Simple but rigorous procedure for ultrasound quality assurance.

backgroundMedical ultrasound (US) system quality assurance (QA) is employed only modestly by users because of the large number of variables recommended for measurement and the relatively low use of all that information when it is acquired. By 2011 it was again clear that the equipment was not as reliable as often contended; that failures were mainly transducer, cable, and visible mechanical effects, easily addressed by uniformity tests and visual inspections. Monitor degradation also became more widely recognized. Recently it was proposed that system imaging performance QA could be accomplished by simple, cine loop acquisitions with a random hypoechoic sphere phantom (RHSP). With automated analysis, this should detect almost all B mode image quality difficulties and sources of most defects in other modes. The measures, based on the spheres' mean lesion signal-to-noise ratios (LSNRms), also would be absolute tests in 3D and valid for acceptance testing and comparisons, providing intuitive results for users to anticipate a transducer's utility in various depths and applications. This challenge to traditional phantoms has stimulated improvement in more traditional polymer filament or wire phantoms, making it easier to see poor slice thickness, but with considerable acquisition effort and cost. PURPOSE: To evaluate the potential and ease of use of the commercially available RHSPs, we compared them with a new wire phantom and an established uniformity phantom for detecting physically simulated, common transducer defects, using commercially available automated analysis.

methodsThe three types of transducer defects were simulated physically for comparison with the same, undisrupted, transducer. A layer of Styrofoam was taped over either one end of the array, one side of the array, one part of one side of the array, or not at all, essentially blocking all of the elements on one end or one side, or just part of the elevational extent of the involved elements. Scanning methods and analysis, generally described in IEC TS62736, provided median lesion signal-to-noise ratio (LSNRm) and clarity index (Ci) on the spheres, -6 dB pulse-echo response widths on the wires and median signal level on the uniformity phantom and RHSP. Several defective transducers were evaluated as well.

resultsThe RHSP phantom/analysis gave consistent results, with delineation of the lateral and the full elevational disruptions, but only meager delineation of the partial elevational disruption given chosen settings. Beam width measurements on the 90°/45° wire phantom delineated all three disruptions but were more difficult to perform and function over limited settings. The uniformity measurements revealed only the lateral disruption.

conclusionsThe RHSP phantom is now available at a considerably lower cost than wire-based phantoms. It demonstrated good detection of all but the subtlest disruption. We preliminarily recommend this phantom for US system user QA and performance evaluation, after initial verification on each model of the system of the grayscale characteristic curve and spatial measurements. Quantitative expectations for different US systems will develop rapidly, as more users will be implementing this simpler QA. Software or phantom modifications can add evaluation of spatial measurements.