Blood vessel detection in navigated ultrasound: An assistance system for liver resections

Open liver resection is a difficult and dangerous surgical operation, which requires a lot of experience and knowledge of the patient's anatomy. Ultrasound, as a predominantly used interoperative image modality, helps the surgeon guiding the resection process. Until present time there is no convincing solution to reduce the risk of accidental injure of liver vessels or to guide the targeting of structures like liver tumours. In this work a system is presented, which provides the surgeon with a guidance tool to position the resection instrument relatively to vessel and tumour structures in the ultrasound image. An important step for future navigation concepts in this domain is the development of a stable segmentation algorithm for liver vessels. In this work a fast segmentation algorithm is applied and tested with phantom and patient ultrasound images. The segmentation algorithm performs well and is integrated in a planning and surgery assistance system (MiMed Liver). Moreover, diameters of vessel structures are calculated and displayed. The distance of the navigated resection instrument (CUSA) to the ultrasound image plane is visualized in the ultrasound image.