Which Company's Case Study?
Johnson & Johnson MedTech is the medical device division of Johnson & Johnson, one of the world's largest healthcare companies. Its products are used in 80% of operating rooms globally, supporting healthcare professionals in areas such as general surgery and orthopedics. This initiative integrates NVIDIA's cutting-edge AI technology into J&J MedTech's extensive network and digital ecosystem.
Challenges to be Solved
In the operating room, vast amounts of data are generated in real-time from endoscope video feeds, vital sign monitors, and various medical devices. However, it has been challenging to integrate and analyze this data comprehensively and utilize it immediately to support surgeons' decision-making during operations. Furthermore, standardizing surgical procedures, providing effective training for junior surgeons, and improving the overall efficiency of the surgical workflow were also significant challenges.
How AI Was Used
To solve these challenges, the company decided to leverage NVIDIA's edge AI platform for healthcare, 'NVIDIA IGX', and its software development kit, 'NVIDIA Holoscan'. These are designed to process camera footage and sensor data with low latency at the surgical site (the edge) and provide AI-driven analysis results in real-time. This establishes a foundation for safely running AI applications within the hospital without transferring sensitive patient data externally.
Estimated Architecture
The following is an estimated architecture derived from publicly available information and general configurations. It does not definitively represent the actual internal structure.
- ▸Input Data: Video footage from endoscope cameras installed in operating rooms, sensor data from various medical devices, vital sign monitor data, etc.
- ▸AI Processing/Model or Search Layer: AI applications built using the NVIDIA Holoscan SDK run on the NVIDIA IGX platform. Computer vision AI models perform real-time identification of anatomical features, tracking of surgical instruments, and analysis of workflows.
- ▸Output to Business Systems: Real-time display of AI-driven analysis results and guidance on operating room monitors. Generation of post-operative analysis reports and utilization as training data.
- ▸Monitoring/Governance: Model validation compliant with medical device regulations (e.g., FDA), and operation ensuring data privacy and security.
Implementation Effects and Key Points to Note
- ▸Expected improvements in surgical precision and safety through real-time decision-making support for surgeons.
- ▸Potential contribution to hospital management by standardizing procedures and improving efficiency through data analysis of surgical workflows.
- ▸Readers should note the utilization of 'edge AI' in addition to cloud AI. This is a crucial choice for achieving real-time processing in the operating room, where low latency is an absolute requirement.
- ▸This initiative aims not only for a single AI function but also for an open platform where third-party developers can participate, envisioning the construction of a surgical AI ecosystem.
What Japanese Companies Can Learn
This case study offers significant insights for AI utilization in 'on-site' environments that demand real-time processing and high reliability, not limited to medical settings but also applicable to quality inspections on manufacturing lines, infrastructure equipment maintenance, and construction sites requiring advanced operations. The 'collaborative AI' adoption approach, where AI augments experts' capabilities and supports their judgments rather than replacing them, is applicable across many industries. Furthermore, the strategy of accelerating innovation through the creation of an open platform can also serve as a valuable reference.