AI-Guided Robot Performs Gallbladder Surgery Autonomously

On July 9th, researchers announced the development of an experimental AI-guided robot capable of autonomously performing a complex phase of a common gallbladder operation. This marks a significant advancement towards automated medical procedures.

Unlike existing surgical robots that are remotely controlled by surgeons, the new system utilizes artificial intelligence to make independent decisions and adapt to unforeseen complications during operations. The lead researcher, Axel Krieger from Johns Hopkins University, compared its functionality to an autonomous vehicle, capable of navigating various conditions and responding intelligently to encountered situations. This advancement signifies a shift from robots executing specific tasks to those that truly comprehend surgical procedures.

The SRT-H robot was trained using an AI framework called language-guided imitation learning. This involved analyzing videos of surgeons performing gallbladder removal surgeries on pig cadavers. The robot was then tested on eight different sets of pig gallbladders and livers that had been removed from the animals.

The task of separating the gallbladder from the liver, which takes several minutes, involves a variety of actions, including grabbing, clipping, and cutting. These skills are common in real surgical procedures and require both decision-making and adaptation. The pig organs and blood vessels used in the tests varied significantly in appearance and anatomy, mirroring the diversity encountered in human surgeries. While the robot achieved 100% accuracy in the surgeries, it took longer to complete the work than a human surgeon.

The development of autonomous surgical robots holds the potential to address surgeon shortages, reduce human error, and provide consistent, high-quality care in underserved regions.