Multimodal imaging is of crucial importance for the early detection of breast cancer, as the imaging methods differ in their diagnostic features. The challenge in combining the modalities lies in the different position of the patient during imaging and especially in the deformation of the breast. To overcome this challenge, we are developing an image registration method that automatically establishes the spatial correspondence between modalities. Especially in the case of dense breasts, the combination of information from mammography and MRI is essential for a more conclusive diagnosis. However, the registration of MRI and mammography faces the problem of complex non-linear breast deformation in 3D, which is only shown in a 2D projection image.

Digital breast tomosynthesis (DBT) is an emerging technology that provides 3D information about the anatomy of the compressed breast. In this HEIKA-funded project, this new modality was used to gain a deeper knowledge of the complex breast deformation and to investigate the accuracy of registration methods in 3D.

Technology: Medical Image Processing