Automated Analysis of QA Images Acquired with Low-Field MRgRT
R Moskwa1*, B Paliwal2, I Tanumihardjo2, P Yadav2, (1) Medical Physics, University of Wisconsin-Madison, Madison, WI, (2) School of Medicine and Public Health, University of Wisconsin-Madison, Madison, WI
Presentations
PO-GePV-M-172 (Sunday, 7/25/2021) [Eastern Time (GMT-4)]
Purpose: Test the accuracy and reproducibility of Magphan(R) phantom, designed for quality assurance (QA) needs of MRI-guided radiotherapy. It allows automated measurements for image analysis and acquisition of important imaging parameters.
Methods: QA was performed weekly. Magphan(R) RT phantom in a low-field MR-LINAC system was scanned over a period of 10 months. At least 35 measurements for each imaging parameter were acquired. Three MRI sequences were used for QA: TRUFI, T1, and T2. DICOM files were used to calculate the following imaging parameters: geometric distortion, uniformity, resolution, signal-to-noise ratio (SNR), and laser alignment. Point spread function and edge spread function are also calculated for resolution analysis. All data was collected and analyzed for accuracy and reproducibility. Mean, median, standard deviation, maximum and minimum values were used in combination with plots to check for data consistency over the 10-month period.
Results: Data showed small standard deviation and high consistency for each tested imaging parameter with each MRI sequence. Highest variability in data was seen with the TRUFI sequence, which was expected because of low resolution and short scan time (25sec). Point spread function, edge spread function, signal uniformity and SNR measurements remained consistent over 10 months. Laser alignment traditional offsets and angular deviation remained consistent over 10 months, except for one day where phantom was misaligned. Distortion analysis, performed for two different phantom fiducial diameters per imaging sequence, remained consistent for both diameters.
Conclusion: The Magphan(R) RT phantom showed to be both reliable, reproducible, and easy to use. Imaging parameters that are not available with the current accepted low-field LINAC QA phantom, such as laser alignment and geometric distortion, were found to be both important and relevant. These results show promise for a more streamlined, time-saving, and less error-prone low-field MR-LINAC QA workflow when compared to current practices.
ePosters
Keywords
Quality Assurance, MRI, Image Analysis
Taxonomy
IM/TH- MRI in Radiation Therapy: MRI/Linear accelerator combined Quality Assurance
Contact Email