Investigation of A Method for Quantifiying Diffuse Liver Disease Based on Histogra of Ultrasound Signal-to-Noise Ratio

A method for quantifying liver fibrosis and steatosis generation separately to evaluate diffuse liver disease by the ultrasound scanner was investigated. We proposed 4-kernel method to calculate local SNR in the liver for the 2D parametric image of the SNR. We also proposed Area ratio in the SNR histogram to extract information of small structures. RF signals for 273 patients in chronic liver disease were analyzed and compared with amount of fibrosis and steatosis taken by histology. The result shows that the mean value of SNR distinguished generation of either steatosis or fibrosis from the normal, but not distinct when both was large. Area ratio, on the other hand, has significant difference when both of steatosis and fibrosis was large. As a result, the combination of the mean and area ratio of SNR (X-Y plot) can quantify liver steatosis and fibrosis separately.