In order to provide high-quality medical care to its population, approximately 30% of whom are aged 65 or older, Japan strives to Sovereign AI initiatives supporting almost every aspect of healthcare.
AI tools trained on country-specific data and local IT infrastructure are boosting the capabilities of Japanese clinicians and researchers so they can care for patients, amid an expected shortage of nearly 500,000 professionals health by next year.
Groundbreaking technology deployments by the nation’s healthcare leaders, including in accelerated AI drug discovery, genomic medicine, medical imaging and robotics, are highlighted at the NVIDIA AI Japan Summitwhich will take place in Tokyo until November 13.
Powered by NVIDIA AI computing platforms like NVIDIA DGX Tokyo-1 Supercomputerthese applications have been developed using domain specific platforms such as NVIDIA BioNeMo for drug discovery, NVIDIA MONAI for medical imaging, NVIDIA Parabrics for genomics and NVIDIA Holoscan for medical robotics.
Drug discovery AI factories deepen understanding, accuracy and speed
NVIDIA supports the Japanese pharmaceutical market, one of the three largest in the world, with NVIDIA BioNeMo, an end-to-end platform that enables drug discovery researchers to develop and deploy AI models to generate biological intelligence from biomolecular data.
BioNeMo includes a customizable and modular programming framework and NVIDIA NIMs microservices for optimized AI inference. New models include AlphaFold2which predicts the 3D structure of a protein from its amino acid sequence; DockDiffwhich predicts the 3D structure of a molecule interacting with a protein; And RF broadcastwhich designs new protein structures capable of binding to a target molecule.
The platform also offers BioNeMo plansa catalog of customizable reference AI workflows to help developers adapt biomolecular AI models for enterprise applications.
The NIM microservice for AlphaFold2 now includes MMSeqs2-GPU, a scalable information retrieval tool that accelerates the traditional AlphaFold2 pipeline by 5x. Led by researchers from Seoul National University, Johannes Gutenberg University Mainz and NVIDIA, this integration can predict protein structure in 8 minutes instead of 40 minutes.
At AI Summit Japan, TetraScience, a company that builds AI-native scientific datasets, announced a collaboration with NVIDIA to industrialize the production of scientific AI use cases to accelerate and to improve workflows across the life sciences value chain.
For example, choosing an optimal cell line to produce biological therapies such as vaccines and monoclonal antibodies is a critical but time-consuming step. TetraScience’s new Lead Clone Assistant uses BioNeMo tools, including the NVIDIA VISTA-2D base model for cell segmentation and the Geneformer model for gene expression analysis, to reduce the selection of lead clones to hours at place of several weeks.
Tokyo-based Astellas Pharma uses BioNeMo biomolecular AI models such as ESM-1nv, ESM-2nv and DNABERT to accelerate biologics research. Its AI models are used to generate new molecular structures, predict how these molecules will bind to target proteins, and optimize them to bind more efficiently to these target proteins.
Using the BioNeMo framework, Astellas has accelerated the generation of chemical molecules by more than 30 times. The company plans to use BioNeMo NIM microservices to advance its work.
Japanese pharmaceutical companies and research institutes advance drug research and development
Astellas, Daiichi-Sankyo and Ono Pharmaceutical are leading Japanese pharmaceutical companies operating the Tokyo-1 system, a NVIDIA DGX AI supercomputer built in collaboration with Xeureka, a subsidiary of Japanese business conglomerate Mitsui & Co, to create AI models for drug discovery. Xeureka uses Tokyo-1 to accelerate AI model development and molecular simulations.
Xeureka also uses NVIDIA H100 Tensor Core GPU to explore the application of confidential computing to improve the ability of pharmaceutical companies to collaborate on training large AI models while protecting proprietary datasets.
To further support disease research and precision medicine, genomics researchers in Japan have adopted the NVIDIA Parabrics software suite to accelerate secondary analysis of DNA and RNA data.
Among them is the University of Tokyo’s Human Genome Center, the main academic institution working on a government whole-genome project focused on cancer research. The initiative will help researchers identify genetic variants unique to the Japanese population and support the development of precision therapies.
The Genome Center is also exploring the use of Giraffe, a tool now available through Parabricks v4.4 that allows researchers to map genome sequences onto a pangenome, a reference genome that represents diverse populations.
AI scanners and oscilloscopes give radiologists and surgeons real-time superpowers
Japanese healthcare innovators are building AI-enhanced systems to support radiologists and surgeons.
Fujifilm has developed an AI application in collaboration with NVIDIA to help surgeons perform surgeries more efficiently.
This application uses an AI model developed using NVIDIA DGX systems to convert CT images into 3D simulations to support surgery.
Olympus recently collaborated with NVIDIA and a telecommunications company NTT to demonstrate how cloud-connected endoscopes can efficiently run real-time image processing and AI applications. Endoscopes presented NVIDIA Jetson Orin modules for edge computing and connected to a cloud server using the IOWN All-Photonics network of the NTT communications platform, which introduces photonics-based technology to the network to enable power consumption lower, higher capacity and lower latency.
NVIDIA also supports AI-based real-time robotic systems for radiology and surgery in Japan with Holoscana sensor processing platform that streamlines the development of AI models and applications for real-time insights. Holoscan includes a catalog of benchmark AI workflows for applications such as endoscopy and ultrasound analysis.
A neurosurgeon at Showa University, a medical school with multiple campuses across Japan, adopted Holoscan and NVIDIA IGX cutting-edge industrial-grade AI to develop a surgical microscopy application that takes video footage from surgical oscilloscopes and converts them into 3D images in real-time using AI. With access to 3D reconstructions, surgeons can more easily locate tumors and key structures in the brain to improve the efficiency of procedures.
Japanese surgical AI companies, including AI medical service (AIM), Anaut, iMed Technologies And Jmees are investigating the use of Holoscan to power applications providing diagnostic assistance to endoscopists and surgeons. These applications could detect anatomical structures such as organs in real time, with the potential to reduce the risk of injury, identify conditions such as gastrointestinal cancers and brain hemorrhages, and provide immediate information to help doctors to prepare for and carry out surgical procedures.
Evolving healthcare with digital health workers
Older adults have higher rates of chronic diseases and use health services the most. So, to keep up with its aging population, Japan-based companies are at the forefront of developing digital health systems to improve patient care.
Fujifilm launched NURA, a group of medical screening centers with AI-enhanced medical exams designed to help doctors screen for cancer and chronic diseases with faster exams and lower radiation doses for CT scans.
Developed using NVIDIA DGX systems, the tool integrates large language models that create textual summaries of medical images. AI models run on NVIDIA RTX GPUs for inference purposes. Fujifilm is also evaluating the use of MONAI, NeMo and NIM microservices.
To learn more about NVIDIA’s collaborations with the Japanese healthcare ecosystem, watch the NVIDIA AI Summit On-Demand Session by Kimberly Powellvice president of corporate healthcare.