Towards Medical Imaging without images; Advanced Image Reconstruction and Machine learning in PET and Microwave Imaging
2021 NEXUS ONLINE DATA ENLIGHTENING SERIES, (NODES) SEMINAR SERIES
Thursday, March 25
Online Webinar & Q&A
Livestream Link – https://www.youtube.com/watch?v=mScZTePZTPg
Q&A – Sli.do event code: V214
Dr. Stephen Pistorius
Tyson Reimer, Gabrielle Fontaine, Jordan Krenkevich, Stephen Pistorius
Department of Physics and Astronomy, University of Manitoba
Cancer mortality is higher in remote regions in Canada and in developing countries where access to early detection is limited. Mammography, the standard for breast cancer screening, uses ionizing radiation, requires breast compression, has a high false-positive rate, and requires a well-established human and capital infrastructure. Positron Emission Tomography (PET), an essential functional imaging modality, also uses ionizing radiation, which scatters when it interacts with tissue, contributing to patient dose, but traditionally not providing any value from an imaging perspective.
Breast Microwave Imaging (BMI) and Scatter Enhanced PET offer the potential of improved sensitivity and specificity when compared to traditional mammographic x-ray and PET reconstruction techniques and, in the case of BMI, may provide other benefits such as lower cost and portability. Image reconstruction typically uses algebraic methods, Iterative Delay and Sum (for microwave imaging), or Maximum Likelihood Expectation maximization (MLEM) approaches (for PET and BMI), but there are technical and computational constraints that limit their applicability. This presentation is a tour through time and place; starting with simulations and experiments that show the benefits of improved reconstruction techniques for microwave radar imaging for breast cancer detection and scatter imaging for PET and finishing with a demonstration as to how machine learning can be used to improve our results, reconstruct images and to detect breast lesions even when they may not be visible.
Keywords: Microwave radar, Positron Emission Tomography, Breast imaging, Iterative Reconstruction, Scatter Imaging, Machine learning, Deep learning, Convolutional Neural Networks.
Stephen Pistorius is a Professor and Associate Head: Medical Physics in Physics and Astronomy, Professor in Radiology, and Anatomy and Cell Science at the University of Manitoba. He serves as the Vice Director and Graduate Chair of the Biomedical Engineering Program and is a Senior Scientist at the Research Institute in Oncology and Hematology. He holds a B.Sc. (Physics & Geography) from the University of Natal-Kwa Zulu and a Hons. B.Sc. (Radiation Physics), M.Sc. (Medical Science), and Ph.D. (Physics) from the University of Stellenbosch, South Africa, and a Post-Graduate Diploma in Business Management from the Edinburgh Business School, UK. His research interests focus on image processing and reconstruction, medical device development, improving, optimizing, and quantifying various diagnostic and therapeutic techniques and in modelling and understanding radiation transport in clinically useful imaging and treatment modalities, particularly as they apply to provide more equitable health care to persons in disadvantaged communities. He is supported by an excellent and diverse group of graduate and undergraduate students who make this work possible.