Health apps and wearables that measure our pulse and detect heart rhythm disorders (HRS), for example, are already part of everyday life for many. Yet they can be much more than a lifestyle product, as they enable patients to track their own health data and implement therapies. These tools thus offer a novel way to ensure continuous follow-up after a myocardial infarction or in the case of HRS and thus to care for patients more quickly and to change therapies.
The first AI-based programmes for the automatic evaluation of echocardiography or MRI images are also in the clinical validation phase and will be available in the next five years. From this, hospitals hope to be able to continue treating the growing number of patients while the medical profession is short of staff.
Prof. Alan Fraser (UK) is confident that AI will not only help us to better understand diseases, but also to personalise diagnostics and therapies to achieve better outcomes.
We speak of AI when computer programmes are capable of learning. And there are two main ways for this learning to take place. One way can be called "supervised learning": Put simply, the researchers show the computer a large number of similar images and, depending on the question, tell it what is right or wrong, healthy or pathological. The other way is called "unsupervised learning": Here, the computer must independently recognise and classify differences in diseases.
In medicine, supervised learning is mainly used because unsupervised learning requires a much larger amount of data. Moreover, AI only solves individual tasks and cannot analyse complex issues. To claim that AI will take over medical tasks or even replace the medical profession is completely absurd at the current stage of development. Another important point, Prof. Partho Sengupta stressed, would be for society to debate whether faulty AI-based diagnoses or therapies are acceptable.
Artificial intelligence has already revolutionised several fields and it is impossible to imagine the modern world without it. However, the use of AI in medicine is shaping up to be more sensitive and slower, as it is practically a matter of "life" and "death". On the one hand, health data must be much better organised and structured so that the AI algorithms can be trained, and on the other hand, society must decide whether misdiagnoses or incorrect AI-based therapy decisions are legally and ethically acceptable. Until then, AI can certainly provide great support to doctors in their daily tasks.
Source: Great Debate: artificial intelligence will change the way we practice cardiology - ESC Congress 2022 - Topic: Artificial Intelligence (Machine Learning, Deep Learning) - 26 August 2022 - 14:00 - 15:15