trend line was also determined by the software, describing how 3D gamma index analysis results vary throughout the treatment course for ensuring the quality and safety of patient treatment delivery. Material/Methods:

Results:

The validation of the beam model revealed an average 3D gamma passing rate of 99.4%±1.4% for the 3%/3mm criteria and 95.1%±3.6% for the 2%/2mm criteria, across all 50 SOBP-type fields. These results consistently demonstrate a high level of agreement. In the clinical cases, the gamma passing rate achieved a value of 99% ±1.5% for the 3%/3mm criteria. Furthermore, when focusing specifically on the Planning Target Volume structures, the agreement reached a level of 99.7%±0.4%. Additionally, the agreement between the TPS calculated dose distributions and the log-