9. Definition of a criterion to evaluate linac capacity to precisely deliver a VMAT treatment plan

Introduction With dosimetric optimization, VMAT is a treatment technic allowing high complexity planification. This was demonstrated with dosimetric competitions organized by Saudi Arabia in 2016 (beast with nodes case), and in 2017 (head and neck case). The aim of both sessions was to approach a solution defined by very restrictive criteria for covering target volumes and saving the organs at risk in contact with the targets. A strong modulation of intensity enabled us to achieve these objectives. We analyze capacity of linac to deliver this kind of a treatment (see: Dosimetry evaluation for VMAT breast and nodes treatment, oral communication, JS SFPM 2016). We observed that with the increase level of complexity, differences also increase between simulation on Treatment Planning System (TPS) and what linac delivers actually. The objective of this study is to define predictive criteria about linac capacity to deliver a treatment conforming to simulation. These criteria will make possible to elaborate more complex dosimetric simulation, while being more confident about linac capacity for delivering the treatment. Methods Quantification of the patient QA through analysis of measurements containing known errors (shit of phantom position, absolute dose, and gantry). Analysis of patient QA according to different dosimetric parameters defined as new criteria. Results Measurement analysis made possible to establish the nature of error revealed by the gamma index. We have separated measurement uncertainties from uncertainties due to the inability of the accelerator to deliver the TPS simulation precisely. First results show a correlation between the number of monitor units (UM) per arc degree normalized to the prescribed dose of each arc and the value of gamma index. Conclusions This study defines the origin of measure deviations and therefore evaluates deviation from the inability of the accelerator to precisely deliver the simulation of the TPS. The definition of the criterion (number of CUs per degree of arc normalized to the prescribed dose) seems relevant in order to understand the complexity level of the simulated arc. The continuation of this study should make possible to know a priori the capacity of the accelerator to deliver precisely a treatment and thus adapt the dosimetric practice according to this criterion. Other criteria are being studied in order to quantitatively refine our initial results.