KMUH Partners with Foxconn to Launch Colonoscopy AI System; Achievements Step Onto the International Stage
KMUH and Foxconn co-developed the world's first 'Colonoscopy AI Agent' powered by NVIDIA's platform, offering real-time lesion detection and auto-reporting to enhance medical quality.
📋 Article Processing Timeline
- 📰 Published: April 24, 2026 at 18:21
- 🔍 Collected: April 24, 2026 at 18:31 (10 min after Published)
- 🤖 AI Analyzed: April 24, 2026 at 22:23 (3h 51m after Collected)
Central News Agency
(CNA Reporter Lin Chiao-lien, Kaohsiung 24th) In response to the demand for colorectal cancer screening and the development of medical AI, Kaohsiung Medical University Hospital (KMUH) has collaborated with Foxconn Technology Group to announce the world's first 'Colonoscopy AI Agent' powered by the NVIDIA IGX Thor platform, serving as a real-time clinical decision support tool.
KMUH issued a press release today stating that according to 2022 data from the Ministry of Health and Welfare, one person is diagnosed with cancer every 4 minutes, with the number of confirmed colorectal cancer cases remaining high. The gold standard for colorectal cancer is a colonoscopy (full colon endoscopy), where the doctor directly observes the lesions and performs biopsies for pathological examination. The key to diagnosis highly depends on the accuracy of the endoscopy.
KMUH stated that in order to make Artificial Intelligence a real-time assistant for clinical decisions, KMUH collaborated with Foxconn to build an advanced smart endoscope solution. The 'Colonoscopy AI Agent' has already debuted at NVIDIA GTC 2026.
KMUH pointed out that the 'Colonoscopy AI Agent' has three core functions: real-time lesion detection and pathological classification, which completes analysis in milliseconds; simultaneous monitoring of bowel cleanliness and withdrawal speed to improve inspection quality consistency; and 3D intraoperative navigation capabilities, which instantly mark the location and size of polyps to help doctors position accurately.
The hospital stated that the system also introduces AI agent technology, automatically generating inspection reports through the NVIDIA Nemotron model, and combining with AI chatbots to optimize clinical workflows. This helps doctors quickly integrate information, reduce administrative burdens, and allow medical personnel to focus on patient care and diagnostic judgment.
KMUH Superintendent Wang Chao-yuan stated that the hospital combines clinical data with independent computing power to introduce an advanced AI platform. The goal is to make AI the doctor's 'second pair of eyes,' enhancing diagnostic accuracy and medical quality. In the future, they will continue to collaborate with international AI technology partners, deepen cross-domain integration, strengthen smart healthcare applications, and improve patient safety and medical efficiency.
KMUH Vice Superintendent Lu Hung-hsing said the key to the system gaining attention on the international stage is the medical team's dedication to massive, high-quality endoscopic image annotation. This provides the necessary foundation for AI model training, ensuring the diagnostic assistance closely aligns with clinical practice. (Editor: Chen Jen-hua) 1150424
Stand with the facts; every sponsorship from you is the power to protect press freedom.
Download the CNA 'First-hand News' APP to grasp the latest news instantly.
The text, images, and audio/video on this website are not to be reproduced, publicly broadcast, or publicly transmitted and utilized without authorization.
(CNA Reporter Lin Chiao-lien, Kaohsiung 24th) In response to the demand for colorectal cancer screening and the development of medical AI, Kaohsiung Medical University Hospital (KMUH) has collaborated with Foxconn Technology Group to announce the world's first 'Colonoscopy AI Agent' powered by the NVIDIA IGX Thor platform, serving as a real-time clinical decision support tool.
KMUH issued a press release today stating that according to 2022 data from the Ministry of Health and Welfare, one person is diagnosed with cancer every 4 minutes, with the number of confirmed colorectal cancer cases remaining high. The gold standard for colorectal cancer is a colonoscopy (full colon endoscopy), where the doctor directly observes the lesions and performs biopsies for pathological examination. The key to diagnosis highly depends on the accuracy of the endoscopy.
KMUH stated that in order to make Artificial Intelligence a real-time assistant for clinical decisions, KMUH collaborated with Foxconn to build an advanced smart endoscope solution. The 'Colonoscopy AI Agent' has already debuted at NVIDIA GTC 2026.
KMUH pointed out that the 'Colonoscopy AI Agent' has three core functions: real-time lesion detection and pathological classification, which completes analysis in milliseconds; simultaneous monitoring of bowel cleanliness and withdrawal speed to improve inspection quality consistency; and 3D intraoperative navigation capabilities, which instantly mark the location and size of polyps to help doctors position accurately.
The hospital stated that the system also introduces AI agent technology, automatically generating inspection reports through the NVIDIA Nemotron model, and combining with AI chatbots to optimize clinical workflows. This helps doctors quickly integrate information, reduce administrative burdens, and allow medical personnel to focus on patient care and diagnostic judgment.
KMUH Superintendent Wang Chao-yuan stated that the hospital combines clinical data with independent computing power to introduce an advanced AI platform. The goal is to make AI the doctor's 'second pair of eyes,' enhancing diagnostic accuracy and medical quality. In the future, they will continue to collaborate with international AI technology partners, deepen cross-domain integration, strengthen smart healthcare applications, and improve patient safety and medical efficiency.
KMUH Vice Superintendent Lu Hung-hsing said the key to the system gaining attention on the international stage is the medical team's dedication to massive, high-quality endoscopic image annotation. This provides the necessary foundation for AI model training, ensuring the diagnostic assistance closely aligns with clinical practice. (Editor: Chen Jen-hua) 1150424
Stand with the facts; every sponsorship from you is the power to protect press freedom.
Download the CNA 'First-hand News' APP to grasp the latest news instantly.
The text, images, and audio/video on this website are not to be reproduced, publicly broadcast, or publicly transmitted and utilized without authorization.