AI helped guide a brain tumour operation for the first time on a real patient
London surgeons used a real-time AI system to highlight nerves and blood vessels during a delicate pituitary operation, and the patient kept his sight.
What happened: In May, surgeons at the National Hospital for Neurology and Neurosurgery in London removed an 11mm tumour from the pituitary gland of 48-year-old Rhys Hibbert. For the first time on an actual patient, they used an AI system that analysed live camera footage during the operation, colour-coding nearby nerves and blood vessels so the team could avoid them. Hibbert kept his sight, walked unaided within a week, and has returned to work. The case was kept quiet until his recovery was confirmed.
Why it matters: The pituitary gland sits packed tightly against blood vessels and the nerves that control vision, so a millimetre of error can cause blindness, stroke or death. Hibbert's tumour had already worsened his eyesight and hormone levels, and doctors say he faced going blind without surgery. Having a system trained on hundreds of past surgical videos flag critical anatomy in real time offers a second layer of caution in a procedure where surgeons normally rely purely on their own experience and eyesight.
How it works, plainly: The AI doesn't touch instruments or make decisions. It watches the same camera feed the surgeon sees and overlays colour markers on structures like nerves, vessels and tissue boundaries, based on patterns learned from many recorded operations. The surgical team stayed fully in control throughout, using the highlighted view as an extra reference point rather than a replacement for judgment. The hospital had tested the tool as a research aid before this was cleared for use on a real patient.
The rollout: This was one operation, run as part of a clinical trial funded by the National Institute for Health and Care Research, at a hospital with a long history in neurosurgery. Researchers, including the system's technical lead at UCL, describe it as a step toward wider support tools for delicate surgery, but a single successful case doesn't establish how reliably the system performs across different tumours, patients or surgical teams. Broader use will depend on further trials, oversight and validation.