Segmenting bladder wall and estimating wall thickness for magnetic resonance virtual cystoscopy

A level set method for bladder wall segmentation and wall thickness estimates was developed to extract the features of bladder abnormalities in a virtual cystoscopy system, using T1-weighted magnetic resonance (MR) images. The local intensity contrast information is used to construct the image energy with two level set functions to segment the inner and outer borders of the bladder wall. A path integration distance is used to estimate the bladder wall thickness, by mimicing the distribution of the electric field line between two iso-potential surfaces. The bladder wall thickness is then mapped to a pseudo-color space for rendering the reconstructed 3-D bladder model. Ten clinical cases including volunteers and patients were used to test the system. The results show that the system is robust and effective for automatically segmenting the bladder wall, estimating the wall thickness, capturing the abnormal variations of the wall thickness in the patient dataset, and further showing the abnormal variations on the 3-D model to offer reliable information for virtual cystoscopy aided diagnosis.