How AI is Revolutionizing Cardiology with Ioana Cretu

Ioana Cretu completed a bachelor’s degree in biomedical engineering before embarking upon her PhD project in 2020. In her words, ‘the main aim of my PhD was creating this temporary cardiac pacing simulator that can be used as a training tool for clinicians’.  

Ioana spoke about the practice of temporary cardiac pacing devices, which are like pacemakers but are not implanted into patients for life. They’re ‘an external device – so basically a generator that stands on the table next to the patients – that is inserted after cardiac surgery’. However, the problem with these devices comes down to a lack of training among professionals – meaning they sometimes do more harm than good. 

As a result of this, Ioana created a platform that can provide clinicians with better training and knowledge of temporary cardiac pacing devices. ‘By using this platform, [doctors] will have a safe environment to practice as much as they need, until they feel confident. They can also practice without the risks associated with real settings, as the platform can be used at home. They don't have to be in the hospital’.  

Essentially, her project aims to bridge the gap between theoretical knowledge and practical application in this specific area of cardiological treatment.  

The involvement of AI and machine learning  

As well as using AI within her own project, Ioana spoke about the impact AI and machine learning are making in hospitals and how they’re assisting medical professionals in their decision making. 

In response to Glenn’s question, ‘how do you see AI affecting this world? It's only going to get bigger, isn't it?’ Ioana responded:  

‘AI is already making an impact in hospitals and is assisting doctors in making the right treatment decisions. It’s being used in diagnostics prevention and in imaging techniques as well. However – all this progress should really be a lot slower as we need to ensure we’re making responsible decisions. 

‘It's about human life. It's about our life. So, when you make the decision of using an AI system, you have to look very carefully at the ethical side of things. Is the data biased or not? Ultimately, AI is only as good as the data you're training the model on.’ 

Being wary of AI data biases 

The truth is, AI is not free from the tendencies of human prejudice and very often ends up with biases.  

When discussing the topic, Glenn pointed out that ‘we have a responsibility to proactively address these [AI] biases. There’s no excuse now, to not be aware that they exist. Moving forward as an industry, we need to acknowledge that if we’re not getting enough data from certain communities and demographics, we’ve got to try our hardest to fill the gaps’.   

Ioana agreed, adding that it’s also important to look at these groups individually, especially from a healthcare point of view. She gave an example of using specific thresholds in traditional medicine. When a patient goes for a blood test, the doctor will look at the specific thresholds and note whether they are at risk of any conditions. But very often, these thresholds only account for one demographic – and someone from a different background may not receive the right diagnosis.  

Essentially, one person’s normal isn’t necessarily another person’s normal. That’s why it’s so important to have well distributed data sets while looking at individual groups.  

AI prevention techniques in cardiology   

Ioanna then went on to discuss some of the incredible prevention techniques within cardiology that are taking place thanks to artificial intelligence.  

Currently, there’s a group of researchers from the University of Oxford who have created an AI model that can actually analyze the CT (computer tomography) images that are routinely done in patients with chest pain. Their AI model is so advanced that it can predict heart attacks in groups of patients up to five years before they were due to occur. It does this by looking at miniscule details, such as scarring, inflammation and changes in the structure of their coronary arteries.  

The fact that AI is already achieving such feats shows just how remarkable it can be. And as it continues to develop, there’s no question that it’ll continue having a massively positive effect on patient outcomes.  

Using AI as a tool, not a decision maker 

So, will AI eventually replace people in the field of medicine? Ioana concluded the conversation by stating that ultimately no, it should never become a decision maker.  

When it comes to treating patients, you need empathy – and AI simply cannot replicate this level of human touch. Medical decisions must never be made strictly based on data. There’s patient history, feelings, test results and so much more to consider. 

Ioana said, ‘AI is better used as a tool that guides doctors to make the right decisions during surgery, like in highly advanced image processing. It’s already out there and I hope it goes even further in helping doctors along the right path’.  

It’s inspiring to hear about the amazing work done by Life Sciences professionals, and in this case the wonderful Ioana Cretu, whose work has led to the creation of new techniques that are advancing cardiovascular research. It was a pleasure hearing her thoughts on the involvement of AI in cardiology along with the details of her PhD project. 

If hearing Ioana’s story has inspired you to share yours, please feel free to get in touch with our team! We’d be thrilled to host you on a future episode of The Affected.