P072] MRI quality control and system performance using ACR phantom tests and automated software

Abstract

Purpose For monitoring the performance of Magnetic Resonance Imaging (MRI) systems, it is essential to implement regular quality control (QC) protocols. We evaluated the MRI system performance of the Philips Ingenuity TF PET/MR, using the American college of radiology (ACR) QC protocol [1] and automated QC software. Methods A total of 80 QC scans were performed during years 2013 to 2018 using the ACR QC protocol. In the initial analysis, 21 scans were analyzed by a human observer and automatic software implemented in MATLAB2015b [2] , which performs the analysis according to ACR guidelines [1] . Five quantitative tests were included: geometric accuracy (148 ± 2 and 190 ± 2 mm), slice thickness accuracy (5.0 ± 0.7 mm), slice position accuracy ( 82 %) and percent signal ghosting ( Results For geometric accuracy, (M = 147 ± 1), (A = 148 ± 1) and (M = 190 ± 0.5), (A = 191 ± 0.2) were measured for the localizer and T1 series. For slice thickness accuracy, (M = 5.5 ± 0.5), (A = 4.7 ± 0.8) and (M = 5.9 ± 0.5), (A = 4.5 ± 1.4) were measured for the T1 and T2 series. For slice position accuracy, (M = 5.6 ± 3.3), (A = 4.2 ± 2.7) and (M = 5.3 ± 2.8), (A = 4.0 ± 2.5) were measured for the T1 and T2 series. For image intensity uniformity, (M = 85 ± 1), (A = 96 ± 1) and (M = 89 ± 2), (A = 97 ± 2) were measured for the T1 and T2 series. For percent signal ghosting, (M = 0.0010 ± 0.0009) and (M = 0.0015 ± 0.0007) were measured. The time to analyze and report was two minutes with the automatic software. Conclusions Majority of the tests reached the acceptance criteria in the initial analysis. Slice thickness and slice position accuracy were the most common tests to fail. Automated software was beneficial for easing the analysis workflow. The QC software with a test dataset will be provided during the conference.