SU-F-R-23: Texture Feature Analysis for Assessment of Liver Cirrhosis and Normal Liver in CT Image




In this study, texture analysis is proposed to assess liver cirrhosis and normal liver in CT image, and to automatically identify cirrhosis tissue and normal tissue for liver.


Texture feature parameters were extracted from 10 liver cirrhosis and 10 normal liver CT images. To individuate which parameter is available in classify the different tissue, a Wilcoxon signed-rank test between textural indices was performed.


We acquired 9 texture parameters (such as energy, inertia, entropy, correlation, etc.) for each patient. Only correlation showed a statistically significant between cirrhosis liver tissue and normal liver tissue (0.6408±0.0611 vs 0.4108±0.0733, p=0.00).


The results indicate that texture feature analysis for liver parenchyma maybe helpful in identifying cirrhosis tissue and normal tissue. However, a deeper validation of this new approach on a large population is required in the future.