A research project on the development of innovative methods for three-dimensional (3D) image analysis of blood vessels will represent the University of Heidelberg and the German Cancer Research Center (DKFZ) at an exhibition of the German Research Foundation (DFG). Under the title “From an Idea to Discovery” („Von der Idee zur Erkenntnis“), the DFG presents outstanding scientific projects, which the foundation funds, in the Deutsche Bundestag (German Federal Parliament) from 6-30 March 2012. The Heidelberg project “QuantVessel” is headed by associate professor Dr. Karl Rohr (CellNetworks member) and carried out at the BioQuant Center, the Institute of Pharmacy and Molecular Biotechnology (IPMB) of the University of Heidelberg, and the Division Theoretical Bioinformatics of the DKFZ.
As Dr. Rohr explains, a challenge in the accurate measurement – quantification – of blood vessels is to capture the complex and curved 3D anatomy from 2D image slices. „Existing image analysis systems for automated quantification generally do not fulfill clinical requirements regarding their accuracy, and reliable solutions are lacking“, the scientist emphasizes. Dr. Rohr and his research group „Biomedical Computer Vision“ work particularly on the 3D quantification of the aorta to treat pathologic dilations, i.e. aneurysms.
|K. Rohr, Project QuantVessel:
Segmentation of the aorta from a 3D CTA image
The scientists have developed a new 3D approach, which accurately determines the size and shape of blood vessels from 3D tomographic image data, for example, from computed tomography angiography (CTA). They combine so-called 3D intensity models with mathematically well-founded optimization methods. The models describe the variation of images intensity values of blood vessels and are fitted to the image data of a particular person during the computer-based analysis to determine size and shape parameters, in particular, the diameter along vessels. This approach enables quantification of blood vessels with arbitrary orientation, different curvature, and with branches.
As Dr. Rohr explains, is the developed approach more accurate than previous image analysis approaches, in particular, for determining the diameter of blood vessels. At the same time, additional geometric parameters like, for example, the length and curvature of the outer contour of the aortic arch can be determined. Moreover, the approach was extended for automatic analysis of time-resolved dynamic 4D images during the heart cycle. “The quantification results from image analysis are important to select optimal individual vessel prosthesis (stent grafts) for operation planning“, Dr. Rohr says. The research work is carried out in close cooperation with radiologists and vascular surgeons of the University Hospital Heidelberg and the German Cancer Research Center (DKFZ), in particular, with Dr. Hendrik von Tengg-Kobligk as well as with Prof. Dr. Hans-Ulrich Kauczor and Prof. Dr. Dittmar Böckler. A clinical study is planned.
The exhibition “From an Idea to Discovery” was opened on 6 March by the President of the Bundestag, Prof. Dr. Norbert Lammert, and the President of the DFG, Prof. Dr. Matthias Kleiner, and will be shown at the Paul-Löbe-Haus in Berlin until the end of March. The project “Quantification of the Morphology of Human Vessels from 3D Tomographic Image Data” (QuantVessel) will be presented with pictures, text and videos, a 3D print of an aorta, and the developed software program. The project is one of ten projects that can be seen at the DFG exhibition, selected from about 20,000 projects the DFG funds annually and the only one from the state of Baden-Württemberg.
The research group „Biomedical Computer Vision“ of Dr. Rohr develops computer science methods for automatic analysis of medical and biological images. In the project “QuantVessel“ participate particularly Dr. Stefan Wörz and Andreas Biesdorf as well as Simon Eck and Wei Liao.
PD Dr. Karl Rohr
University of Heidelberg
Institute of Pharmacy and Molecular Biotechnology (IPMB)
German Cancer Research Center (DKFZ)
Division Theoretical Bioinformatics
Biomedical Computer Vision Group
Im Neuenheimer Feld 267
Phone: +49 6221 54 51298
k.rohr [ aT ] uni-hd.de, k.rohr [ aT ] dkfz.de