Main healthcare organizations throughout the globe are utilizing agentic AI, robotics and digital twins of medical environments to reinforce surgical precision, enhance workflow effectivity, enhance medical diagnoses and extra.
Bodily AI and humanoid robots in hospitals have the potential to automate routine duties, help with affected person care and deal with workforce shortages.
That is particularly essential in locations the place challenges to optimum healthcare companies are paramount. Such challenges embrace hospital overcrowding, an getting older inhabitants, rising healthcare prices and a scarcity of medical professionals, all of that are affecting Taiwan, in addition to many different areas and nations.
On the COMPUTEX commerce present in Taipei, NVIDIA right this moment showcased how main Taiwan medical facilities are collaborating with high system builders to combine good hospital applied sciences and different AI-powered healthcare options that may assist scale back these points and save tens of millions of lives.
Cathay Basic Hospital, Chang Gung Memorial Hospital (CGMH), Nationwide Taiwan College Hospital (NTUH) and Taichung Veterans Basic Hospital (TCVGH) are among the many high facilities within the area pioneering healthcare AI innovation.
Deployed in collaboration with main system builders reminiscent of Advantech, Onyx, Foxconn and YUAN, these options faucet into NVIDIA’s agentic AI and robotics applied sciences, together with the NVIDIA Holoscan and IGX platforms, NVIDIA Jetson for embedded computing and NVIDIA Omniverse for simulating digital worlds with OpenUSD.
CGMH Boosts AI-Powered Medical Imaging
With a mean of 8.2 million outpatient visits and a pair of.4 million hospitalizations a 12 months, CGMH estimates {that a} third of the Taiwanese inhabitants has sought therapy at its huge community of hospitals in Taipei and 7 different cities.
The group is pioneering good hospital innovation by enhancing surgical precision and workflow effectivity by means of superior, AI-powered colonoscopy workflow options, developed in collaboration with Advantech and primarily based on the NVIDIA Holoscan platform, which incorporates the Holoscan SDK and the Holoscan Sensor Bridge working on NVIDIA IGX.
NVIDIA Holoscan is a real-time sensor processing platform for edge AI compute, whereas NVIDIA IGX gives enterprise-ready, industrial edge AI purpose-built for medical environments.
Utilizing these platforms, CGMH is accelerating AI integration in its colonoscopy diagnostics procedures. Deployed in gastrointestinal session rooms, the AI-powered instrument collects colonoscopy streams to coach a personalized mannequin constructed on Holoscan and offers real-time colonic polyps identification and classification.
CGMH’s AI infrastructure — comprising NVIDIA accelerated computing, NVIDIA DGX programs, the MONAI framework, NVIDIA TensorRT-LLM open-source library, NVIDIA Dynamo inference framework, and the NVIDIA NeMo and Clara platforms — allows accelerated analysis and improvement throughout the group.
CGMH serves almost 50 AI agent fashions that day by day assist the hospital analyze medical imaging, bettering diagnostic accuracy, throughput and real-time inference at scale. For instance, NVIDIA Triton-powered AI sped new child examination document processing by 10x.
Cathay Basic Hospital Improves Diagnostics With AI
Cathay Basic Hospital, a Taipei-based healthcare middle that gives hospital administration and medical companies, has labored with Onyx and software program supplier aetherAI to develop an AI-assisted colonoscopy system that highlights lesions, detects hard-to-spot polyps and points alerts to assist physicians with diagnoses.
Powered by a compact, plug-and-play AI BOX system — constructed with the NVIDIA Jetson AGX Xavier module — the AI system is educated on over 400,000 high-quality, physician-annotated pictures collected from sufferers with various and extreme lesions over 4 years.
Research have proven the system can obtain as much as 95.8% accuracy and sensitivity whereas bettering adenoma detection charges by as a lot as 30%. These enhancements help physicians in lowering diagnostic errors and making extra knowledgeable therapy selections, in the end contributing to improved affected person outcomes.
NTUH Detects Liver Tumors, Cardiovascular Dangers With AI
Within the 100+ years since its founding, NTUH has nurtured numerous professionals in medication and is famend for its trusted scientific care. The nationwide instructing hospital is now adopting AI imaging to extra rapidly, precisely diagnose sufferers.
NTUH’s HeaortaNet mannequin, educated on greater than 70,000 axial pictures from 200 sufferers, automates CT scan segmentation of the guts, together with the aorta and different arteries, in 3D, enabling fast evaluation of dangers for heart problems. The mannequin, which achieves excessive segmentation accuracy for the pericardium and aorta, considerably decreased knowledge processing time per case from an hour to about 0.4 seconds.
As well as, NTUH collaborated with the Good Liver Basis and system builder YUAN to develop a diagnostic-assistance system for liver most cancers detection throughout ultrasounds. It faucets into an NVIDIA Jetson Orin NX module and a deep studying mannequin educated on greater than 5,000 annotated ultrasound pictures to establish malignant and benign liver tumors in actual time.
NVIDIA DeepStream and TensorRT SDKs speed up the system’s deep studying mannequin, in the end serving to clinicians detect tumors earlier and extra reliably. As well as, NTUH is utilizing NVIDIA DGX to coach AI fashions for its system that detects pancreatic most cancers from CT scans.
TCVGH Streamlines Multimodal Imaging and Medical Documentation Workflows With AI
Taichung Veterans Basic Hospital (TCVGH), a medical middle and a instructing hospital administered by the Veterans Affairs Council in Taipei, has partnered with Foxconn to construct bodily and digital robots to reinforce staffing, bettering clinician productiveness and affected person experiences.
Foxconn developed an AI system that may analyze medical pictures and spot indicators of breast most cancers sooner than conventional strategies, utilizing NVIDIA Hopper GPUs, NVIDIA DGX programs and the MONAI framework. By tapping into scientific knowledge and multimodal AI imaging, the system creates 3D digital breast fashions, rapidly highlighting areas of concern in scans to assist radiologists make quicker, extra assured selections.
Foxconn can be working with TCVGH to construct good hospital options just like the AI nursing collaborative robotic Nurabot and tapping into NVIDIA Omniverse to create real-time digital twins of hospital environments, together with nursing stations, affected person wards and corridors. These digital replicas function high-fidelity simulations the place Jetson-powered service robots might be educated to autonomously ship medical provides all through the hospital, in the end bettering care effectivity.
As well as, TCVGH has developed and deployed its Co-Healer system, which integrates the Taiwanese native massive language mannequin TAIDE-LX-7B to streamline scientific documentation processes with agentic AI.
Co-Healer, constructed on the NVIDIA Jetson Xavier NX module, processes and helps summarize medical paperwork — reminiscent of nursing progress notes and well being training supplies — and helps medical examination preparation by offering college students with on the spot entry to nursing pointers and patient-specific protocols for scientific procedures and diagnostic exams. This helps healthcare staff alleviate burnout whereas giving sufferers a clearer understanding of their diagnoses.
Be taught extra in regards to the newest AI developments in healthcare at NVIDIA GTC Taipei, working Might 21-22 at COMPUTEX.