The Children’s Hospital of Philadelphia is using open-source medical imaging software to plan how cardiac devices are placed in children, and says a workflow that once took a skilled researcher up to four hours now completes fast enough for routine clinical use.
The tooling is open source. MONAI, the medical imaging framework, handles the imaging pipeline, with MONAI Label and Auto3DSeg for segmentation and SlicerHeart for cardiac modelling; NVIDIA’s Warp and its Newton physics engine run the simulation step. The clinical problem is that children’s hearts vary far more than adult sizing charts allow for, so fitting a device means modelling the individual anatomy. Testing several device configurations previously meant leaving a job to run overnight.
On scale, the write-up published on 15 September says more than 20 children’s hospitals in the United States now run cardiac modelling programmes. Boston Children’s Hospital’s modelling supports roughly 500 cardiac surgeries a year, and CHOP expects about 200 modelled cases in 2026.
Worth being clear about the source: these figures come from NVIDIA’s own account of a hospital using its software, and the timings are the hospital’s. No independent evaluation has been published on whether the simulations change surgical decisions or patient outcomes. The speed-up is a claim about workflow, and a plausible one, but it is not yet evidence about results.