At the National Hospital for Neurology and Neurosurgery at University College London (UCL), the world's first brain tumor removal operation took place, during which surgeons used an artificial intelligence system to analyze what was happening in real time. This procedure marks a new stage in the development of computer-assisted surgery, where AI acts not as an autonomous executor, but as an intelligent assistant capable of preventing critical errors.
Background: when vision is at stake
The patient was 48-year-old Rhys Hibbert, a father of two from Bedfordshire. A year and a half ago, he began experiencing sudden bouts of dizziness and severe fatigue, and then his peripheral vision began to noticeably narrow. An MRI revealed a benign pituitary tumor measuring 11 millimeters.
Although the growth was not malignant, its location posed a serious threat. The pituitary gland sits at the base of the skull in close proximity to the carotid arteries and optic nerves. Any careless movement by the surgeon could lead to irreversible vision loss or serious vascular damage. According to doctors' estimates, the risk of incomplete tumor removal ranged from 25% to 50%, and the risk of damaging a major vessel ranged from 0,5% to 2%.
How the surgeon's "second pair of eyes" works
The key element of the intervention was an innovative AI system developed by UCL researchers (including the Hawkes Institute) with financial support from the National Institute for Health and Care Research (NIHR) and Google.
During the procedure, an endoscope was guided through the patient's nose to the base of the skull. In parallel with the main monitor, on which surgeons saw the operative field, a second screen with the AI interface worked. The system analyzed the video stream in real time, recognizing surgical instruments and overlaying color coding on critical anatomical structures hidden beneath layers of bone and membranes.
The algorithm was pre-trained on hundreds of video recordings of previous pituitary tumor removal operations. According to Professor Hani Marcus, one of the leading neurosurgeons involved in the procedure, this system is "trained on more operations than most surgeons will see or perform in their entire career."
Technology in service of humans, not instead of them
Developers and medical staff emphasize a crucial nuance: artificial intelligence does not make decisions or control instruments. Full control over the course of the operation remains with the human. AI only provides recommendations, working like an advanced navigation system or, as Professor Marcus figuratively put it, like a "surgical ChatGPT" that can be heeded or ignored in favor of the doctor's own professional judgment.
"Inside our head there are no road signs," noted Rhys Hibbert himself. "From a patient safety perspective, I clearly see the benefits of this technology."
Results and prospects
The operation, performed in May of this year, was a complete success. The surgeons managed to safely remove the maximum amount of tumor without touching any nerves or vessels.
A few weeks after the intervention, the patient reported significant improvement: "When I opened my eyes after surgery, it felt like I had regained full, 360-degree vision – I hadn't seen this well for a whole year." By now, Rhys Hibbert has fully recovered, got rid of the associated symptoms, and returned to work.
The success of this intervention opens the way for broader clinical application of the technology. The UCL research team is already planning large-scale trials to prove the effectiveness and safety of the method on a larger patient sample. If the trend continues, real-time AI navigation systems could become the gold standard in complex neurosurgical interventions worldwide, making them predictable and maximally safe for patients.
