• Posted November 22, 2023

AI Could Predict a Nonsmoker's Risk for Lung Cancer

Artificial intelligence (AI) can help assess lung cancer risk in nonsmokers, a new study shows.

The "CXR-Lung-Risk"AI program evaluates routine chest X-ray images, looking for patterns associated with lung cancer, researchers said.

People whose chest X-rays were flagged by the AI program had twice the risk of developing lung cancer as those whose images were deemed low-risk, according to findings to be presented Wednesday at the Radiological Society of North America's annual meeting in Chicago.

"This AI tool opens the door for opportunistic screening for never-smokers at high risk of lung cancer, using existing chest X-rays in the electronic medical record,"said senior researcher Dr. Michael Lu, director of artificial intelligence and co-director of the Cardiovascular Imaging Research Center at Massachusetts General Hospital in Boston.

"Since cigarette smoking rates are declining, approaches to detect lung cancer early in those who do not smoke are going to be increasingly important,"Lu added in a meeting news release.

Lung cancer is the most common cause of death by cancer. There will be nearly 240,000 new cases of lung cancer in the United States this year, and more than 127,000 lung cancer deaths, according to the American Cancer Society.

Between 10% and 20% of lung cancers occur in so-called "never-smokers"-- people who have never smoked at all or have smoked fewer than 100 cigarettes in their lifetime, researchers said in background notes.

Because smoking is so strongly linked to lung cancer, the U.S. Preventive Services Task Force recommends low-dose CT scan screening only for adults between 50 and 80 with a current or longstanding history of smoking.

Screening is not recommended for those who've never smoked, or who have smoked very little, researchers said.

One reason these guidelines exclude never-smokers is that it's difficult to predict their lung cancer risk, researchers said.

"However, lung cancer is increasingly common in never-smokers and often presents at an advanced stage,"lead researcher Anika Walia, a medical student at Boston University School of Medicine, said in a meeting news release.

The research team developed the CXR-Lung-Risk model using more than 147,000 chest X-rays of nearly 41,000 smokers and never-smokers who participated in a cancer screening trial.

After the AI learned lung cancer patterns from those chest X-rays, researchers then tested its abilities on a set of 17,400 never-smokers. These were patients who'd all had routine outpatient chest X-rays between 2013 and 2014.

The AI flagged about 28% of the never-smokers as high risk for lung cancer, and 2.9% later were diagnosed with lung cancer.

That well exceeds the 1.3% six-year risk threshold by which lung cancer screening CT is recommended under National Comprehensive Cancer Network guidelines, the researchers argued.

"A major advantage to our approach is that it only requires a single chest-X-ray image, which is one of the most common tests in medicine and widely available in the electronic medical record,"Walia said.

Because the findings were presented at a medical meeting, they should be considered preliminary until published in a peer-reviewed journal.

More information

The American Cancer Society has more about lung cancer.

SOURCE: Radiological Society of North America, news release, Nov. 22, 2023