GE Healthcare’s Chief Technology Scientist for Cardiology Solutions, Eigil Samset, shares what it takes to keep progressing AI under the pressure of an ongoing global health crisis while overcoming roadblocks on the development pathway.
“I believe it’s not so much about AI but what it can bring,” says Eigil Samset, Chief Technology Scientist for Cardiology Solutions at GE Healthcare. “We want to improve patient outcomes, their quality of life and minimize disease and death. We also want to make healthcare more efficient by offering better workflows for clinicians while saving costs. These are the objectives in our minds when we introduce AI into cardiology.”
To that end, last June, GE Healthcare announced a new collaboration with the American College of Cardiology (ACC) to participate and support its Applied Health Innovation Consortium to build a roadmap of AI and digital technology in cardiology.
As a silver sponsor for the upcoming AIMed Global Summit in California, GE Healthcare hopes to take the opportunity to engage with fellow clinicians and C-suite leaders, gather feedback to refine the roadmap, and immerse themselves in an atmosphere where new partnerships can be formed to co-create meaningful AI applications. “We are not able to make everything ourselves and we can’t accelerate the development of cardiology AI without help from developers and healthcare providers,” Samset explains. “The mission of AIMed is fully aligned with our mission in healthcare. We both wish to create efficient workspaces for clinicians with advancing patient outcomes at its core.”
When it comes to the significance of a cardiology AI roadmap, Samset believes it’s more appropriate to describe it as a journey. “We continuously set ourselves targets and overcome challenges as we travel,” he adds. “We have witnessed some preliminary success of AI in the field – companies and researchers who have brought us fantastic, singular AI algorithms that perform image recognition or automate tedious and repetitive tasks. To continue heading in the right direction, we need to develop sound and robust products and adopt and integrate them into clinical workflows. To achieve real progress, we need what we call, ‘explainable AI’.
“Sometimes, we may feel there’s some magic going on inside the Blackbox. In reality, there’s a lot of work to be done in interrogating these AI models to make them understand what’s going on and use them in clinical contexts. I am excited about a future where cardiologists can have all the relevant information about a patient at their fingertips and be able to make decisions based on a set of available algorithms. In addition, we can combine data acquired at home and clinical settings to influence patients’ management of their diseases and detect abnormalities at an earlier stage.”
To arrive at that future, Samset believes it’s important for everyone to use and understand an AI-driven tool. “People shouldn’t fear AI and they should not trust AI blindly too,” he says. “We don’t need to go into the nitty gritty. Instead, people should know AI’s potentials, limitations, and possible biases. We should demand explainability and solid evidence for whatever they are built for. One of my biggest concerns is that we fail to realize when a model is biased.
“Today, most of the development in AI is happening in first-world countries. As much as we are targeting a future where everyone can benefit from AI, especially those from developing countries, we can’t deny that our data and the models we developed today are more catered to the first-world countries. That’s why I can’t stress enough about the need to work together – technologists, healthcare providers, and patients – to build AI models that we can trust and be generalized for a wider population.”
Samset also feels for those who find AI disruptive. “Ideally, we should have AI models that can seamlessly fit into the existing workflow. A new workflow may sound painful for some but if we do it right, it brings about positive changes that are fulfilling. We need the right education, training and co-creation between developers and clinicians to give people the courage to take that leap and examine how they can best benefit from AI. At GE Healthcare, our vision is to build a world that works. We are here to help in the journey towards that future.”
When asked if the ongoing COVID-19 pandemic has changed people’s perception of AI, Samset is adamant that the global health crisis has provided many valuable lessons. “It’s not limited to AI,” he explains. “There’s an enormous wealth of data accumulated from those who suffered from the disease and everything around it. We worked together with many scientists and hospitals on the acceleration of respirator manufacturing and setting up tele-ICUs. All these showed that if we put our minds to something, we can rapidly accelerate technology development.
“As we recover from the pandemic, the main challenge becomes resilience. New waves of infections and new diseases are likely to strike the community again. AI-driven clinical decision support will surely play a major role, from hospitals to home-based diagnostic units, because we need to be more efficient, eliminate backlogs and bring better outcomes in the post-pandemic era. Even though the field is still in its early days, I am optimistic. There’s so much investment and effort being put into this by medical scientists and companies that will surely bring positive results. We just need to be careful about biases, and some of the other concerns around AI. If we keep the mindset right and keep the investment going, there will be many things we can look forward to in the next few years thanks to AI.”
GE Healthcare is the silver sponsor for the AIMed Global Summit 2022.