Fatty liver evaluation with double-Nakagami model under low-resolution conditions

In previous studies, the double-Nakagami (DN) model has been proposed for fatty liver assessment and applied to in vivo rat livers and clinical data sets. The healthy liver structure filter (HLSF) method, which extracts non-healthy areas using two DN parameters, has also been proposed. In this paper, we first verify the accuracy of the DN model and the HLSF method for acoustic fields at 15 and 5 MHz, which were reproduced using numerical simulation. We then apply the method to clinical data sets of livers observed using a frequency of 3 MHz and investigate the method's clinical usefulness. A positive correlation (r=0.28) was found between the ratio of the non-healthy area and fat mass. Although the results were inferior to the results produced using 15 MHz ultrasound (r=0.96), we found that it was possible to detect the difference between a normal liver and a fatty liver even at a lower frequency.