The U.S. healthcare system is rapidly adopting digital health agents powered by artificial intelligence, spanning from research laboratories to clinical environments. This development was highlighted at the NVIDIA AI Summit in Washington, D.C., where the company showcased its latest AI-accelerated tools, including NVIDIA NIM microservices and NIM Agent Blueprints.
These technologies are already being utilised in the public sector to advance medical image analysis, aid in drug discovery, and extract information from complex PDF databases. For instance, researchers at the National Cancer Institute are using AI models built with NVIDIA MONAI for medical imaging, including the VISTA-3D NIM foundation model for 3D CT image segmentation and annotation.
At the National Center for Advancing Translational Sciences (NCATS), part of the National Institutes of Health, researchers are employing the NIM Agent Blueprint for generative AI-based virtual screening to expedite and reduce costs in developing new drug molecules.
NVIDIA has introduced several NIM Agent Blueprints tailored for healthcare applications. These include a blueprint for generative virtual screening in drug discovery, a multimodal PDF data extraction tool, and a digital human blueprint for creating AI-powered avatars with potential telehealth applications.
Two new NIM microservices for drug discovery have been made available: AlphaFold2-Multimer for rapid protein structure prediction, and RFdiffusion for designing novel proteins as potential drug candidates using generative AI.
The ASPIRE research laboratory at NCATS is evaluating the NIM Agent Blueprint for virtual screening and using NVIDIA's RAPIDS software libraries to accelerate drug discovery research. According to the NCATS informatics team, processes that previously took hours on CPU-based infrastructure now complete in seconds with NVIDIA AI.
Sam Michael, chief information officer of NCATS, highlighted the potential of AI-powered PDF data extraction in making valuable information more accessible from previously unsearchable databases, particularly for rare disease research.
Several startups and major tech companies are integrating NVIDIA's NIM microservices and NIM Agent Blueprints into their platforms. Abridge, an NVIDIA Inception startup, recently won a contract with the U.S. Department of Veterans Affairs to help transcribe and summarise clinical appointments. Amazon Web Services is partnering with NIH to make NIM Agent Blueprints more accessible to the biomedical research community.
From accelerating drug discovery to improving patient care, the adoption of AI agents in the U.S. healthcare system offers much promise.