RESUMEN
INTRODUCTION: New radiotherapy machines such as Halcyon are capable of delivering dose-rate of 600 monitor-units per minute, allowing large numbers of patients treated per day. However, patient-specific quality assurance (QA) is still required, which dramatically decrease machine availability. Innovative artificial intelligence (AI) algorithms could predict QA result based on complexity metrics. However, no AI solution exists for Halcyon machines and the complexity metrics to be used have not been definitively determined. The aim of this study was to develop an AI solution capable of firstly determining the complexity indices to be obtained and secondly predicting patient-specific QA in a routine clinical setting. METHODS: Three hundred and eighteen beams from 56 patients with breast cancer were used. The seven complexity indices named Modulation-Complexity-Score (MCS), Small-Aperture-Score (SAS10), Beam-Area (BA), Beam-Irregularity (BI), Beam-Modulation (BM), Gantry and Collimator angles were used as input to the AI model. Machine learning (ML) and deep learning (DL) models using tensorflow were set up to predict DreamDose QA conformance. RESULTS: MCS, BI, gantry and collimator angle are not correlated with QA compliance. Therefore, ML and DL models were trained using SAS10, BA and BM complexity indices. ROC analyses enabled to find best predicted probability threshold to increase specificity and sensitivity. ML models did not show satisfactory performance with an area under-the-curve (AUC) of 0.75 and specificity and sensitivity of 0.88 and 0.86. However, optimised DL model showed better performance with an AUC of 0.95 and specificity and sensitivity of 0.98 and 0.97. CONCLUSION: The DL model demonstrated a high degree of accuracy in its predictions of the quality assurance (QA) results. Our online predictive QA-platform offers significant time savings in terms of accelerator occupancy and working time.
RESUMEN
PURPOSE: Full integration of EPID-based dosimetry in a global quality control workflow is still complicated. All the actual solutions are based on a relation between image gray-level signal and total linac-delivered dose. In this study, we propose a simple algorithm relying pixel gray-level of EPID image with average linac delivered dose per acquisition frame. METHODS: Calibration models are constructed for Varian and Elekta linacs including scattering conditions and EPID-arm backscatter-specific corrections. Only simple homogeneous fields are required to establish the EPID dose conversion model for each x-ray beam. Then, the model was evaluated by comparing calculated and converted dose distributions for homogeneous and modulated beams using gamma maps. RESULTS: To fit average dose per frame (Dfnorm ) vs pixel gray value (Ngnorm ) of each EPID image, a logarithmic curve Dfnorm=A+B∗lnNgnorm-C, has been chosen where A, B and C are constants depending on beam energy. Gamma comparison (2%, 2 mm, threshold 15%) between converted images and calculated dose distributions for linac control and pretreatment patient fields led to a gamma pass rate higher than 97% for all the analyzed fields. CONCLUSIONS: Without a prior irradiation settings knowledge except the incident energy beam, we use EPID as a reliable dose to water detector for both homogeneous and modulated beams.