Author: Ted Lynch 👨🔬
Affiliation: Sun Nuclear, a Mirion Medical Company 🌍
Purpose: A novel ultrasound quality assurance method is presented that constructs “synthetic spheres” at arbitrary locations within the image frame by combining correlation length measurements from a uniform tissue mimicking phantom with noise measurements obtained using an in-air scan. Calculation of lesion signal-to-noise ratio (LSNR) for each simulated lesion can then be used to map ultrasound system performance over the full image frame.
Methods: Correlation length measurements from the uniform phantom were used to estimate speckle fill-in at the border of simulated voids, while in-air image data provided the noise signal within the rest of the simulated void. These synthetic spheres were then combined with background signal levels from the phantom image to calculate LSNR over an MxN grid within the image field. Three transducers were tested to assess the method’s ability to detect changes in system performance due to changed imaging presets, simulated dead transducer elements, and simulated lens damage. Additional test results from transducers with damaged lenses are also presented.
Results: Turning off harmonic imaging reduced the integrated LSNR over the full image depth by an average of 28%, while decreasing the system dynamic range from 86 dB to 30 dB decreased the integrated LSNR by over 200%. Two-dimensional maps of LSNR were generated that demonstrate the ability to detect localized regions of reduced image contrast due to dead transducer elements, while plots of the mean LSNR vs. depth show promise for detecting changes in the elevational focus due to lens damage.
Conclusion: The presented results demonstrate successful proof-of-concept of the “synthetic spheres” methodology for detecting changes in ultrasound system performance for quality assurance testing. Additional work comparing the accuracy of the method versus standard quality assurance measurements is needed prior to clinical implementation.