An initial quantitative comparative study of the contrast resolution of different ultrasound imaging transducers

Abstract

Background: The current established guidelines to assess the contrast resolution in ultrasound (US) systems are the American Association of Physicists in Medicine (AAPM) and the Institute of Physics and Engineering in Medicine (IPEM). These suggest that tests are done qualitatively. Therefore, the need to investigate contrast resolution objectively is crucial to limit subjectivity as much as possible. Objectives: The study’s objectives revolve around finding quantitative methods to analyse contrast resolution in US and using them to investigate the contrast resolution behaviour across different US probes/transducers, frequencies and depths. Research Methodology: Data was gathered from 21 US systems equipped with 17 linear, 20 curvilinear and 15 sector probes. For data to be collected, the CIRS Model 040GSE tissue equivalent test object (TETO)/phantom, was used to obtain contrast resolution images at two different depths: 3 cm and 11.5 cm. Quantitative data was extracted through the open-source software, ImageJ. Three methods were then considered: the index contrast (IC), the gamma of the system, and the contrast-to-noise ratio (CNR). Any graphs and tables were generated by Microsoft Excel Version 16.86. Results: From the methods assessed, it was found that the IC increases with hyperechoic structures, for all transducer types. All probes had varying gamma values. However, it was found that it depends on the type of transducers and not on the frequency. All transducer types established characteristic U-shaped scatter plots for the CNR. Considering all probe types, it was found that the sector probes established a better performance at the near and far field zones. Lastly, from most data plotted, it was found that the contrast resolution is not frequency-dependent. Conclusions and Recommendations: In this study, the results drawn were overall consistent with what the literature suggests. However, this research could be improved by using a larger pool of data so the full ultrasonic frequency range can be investigated. Moreover, the way of how data was extracted from ImageJ could be improved by using equal circular selections when working on the different images. Eventually, the results drawn from this study can allow the current standards to test the contrast resolution to be possibly updated and used in clinical testing procedures.