An algorithm for the surgical planning of hepatic resections

Abstract

The aim of this work is to determine the optimal location for the hepatic resection. Initially, a hepatic segmentation method based in region growing was implemented from CT images with and without contrast. This algorithm was evaluated comparing its contours with the manual contours drawn by two physicians. The identification of the hepatic vessels is a fundamental step for surgical planning. The developed method is based in two segmentation techniques: region growing and threshold. By successive application of region growths, was computed the distribution of the number of voxels in terms of intensity. From this distribution, the information about the optimal threshold value for vessel segmentation was obtained. Finally, the eight liver segments were located, according Couinaud's classification, by a method that includes knowledge of the anatomy and differential geometry. The results show similarities between the volumes computed by the algorithm and the volumes used as reference.