The gall! Johns Hopkins AI robot performs gallbladder surgery with 100% accuracy

Johns Hopkins University researchers have developed an AI-powered surgical robot that successfully performed gallbladder removal surgery on pig organs with 100% accuracy. The system, called SRT-H (Surgical Robot Transformer), uses ChatGPT-like transformer models to control a standard DaVinci robot, marking a significant advance from pre-programmed surgical automation to AI that can learn from demonstrations and adapt to real-time conditions.

How it works: The SRT-H system employs two transformer models working together to perform complex surgical procedures.

• A high-level policy module handles task planning and ensures the procedure progresses smoothly, while a low-level module translates those instructions into specific movements for the robotic arms.
• The AI was trained on over 17 hours of video footage from DaVinci endoscopes and cameras, combined with kinematics data (precise arm movement information) and natural language annotations.
• Unlike previous systems that required specially marked tissues and predetermined plans, SRT-H can accept real-time feedback in natural language, such as “move your arm a bit to the left” or “put the clip a bit higher.”

The training process: Researchers used an imitation learning approach similar to mentoring a novice human doctor.

• The team broke down cholecystectomy (gallbladder removal) into 17 distinct steps, a procedure performed roughly 700,000 times annually in US hospitals.
• A trained research assistant repeatedly performed the surgery on porcine gallbladder and liver samples while the AI observed and learned from the demonstrations.
• The robot proved robust against anatomical differences between samples, tissue obstructions, and imperfect imagery while maintaining precision equal to expert human surgeons, though operating somewhat slower.

What they’re saying: Lead researcher Ji Woong Kim emphasized the system’s flexibility compared to previous approaches.

• “Our current work is much more flexible. It is an AI that learns from demonstrations,” Kim explained, contrasting it with earlier robots that “worked like in these Kuka robotic arms, welding cars on factory floors.”
• “You can take any kind of surgery, not just this one, train the robot in the same way, and it will be able to perform that surgery,” Kim noted about the system’s potential applications.

The corporate roadblock: Progress toward human trials faces a significant data access challenge with Intuitive Surgical, maker of the DaVinci robot.

• While Intuitive Surgical releases video feed data from DaVinci robots, it refuses to provide kinematics data that Kim says is essential for training the algorithms.
• “I’ve been begging them to give us the data. They did not agree,” Kim revealed, citing the company’s concerns about competitors reverse-engineering their robot mechanics.
• Kim attributed the resistance to corporate conservatism: “It’s really the upper management who is not up to speed with AI. They don’t realize the potential of these things.”

Looking ahead: Researchers are exploring alternative approaches to overcome data limitations and expand surgical AI capabilities.

• Kim’s team plans to attach motion-tracking sensors to manual surgical tools to capture kinematics data directly from expert human surgeons’ movements.
• Future applications could involve humanoid robots in operating rooms, with Kim currently working on a general-purpose humanoid robotics project at Stanford University.
• The breakthrough represents a crucial step toward fully autonomous surgical systems that could eventually operate on live patients.