Researchers at Johns Hopkins University have developed a robotic system that autonomously performed a major component of a gallbladder removal procedure without human intervention. The procedure was conducted on a lifelike anatomical model and marks the first time a surgical robot has carried out such a complex task with full autonomy.
The system, named Surgical Robot Transformer-Hierarchy (SRT-H), was trained using video data of human-performed surgeries. Unlike previous robotic systems, which relied on tightly controlled environments and pre-programmed instructions, SRT-H operated using real-time adaptability. It responded to spoken instructions and corrections during the procedure, akin to a surgical trainee working under the supervision of experienced clinicians.
The findings, published in Science Robotics, highlight a shift in the capabilities of autonomous surgical systems. Previous work by the same team in 2022 demonstrated autonomous surgery on live animal tissue using a robot that followed rigid plans. In contrast, SRT-H performed 17 sequential tasks necessary for gallbladder removal, including identifying anatomical structures, placing surgical clips, and making incisions, all without manual input during the procedure.
The robot’s actions were based on a combination of video analysis and text annotations that allowed it to generalize its learning across variable conditions. It adapted to changes such as altered camera angles and visual obstructions designed to simulate bleeding, completing the tasks with outcomes comparable to those of expert human surgeons, though at a slower pace.
The work was led by Johns Hopkins medical roboticist Axel Krieger and included collaboration with researchers now based at Stanford University and industry partners at Optosurgical. According to the team, the project demonstrates the viability of modular, learning-based surgical automation. Future plans include expanding the robot’s training to additional surgical procedures and pursuing the long-term goal of full-operation autonomy. The research was federally funded.
Photo credit: Juo-Tung Chen/Johns Hopkins University
