PhD Seminar - Information fusion of mammography and MRI scans for improving diagnostic management of breast cancerExport this event to calendar

Friday, October 2, 2020 — 12:00 PM EDT

Candidate: Hossein Soleimani

Title: Information fusion of mammography and MRI scans for improving diagnostic management of breast cancer

Date: October 2, 2020

Time: 12:00 PM


Supervisor(s): Michailovich, Oleg



Medical imaging is critical to non-invasive diagnosis and treatment of many medical conditions. However, different modalities of medical imaging exploit different contrast mechanisms and, consequently, they provide different depictions of bodily anatomy. As a result, situations are frequent in which the same pathology can be detected by one type of medical imaging, while being missed by others. This problem brings forward the importance of the development of image processing tools for integrating the information provided by different imaging modalities via the process of information fusion. One particularly important example of clinical application of such tools is in diagnostic management of breast cancer, which is a prevailing cause of cancer-related mortality in women. Currently, the diagnosis of breast cancer relies mainly on X-ray mammography and Magnetic Resonance Imaging (MRI), which are both important throughout different stages of detection, localization, and treatment of the disease. The sensitivity of mammography, however, is known to be limited in the case of breast tissues, while contrast-enhanced MRI tends to yield frequent “false alarms” due to its high sensitivity. This situation makes it critical to find reliable ways of fusing the mammography and MRI scans so as to improve the sensitivity of the former, while boosting the specificity of the latter.


Unfortunately, fusing the above types of medical images is known to be a difficult computational problem. Indeed, while MRI scans are usually volumetric (i.e., 3-D), digital mammograms are always planar (2-D). Moreover, the mammograms are invariably acquired under the force of compression paddles, thus making the breast anatomy undergo sizeable deformations. In the case of MRI, on the other hand, the breast is rarely constrained and imaged in a pendulous position. Finally, X-ray mammography and MRI exploit two completely different physical mechanisms, which produce distinct diagnostic contrasts which are related in a non-trivial way. On such conditions, the success of information fusion depends on one’s ability to establish spatial correspondences between mammograms and their related MRI volumes in a cross-modal cross-dimensional (CMCD) setting in the presence of spatial deformations (+SD).


It goes without saying, solving the problem of information fusion in the CMCD+SD setting is a very challenging analytical/computational problem, efficient solutions to which are still to be found. In the literature, there is a lack of a generic and consistent solution to fuse mammograms and breast MRIs and use their complementary information. Most of the existing MRI to mammogram registrations techniques are based on biomechanical approach which builds a specific model for each patient to simulate the breast deformation. We speculate that the substantial part of breast deformation is common among most of the patients, and as a result, it can be computed without considering the breast geometry and properties. Accordingly, our first goal is to investigate and understand the computation of this global deformation. The remaining part of the deformation (residual deformation) is patient-specific, hence, it requires to take to account the breast's properties. We will compute the residual deformation by a nonrigid and intensity-based registration algorithm in a multi-grid manner. Our goal is to come up with a reliable mathematical model which is generalized over all cases. Considering our assumption above, we did a few experiments on clinical MRI and X-ray mammography cases acquired at the Princess Margaret Cancer Center (Toronto, Canada).



  1. 2022 (12)
    1. February (1)
    2. January (11)
  2. 2021 (131)
    1. December (11)
    2. November (18)
    3. October (10)
    4. September (4)
    5. August (12)
    6. July (4)
    7. June (5)
    8. May (10)
    9. April (25)
    10. March (14)
    11. February (10)
    12. January (9)
  3. 2020 (246)
  4. 2019 (282)
  5. 2018 (150)
  6. 2017 (212)
  7. 2016 (242)
  8. 2015 (242)
  9. 2014 (268)
  10. 2013 (190)
  11. 2012 (31)