Sitting in an office on the campus of the Massachusetts Institute of Technology, researcher Peter Mikhael is scrolling through hundreds of CT scans.
He can't see the lung cancer that will develop years later on the slides, but a new form of artificial intelligence can.
"It is actually predicting that three years later this person is going to get cancer," Mikhael said as he clicked through a series of CT scans.
Mikhael is part of a team of researchers with the Abdul Latif Jameel Clinic for Machine Learning in Health at MIT. They have spent the last few years teaming up with specialists at Mass General Cancer Center and Chang Gung Memorial Hospital. What they have developed is a new form of AI called Sybil. Much like the Greek oracles she's named after, Sybil's job is to see into the future and detect lung cancer years before doctors or radiologists can.
And it seems to be working.
"AI in health care is one of the most underutilized areas of applications of AI," said Regina Barzilay, another researcher on the project.
In a nod to JFK, President Biden pushing 'moonshot' to fight cancer
The president rallied the nation around developing treatments and therapeutics for the pervasive diseases.
Lung cancer is the deadliest cancer in the world. In 2020, the disease killed 1.7 million people. In the U.S. an estimated 238,000 people will be diagnosed with the disease this year.
Currently, the CDC recommends adults at risk for lung cancer get CT screens each year. But even with regular screenings it's impossible for radiologists and doctors to spot every potential problem with the naked eye.
Sybil, though, can see what humans can't while looking at millions of tiny pixels to spot patterns.
In this case, those patterns show the possibility of lung cancer developing years before cancer shows up in a traditional screening.
"What Sybil tries to do is say how likely it is this person will develop cancer," Barzilay said.
Barzilay and her team don't see Sybil as replacing doctors or radiologists. Instead, they view her as another tool that can be used to help doctors develop a more specialized treatment for an individual's cancer.
Large-scale, clinical deployment of Sybil is still years away. But for the hundreds of thousands of Americans who will be diagnosed with lung cancer each year, artificial intelligence could someday be a game changer.
"When it's deployed in large scale, it will save lives," Barzilay said.
