I think that AI applications in surgery are a little bit more safe and filled with promises compared to what is going on in this enormous market advanced for AI applications in medicine, where the ability of diagnosing is made faster by applications that you cannot control because they are based upon a black box of reasoning with any confidential, you know, ability of the machine to explain to you how they made their decision...
I think that AI applications in surgery are a little bit more safe and filled with promises compared to what is going on in this enormous market advanced for AI applications in medicine, where the ability of diagnosing is made faster by applications that you cannot control because they are based upon a black box of reasoning with any confidential, you know, ability of the machine to explain to you how they made their decision. In surgery, surgeons make the difference because in surgery we have to sense and understand what are, in a video-based surgical operation, what are the main points that should be ameliorated and made more precise. Then we have to select this and give to the machine learning algorithm what we want to understand in terms of wrong or correct. And then, as things go by, we will learn to work together in providing big data, and this big data will come back to us and act at the end like an AI visual companion while we are operating. So it’s totally different compared to other kinds of applications where you delegate to the machine a diagnosis that should remain in your hands. Surgery will remain in your hands, but with the help of a special assistant that knows much more than you know, because he’s been assisting thousands of other people. Right now, we have a lot, really a lot of, we call them AI tools. They are not, you know, integrated systems of AI. They are just tools that enhance some specific task. So we can see better the recurrent laryngeal nerve in the thorax. We can identify the dangerous points and maintain a critical view of safety during a cholecystectomy or a major operation inside the abdomen or the thorax. In some cases, all because surgeons have been giving their major worries to the machine, we can highlight a plane of division in order for the surgeon to dissect without making any specific damage to the surrounding tissue. But these are separate tools. We are aiming to get much more. We would like a companion that in surgical oncology will accompany the entire patient pathway from the beginning, so the diagnosis, to the organization of his multidisciplinary discussion. The selection of opportunities in order to choose the best treatment in an era where there are so many new parameters making the difference, like molecular subtype of the tumor or augmented risk for patients that we call fragile. Not anybody should receive the same thing. And this is the way, based on data, in which we can receive assistance from the same system that then will accompany us during the operation, like I said before. Well, what we are afraid of is a total delegation of our task to the machine. But fortunately, in Europe, level five of confidence is not allowed for clinical use. So we should not be scared about a day where the robot is doing the operation without our participation. But I think that we would be participating in this process, guiding not the market, but the patient needs in using algorithms that will support us in making better choices that sometimes are not just in those challenging cases where you don’t know exactly what to do, but also in ordinary cases because there are every day more parameters and more factors that sometimes would totally change your decision-making process in deciding what is the treatment plan for a patient. So I think this will remain and this will be our best achievement if we learn how to interact with the big companies that are producing these tools.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.