New Episode of ADAPT Radio’s ‘HumanAIse’ on AI and Healthcare

30 November 2022

ADAPT Radio’s HumanAIse series continues this month with a discussion on the groundbreaking advances in AI and healthcare with Dr. Lucy Hederman and Prof. Mark Little.

From wearable devices to digital hospitals, modelling disease patterns to training medical professionals, Artificial Intelligence (AI) can have a transformative impact on healthcare. But the challenge remains to develop and deploy innovative and agile solutions that can unlock the data needed for precision medicine. ADAPT Academics based at Trinity College Dublin, Professor Mark Little and Dr. Lucy Hederman, join this month’s discussion to give their professional insight.  HumanAIse, is an ADAPT podcast that provides an in-depth look at the future of AI, automation, and the implications of entrusting machines with our most sensitive information and decisions. This series will feature conversations with a range of ADAPT experts who explore the future of AI and human input.

Prof. Little is Professor of Nephrology at the School of Medicine, Trinity College Dublin and Consultant Nephrologist in Tallaght and Beaumont Hospitals. He is a clinician scientist with an interest in translational immunology as applied to autoimmune disease. Dr. Hederman is an assistant professor in the School of Computer Science and Statistics, a member of the SFI funded ADAPT Centre, and the Director of the Centre for Health Informatics, at Trinity College Dublin. 

Prof. Little and Dr. Hederman kicked off the podcast with a discussion on the potential power of AI in healthcare. The example they provided included utilising AI to predict when a patient is going to have a disease flare so that they can receive an appropriate amount of medicine to match the potential of the projected flare. AI can look at patterns of what has happened to a patient before and predict what will happen in the future. A doctor can see not only the features of their patient but also previous patients who have also been affected, build models, and then use that information to do something specific for the patient. 

This type of care is not without its challenges, for instance, when predicting disease flares the volume of data available can hamper findings. For example, vasculitis, an inflammation of the blood vessels, is a rare disease meaning that the number of patients from which to draw data is significantly reduced, particularly in Ireland where we already have a relatively small population. As a result, there is not enough data available to be able to effectively detect patterns. The Vasque project, as described by Dr. Hederman, aims to address this by integrating data from vasculature patients across Europe. Currently, Vasque is integrating data across seven countries in Europe who have registers of patients with vasculitis and is integrating that data so that clinical researchers and AI techniques can be applied across much greater scale of data. 

Prof. Little also discusses robotics merging into the surgical sphere and how some hospitals can implement this. Robots can be trained on numerous types of surgeries, such as abdominal or bowl surgeries, and can perform them very precisely once trained. However, according to Prof. Little, this will not replace the kind of close patient contact, such as a hand on an abdomen to try and palpate a liver, because patient contact is so important in healthcare. An example of how effective utilising robotics in a healthcare setting can be is with the success of Akara Robotics, a spin-out from Trinity College Dublin now based at The Digital Hub in Dublin’s Liberties. Akara developed ‘Violet’, the first clinical-grade fully autonomous UV disinfection robot. Violet uses advanced sensing and AI abilities that allow it to traverse complex hospital environments and closely control the parts of the room that are irradiated.

Another area where AI is making a big impact is in the development of new drugs. Traditionally, developing a new drug is a long process with a 10-12 year multibillion euro pathway. AI can be used to access massive databases of compounds to look for patterns to treat conditions and massively accelerate that aspect of drug development. For example, the Deep Mind group have been able to design an AI algorithm that can predict the 3D structure of nearly every protein. Previously, medical professionals would use an X-ray to take a picture of proteins to see how they diffracted so as to get an estimate of what they look like. It was estimated that about 1% of protein that had their structure worked out but now with this new approach it has been brought up to about 80%.

Precision medicine allows for deep learning methods and modern models to assist medical professionals in treating each patient as closely as possible to what is precisely appropriate for them because they have more data to work with. Prof Little stresses that the expertise of a medical professional utilising the power of instinct and experience is something that AI cannot replicate. Therefore both professional expertise and the ability of AI can both be utilised to provide a high level of care to the patient. AI can provide the doctor or nurse with insights into what’s going on with that patient that they may not have known otherwise.

For further insight into How AI can provide decision-making support, disease treatment and precision medicine delivery, catch HumanAIse on SoundCloud, iTunes, Spotify, and Google Podcasts.

ADAPT Radio: HumanAIse is ADAPT’s newest podcast series providing an in-depth look at the future of AI, automation and the implications of entrusting machines with our most sensitive information and decisions.