AI Guides World-First Brain Surgery, Saves Sight

Surgeons in London removed a brain tumour from a 48-year-old man while an artificial intelligence system read the live camera feed and told them where not to cut — the world's first AI-assisted brain surgery on a patient, doctors said Thursday.

The patient, Rhys Hibbert, a customer services manager and father of two from Bedfordshire, was facing blindness. An 11mm non-cancerous tumour on his pituitary gland was pressing on the optic nerves.

The operation was performed in May at the National Hospital for Neurology and Neurosurgery, part of University College London Hospitals, and kept under wraps until Thursday while Hibbert recovered, The Guardian reported.

How It Worked

Surgeons fed an endoscope — a thin camera on a stick — through Hibbert's nose to the base of his skull, where the tumour sat hidden behind bone and membrane.

On a second screen, the AI analysed the live video, tracked the surgical instruments and colour-coded the vessels and nerves most likely to be in the way, marking the safest places to cut. The BBC compared it to the facial recognition on a phone, except it recognises hidden anatomy.

The pituitary gland sits beside the carotid arteries and the optic nerves. Health officials said going a millimetre wrong can mean blindness, stroke or death. According to the BBC, the procedure carries a 25 to 50 percent chance of not getting the whole tumour out and a 0.5 to 2 percent chance of injuring a major blood vessel.

The system was trained on hundreds of videos of pituitary tumour removals, with the vessels and nerves drawn in by hand on each one. The average UK surgeon performs about 10 to 20 of these operations a year.