Markerless navigated surgery

Surgical resection is one of the primary types of treatment for tumors in the abdominal cavity. It is the objective to remove the tumor including a safety margin with maximum protection of the surrounding tissue. Computer-assisted planning systems can use preoperative image data to calculate optimal resection recommendations while considering the individual anatomy of the patient. However, the reliable implementation of the plan is problematic since this requires a substantial ability to conceptualize in three dimensions. This is especially true of laparoscopic surgery which - although especially protective of the patient - places extremely high demands on the surgeon who cannot palpate the tumor and must deal with the absence of depth information.

During the past decade, computer-assisted information and navigation systems for medical interventions have experienced a rapid spread in routine clinical use. In the fields of neurosurgery, orthopedics, dental surgery as well as maxillofacial surgery, navigation systems today support a majority of all operations with important information about patient anatomy. The systems used usually work with a preoperatively created patient model which at the beginning of surgery is transferred to the presently available patient anatomy (registration). By means of optical or magnetic systems installed in the operating room to localize the surgical instruments, these can be visualized in relationship to critical anatomical structures. It is required that the pose of the organs in the target region remains the same. However, since organs composed of soft tissues are constantly subject to movements and deformations - caused by respiration and the patient's heartbeat as well as surgical manipulations - established navigational techniques are not suited for use in the abdominal cavity.
This project should develop concepts for computer-assisted open and laparoscopic tumor resection which facilitate a markerless, accurate transfer of preinterventional operation planning to the surgery site and thus significantly improve patient prognosis.
The workflow proceeds as follows:
1. Before surgery, a resection recommendation is calculated from the preoperative planning image while considering the individual anatomy of the patient.
2. During surgery, the present location of the target organ is continuously monitored via an intra-operative imaging modality. For example, this can be a Time-of-Flight (ToF) camera [Seitel10] which determines distances by means of intensity modulated light. During laparoscopy, the surface can be detected via classical 3D reconstruction techniques such as stereoscopy or structure from movement [Groch10].
3. To make the operation planning available to the surgeon during surgery, it is necessary to perform an intraoperative registration of the planning image with the patient. To this purpose, the organ surface acquired during surgery is first matched onto the surface extracted from the preoperative planning image with a graph-based approach [dosSantos10, Maier-Hein10]. To compensate for intraoperatively occurring deformations, a model of the liver created from the planning data is iteratively adapted to the surface data by means of the Finite Element Method (FEM). If the imaging modality and the instruments are tracked, they now also can be visualized relative to the preoperative planning data.

Selected publications

• [P32-10] Dos Santos T, Seitel A, Meinzer HP, Maier-Hein L. Correspondences Search for Surface-Based Intra-Operative Registration. Proceedings of the 13th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2010 (2), volume 6362, pages 660-667, Beijing, China, September 2010. Springer Heidelberg
• [P30-10] Maier-Hein LL, Schmidt M, Franz AM, Dos Santos TR, Seitel A, Jähne B, Fitzpatrick JM, Meinzer HP. Accounting for Anisotropic Noise in Fine Registration of Time-of-Flight Range Data with High-Resolution Surface Data. Proceedings of the 13th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2010 (1), volume 6361, pages 251-258, Beijing, China, September 2010. Springer Heidelberg
• [P10-10] dos Santos TR, Gergel I, Meinzer HP, Maier-Hein L. Fast correspondences search in anatomical trees. Proc. SPIE Vol. 7623, Medical Imaging 2010: Image Processing, 762332; DOI: 10.1117/12.844192.
• [P05-10] Groch A, Baumhauer M, Meinzer HP, Maier-Hein L. Automatic feature detection for 3D surface reconstruction from HDTV endoscopic videos, Proc. SPIE, 7625, 76251P, 2010.
• [Maier-Hein10] Maier-Hein L, Schmidt M, Franz AM, dos Santos TR, Seitel A, Jähne B, Fitzpatrick JM, Meinzer HP. Accounting for anisotropic noise in fine registration of Time-of-Flight range data with high-resolution surface data, 13th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2010, to appear
• [P06-10] Seitel A, dos Santos TR, Mersmann S, Penne J, Tetzlaff R, Delorme S, Meinzer HP, Maier-Hein L. Time-of-Flight Kameras für die intraoperative Oberflächenerfassung. In Deserno TM, Handels H, Meinzer HP, Tolxdorff T (eds). Bildverarbeitung für die Medizin (2010). Heidelberg: Springer (2010) 11-15. Download PDF file

Project team