Children’s Hospital Taps NVIDIA AI to Transform Cardiac Care

CHOP uses NVIDIA-powered open source AI to model pediatric hearts in seconds, revolutionizing congenital heart disease treatment.

The Children’s Hospital of Philadelphia (CHOP) is leveraging open source AI tools, powered by NVIDIA, to revolutionize cardiac care for children with congenital heart disease. Using the MONAI medical imaging framework, CHOP has developed a system capable of creating precise 3D heart models in seconds—a process that previously required four hours of manual effort.

Congenital heart defects affect about 1% of all live births, presenting unique challenges due to the highly variable anatomy of each patient. Traditional treatment relies on generic devices not tailored to individual cases. Dr. Matthew Jolley, a cardiologist and researcher at CHOP, emphasized the mismatch: “You’ve got a one-of-a-kind kid and an off-the-shelf device.” The new AI-driven modeling system allows surgeons to better prepare for procedures by visualizing and testing devices in a virtual environment before entering the operating room.

CHOP’s cardiac modeling service integrates imaging data from CT scans, MRIs, and 3D ultrasounds to create anatomically accurate heart models in near real time. This advancement isn’t isolated: more than 20 U.S. children’s hospitals have adopted similar programs, with Boston Children’s Hospital now using modeling to support over 500 cardiac surgeries annually. At CHOP, roughly 200 cases will be modeled this year, a number expected to grow as the hospital expands its use of AI tools across disciplines through its Morgan Center for Research and Innovation’s IDEA Lab.

CHOP’s journey into AI-powered cardiac care began in 2015, when Dr. Jolley’s team started adapting open source imaging tools like 3D Slicer for pediatric cases. Initial workflows were labor-intensive, but the integration of machine learning through NVIDIA’s MONAI framework has dramatically streamlined the process. For example, the team uses MONAI Label and NVIDIA’s Auto3DSeg to train AI models on existing image-model pairs, enabling the creation of high-quality 3D models in seconds.

The clinical impact is significant. In one notable case, the AI tool helped locate a complex ventricular septal defect that had eluded surgeons in two prior repair attempts. With the 3D model, the surgical team successfully completed the repair on the first try. Such breakthroughs have elevated cardiac modeling from a research experiment to a standard component of care for many complex cases.

CHOP isn’t stopping at visualization. The hospital is working with NVIDIA to implement real-time physics simulations using the Newton engine, built on NVIDIA’s Warp framework. These tools allow doctors to predict how medical devices will interact with a child’s heart tissue before any procedure begins. For example, Newton can simulate how a closure device used to seal heart defects will behave under varying conditions, reducing the time required for such simulations from hours to minutes.

Future developments include integrating these simulations with NVIDIA’s Omniverse digital twin technology, powered by the OpenUSD standard. This would enable clinicians to interact with 3D models of a child’s heart in virtual reality, offering unprecedented insights into treatment planning.

Congenital heart disease affects 2.4 million people in the U.S., a population too small and diverse to attract large-scale commercial investment. Open source tools like SlicerHeart and MONAI have become vital for overcoming these economic barriers. By fostering collaboration among institutions like CHOP, Boston Children’s Hospital, and Stanford University, the open source community is accelerating progress in pediatric cardiac care.

Dr. Jolley highlighted the unique value of open source in this context: “It’s too small a population to support traditional commercial development by normal economics. But it’s such an important problem that between the research community and philanthropy, people are getting behind it.” NVIDIA’s commitment to open platforms ensures that hospitals like CHOP can access and build on cutting-edge tools without the overhead of developing them from scratch.

As CHOP continues to refine these technologies, the hospital is setting a new standard for personalized medicine. With AI-driven modeling and real-time simulations, children with congenital heart disease are receiving care tailored to their unique needs—marking a significant step forward in pediatric healthcare.