Reducing the effects of specular scatterers on QUS imaging using thegeneralized spectrum

The effective scatterer diameter (ESD) and effective acoustic concentration(EAC) are quantitative ultrasound (QUS) imaging parameters that employscattering models and spectral fit methods to characterize tissuemicrostructure. These methods work best when the region of interest (ROI) fromwhich the parameters are derived contains uniform diffuse scatterers. In sometissues, specular scatterers (e.g., calcifications, blood vessels, etc.) canexist and cause decreases in the accuracy and precision of QUS parameterestimates based on diffuse scattering. In this study the generalized spectrum(GS) intercept parameter was used to detect echoes from specular scatterers. Thesignals corresponding to the specular scatterers were then removed in order toreduce the effect of specular scatterers on QUS estimates. Backscatter data froma simulated phantom, rat mammary tumors, and fresh beef liver samples thatunderwent elevations in temperature were analyzed to evaluate the effectivenessof using the GS intercept parameter. The ESD and EAC were estimated assuming aspherical Gaussian scattering model for each data block outlined in an ROI inthe sample. The GS intercept parameter was estimated for each data block andused to sort data blocks and their corresponding QUS estimates were sorted intodiffuse and specular scattering groups. Modified parametric images were thenformed by using only the data blocks in the diffuse scattering group. For thesimulated phantom, the exclusion of specular scatterers from the QUS estimatesresulted in a reduction in ESD standard deviation of 66.4%. For the rat mammarytumors, the average reduction in ESD and EAC standard deviation was 17.1% and24.8%, respectively. When monitoring the changes in ESD and EAC in beef liversamples versus temperature over the temperature range of 37 to 50 C, the meanESD and EAC values changed monotonically with temperature. By excluding thespecular scatterers, ESD and EAC were observed to change by 25.4% and 40.3%respectively as opposed to 14.8% and 30.7% respectively when including specularscatterers. When excluding specular scatterers from QUS analysis, the precisionof QUS estimates was improved and the sensitivity of QUS estimates totemperature changes was increased. These results suggest that the GS interceptparameter has the potential to reduce the effects of specular scatterers ondiffuse scattering estimates and to improve QUS imaging.