By Silas Gemma ’26
Science and Environment Editor
In an increasingly automated world, Artificial Intelligence is becoming commonplace in almost every area of our lives.
According to the Pew Research Center, many Americans have misgivings about its effects on privacy, customer service and public safety. Forbes lists job loss, reduced human interaction, misinformation and an uncertain future for humanity as primary concerns of AI’s rise. However, AI also illustrates great promise for the future of industries like medicine.
BBC News and University College London have reported that scientists in the United Kingdom are training AI to visualize and navigate neural structures in the hopes that it will aid neurosurgeons in successfully performing operations prone to human error.
One article by the National Health Service of the United Kingdom explained that scientists at University College London are currently developing an AI program to permit safer brain surgery by allowing more precise movement and better visualization of neurovasculature and subcortical structures, or areas beneath the cortical surface.
In other words, it can provide an interactive image of the network of blood vessels and arteries within the brain, as well as deeper structures that are more difficult to navigate during surgery. More specifically, the program could be specialized to aid in removing tumors from central structures of the brain, such as the pituitary gland.
The NHS explained that navigating the complex labyrinth of blood vessels near the pituitary gland, which regulates hormone function, is painstaking work — even the slightest error could inflict irreversible damage. This technology would more clearly identify the boundaries of a tumor to ensure its complete removal while also keeping healthy structures intact.
In a video by NowThis News, Dr. Hani Marcus, one of the foremost neurosurgeons developing this technology, explained the process of preparing the AI.
“What this [technology] practically entails is us training the AI with hundreds of videos telling it what structures are what,” Marcus said.
This extensive experience then allows the technology to support training neurosurgeons, as well as provide more confidence to practiced neurosurgeons while performing risky surgeries.
Dr. Marcus compared the technology to having an experienced surgeon beside you, offering guidance when needed. “[The AI technology] might just sit quietly, but if he sees you doing something he doesn’t quite agree with, he might gently suggest that there’s another way of doing things,” he said. This error correction exemplifies one of the many advantages of using this AI for teaching purposes.
As the NHS has pointed out, this technology is still only being used in simulation labs. However, the technology could be available to patients within two years.
As the University College London reported, PreSize Neurovascular is another AI-powered technology and concurrent neurosurgery innovation that permits safer and more accurate surgeries. Aneurysms describe swelling in the wall of a blood vessel somewhere in the body, and, according to the Brain Aneurysm Foundation, have a 50% fatality rate when they occur in the brain. Ignited by Dr. Katerina Spranger, CEO of Oxford Heartbeat — the founding company of this technology — PreSize Neurovascular focuses on brain surgeries involving blood vessels, particularly those that involve placing stents for aneurysm treatment.
Founded in 2016, this technology has already garnered millions of British pound sterlings in grants and esteem from national entities in the United Kingdom, such as the Royal Academy of Engineering. Unlike the technology headed by Marcus and his colleagues, this software is already used in clinical settings.
In a lecture at the Falling Walls Venture in 2017, Spranger gave an overview of the technology and its applications. She explained that without the technology, there is much more guesswork based on visual observation.
“There is no objective way to choose the best device for a given patient,” she explained, as everyone “has a unique structure of blood vessels.”
Additionally, the morphology of aneurysms differs for every patient, even within the same blood vessel. Furthermore, stents — small medical tubes that open up arteries and other bodily passageways — differ in shape, width and durability, making it more difficult to predict outcomes.
As Spranger pointed out, these points of uncertainty partially explain why about 20% of these stent surgeries are unsuccessful. In other words, one in five patients has to undergo subsequent surgeries to ensure that the stent is safe and effective. These increase the risk of complications and require more time, energy and financial resources. In Germany alone, these unsuccessful surgeries lead to an additional 10,000 follow-up surgeries per year.
Carly Werner of Medical News Today reported that stents are often used to prevent the rupture of brain aneurysms, which occur when blood vessels in the brain weaken and cause blood to clot in the area. While Werner pointed out that only 1% of these aneurysms ultimately rupture, stent placement is common as a preventive measure.
The stent, a small metal or plastic tube, is used to reroute blood flowing into the area, therefore reducing pressure in the area and greatly reducing the risk of rupture. The first neural stent was only approved by the Food and Drug Administration in 2011, making this a relatively novel procedure. However, it is often considered a minimally invasive operation.
Nonetheless, precision is key for a surgeon navigating the tangled mesh of blood vessels within the brain. Presize Neurovascular drastically reduces error rates and consequently increases long-term success rates. According to the Oxford Heartbeat, to use the technology, a clinician begins by importing scans of a patient’s neurovasculature into the AI program, which then generates an interactive three-dimensional model of their vascular geometry. Based on the images provided and the location selected by the clinician, the technology computes the best type of stent based on characteristics such as length and diameter.
As Spranger described in her 2017 lecture, this software uses mathematical functions to visualize and predict outcomes of different stents, which “gives surgeons crucial information that is not available at the moment — for example, that a particular stent is not going to occlude a vital blood vessel responsible for eyesight.”
Importantly, the technology allows the clinician to visualize alternative stent options, allowing them to choose the optimal device. In this way, AI is used to minimize errors and assure doctors rather than to replace them entirely.
There are additional benefits to this technology beyond the operating theater. According to Oxford Heartbeat, increased accuracy in stent placement reduces waste because there is less guesswork with more than one stent to determine the best fit.
Additionally, this technology’s success depends on interdisciplinary networking and the fusion of diverse perspectives. The team comprises specialists such as engineers, business strategists, medical paralegals and software developers of various academic backgrounds.
On the Oxford Heartbeat website, there is a page dedicated to profiles of the team members responsible for creating and maintaining this technology. This team consists of highly educated individuals, including six with PhDs, demonstrating how interdisciplinary networking creates a more robust and reliable technology.
These pioneering technologies are two prominent examples of how AI is already changing the face of medicine, particularly in the realm of neurosurgery. They offer a glimpse at the potential of AI to educate and advise medical professionals, adapt to novel situations and improve patient outcomes to save lives and lessen the burden on healthcare systems.