Trending
Feature - -
Rhys Hibbert, right (with neurosurgeon Hani Marcus, left) was the first patient to have a brain tumour successfully removed in an AI-assisted operation. He was operated on at the National Hospital for Neurology and Neurosurgery in London. | Photo: UCLH/PA
By Beyond Headlines Desk
A surgical team at University College London Hospital prepared to cut into the base of a man's skull inside an operating theatre. The target was an 11-millimetre tumour lodged on the pituitary gland of Rhys Hibbert, a 48-year-old customer services manager and father of two.
The odds were not simple. In conventional neurosurgery like this, there is a 25 to 50 per cent chance surgeons do not get the whole tumour out, and a 0.5 to 2 per cent risk of puncturing a major blood vessel. This mistake could trigger a stroke, permanent disability, or death.
Before the operation, Hibbert was living with debilitating dizziness, sudden seizures, and a peripheral vision that was fading fast. Still, he was not going to let the tumour stop him from living his life. So, he let the experts take their chance.
What made this operation different was not just the skill of the surgical team; it was the silent co-pilot working alongside them.
An artificial intelligence system was processing the surgical field in real time, frame by frame, via a microscopic camera mounted on an endoscope threaded through Hibbert's nasal passage.
As the surgeon navigated the dense tissue, the AI picked out hidden anatomical structures in real time, drawing visual safety boundaries over nerves and blood vessels buried beneath the surface.
As a result, Hibbert got his sight and energy back, fully.
The operation, carried out in May, is being called the world's first live AI-assisted brain tumour removal, with clinicians. Researchers and global health experts are calling it a fundamental shift in how healthcare gets done.
Over the past few decades, the operating room has modernised through laparoscopic tools and robotic platforms. These gave surgeons sharper dexterity, steadier hands and high-definition 3D vision, but they were still, fundamentally, mechanical extensions of the human hand.
The core cognitive work – reading distorted tissue, navigating anatomical surprises, predicting fluid movement under pressure – stayed entirely on the surgeon.
Pre-operative scans like MRIs and CT scans map the terrain before surgery begins. But the moment an incision is made, soft tissue shifts, fluid changes the picture, and those static scans start losing real-time accuracy.
Live surgical AI changes that. It brings real-time computer vision into the operation itself.
Trained on thousands of hours of annotated surgical footage, the neural network reads the live video feed, tracks instrument movement, and differentiates malignant tissue from healthy tissue as the surgery unfolds.
Professor Hani Marcus, the lead neurosurgeon on Hibbert's case, says the system has been trained on more operations than most surgeons will see in a lifetime, and it acts like an expert second pair of eyes.
Professor Hani Marcus, the lead neurosurgeon on Hibbert's case, says the system has been trained on more operations than most surgeons will see in a lifetime, and it acts like an expert second pair of eyes.
It does not cut or take control; it stays firmly in an advisory role. Marcus sees it eventually growing into a continuous layer of clinical intelligence running through complex procedures.
In medical history, "revolution" is not a word to throw around lightly. It's earned only when something fundamentally changes clinical outcomes, safety standards, and the baseline of care worldwide. AI-assisted surgery clears that bar on three fronts.
First, it addresses human visual limits. Micro-neurosurgery, structural cardiology and complex oncology all operate on margins measured in fractions of a millimetre. AI surfaces hidden structures by processing diagnostic data faster than the eye can process them, turning guesswork into verified navigation.
Second, it levels the playing field. Surgical skill has always varied by years of experience and the institution behind a surgeon. AI narrows that surgical judgment by distilling global surgical data into real-time guidance.
Third, it keeps humans in charge while expanding capability. Unlike autonomous systems used in industry, surgical AI runs on a strict "human-in-the-loop" model. The surgeon stays the absolute authority, combining empathy and intuition with the machine's memory and precision.
The implications of this shift stretch well beyond any single operating theatre.
In structural cardiology, AI paired with intracardiac echocardiography is letting doctors perform complex heart valve repairs through catheters, avoiding open-chest surgery.
In oncology, real-time tissue classification helps surgeons keep margins clean while preserving vital organs.
Moreover, for developing healthcare systems facing shortages of specialised surgeons, AI offers a path to bridge the gap between rural clinics and major academic centres.
Fewer complications mean shorter hospital stays, fewer readmissions, and lower operational costs for public health systems already stretched thin.
However, as this technology moves from clinical trials into everyday use, experts are cautious.
Regulatory oversight needs to keep pace, and questions around algorithmic bias, data privacy, and model performance across different patient groups remain critical areas of scrutiny.
James Frith, the UK's health innovation minister, welcomed Hibbert's "life-changing surgery" as part of government-funded research. "AI needs proper safeguards, and we will always ensure that safety is taken seriously," he said. "But we will also make sure we benefit from the opportunities it brings."
Surgical AI is not like ordinary consumer software, it demands constant validation to catch edge-case errors, like misinterpreting dynamic fluid reflections, surgical smoke, or unusual anatomical anomalies for healthy tissue or a safe boundary.
Furthermore, medical ethicists highlight the risk of "Automation Bias". A cognitive trap where clinicians might subconsciously defer to machine recommendations over their own hard-won instincts.
There’s also a subtler risk that medical ethicists highlight as “automation bias” -- the tendency for clinicians to quietly start trusting the machine's read over their own hard‑earned instincts.
To prevent over-reliance, emerging clinical protocols mandate that surgical teams has to go through simulated training to handle unexpected situations, system outages, or algorithmic discrepancies mid-operation.
For all the complex code, neural networks, and clinical metrics, the real test of any medical revolution is what it does for a human life.
Two months after his surgery, Rhys Hibbert is back home, resuming the life he feared he might lose.
"I have my energy back, and all of the side effects I was suffering before surgery have gone," he said. "It's given me my life back."
As AI co-pilots take their place in operating rooms around the world, medicine is crossing a threshold -- one where the room for human error keeps shrinking, and the list of what's curable keeps growing.
Walk into hospitals across Bangladesh’s Khulna division, and you will find patients lying on the floor, sitting on mats, and waiting in crowded corridors. Their only hope is to find a bed and get treatment.
- Feature
America’s streets are being taken over by self-driving robotaxis operated by Waymo, which claims they are safer than human drivers. But what does “safe” actually mean to them?
- Feature
What began as a wedding celebration in southern Iran ended in devastation after a reported US strike hit a residential area in Kuhestak, in Sirik County, killing civilians and injuring dozens as fighting between Washington and Tehran escalated once again.
- Feature
