Solutions for Multimodal Breast Imaging

Clinical Challenges


Modern breast cancer diagnosis comprises a variety of different and complementary imaging techniques. Whereas additional modalities and combinations can improve the detection and diagnosis of breast cancer, they also require clinicians to handle more data and correlate and combine multimodal information. To reduce complexity and help clinicians focus on decision making, automatic solutions for multimodal position correlation, image normalization and workflow support are needed.
Fraunhofer MEVIS develops software solutions for multi-modal breast diagnosis supporting a clinician‘s work and increasing productivity and precision from screening to therapy. A high emphasis is placed on seamless integration into modern clinical workstations and workflows.

 

Solutions & Features

 

  • Contrast homogenization of digital Mammograms
    During screening, patients may undergo examinations with mammography systems from different vendors. This can yield large variations in the multiscale contrast of follow-up images. Our automatic wavelet-based contrast homogenization facilitates image comparison tasks by adjusting the contrast differences on multiscale levels while preserving all relevant image information [1]
  • Multimodal position correlation
    We support the diagnostic workflow by automatically synchronizing cursor positions between different views and modalities of the breast in real time. Our solution is based on a nonlinear compression simulation and is designed to integrate into existing workstations and workflows [2].
  • Breast MRI landmark segmentation
    Often, computer-aided breast cancer diagnosis requires precise segmentation of the breast as a fundamental step to facilitate further diagnostic tasks. We have developed a fully automated segmentation method which precisely identifies the breast-air boundary and the pectoral muscle boundary and subsequently yields the breast segmentation in MR images [3].
  • DCE-MRI motion correction
    Patient motion during breast DCE-MRI acquisition may hamper the detection and analysis of findings in the breast. Automatic motion correction improves image reading and reduces artifacts which could mimic contrast agent uptake.
  • Research from screening to therapy
    We cooperate in large european projects focusing on improving today‘s clinical workflows by developing methods to personalize breast cancer screening and integrate breast radiology with pathology and surgery [4].

 

Highlights

 

  • Solutions integrated into clinical study software tools
  • Fast and robust multimodal segmentations
  • Multi-modal real-time spatial correlation and registration
  • Solutions integrated into multiple vendor workstations
  • Coordinator of €6.8M EU projects VPH-PRISM and HAMAM
  • Pioneer in establishing digital mammography in European breast cancer screening
  • Research ranging from screening to therapy
  • Finite-element breast deformation simulation

[1] Zoehrer et al., IWDM 2010. [2] Georgii et al., SPIE 2013. [3] Wang et al., ISBI 2012. [4] EU Projects ASSURE: www.assure-project.eu & VPH-PRISM: www.vphprism.eu/home/