AI adds a visual layer to brain surgery
A UCLH team used live AI to highlight critical anatomy during a pituitary-tumour operation; the surgeon stayed in control and the evidence is still a first-patient milestone.
The 60-second version
A live AI system highlighted critical anatomy during a pituitary-tumour operation while the neurosurgeon retained control.
Key points
- The system analysed camera footage rather than making autonomous surgical decisions.
- The first patient reported restored vision and rapid functional recovery.
- Broader trials must test errors, complications and long-term outcomes.
Verdict. A meaningful feasibility milestone, not proof that AI is safer than surgeons.
A narrow clinical roleWhat the first case does — and does not — show
A London neurosurgical team has reported the first live use of an AI system to support a brain-tumour operation. The system analysed the camera feed during surgery and highlighted critical anatomy near an 11-millimetre pituitary tumour, while the surgeon retained full control.
The operation took place at the National Hospital for Neurology and Neurosurgery, part of UCLH, in a clinical trial funded by NIHR and supported by research partners. The patient, 48-year-old Rhys Hibbert, had worsening vision and hormone problems and volunteered to be the first participant.
The location is unusually unforgiving. The pituitary gland, blood vessels and nerves involved in vision are packed into a small space. A difference of a millimetre can change the risk profile. The AI did not cut tissue or make an autonomous decision; it provided colour-coded, real-time assistance intended to help the team recognise structures to avoid.
Researchers trained the system on hundreds of annotated endoscopic pituitary-surgery videos. It had previously been used as a training tool, but this was the first reported live patient use. The platform ran on NVIDIA Clara IGX hardware and was developed by a multidisciplinary UCL team.
Hibbert reported that his sight improved when he woke and that he was walking without glasses or sticks within a week. Those are the outcome of one patient, not a comparative trial showing that AI is safer than experienced surgeons.
The story matters because it demonstrates a narrow, practical role for medical AI: watching a live feed and surfacing anatomy at the moment it matters. It also shows why safeguards matter. The evidence still comes from a first patient and a research setting, so future participants and longer follow-up will determine whether the benefit generalises.
The bottom line is not that AI has replaced the neurosurgeon. It is that a trained system can act as an additional visual layer while responsibility stays with the clinical team.
| AI does | Analyses live video and highlights critical anatomy |
|---|---|
| AI does not | Control instruments or replace surgical judgement |
| Next evidence | Comparative trials and longer follow-up |
The credible promise is an extra set of eyes, not a robot surgeon.
Primary sourcesUCLH·The Guardian·BBC News