Method for realizing automatic dose verification based on accelerator log file

Abstract

The invention discloses a method for realizing automatic dose verification based on an accelerator log file. The method comprises the steps: collecting 135 chest intensity modulation plans on a Villian Eclipse planning system, and the chest intensity modulation plans comprise 584 radiation fields; and before a radiotherapy plan is implemented, implementing intensity modulated verification by using a linear accelerator delivery plan, and besides, collecting log files in the linear accelerator plan delivery process. A deep learning algorithm is used, the log file in the accelerator delivery process is used as the input of a model in a flux graph form, a gamma passing rate prediction model under different threshold standards is established, and the accuracy of the prediction model is verified. The flux graph formed based on the accelerator log file serves as input to train the gamma passing rate prediction model, real delivery parameters are considered, the individualized dose verification result can be accurately predicted, automation of dose verification before treatment is achieved, the dose verification efficiency is improved, and a physician is allowed to have more time to pay attention to the reason of dose verification failure.

Description

technical field [0001] The invention relates to the technical field of medical radiation therapy, in particular to a method for realizing automatic dose verification based on accelerator log files. Background technique [0002] As the complexity of modern radiotherapy techniques continues to increase, so do the requirements for the quantity and quality of dose verification. At present, the progress of modern radiotherapy technology is mainly reflected in: 1) The conformity is continuously increased to increase the dose of the tumor target area, and at the same time significantly reduce the radiation dose of surrounding normal tissues and organs; 2) The physical dose distribution in the target area is purposefully allocated according to the requirements ; 3) The dose of a single split is increased. The development of these technologies has caused a significant increase in the complexity of plan design and implementation, and at the same time, the uncertainty of dose error ha...

AI Technical Summary

Problems solved by technology

The measurement-based dose verification process is cumbersome and time-consuming and highly dependent on equipment, equipment resolution, probe angle response, etc. will affect the results

At present, the individualized dose verification methods generally have problems such as low efficiency and relatively large limitations of verification points and surfaces.

This makes routine individualized QA time-consuming and laborious, making it more difficult for treatment centers with heavy tasks to achieve individualized QA. Therefore, it is very important and urgent to find a more time-saving, labor-saving and comprehensive effective dose verification method

Although the existing automated individualized dose verification method uses machine learning methods to establish a two-dimensional pass rate prediction model based on planning complexity parameters, this method does not take into account the impact of real parameters on dose verification during accelerator execution. Therefore, it is not suitable for dose verification under real exposure conditions of patients

Images