No abstract

Introduction: The aim of this study was to evaluate the new 2-Dimensional diode array SRS MapCHECK (SunNuclear, Melbourne, USA) with dedicated phantom StereoPHAN (SunNuclear, Melbourne, USA) for the pre-treatment verification of the stereotactic body radiotherapy (SBRT). Material and methods: For the system, the short and mid-long stability, dose linearity with MU, angular dependence, and field size dependence (ratio of relative output factor) were measured. The results of verification for 15 pre-treatment cancer patients (5 brains, 5 lungs, and 5 livers) performed with SRS MapCHECK and EBT3 Gafchromic films were compared. All the SBRT plans were optimized with the Eclipse (v. 15.6, Varian, Palo Alto, USA) treatment planning system (TPS) using the Acuros XB (Varian, Palo Alto, USA) dose calculation algorithm and were delivered to the Varian EDGE® (Varian, Palo Alto, USA) accelerator equipped with a high-definition multileaf collimator. The 6MV flattening-filter-free beam (FFF) was used. Results: Short and mid-long stability of SRS MapCHECK was very good (0.1%-0.2%), dose linearity with MU and dependence of the response of the detector on field size results were also acceptable (for dose linearity R2 = 1 and 6% difference between microDiamond and SRS MapCHECK response for the smallest field of 1 × 1 cm2). The angular dependence was very good except for the angles close to 90° and 270°. For pre-treatment plan verification, the gamma method was used with the criteria of 3% dose difference and 3 mm distance to agreement (3%/3 mm), and 2%/2 mm, 1%/1 mm, 3%/1 mm, and 2%/1 mm. The highest passing rate for all criteria was observed on the SRS MapCHECK system. Conclusions: It is concluded that SRS MapCHECK with StereoPHAN has sufficient potential for pre-treatment verification of the SBRT plans, so that verification of stereotactic plans can be significantly accelerated.

The use of machine learning algorithms (ML) in radiotherapy is becoming increasingly popular. More and more groups are trying to apply ML in predicting the so-called gamma passing rate (GPR). Our team has developed a customized approach of using ML algorithms to predict global GPR for electronic portal imaging device (EPID) verification for dose different 2% and distance to agreement 2 mm criteria for VMAT dynamic plans. Plans will pass if the GPR is greater than 98%. The algorithm was learned and tested on anonymized clinical data from 13 months which resulted in more than 3000 treatment plans. The obtained results of GPR prediction are very interesting. Average specificity of the algorithm based on an ensemble of 50 decision tree regressors is 91.6% for our criteria. As a result, we can reduce the verification process by 50%. The novel approach described by our team can offer a new insight into the application of ML and neural networks in GPR prediction and dosimetry.

No abstract