Could AI & a Blood Test Help Spot Lung Cancers Early?

June 10, 2024
Carole Tanzer Miller HealthDay Reporter

Researchers have shown that artificial intelligence (AI) can identify people who are at risk for lung cancer based on genetic markers in their blood.

"We have a simple blood test that could be done in a doctor's office that would tell patients whether they have potential signs of lung cancer and should get a follow-up CT scan," said corresponding author Dr. Victor Velculescu, co-director of the Cancer Genetics and Epigenetics program at Johns Hopkins Kimmel Cancer Center in Boston.

This so-called liquid biopsy could help identify patients who would need further screening, potentially boosting overall screening rates and preventing deaths, computer modeling shows.

Lung cancer is the world's deadliest cancer, according to the World Health Organization.

For the new study, researchers at Johns Hopkins and other institutions used AI to identify patterns in the blood of DNA fragments linked with lung cancer.

It included more than 950 people in 23 states who met criteria for screening with low-dose computed tomography (CT) scans.

Yearly screening, which can help detect cancers early when they are most treatable, is recommended for high-risk individuals, including current and former smokers. Only between 6% and 10% of Americans who are eligible are screened each year.

Velculescu said people may avoid screening due to the time it takes and the low doses of radiation that the scan exposes them to during screening.

So, he and his colleagues spent five years developing a test that relies on AI to recognize patterns of DNA fragments found in patients with lung cancer.

In normal cells, DNA is neatly folded, almost like a rolled up ball of yarn. In cancer cells, DNA is disorganized. When any type of cell dies, fragments of DNA wind up in the blood. Those fragments tend to be more irregular and haphazard in patients who have cancer than in others.

Researchers trained AI software to identify patterns in DNA fragments in blood from 576 people with and without lung cancer. They double-checked their method in a second group of 382 people.

The test was so accurate, the study found, that only 2 in 1,000 individuals with lung cancer may be missed.

If the new test boosted the lung cancer screening rate to 50% within five years, it could quadruple the number of cancers detected, computer simulations showed. That could prevent an estimated 14,000 cancer deaths over five years.

"The test is inexpensive and could be done at a very large scale," Velculescu said in a Hopkins news release. "We believe it will make lung cancer screening more accessible and help many more people get screened. This will lead to more cancers being detected and treated early."

The test is already available for laboratory use, and the team plans to seek U.S. Food and Drug Administration approval to use it for lung cancer screening.

Velculescu's team also plans to investigate whether a similar strategy could help detect other cancers.

The findings were published June 3 in the journal Cancer Discovery.

More information

The American Cancer Society has more about preventing lung cancer.

SOURCE: Johns Hopkins Medicine, news release, June 6, 2024

Cancer: Lung