TH-CD-304-03: A Novel and Intelligent Strategy for Automated On-Board Imager Quality Assurance

Abstract

Purpose: Traditionally, on-board imager (OBI) quality assurance (QA) tests are manually and subjectively performed for the kV/MV imaging system, lasting several hours. To improve test efficiency and accuracy, eliminate the inflexibility of using third party tools, and minimize potential human errors, we design a novel, automated, and intelligent QA strategy. Methods: The proposed strategy includes compact phantom design, automated EPID-image acquisition, and a modular computational platform for image analysis. For reduction of setup errors and human interaction, two compact phantoms are designed to encompass X-ray tube output, geometric, and image quality tests. QA tests and image acquisitions are automatically completed via XML controlled machine operation following the phantom setup. A compact modular computational platform is developed to automatically analyze the acquired images, generate reports, manage the QA database, and provide information for strategy self-optimization. Results: The entire strategy only requires one initial manual phantom setup, significantly minimizing the time of user intervention, avoiding potential setup errors, and eliminating the necessity of using various QA tools. Compared to traditional tests often requiring several hours, the proposed intelligent strategy considerably speeds up the QA process, requiring only 10–15 minutes. With the in-house implemented computational platform, the test results can be clearly reported in both graphical and textual representations within 2 minutes. The analysis of images acquired from three Varian Truebeam machines shows sub-millimeter detector positional accuracy; kVp accuracy of +/−2 kVp interval, and exposure accuracy and linearity within 2%. Conclusion: We introduce an automated and intelligent OBI QA strategy that enables cross-calibration, robust data analysis, and QA test validation. The strategy eliminates the inflexibility and incompatibility obstacles of using third party QA devices. It is capable of collecting long-term multi-machine/multi-institution QA data, and thus has great potential for making the self-optimized intelligent QA environment a reality. Research Funding from Varian Medical Systems Inc. . Dr. Sasa Mutic receives compensation for providing patient safety training services from Varian Medical Systems, the sponsor of this study.