gms | German Medical Science

60. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC)
Joint Meeting mit den Benelux-Ländern und Bulgarien

Deutsche Gesellschaft für Neurochirurgie (DGNC) e. V.

24. - 27.05.2009, Münster

lmprovement of intraoperative 3D-contrast enhanced ultrasound angiography (3D-iUSA) based on image processing tools

Meeting Abstract

  • D. Lindner - Neurochirurgische Klinik, Universitätsklinikum Leipzig
  • C. Chalopin - ICCAS Innovation Center Computer Assisted Surgery Leipzig
  • C. Renner - Neurochirurgische Klinik, Universitätsklinikum Leipzig
  • D. Fritzsch - Klinik für Diagnostische Radiologie, Abteilung für Neuroradiologie, Universität Leipzig
  • C. Trantakis - Neurochirurgische Klinik, Universitätsklinikum Leipzig
  • J. Meixensberger - Neurochirurgische Klinik, Universitätsklinikum Leipzig

Deutsche Gesellschaft für Neurochirurgie. 60. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC), Joint Meeting mit den Benelux-Ländern und Bulgarien. Münster, 24.-27.05.2009. Düsseldorf: German Medical Science GMS Publishing House; 2009. DocMO.06-08

doi: 10.3205/09dgnc034, urn:nbn:de:0183-09dgnc0349

Veröffentlicht: 20. Mai 2009

© 2009 Lindner et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen ( Er darf vervielfältigt, verbreitet und öffentlich zugänglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.



Objective: Ultrasound angiography (iUSA) has a high intraoperative potential. 3D reconstruction of intracranial blood flow offers a new perceptivity for visualisation of vascular surgery. However, ultrasound is known to be noisy and the use of contrast agent is responsible for artefacts. For this presentation, we focus on the assessment of the extraction of cerebral arteries in 3D-iUSA.

Methods: For 3D navigation, the Brain Navigator™ (LOCALITE) consisting of a PC which includes a video grabber card, a tracking system (NDI Polaris) and the Sonoline Elegra device (Siemens) with a 2.5-MHZ phased-array transducer were used. Ultrasound angiography (iUSA) was acquired with a power of 1–3% and a frequency of 1.1 MHZ. During vascular surgery ultrasound datasets were obtained after craniotomy and after clip placement. SonoVue® (Bracco) contrast agent was used as an intravenous bolus of 2.4 ml.

Vessel segmentation process:

Input data are first pre-processed in order to enhance vascular structures. A 3D vesselness map, which represents the probability for a Voxel of the original 3D-iUSA data to belong to a vascular structure, is computed based on the Hessian matrix and its Eigen values.
A region growing process aggregates to a given seed point all the connected Voxel whose vesselness value is within a range. Seed points are specified manually while the range is automatically computed from the 3D vesselness map values. An output volume is finally provided which includes the segmented objects.

Results: The segmentation tool was first tested on a vascular phantom. Then the image processing package was tested offline on 3D-iUSA data of four patients. One example is described in Figure 1 [Fig. 1]. The original 3D-iUSA data includes the anterior part of the circle of Willis (Figure 1a [Fig. 1]). The computed 3D vesselness map (Figure 1b [Fig. 1]) shows in high intensity the enhanced vessel structures. Thus a user indicates the blood vessels of interest in the original 3D data by a couple of seed points. In Figure 1c [Fig. 1], the extracted brain arteries (white intensity) overlay the original 3D-iUSA data. Finally, a volumetric view is represented in Figure 1d [Fig. 1]. We also obtained correct results in all patients requiring a few seed points. Moreover, the computing time was less than one minute

Conclusions: The improvement of 3D-iUS by image processing tools offers an image quality comparable to x-ray-angiography with high anatomic resolution. Now the tool is suitable for the operating room.