Dr. Dafna Sussman leads the Maternal Fetal Imaging Laboratory at St. Michael’s Hospital’s Institute for Biomedical Engineering, Science and Technology (iBEST). Her team develops MRI sequences and artificial intelligence algorithms to improve diagnosis and quality of life from birth.
What initially inspired you to research biomedical engineering & biomedical physics?
I was always interested in physics and math, which drew me to a Bachelor's in Engineering Science with a physics specialty. After working as an optical engineer and designing telescope components, I felt that I needed to do something that had a more direct, tangible, and imminent impact on human lives. I then decided to explore the biological applications of physics as part of my Master's degree. That is when I was introduced to the field of developmental physiology. At that point, I knew that I found my passion: using physics and engineering to understand human development and physiological processes.
How would you describe your work at the Maternal-Fetal Imaging Lab?
Our work at the MFI lab is highly multidisciplinary and merges biomedical physics and engineering with obstetrics, gynecology, fetal medicine, and public health. Projects range from developing novel MRI sequences for imaging fetal metabolism to developing imaging phantoms, conducting clinical studies on fetal development, and ending in automatic image analysis (MRI, Ultrasound, Pathology slides) and disease prognostication.
What's the biggest challenge of your work?
I would say that the biggest challenge of my work is gaining access to large enough clinical datasets to be able to create robust, accurate, and generalizable AI algorithms. For several recent studies, we ended up aggregating medical data from a variety of national and international sources. This was time-consuming not only due to the multiple legal agreements and ethics boards approvals but also because we had to consolidate medical parameters which sometimes had the same names but were used differently depending on the site or country. Having a large network of wonderful collaborators, including clinical practitioners, was certainly key to being able to complete this task both quickly and accurately.
How did you become interested in AI?
I became interested in AI when I first established the MFI lab. We were working on medical image processing algorithms and were looking for a robust approach for automating our analysis and creating disease prognostication tools that were also clinically usable and user-friendly. That's when I came across various AI tools which sparked my interest.
What’s the number one piece of advice you’d give to students following in your footsteps?
One piece of advice I'd give students is to always keep an open mind and not be scared to explore. AI in medicine is rapidly evolving and our work and curiosity help the field evolve. If nobody else has done what you would like to do, that is just the more reason you should do it yourself!
Are there any projects or initiatives you’re working on right now that you're really excited about?
Honestly, I am excited about all of our projects! To mention a few, we are currently developing a COVID-19 disease prognostication tool for pregnant individuals (the PROTECT study), a placental disease classification and clinical outcome prediction algorithm, and a gynecological cancer diagnostic algorithm.
What excites you the most about the possibilities of AI in healthcare?
What excites me most about AI in healthcare is the potential for forecasting disease progression or even predicting disease. These, in turn, could be used in treatment planning or to offer preventative treatments, thereby improving the quality of life of patients as well as potential patients.
