Neighborhood-pixels algorithm combined with Sono-CT in the diagnosis of cirrhosis: an experimental study

The goal was to investigate the role of neighborhood-pixels algorithm (NPA) in analyzing the echogram of experimental cirrhosis and the value of high frequency real-time compound imaging (Sono-CT) in improving texture analysis. A cirrhosis model was established by subcutaneously injecting CCl(4) in 80 rats. The total group of rats were divided into a control group and four treatment groups (treated for 6, 8, 10 and 12, weeks respectively). The texture of hepatic-echograms was analyzed using a "neighborhood-pixels" algorithm. Images were obtained under conventional imaging mode and Sono-CT, respectively. The second texture parameter (FP(2)) was estimated and compared in different groups and under different modes. FP(2) increased gradually with the time of treatment and group differences were significant (p < 0.01). In these groups, FP(2) was higher under Sono-CT than under conventional condition (p < 0.01) and group differences in FP(2) under both conditions were significant (p < 0.01). Thus, FP(2) measured by neighborhood-pixels algorithm can reflect the dynamic change of the texture of echogram of cirrhosis in rats and Sono-CT can improve texture analysis by neighborhood-pixels algorithm, thus facilitating the early diagnosis of cirrhosis.