T-CAIREM recently appointed Dr. Benjamin Haibe-Kains as an Infrastructure co-lead. In this role, he'll oversee the development of our Health Data Nexus (HDN) platform. The HDN makes it easy for accredited researchers and clinicians to access real-world health data to develop algorithms and encourage innovative data analyses.
Dr. Haibe-Kains is a Senior Scientist at the Princess Margaret Cancer Centre (PM), University Health Network, and a Professor in the Medical Biophysics department of the University of Toronto. His research team analyzes large-scale radiological and (pharmaco)genomic datasets to develop new prognostic and predictive models to improve cancer care. We recently caught up with him to learn more about his experience.
What inspired you to take on your new role as the Infrastructure Co-Lead with T-CAIREM?
I have seen so many researchers working very hard at generating data and spending lots of time and resources to organize these data for analysis. But after the main study is published, these data are often shelved or archived, making it extremely challenging to leverage for further studies. So much investment for so little return! If a dataset is properly organized, annotated and shared with the scientific community, its value will be dramatically increased and will allow new avenues of research to be explored more rapidly and efficiently.
What initially sparked your research interests in Bioinformatics and Computational Genomics?
Trained as a computer scientist, I initially wanted to work in robotics. But my thesis supervisor advised me to work with a clinical researcher as they began generating large genomic data for breast cancer. This experience completely changed my perspective: I loved doing research that could ultimately help patients and was super impressed by the dedication and dynamism of the multidisciplinary team I was working with. At that point, I knew I wanted to apply my computer science skills to cancer research.
What excites you most about the possibilities of AI in healthcare?
From a patient perspective, the introduction of AI tools to guide us through the complex journey of health care is an exciting one: optimizing schedule, reporting results, communicating with the healthcare team, managing adverse side effects are some of the many benefits. From a clinician’s perspective, the automation of data collection and analysis will open new avenues for more in-depth relations with the patients and focus on more complex cases. From a researcher’s perspective, AI will streamline the acquisition and integration of multiple data modalities to characterize and study disease in a more comprehensive way.
What are some AI in medicine projects you’re currently working on that you’re really excited about?
There are three projects. The first one is the automation of tumour and organs-at-risk segmentation from radiological images to optimize treatment planning. The second project focuses on predicting tumours’s genomic “signatures” from radiological data to provide a non-invasive technology for precision oncology. The third project aims to develop a platform for systematic matching of cancer patients to clinical trials based on clinical and genomic data. The fairness assessment of these AI models is also something that we pay close attention to.
What’s the number one piece of advice you’d give to students interested in AI in health?
AI technologies are moving at a fast pace. But they often rely on large, complete datasets that are rarely available in healthcare. Understanding the limitation of health data and developing and using the right AI approach for solving these healthcare challenges is key. So it’s important for students to learn both the health and AI sides to develop the best modelling strategy.
