Geometry model extraction for perfusion modeling in liver surgery

Background: Planning of liver surgery especially in changed liver parenchyma (cirhosis, steatofibrosis) needs to know liver perfusion status. This can be described by liver perfusion modeling, which basic elements are models of organ geometry and vascular system. However, nowadays Methods does`t produce the model suitable for flow modeling. Due to the imperfections of the scanning Methods, the model’s vascular wall is not smooth enough to faithfully imitate the real blood vessels. Based on computed tomography images, it is necessary to create a model that is suitable for flow modeling. Methods: The model of parenchyma was created by delineation of CT data using 3D Graph-Cut method and user interaction. Liver density is modeled by the Gaussian Mixture as well as the density model of other tissues. Model of the vascular system geometry was obtained by the Connected Threshold method with pre-processing using Gaussian Filter and post-processing based on Binary Morphological Operations. Branch analysis was performed on skeletonized image to determine parameters of vascular segments: length, diameter and orientation. Results: For basic parameters of the model quality we considered the uniformity of hepatic parenchyma coverage by the vascular system. Uniformity was measured as the ratio of the maximum distance of any voxel and the average distance of all voxels. A new geometry model for blood flow modeling has been established. Conclusions: Extraction of the liver blood vessel tree geometry is the starting element for perfusion modeling, that are absolutely crucial for lanning of extended liver resections in diffuse liver parenchyma diseases. The proposed modeling algorithm meets the basic requirements for uniformity of stromal tree distribution in the parenchyma. The Results will be used in preparation of liver surgery analysis and planning. Acknowledgment: This study was supported by the Centrum of Clinical and Experimental liver surgery (UNCE/MED/006) and National Sustainability Program I Nr.LO1503.