Artificial Intelligence Could Revolutionize Brain Surgery, Says Leading Neurosurgeon

Brain surgery using artificial intelligence (AI) could become a reality within the next two years, leading to safer and more effective procedures, according to a prominent neurosurgeon. Trainee surgeons are currently utilizing AI technology to learn to perform more precise keyhole brain surgery. Developed at University College London (UCL), the AI system can identify small tumors and critical structures, such as blood vessels, located in the center of the brain.

Brain surgery requires extreme precision, as even a slight error could have fatal consequences. It is crucial to avoid damaging the pituitary gland, which is the size of a grape and controls the body’s hormones. Consultant neurosurgeon Hani Marcus of the National Hospital for Neurology and Neurosurgery explains that finding the right balance is crucial: “If you go too small with your approach, then you risk not removing enough of the tumor. If you go too large, you risk damaging these really critical structures.”

The AI system has analyzed over 200 videos of pituitary surgery and, in just 10 months, has reached a level of experience that would take a human surgeon 10 years to gain. The system’s capabilities offer significant advantages to surgeons. Mr. Marcus states, “Surgeons like myself – even if you’re very experienced – can, with the help of AI, do a better job to find that boundary than without it. You could, in a few years, have an AI system that has seen more operations than any human has ever or could ever see.”

The potential of AI in healthcare is recognized by the UK government, which considers it a “game-changer.” AI could enhance outcomes and productivity in healthcare, according to AI government minister Viscount Camrose. The collaboration between UCL, engineers, clinicians, and scientists at the Wellcome / Engineering and Physical Sciences Research Council (EPSRC) Centre for Interventional and Surgical Sciences aims to revolutionize healthcare in the UK with the help of AI.

– University College London