Fraunhofer Institute for Digital Medicine MEVISPress Release
The procedure BRIDGE maps clincial breast images from mammography (MG), tomosynthesis (Tomo), magnetic resonance imaging (MRI) and ultrasound (US) into a breast model.
A software procedure that will accelerate and improve breast cancer diagnostics is ready for the market. Fabian Zöhrer, physicist at the Fraunhofer Institute for Medical Image Computing MEVIS, will receive a distinguished award. In Chicago on December 2nd at RSNA 2014, the world’s largest radiology conference, he will be bestowed the GHTC® – the German High Tech Champions Award. This will grant him the unique opportunity to introduce his software to the world’s leading medical technology enterprises and to subsequently begin its practical application. A team of three experts is responsible for the development of the software: Fabian Zöhrer is joined by computer scientist Joachim Georgii and MEVIS Institute Director Horst Hahn.
Today, there are various methods of breast cancer diagnostic – via mammography, ultrasound, or magnetic resonance imaging (MRI). Often, when only one method is used, no reliable diagnosis can be made, especially for women with very dense breast tissue. In such cases, a combination of diagnostic methods will give different, supplementary information.
Problematic here are the different positions of the woman in all three methods. In the MR scanner, she lies on her stomach, during the ultrasound she lies on her back, and during the mammography, she stands upright. These different positions can lead to drastic changes in tumor position or suspicious areas and complicate image comparison.
The software method introduced by Zöhrer and his team can correct this flaw. ‘Multimodal position correlation’ can automatically transfer the position of a tumor from one image dataset to another with the help of an elaborate algorithm that copies the entire breast as a three-dimensional object. This virtual model is subsequently divided into many small boxes. The computer simultaneously calculates how far the position of every box will differ if the tissue changes position.
Using this procedure, a doctor can select a certain critical area in the tissue on an X-ray image. The same monitor shows an ultrasound image of the patient. A small circle automatically without delay emerges in the latter image showing the same critical area identified by the MEVIS software on the X-ray. “Doctors no longer need to reconstruct the tissue mentally to estimate where it should be seen in another image. Our software does that for them”, says Zöhrer. “It simplifies the procedure, saves time, and sometimes helps prevent errors.”
The software can be integrated in the so-called PACS viewer. These are common, commercial image storage and display programs that enable images from different methods to be displayed on a single monitor. However, workstations customized for single imaging procedures, such as mammography or MRI, can also benefit from multimodal position correlation. The method also holds promise for other fields of application. In clinical trials, it could automatically identify particular tissue areas in the images of different participants. In future computer-aided diagnosis (CAD) software, the method could also support automatic image recognition.
The award gives Zöhrer’s team the opportunity to showcase their innovation to the world’s most important medical technology companies. The prize will be bestowed at the RSNA 2014 radiology conference in Chicago. During the award ceremony and business-networking afterwards, the MEVIS researcher will have a chance to present his project to selected industry representatives and discuss promising applications and business ideas.
GHTC® – the German High Tech Champions Award is a constituent part of the collaborative project “International Research Marketing” which is a joint initiative by the Alexander von Humboldt Foundation, the German Academic Exchange Service, the Deutsche Forschungsgemeinschaft and the Fraunhofer-Gesellschaft.
All the activities within the project are part of the “Promote Innovation and Research in Germany” initiative under its brand “Research in Germany.” The initiative is funded by the German Federal Ministry of Education and Research. More information: www.research-in-germany.de
Material
Related Topics
- Position Correlation between Mulitmodal Breast Images (mevis.fraunhofer.de)