Harnessing the Power of AI for Pediatric Cardiology: An Interview with Matheus Araújo
The intersection of technology and healthcare continues to evolve, with artificial intelligence (AI) playing a transformative role. One of the key individuals driving this innovation is Matheus Araújo, One Heart Health’s Technology Director. His work with One Heart Health is helping to revolutionize pediatric cardiology through advanced AI algorithms. In a recent interview, Matheus shared insights into his journey, challenges, and the exciting future of AI-powered healthcare.
Matheus’s passion for combining technology and healthcare began long before his current role. “I’ve been involved with One Heart Health since 2019, which predates my work in a hospital,” Matheus explained. “At the time, I was doing my PhD at the University of Minnesota, and I always had this willingness to use the knowledge I created in the computer world to help those who need it.”
His involvement with One Heart Health started after meeting Quan Ni, One Heart Health’s founding leader. “Quan showed me the project and his vision for One Heart Health. I immediately saw a unique role I could perform, especially in building algorithms,” Matheus said.
Breaking Down the CHD Screening Process to Rebuild It
To improve the diagnosis of congenital heart disease (CHD), Matheus and his team analyzed how pediatric cardiologists achieve accurate diagnoses. It became clear that this multi-step process, intuitive for doctors, needed to be replicated digitally. This led to the development of two distinct AI algorithms: a heart sound vs noise classification algorithm and a heart murmur detection algorithm.
To address this, Matheus and his team developed a two-step AI process. “We created an algorithm that solves one specific problem,” Matheus explained. “When our partners collect heart sounds, sometimes they accidentally record noise instead of actual heart sounds. This algorithm ensures only valid heart sounds are used in the murmur detection model.”
Matheus emphasized the importance of this step: “If we feed noise into the murmur detection algorithm, the results can be unpredictable. Ensuring the input data is clean is crucial for accurate diagnosis.” By mimicking the critical listening skills of a doctor, these algorithms enable the One Heart App to perform accurate diagnostics in real time.
The road to developing these algorithms has not been without its challenges. “There are two big challenges,” Matheus shared. “The first is data. We need thousands of samples, both with and without murmurs, to train the algorithm.” To address this, One Heart Health has partnered with collaborators in West Virginia, Colombia, Mongolia, and Nigeria, building a diverse and comprehensive dataset.
The second challenge is the evolving nature of AI technology. “When I started in 2018, the technology for processing heart sounds was just beginning to develop,” he said. “It wasn’t clear which models would work, so we conducted a large number of experiments to find the most effective one.”
Real-Time Diagnostics and Expanded Partnerships: AI is Opening Doors
The ultimate goal of Matheus’s work is to integrate these algorithms into the One Heart app, enabling real-time diagnostics in low-resource areas. “We’ve created an API framework that allows the app to communicate directly with the algorithm,” he explained. “Healthcare professionals will be able to upload heart sounds and receive diagnostic results in real time. Seeing this system work smoothly in practice will be a major milestone for me.” The plan is to have these features fully integrated into the app before the end of 2025, providing low-resource healthcare settings with a powerful tool for early CHD detection.
Looking ahead, Matheus sees One Heart Health’s work as just the beginning of what AI can achieve in pediatric cardiology. “We have collaborators in West Virginia, Bolivia, Mongolia, and Nigeria, all contributing data,” he said. “I truly believe this is just the start. As more people see the potential of what we’re doing, we’ll attract new partners, which will help us create better datasets and improve our models over time.”
He believes that AI’s role in healthcare will continue to grow, with increasingly accurate and sophisticated models. “This is just the first spark of what we can do,” Matheus said. “The goal is to always keep improving our model as we get more data from around the world.”
Matheus’s work speaks to the power of innovation, collaboration, and determination. “Our team has been working to optimize everything to make this possible,” he said. “Creating new algorithms isn’t easy—it’s not cheap, and it comes with risk factors—but we’re doing our best to deliver the best possible solution for our organization and to meet our targets.”
As One Heart Health advances its mission to deliver life-saving diagnostics to children in need, Matheus and his team play a vital role in driving this effort forward. Their work highlights the transformative potential of AI in healthcare. Our team is looking forward to reaching these 2025 goals.
Stay tuned as One Heart Health and Matheus’s team continue their groundbreaking work to transform pediatric cardiology through AI.
