The two images from an MRI scan of an injured knee look similar, but there is one big difference: one was made faster using artificial intelligence.
Facebook researchers and doctors at NYU Langone Health say they were able to speed up MRI scans using AI. The images were so clear that most doctors couldn't distinguish an AI-generated scan from a traditional one.
For more of it
Subscribe to the CNET Now newsletter to receive the most important stories of the day from our editors.
Doctors use detailed images of organs, muscles, and other soft tissues from MRI scans to diagnose problems in the human body from tumors to knee injuries. However, a scan can be uncomfortable for some patients and children as they sometimes have to stay in a confined space for more than an hour. Faster MRI scans would not only improve a patient's experience and potentially increase the use of the technique, but would also reduce waiting times in the doctor's office and the cost of medical bills.
Facebook doesn't make MRI machines, but partnering with NYU is a way for the company to advance AI research. Artificial intelligence plays a bigger role in healthcare, including robotic surgery, cancer risk prediction, and medical image analysis.
Larry Zitnick, head of Facebook AI research, said the company's scientists are interested in working in the medical field because they want to tackle difficult and powerful problems that AI could solve. To cut the time it takes to complete an MRI scan, the AI model developed by Facebook used less data to create an image. However, these images must be accurate, otherwise it could lead a doctor to misdiagnose a patient.
Dr. Michael Recht, professor and chairman of the Department of Radiology at NYU Langone Health, worked with Facebook on a study that included 108 patients who received an MRI scan of their knees last year. Using the traditional approach, her MRI scans took between approximately 8 minutes and 11 minutes. With the help of AI, Facebook and NYU were able to cut the time for these scans to 4 to 6 minutes. This comes from a study to be published in the American Journal of Roentgenology.
These time savings don't seem like much, but they add up. The time it takes to perform an MRI scan depends on which parts of the body doctors want to examine and what condition they are trying to diagnose. When a patient suffers a stroke - when their brain's blood supply is cut - minutes can mean the difference between life and death, said Recht, who focuses on musculoskeletal imaging.
"We believe that we can accelerate a lot faster in the brain than in the knee. And if we can accelerate ... three times as fast or four times as fast, then you really save a lot of time," he said.
An MRI, which stands for magnetic resonance imaging, uses strong magnetic fields and radio waves to create images of organs and other soft tissues in a person's body. Most of the human body is made up of water. A water molecule contains two hydrogen atoms and one oxygen atom. The hydrogen atoms respond to the magnetic field and send out a specific high frequency signal, which generates data collected by the MRI scanner. This data is then used to create an image of a human body part.
The study also showed that the faster MRI scans were also accurate, despite less data than the traditional approach. Most radiologists couldn't tell which images were taken by AI, and they also thought these scans resulted in a clearer picture of the knee. For example, moving a patient can blur an image from an MRI scan.
Facebook compared the process carried out with the help of the AI to the exact completion of a 1,000-piece puzzle with only 250 pieces.
The algorithms, Zitnick said, have looked at thousands of images of knees so they know how a knee is structured and are able to create an image of that body part "in a way that is clear".
MRI is just one way doctors can look inside the human body without cutting open a patient. Unlike other methods such as CT scans and X-rays, MRIs do not expose patients to radiation. By reducing the time and cost of doing MRIs, Recht said doctors, even in developing countries, could use MRIs over other imaging modalities. Researchers still need to ensure that AI-accelerated MRI scans of the brain, liver, and other parts of the body are accurate.
"We get the same or more information in less time. Therefore, we should be able to use this sequence for all types of pathologies if it persists," said Recht.
The information contained in this article is for educational and informational purposes only and is not intended as health or medical advice. Always consult a doctor or other qualified health care provider with any questions about a disease or health goals.
