Hello BRIGHT Run Family,
It seems to me that 2025 marched into March at a pretty fast rate.
We saw some severe snowstorms in Ontario. Right after one of those ended in Hamilton (in February), I found myself under the bright sunlight in San Diego, CA, where I attended a conference.
The conference is named SPIE (an international society for optics and photonics) Medical Imaging 2025.
I presented my work on the sensitivity of AI models’ performance towards the type of the cancer when we analyze segment tumours in PET/CT scan images of patients with advanced cancer for different types of cancer.
PET/CT is used in different types of cancer for determining the cancer spread. Therefore, sensitivity of the AI models towards the type of cancer suggests that caution is necessary.
The work is done by my student Mahan Pouromidi (graduated last year from McMaster) in collaboration with my colleagues Dr. Katherine Zukotynski and Dr. Thomas Doyle. Though the dataset we used didn’t include breast cancer, our proof-of-concept study basically shows that we can perform it for breast cancer and other types of cancers. However, it will take time and resources to gather relevant annotated (all tumours outlined in the image by experts) imaging data from our Juravinski Cancer Centre.
And yes, one of the AI models we studied used transformers (mentioned in my previous letter to you). At the conference, there were several studies on AI-based mammogram analysis, including determining the presence of cancer in it, assessment of breast tissue density, aiding less experienced breast radiologists to improve analysis. Interestingly, we are also working on some of these topics, as well as on breast ultrasound. Those research stories will unfold in some of my upcoming newsletters.
As I am sharing research information, allow me to also share a recent breakthrough in AI, called DeepSeek-R1. This model enables you to chat (like ChatGPT), but it was reported that the model can be trained at much lower cost, which is a great leap for the AI community.
Getting back to March again. The forward march of AI continues. Though there have been hiccups and issues, I was elated to see immense interest and progress in AI medical imaging research, to go beyond the existing barriers.
There is a famous quote that says “Life is what happens to you when you are busy making other plans.” As I try to align that life with the life of a researcher, it turns out as “Research in life is what also happens around when you are busy doing your own research.”
Stay safe and happy!
Best,
Ashirbani
Dr. Ashirbani Saha is the first holder of the BRIGHT Run Breast Cancer Learning Health System Chair, a permanent research position established by the BRIGHT Run in partnership with McMaster University.
