During the operation, surgeons inserted an endoscope through Hibbert’s nose to reach the tumor. A separate screen showed output from the AI system, which analyzed the live surgical video, tracked instruments and identified where critical structures were likely to be hidden beneath tissue. The goal was to help surgeons judge where tumor removal could be performed more safely. The anatomy in that part of the brain varies between patients, and the tumor itself can sit close to vessels supplying blood to the brain and nerves responsible for vision.
“There are no road signs inside our head,” Hibbert said. “From a patient safety perspective, I can see absolutely the benefit.”
The system was developed by researchers at UCL and trained and evaluated using hundreds of videos from previous pituitary tumor surgeries. Researchers manually marked important vessels and nerves in those recordings so the model could learn to recognize anatomical structures during a live procedure.
Prof Hani Marcus, one of the neurosurgeons involved in the operation, said the volume of training material gives the AI exposure to far more examples than an individual surgeon would typically encounter. “It is trained on more operations than most surgeons see or do in a lifetime and can act like an expert second pair of eyes,” Marcus said.
The system is designed as an assistive tool rather than an autonomous surgeon. Doctors remain responsible for every decision during the operation and can choose whether to follow the AI’s guidance. Marcus said the longer-term aim is to create something resembling a surgical version of ChatGPT, giving clinicians access to another source of guidance during complex procedures.
The surgery came after Hibbert experienced worsening symptoms following the discovery of an 11mm non-cancerous tumor. He had initially collapsed and suffered a seizure during a lunchtime walk. Over time, he also developed severe fatigue, dizziness and problems with his vision.
The operation carried meaningful risks. Surgeons estimated there was a 25% to 50% chance that the entire tumor could not be removed, along with a 0.5% to 2% risk of damaging a major blood vessel. Hibbert agreed to participate when offered the opportunity to undergo the procedure with the experimental AI system. “If patients are not prepared to join research how can doctors ever learn and how can medicine ever progress?” he said.
Eight weeks after surgery, Hibbert said his eyesight and energy had improved and the symptoms he experienced before the procedure had disappeared. “When I opened my eyes after the operation it felt like I had got 360-degree vision - I hadn't seen that clearly for a year,” he said.
The project was funded by the National Institute for Health and Care Research and Google. Researchers spent years testing the technology in laboratory settings before using it during a live operation and are now preparing a larger trial.
James Frith, the U.K.’s health innovation minister, said the procedure demonstrated the potential benefits of AI in medicine while emphasizing the need for safeguards. “AI needs proper safeguards and we will always ensure that safety is taken seriously,” Frith said. “But we will also ensure we benefit from the opportunities it brings.”
This analysis is based on reporting from BBC.
Image courtesy of University College London Hospitals NHS Foundation Trust.
This article was generated with AI assistance and reviewed for accuracy and quality.
