Imagine trying to guide a tiny tool through a maze while only looking at a flat picture of it. That is a little like what doctors face during many minimally invasive surgeries.

In these procedures, doctors may insert thin tools, such as catheters or cameras, through a small cut in the skin. They often use live X-rays to help guide those tools. But an X-ray is flat—it shows height and width, but not depth. That can make it hard to tell exactly where a tool is inside the body.

Now, researchers at MIT and partner hospitals say they have built an AI tool that could give doctors a clearer map.

The new system is called xvr, short for “X-ray volume registration.” It matches a live 2D X-ray with a patient’s earlier 3D CT scan or MRI, per the National Library of Medicine. In plain English: it helps turn a flat X-ray into a more useful picture of where things are inside the body.

How the AI works

Before surgery, a patient may already have a CT scan or MRI. Those scans make a detailed, 3D picture of the body.

Xvr uses that person’s own scan to create many practice X-rays from different angles. The system can make about 1,000 pretend-but-realistic X-ray images every second, according to MIT. It then learns how the patient’s body looks from many viewpoints.

During surgery, xvr compares the real X-ray on the screen with the 3D scan. It can line them up in seconds and locate the images with sub-millimeter precision—that means accuracy smaller than 1 millimeter.

Think of it like matching a photo of a neighborhood street to a 3D map. Once the computer knows where the photo was taken, it can help show where a car—or in this case, a surgical tool—is located.

“We want to make these 2D X-rays more informative, so it becomes safer and easier to do these life-saving procedures,” Vivek Gopalakrishnan, the study’s lead author, said in MIT’s announcement.

Why this matters for minimally invasive surgery

Minimally invasive surgery can often mean smaller cuts, less pain, and a quicker recovery than traditional open surgery. But doctors need excellent guidance to move tools safely around bones, blood vessels, and organs.

Today, medical teams may have to manually line up the live X-ray with a patient’s CT scan or MRI. That can take time and may require a highly trained specialist to interpret fuzzy, flat X-ray images.

Xvr is designed to make that matching job quicker and more reliable for each individual patient. Instead of hoping one AI program works equally well for everyone, this system takes time to adapt to the specific person receiving care.

“We tailor this one specific model for this one specific patient,” Gopalakrishnan said. “It doesn’t matter if it works on other people because there will be different models for those people.”

That patient-by-patient approach is important. Bodies are not all shaped the same. People can also have injuries, implants, tumors, or other differences that make a one-size-fits-all AI system less dependable. The xvr research team says its method uses the patient’s own scan to create training images, avoiding the need for large sets of manually labeled surgical images.

What the research found

The MIT-led team tested xvr using data from five hospitals. The research included adults and children, multiple body areas, and several kinds of medical procedures.

The researchers reported that xvr:

  • Adapted to a new patient in about five minutes.
  • Matched live X-rays to 3D scans in seconds.
  • Reached sub-millimeter accuracy in the study.
  • Performed about 10 times better than existing AI approaches in their evaluations.

The study appears in Nature under the title “Rapid patient-specific neural networks for X-ray to volume registration.” The authors also released xvr as open-source software, meaning other researchers can inspect and build on the code.

Could this help in emergencies?

Possibly. Speed matters in emergencies, especially when doctors need to treat problems such as blocked blood vessels in the brain.

MIT’s system uses a kind of pre-trained AI, sometimes called a foundation model, so it does not have to learn everything from scratch for each patient. The research team says this reduced the patient-specific preparation time from roughly 12 hours to about five minutes.

“An hour in stroke time is incredibly substantial,” Gopalakrishnan said. He added that combining 2D and 3D information could make specialized, life-saving procedures easier to offer in more places.

That does not mean xvr can instantly bring advanced surgery to every hospital. Hospitals would still need trained teams, proper imaging equipment, careful testing, and the right approvals. But a tool that helps doctors see more clearly could eventually make complex procedures easier to perform.

Important: It is not ready for every operating room

The headlines are exciting, but this is still research—not a promise that every surgery will become safer tomorrow.

The MIT team says its next steps include making xvr faster for real-time use, testing it in more situations, and improving how it handles body parts that move, such as those affected by breathing or heartbeat.

Any AI software used as a medical device in the United States must also meet FDA requirements for safety and effectiveness before it can be marketed for a specific medical use. The FDA maintains a public list of AI-enabled medical devices that have gone through its applicable premarket process.

So, xvr should be viewed as a promising navigation helper for clinicians—not a replacement for surgeons, radiologists, or hospital safety checks.

The big picture

AI in healthcare is often talked about as if it will “replace doctors.” This MIT project tells a more useful story: AI can act more like a smart map reader.

The doctor remains in charge. The AI’s job is to quickly connect a live X-ray with a patient’s 3D body scan, helping the medical team better understand where a tool is and what is nearby.

If future studies confirm the system works safely in real operating rooms, xvr could become one more way AI helps make minimally invasive surgery, image-guided surgery, surgical navigation, medical imaging AI, and even robotic surgery more precise.

For patients, the hoped-for result is simple: doctors may be able to make smaller cuts while getting a clearer view of the path ahead.

This content is provided by AIxAverage.com for general educational and informational purposes and does not constitute any substantive advice. It was developed with support from AI and reviewed by our editorial team for accuracy. Even so, AI-assisted material can contain errors, omissions, or information that is no longer current. We always encourage our readers to confirm any important details with trusted, independent sources.