Feasibility of An Automated Quality Assurance Tool for Ultrasound Shear Wave Elastography Images
D Gomez-Cardona1*, SF Stekel2, DJ Tradup2, NJ Hangiandreou2, Z Long2, (1) Gundersen Health System, WI;(2) Mayo Clinic, Rochester, MN
Presentations
WE-IePD-TRACK 1-4 (Wednesday, 7/28/2021) 5:30 PM - 6:00 PM [Eastern Time (GMT-4)]
Purpose: Ultrasound shear wave elastography (SWE) can estimate tissue Young’s modulus in selected regions of interest (ROI). It relies on careful acquisition and measurement by the operator. This study aims to develop an automated quality assurance (QA) tool for clinical liver SWE images on GE LOGIQ E9 scanners, which do not have an onboard quality indicator.
Methods: Settings on LOGIQ E9 scanners which could affect SWE quality and quantitative measurement were investigated. A suite of custom MATLAB programs was built to check these settings and institutional protocol details, which were designed to increase the frame rate on this scanner model and minimize motion artifacts as reported before, from liver SWE DICOM images. Specifically, settings such as transducer model (C1-6D), acquisition parameters such as elastography gain, width of the SWE acquisition box, distance between the liver capsule and the SWE acquisition box, push and track beam output, percentage of color filling of the acquisition box, and diameter of the measurement ROI. The distance from rib shadowing artifact, if any, to the SWE box at transducer face level was also checked. Color-coded QA outputs were overlaid on the original SWE images off the scanners. Suggestions for optimization strategies were provided when certain conditions were met.
Results: The QA tool has been established and retrospectively processed 1455 images acquired since 2018 by clinical sonographers. 1.1% of all parameters and settings was incorrect or could have been improved. The default elastography gain was found to be modified in 36 earlier images. This gain relates to cross correlation threshold for selecting pixels to be included in the quantitative measurement and therefore should be unchanged.
Conclusion: Feasibility of an automated QA tool for liver SWE images that could provide users near real-time QA results and feedback was demonstrated. Artifact recognition will be included in the future.
ePosters
Keywords
Not Applicable / None Entered.
Taxonomy
Contact Email