Characterization of Texture Features Between Carcinomatous and Normal Wall Tissue in MR Bladder Imaging

Objective: To distinguish bladder wall tissue from bladder cancer tissue using the texture analysis in MR images. Method: with collected MRI scans for bladder imaging, the ROI (region of interest) area was first selected manually. Second, the texture features of bladder cancer and bladder wall tissue, such as entropy, uniformity et al, were calculated within the ROI. Third, statistical analysis was applied to these features to reflect their significance. Results: All the texture features were analyzed by t-test. Entropy, uniformity, smoothness, relation and contrast show significant differences between the two groups, while the differences for others are not significant. Conclusion: The preliminary results demonstrate significant differences in some texture features of bladder cancer compared to normal wall tissue. Additionally, more properties of bladder cancer might be obtained directly from patient´s MRI scanning.