Multi-Port bone surgery using the example of the temporal bone

DFG-Forschergruppe FOR 1585

The DFG research project “Multi-Port bone surgery using the example of the temporal bone” (MUKNO) is a cooperation between the ORL-clinic in Düsseldorf, the “WZL” of the RWTH Aachen and the “Interactive Graphics System Group” of the TU Darmstadt. The aim of the project is to investigate how minimally traumatic surgical procedures at the otobasis can be accomplished. The main challenge here is not to damage any important structures such as the facial nerve or the carotid artery. So far, the surgeon dissects all collision structures in the surgical field which means a high risk for the patient. In future, only three drilling canals should be used. Through these canals the instruments are inserted and the computer supports the surgeon during the navigation. In order to keep the complication rate under 0.5 %, the process chain has to be controllable. This means in every step of the chain – from the image acquisition over the segmentation, registration and planning to the drilling – the possible error should be kept under a feasible threshold.

Particular projects of MUKNO

  • Image acquisition
  • Segmentation
  • Registration
  • Planning

Image acquisition

In order to keep the complication rate under 0.5%, the generated three dimensional images should have an isotropic resolution of 0.1-0.2 mm. To achieve this aim two approaches are studied:
- Image acquisition using Cone Beam CT which has high special resolution with isotropic voxel size.
- Providing of two orthogonal sets of image volumes by means of conventional CT. These highly anistropic data sets can be then processed using a super-resolution reconstruction algorithm to reconstruct an isotropic volume data while increasing the image quality.

(Wissam El Hakimi)


In order to be able to further process the acquired images, the structures have to be segmented. Often the important structures like facial nerve and carotid artery are difficult to distinguish in the images. Therfore, we integrate a-priori knowledge in the segmentation. Here, it is also important to quantify the accuracy of the segmentation.

Picture: Extract of the temporal bone, where the facial nerve has been segmented manually.

(Meike Becker)


The position of the instrument relative to the patient´s anatomy should be controlled during the operation. For this purpose appropriate verification images are to be acquired. After registration of the intra- and proxy-operative image data the actual instrument position can be determinated and in case of error corrected. The main challenge here is to develop a “real time” capable registration method.

(Wissam El Hakimi)


We delevoped a tool which analyses – based on pre-operatively acquired CT data – if it is all possible to reach the target of the intervention using linear drill canals without damaging any important anatomical structures. Here, we first compute all feasible paths, i. e. all paths which do not damage any critical structure. From this set, the surgeon can then manually choose a combination of three paths.

Path combination chosen by a clinician for the surgical target of the petrous apex.

We use the SOFA-Framework (SOFA) as a basis for our simulation software.

Contact: Meike Becker


  • Ein Protoyp zur Planung von Bohrpfaden für die minimal-invasive Chirugie an der Otobasis.
    Ralf Gutbell, Meike Becker and Stefan Wesarg In: Bildverarbeitung für die Medizin 2012, Berlin, Germany, 2012. Ralf Gutbell, Meike Becker and Stefan Wesarg
    In: Bildverarbeitung für die Medizin 2012, Berlin, Germany, 2012.
  • Confidence map based super-resolution reconstruction.
    Wissam El Hakimi and Stefan Wesarg
    In: SPIE Mediacal Imaging 2012: Image Processing, San Diego, California, 2012