Fully Automated three-dimensional Segmentation of Liver

The shape and composition of the liver is a vital component in effective diagnosis, surgery planning and therapy control. Manual measurement of three dimensional structures is extremely time consuming, cost intensive and subjective, which means that automated methods are required in today’s challenging clinical environment.

However, due to the large variability in appearance, shape, and potential damages (such as alcoholic cirrhosis and liver tumor), reliable automated liver segmentation is an extremely challenging task. Further difficulties arise because of insufficient separation from adjacent organs such as kidney, heart, and muscles, and due to time and liver state dependent effects of the contrast agent.

In a research study an application was developed - based on Definiens Cognition Network Technology® - to conduct a fully-automated liver segmentation by means of a data pool of CT scans covering a large variation of pathological cases. Manual annotations of two medical experts served as reference during the development and quantitative evaluation.

For further details, please see the poster Definiens Cognition Network Technology® for fully automated 3D Segmentation of Liver.