Room: AAPM ePoster Library
To assess whether an automatic American College of Radiology (ACR) phantom analysis toolbox could be used to facilitate ACR quality assurance (QA) on low-field MR guided radiotherapy systems.
An open-source toolbox was modified for ACR QA at low field strength of 0.35T . In particular, T2 image without intensity normalization was used for the percent signal ghosting test at the low field system per vendor recommendation. A total of 12 ACR datasets were evaluated, including 6 datasets acquired at our center from 2015 to 2020, and 6 datasets from 6 different ViewRay MR-Lianc systems. Geometric accuracy (GA), slice thickness accuracy (ST), slice position accuracy (SP), percent integral uniformity (PIU), and percentage signal ghosting (PSG) were assessed manually and using the toolbox automatically. Measurements between manual and automatic analysis were compared.
The software took less than 2 minutes to complete all analyses. The pass/fail results from automatic measurement agreed with manual results for all tests except one T2 slice thickness measurement, where the manual T2 slice thickness was 5.57 (pass) whereas only 2.88 (fail) using the automatic tool, and one PSG measurement, where manual value was 0.0247 (pass) and automatic output was 0.0258 (fail). Overall, good agreements between manual and automatic measurements were observed. The absolute difference between the two measurements were 0.79±0.74 (mm), 0.51±0.54 (mm), 0.40±0.41 (mm), 1.12±1.13 (%), and 0.0019±0.0022 for GA, ST, SP, PIU, and PSG respectively.
An automatic ACR QA tool was evaluated for the low-field MR guided radiotherapy system. Overall, the software provided comparable results as manual analysis, and reduced the processing time from about 1 hour to less than 2 minutes. Further study is needed to improve the T2 slice thickness calculation.
Funding Support, Disclosures, and Conflict of Interest: The authors would like to acknowledge the research support from ViewRay Inc.