MIDL Vision 2022

Yeah, it's for the planning of the needle during thermal ablation of liver cancer. It's difficult to find the best trajectory for the needle to go through each patient's body because each patient is different. So the conventional optimization algorithm was taking the anatomy of the patient and trying to optimize this trajectory by minimizing a cost function subject to several constraints. We tried to replace the optimization with a deep reinforcement learning framework, so that an agent is trained to find the best trajectory. You are not a doctor. You are not a surgeon. But, this needle and the software which guides it make you very close to being one. Yeah, I try to put my feet into their everyday life. I assist to a lot of those interventions to try to see how they do it. I don't think I could ever do it. I think it needs a lot of expertise, but yeah, I try to, at least on my computer. I would love to hear your message to the medical AI community, if you have any. It's a great field. I hope we continue pushing the frontier of this all together as a community. It's really important to have academic colleagues pushing the research. At the same time, it's important to listen to the medical professionals, and also having the industry - either a big company or startup - to try to move those things forward. What would be a great success ten years ahead? For example, in these liver treatment cases, I will feel so proud if one of the algorithms that we develop makes it to the clinics and actually makes patients live longer and really treat patients with liver disease. More than 100 inspiring interviews with successful Women in Science await for you in our archive! 22 VISION MIDL Women in Science