MO-D-218-02: Ultrasound Phantoms in Image Quality Measurements and Performance Assessment

Agreement has recently emerged for measurement of only three parameters in routine quality assurance. These include element or channel failure (EOCF), maximum depth of penetration (MDOP) and distance measurement accuracy (DMA). The related study leading to this agreement was done at the Mayo Clinic ["Four-year experience with a clinical ultrasound quality control program," Hangiandreou et al. Ultrasound Med & Biol, vol. 37, pp. 1350-1357 (2011)]. A variety of phantoms have been proposed for performing these measurements. These include: 1) a phantom with both flat and cone-shaped scanning windows, tissue-mimicking material and parallel nylon fibers for measuring all three parameters; 2) a starch suspension in water phantom for assessing ECOF; 3) polyurethane phantom for determining ECOF and estimating MDOP; 4) a silicone phantom for determining ECOF. Determination of ECOF and MDOP requires acquisition of image cine loops including many statistically independent versions; analysis of the data requires averaging of the images in the cine loops, and software is being tested at the time of this writing for user-friendly downloading and averaging of the cine images. The phantoms will be described and examples of procedures and results for making measurements. One method for assessing imaging performance involves quantifying the detectability of small low-echo spheres as a function of depth; detectability will depend on sphere size and on depth. It can be argued that the smaller the sphere and the greater the detectability, the better the scanner can delineate the boundary of an inclusion such as a spiculated cancer. Phantoms and software will be described which allow quantitative determination of detectability of low-echo spheres ( 40 dB) as a function of depth; one phantom has a spatially random distribution of 4-mm diameter low-echo spheres for lower frequencies, and the other 2-mm diameter spheres for higher frequencies. The parameter that is quantified is the mean lesion signal-to-noise ratio (LSNR) as described for cylinders by Lopez H et al. IEEE Trans Med Imaging, 1992 and for spheres by Kofler et al. Ultrasound Med & Biol, 2005. The phantoms accommodate any size and shape transducer. Note that spheres have no preferred orientation; therefore, phased arrays and convex (curved) arrays with a broad range of radii of curvature and sector angles can be can be assessed. Phantoms for assessing other aspects of performance such as quantitative determination of the contrast of large objects in a background will also be discussed.