Ultrasonographic tissue signature for shistosomal liver and other related liver pathologies

Ultrasound imaging is a non-invasive, sensitive method in the diagnosis of hepatic shistosomiasis, yet it is a subjective one. It is difficult to differentiate properly diffuse liver diseases during the early changes from normal pathology by visual inspection from the ultrasound images. In this study, 180 cases of bilharzial, cirrhotic, normal, and mixed liver pathologies are analyzed for clinical investigations, ultrasound measurements, pathological, and biochemical measurements. The B-mode images are captured and analyzed for first and second order textural parameters, speckle parameters, attenuation parameter. The statistics done for various groups of pathologies indicated a significant differences between different groups. An unsupervised clustering algorithm is applied to the cirrhotic group to find subclasses of cirrhosis. The results of clustering showed a good match with that of pathological findings. This work can assist the sonographer to quantitatively and efficiently differentiate between diffuse liver diseases.