UCL's AI-assisted brain tumor surgery achieves 98% removal, 30% blood flow cut

University College London (UCL) neurosurgery team performed the United Kingdom’s first AI‑assisted brain operation this month at the National Hospital for Neurology and Neurosurgery on a 48‑year‑old patient who was diagnosed with a brain tumour in 2024. The procedure relied entirely on a deep‑learning image‑analysis system developed by UCL, which helped the surgeons locate the tumour margins in real time and plan the resection pathway. Post‑operative assessment showed a tumour removal rate of 98%, and intra‑operative blood loss was 30% lower than in conventional surgery. The findings were published the same day in *Lancet Neurology*, marking a milestone for artificial intelligence in high‑risk neurosurgical practice.

AI‑assisted surgery has been trialled for years in minimally invasive laparoscopy, plastic surgery and orthopaedics, but the brain’s complex anatomy and the need for instantaneous intra‑operative decisions have kept widespread clinical adoption elusive. UCL’s brain‑imaging platform combines three‑dimensional reconstruction with live video streams and employs a convolutional neural network trained on thousands of historic surgeries to predict the relative positions of tumours and functional areas within milliseconds. In animal‑model testing last year, the system demonstrated a reduction in inadvertent cuts, and after receiving “clinical trial approval” from the UK Medicines and Healthcare products Regulatory Agency (MHRA) earlier this year, it entered human trials.

Founded in 1826, UCL is England’s first fully secular university that does not restrict admission on religious grounds and has long been renowned for interdisciplinary research. Its medical and engineering schools collaborate on a “Medical Artificial Intelligence” programme that focuses on imaging diagnostics, surgical navigation and personalised medicine. In recent years, UCL has commercialised several AI algorithms through technology‑transfer partnerships, publishing more than a thousand related papers in global medical journals and cementing its leadership in medical AI research and development.

Following the operation, the lead surgeon reported that the AI system’s real‑time visual information allowed more precise risk assessment at each step of tumour removal, reducing the chance of accidental damage to eloquent brain regions. The patient’s family expressed gratitude for the outcome, noting that the new technology heightened their hope for recovery. Nevertheless, medical ethicists cautioned that AI remains an assistive tool; ultimate clinical decisions must stay with physicians, and transparent data‑governance and accountability frameworks are needed to prevent misuse or errors that could lead to medical accidents. Regulatory bodies are currently drafting safety standards and post‑market surveillance procedures for AI‑assisted surgical systems.

The case demonstrates the feasibility of AI in high‑precision surgery and is expected to drive a transformation of operative approaches for complex conditions such as brain tumours and vascular malformations, while accelerating commercialisation of related medical devices and software. Academia is evaluating the extension of such technologies to remote surgery and multinational collaboration platforms to address neurosurgical workforce shortages in underserved regions.

For Taiwan, the successful demonstration of AI‑assisted neurosurgery offers a concrete reference for the local medical‑technology sector. Taiwan’s global competitiveness in semiconductor and medical‑imaging chip manufacturing positions it well to partner with British academic institutions to develop surgical navigation systems that comply with domestic regulations. Such collaboration could improve the safety and efficiency of advanced neurosurgical procedures in Taiwan, support the refinement of medical‑AI regulatory frameworks, and broaden the application space for future smart‑healthcare initiatives.