A novel ultrasound phantom to quantify the effect of slice thickness on imaging performance

Abstract

Whilst it is well understood that linear and linear matrix probes have poorer resolution in the elevation plane than in the scan plane, the impact this has on diagnostic imaging of low contrast anechoic structures is yet to be fully quantified. Previous work [1] has introduced the concept of the resolution integral as a clinically relevant method of quantifying ultrasound imaging performance. This method combines resolution measurements made at different depths to produce a dimensionless figure of merit (R), a characteristic resolution (DR) and a depth of field (LR). Over 350 probes have been characterized this way using a standard set of anechoic pipes embedded in an agar based tissue mimicking material (TMM) known as the Edinburgh Pipe Phantom (EPP) [2].In this study, a modified version of the EPP was constructed using six flat anechoic bars embedded in TMM. The bars were 15 mm wide, assumed to be greater than the slice thickness of three linear (Siemens VF13-5, SonoSite L38x/136 and L38x/10-5) and two multi-row array (Siemens VFX13-5 and VFX9-4) probes. The ability of these probes to image the bars was therefore not dependent on slice thickness. Resolution integrals were measured using this new phantom and compared to results obtained using the EPP.When slice thickness effects were eliminated, R increased by a factor greater than 2.5 for all probes, and DR improved by 50-70% compared to values obtained using the EPP. Slice thickness is therefore a significant limiting factor in imaging small anechoic structures using linear and linear matrix probes. This highlights the need for continued use of quality assurance (QA) phantoms such as the EPP that test slice thickness, and for better elevational focusing in clinical applications where the detection of small anechoic structures is a priority.