
Medical Imaging Product Developer
In a packed conference hall in Harbin, China, surgeons watched a large screen display real-time angiography as an AI-powered robot performed vascular intervention surgery on a pig — without human hands guiding the instruments.
The scene at the 21st Cold Region (Ice City) Conference of Cardiology in June 2026 marked a turning point. For the first time anywhere in the world, an AI embodied intelligence surgical robot completed a full vascular intervention procedure on a large animal, moving from mechanical assistance to autonomous execution.
The achievement belongs to Shenzhen Raysightmed Co, Ltd., known as RaysightMed, working alongside the Second Affiliated Hospital of Harbin Medical University. The collaboration represents what Chinese medical device makers have been chasing: a leap from following Western innovation to defining the frontier themselves.
From Assistance to Autonomy
Traditional surgical robots do what they're told. A surgeon manipulates controls; the machine translates those movements into precise instrument actions. The robot assists, but the human decides.
RaysightMed's system works differently. Operators issue commands through voice or text. The robot then identifies vascular anatomy on its own, plans the optimal intervention path, delivers a guidewire to the target vessel, and performs stent delivery and balloon dilation — all without point-by-point human control.
The technology combines three capabilities: large language model semantic understanding, real-time multi-modal image recognition, and precision motion control. The result is a system that hears instructions, understands intent, plans autonomously, and executes precisely.
The Clinical Backing
The animal experiment was led by Professor Yu Bo, dean of the cardiovascular hospital at the Second Affiliated Hospital of Harbin Medical University, and Professor Hou Jingbo from the cardiology department.
"The deep integration of AI and surgical robots is the inevitable direction for cardiovascular intervention," Professor Yu said at the launch ceremony on June 6, 2026. He framed the technology as a path to narrowing regional healthcare gaps and standardizing surgical quality across China.
Professor Hou pointed to a persistent bottleneck: traditional vascular intervention surgery depends heavily on operator experience and hand-eye coordination. Training takes years. Scaling to grassroots hospitals remains difficult.
An AI embodied intelligence surgical robot, she said, could reconstruct the operational logic of intervention surgery from the ground up.
The Results
In the experiment, pigs were randomly assigned to either a manual surgery group or an AI robot group. The AI group outperformed on target task completion rate, radiation dose, and fluoroscopy time.
Algorithm iteration speed exceeded expectations. The system adapted to various levels of vascular anatomical complexity and completed procedures stably.
Professor Yin Li presented the findings at the conference. The robot, he noted, reduces radiation exposure and lead apron burden for operators. Future versions could integrate anatomical, functional, and microcirculation data across the entire care cycle — from preoperative diagnosis through intraoperative procedures to postoperative follow-up.
What Comes Next
The animal experiment is a milestone, not a finish line. Professors Hou Jingbo and He Yuquan, both commenting on the results, agreed on the next step: multi-center clinical data.
Real-world validation of safety and effectiveness must come before large-scale clinical application. The conference hall was full. Many attendees arrived early to secure seats. The interest reflects where the industry is looking.
RaysightMed plans to accelerate clinical registration and deployment with top-tier medical institutions. The goal is to bring the technology to patients — first in China, then globally.
The question now is whether autonomous vascular intervention can cross the same regulatory and clinical trust barriers that earlier surgical robots faced. The pig in Harbin suggests it can. The patients will decide.