The Doctor’s Digital Assistant: AI in Future Medicine

Learn how medical genomics has transformed in recent years and its potential to help patients. I will describe some of the bottlenecks in realising the enormous potential of genomics in healthcare, and how AI tools are helping to overcome these problems and directly benefit patients.

Nobody likes to be ill, and often when we get a cough and a temperature we want to be reassured it’s not something serious. Respiratory tract infections (RTIs) are one of the most common kinds of infections. Most people get better after 5 to 10 days, but for some people, the illness can become more serious or lasting. Our research aims to use the sounds of our speech and breathing, combined with the power of artificial intelligence (AI) tools, to track RTIs and predict if they are getting better or worse.

We know that when we are ill, our voice can change and our breathing can become noisy, and we believe there are clues in those sounds about how an RTI is changing day to day. Sounds like these are easy to record and analyse using smartphones, so people could get a really quick answer about the best treatment. If we can build a tool that reassures people who are going to get better by themselves, those with more serious conditions that need medical treatment will be able to get a GP appointment more easily.

All cancer patients are discussed at an expert multi-disciplinary team (MDT) meeting, where complex data is used to make the best treatment decisions. These meetings are time-constrained and over-burdened, with inherent bias. We have created an explainable and trustworthy AI decision support tool for patients with oesophageal (gullet) cancer. Tim will tell you how we did it and why we trust it, and consider the future of AI-driven treatment planning.