Surgical robot performs gallbladder surgery independently
A surgical robot autonomously removes a gall bladder. It learned the procedure from surgical videos.
(Image: XinHao Chen/Johns Hopkins University)
Researchers at Johns Hopkins University have had a surgical robot remove a gall bladder from a dead pig largely autonomously. The robot had previously learned the procedure from videos. During the operation, the robot also received instructions by voice command from experienced surgeons.
Axel Krieger, lead researcher of the surgical robot project, performed the first autonomous operation on a living animal using the Smart Tissue Autonomous Robot (STAR) back in 2022. The surgical robot performed a laparoscopy on the pig. However, the robot still required some assistance. For example, the operation was performed in a controlled environment. The operation also followed a rigid, fixed surgical plan that could not be deviated from. In addition, the tissue was marked beforehand to give the robot an indication of where to start.
Empfohlener redaktioneller Inhalt
Mit Ihrer Zustimmung wird hier ein externes YouTube-Video (Google Ireland Limited) geladen.
Ich bin damit einverstanden, dass mir externe Inhalte angezeigt werden. Damit können personenbezogene Daten an Drittplattformen (Google Ireland Limited) übermittelt werden. Mehr dazu in unserer Datenschutzerklärung.
The new surgical robot programmed by Krieger and other researchers with the Surgical Robot Transformer-Hierarchy (SRT-H) framework no longer requires this type of support. The scientists trained the system with artificial intelligence (AI) for gallbladder surgery using videos. These 17-hour videos show how researchers at Johns Hopkins University remove the gallbladder in animal cadavers, scientists write in the study “SRT-H: A hierarchical framework for autonomous surgery via language-conditioned imitation learning,” which has been published in Science Robotics. The videos were also provided with descriptive subtitles to make it easier for the system to learn the individual steps of the operation.
Self-correcting system
The removal of the gallbladder involves a total of 17 tasks. These include identifying specific ducts and arteries, precise grasping and strategic placement of staples, and cutting with scissors. The robot learned how to perform the individual steps from the videos by imitation and was then able to carry them out autonomously. The robot turned out to be adaptable and capable of understanding. It adapts to anatomical features in real time. The robot can also correct itself if something does not go quite right. It can also respond to calls from an experienced surgeon, as the system is based on the same machine learning architecture as ChatGPT. It can therefore implement spoken corrections and learn from them.
The robot took longer to perform the gallbladder operation than an experienced human surgeon, but the results are said to have been comparable and 100 percent correct.
Videos by heise
“For me, this really shows that it is possible to perform complex surgical procedures autonomously,” says Krieger. “This is a proof of concept that it is possible and that this imitation learning framework can automate such complex procedures with such a high degree of robustness.”
The research team now wants to teach the surgical robot further operations and test them. The researchers want to expand and improve the system's capabilities to be able to perform fully autonomous operations with it.
(olb)
