Arthritis and breast cancer may soon be detectable decades in advance
AUSTIN (KXAN) — One of the most common medical issues in the country, osteoarthritis, could soon be diagnosed years before it becomes problematic. Affecting one in 10 Americans, preventing osteoarthritis has been nearly impossible. However, new developments with artificial intelligence (AI) may change this.
"It's impossible to tell at age 20, when everyone is everything is fine and dandy, and you're young, to know what's going to happen in your 40s and your 50s," said Vagheesh Narasimhan Ph.D., an assistant professor in the department of integrative biology at the University of Texas.
In September, Narasimhan published new research in the scientific journal Nature Digital Medicine showing how artificial intelligence could predict arthritis years in advance. He, alongside co-author Prakash Jayakumar with Dell Medical School, are now embarking on a new research project that could show how AI could assist doctors in treating arthritis.
"AI has this ability to quantify and examine images at scale," Narasimhan said. The project will study 300 patients, with half having their X-rays analyzed by AI. The scientists will explore outcomes of joint replacements, comparing AI recommendations to those of doctors.
The project will also examine postoperative outcomes: When will the patient walk again? How quickly did they walk? How much pain are they experiencing? These questions will help determine the effectiveness of AI recommendations versus those of doctors.
Why arthritis is perfect for AI
Arthritis is thought to be caused by humans walking upright, with the weight of our bodies grinding our knees over time. "You get to a certain point where your bones start rubbing against each other. And this leads to pain. Thus far, there hasn't been a cure for this disorder," Narasimhan said.
Detecting arthritis early is challenging. Patients typically seek medical attention when they experience pain or when signs are detected in an X-ray. These signs include loss of gaps between joints or bone loss.
Arthritis has an advantage over many other medical conditions because of the abundance of X-rays taken over several decades. This provides a large dataset for AI to analyze. These X-rays span different ages, showing changes over time. X-rays are also relatively cheap, providing data from diverse groups often overlooked in large-scale data collection.
Breast cancer detection and AI
Arthritis isn't the only medical condition benefiting from mass image detection. "We've literally built systems that can detect and predict cancer risk," said Ron Green, co-founder and CTO of Austin-based KUNGFU.AI.
Green mentioned that breast cancer, which also has extensive imaging data available, could soon be detected years in advance. "These computer vision systems can now pick up biological signals that we as humans cannot detect," Green said.
Narasimhan emphasized that these advancements are only possible because of AI's ability to process large amounts of data. "This is something that's not possible purely by a team of doctors, simply because it's not feasible to diagnose tens of thousands of people simultaneously."