Key takeaways:

• The EGFR T790M variant elevated lung cancer risk 25-fold overall and 60-fold in individuals who never smoked vs. those who did not carry the variant.
• The variant only had an association with lung cancer.

A single inherited EGFR mutation may increase lung cancer risk more than 60 times among individuals who never smoked.

An evaluation of more than 3 million people showed the EGFR T790M variant increased risk for lung cancer 25 times overall, with a substantially larger impact on patients who reported never smoking cigarettes.

Data derived from LoPiccolo J, et al. Science. 2026;doi:10.1126/science.aec0473.

The variant had no association with any other malignancy.

“We expected a strong effect, but certainly not as strong as this,” Jaclyn LoPiccolo, MD, PhD, attending physician and lung cancer researcher at the Lowe Center for Thoracic Oncology and Center for Cancer Genetics and Prevention at the Dana-Farber Cancer Institute, told Healio.

“This single variant confers an extraordinarily large amount of risk.”

How ‘strong’ is the association?

Smoking remains the primary driver of lung cancer, as 87% of cases in men and 84% in women can be attributed to cigarettes, according to American Cancer Society’s Cancer Statistics 2026 report.

Current U.S. Preventive Services Task Force screening guidelines revolve around smoking. Adults aged 50 to 80 years who have at least a 20 pack-year smoking history, and who actively smoke or stopped within the past 15 years, are recommended to get low-dose CT.

However, Healio previously reported that current screening guidelines may exclude approximately 65% of patients who develop lung cancer.

Between 20,000 and 40,000 individuals who never smoked or smoked fewer than 100 cigarettes during their lifetime are diagnosed with lung cancer annually, according to CDC.

Approximately 10,000 of these cases are attributable to secondhand smoke and radon exposure.

About 20 years ago, researchers also found the EGFR T790M variant could have a significant association with lung cancer based on high incidence in a single European family, according to study background.

“It was more commonly described as a resistance mutation, found in tumors only, to earlier generation EGFR inhibitors,” Pasi A. Jänne, MD, PhD, senior vice president for translational medicine and director of Robert and Renée Belfer Center for Applied Cancer Science at Dana Farber Cancer Institute, told Healio.

“In 2005, a group identified that it also exists in the germline, in the normal DNA of a family that had multiple members with lung cancer. That was the first description of it, and there have been subsequent studies [conferring that association].”

“What we didn’t have was a population large enough to determine just how strong that association is, the frequency of the mutation, and how risk varies across different groups,” LoPiccolo added.

LoPiccolo, Jänne and colleagues partnered with 23andMe Research Institute to further elucidate the data.

The study included 3,372,531 individuals with European ancestry; 9,799 who had been diagnosed with lung cancer and 3,362,732 who had not.

The association between EGFR T790M and lung cancer served as the primary endpoint.

Variant ‘clearly’ impacts risk

Individuals with the EGFR T790M variant had a significantly higher likelihood of developing lung cancer (OR = 25.18; 95% CI, 13.5-46.78).

Comparatively, smoking alone conferred about seven times lower risk (OR = 3.79; 95% CI, 3.61-3.97).

“This clearly identifies a unique genetic variant that puts an individual at risk of developing lung cancer,” Jänne said.

Researchers observed a substantially larger association among individuals who never smoked (OR = 61.7; 95% CI, 28.9-131.73) than those who had a smoking history (OR = 10.59; 95% CI, 2.39-47.08).

“We’re not interpreting this as saying smoking somehow diminishes the effect of the mutation,” LoPiccolo said. “It’s because people who have never smoked start with a much lower baseline risk of lung cancer, so the effect of this mutation is larger in relative terms. We know that smoking itself greatly increases lung cancer risk, and the mutation also increases lung cancer risk, so having both certainly is not protective, but we did find that the association was particularly striking among people who have never smoked.”

LoPiccolo and colleagues also evaluated whether EGFR T790M increased among 17 other common malignancies, but they did not find an association.

“That’s biologically fascinating, and something that we want to continue to understand moving forward,” Jänne said. “Why is this solely a lung cancer-associated risk?”

Overall, one in 15,850 individuals had the EGFR T790M variant, but researchers found the mutation to be substantially more common (1 in 2,078) in some areas of the Southern Appalachian region of the U.S.

Most carriers (80.56%) had a British or Irish primary ancestry.

“It’s a rare genetic alteration, but the impact is significant,” Jänne said. “Carriers of this mutation are at a higher risk for lung cancer and should be referred for consideration of screening, which is something we’re currently doing.”

Researchers acknowledged study limitations, including participation bias in direct-to-consumer genetic testing.

LoPiccolo highlighted several research questions that future studies should address.

“One of them is lifetime risk estimates from studying carriers and their families,” she said. “The other ones are things that modify this risk. Not all carriers of this mutation develop lung cancer. Genetic and environmental modifiers of this risk are something that we’re looking at. And then, obviously, the ideal CT screening paradigm for these patients, so that we can aim to intervene on these lung cancers when they can be cured.”

Both LoPiccolo and Jänne also noted the importance of 23andMe in the research, and how these types of partnerships could benefit investigations into genetics and other cancers.

“We wouldn’t have been able to quantify this risk without this kind of population-level data, and there is no way to get that data,” LoPiccolo said.

“Even if we had partnered with other academic institutions, I don’t think we would have come up with over 3 million individuals,” Jänne added. “The power is in the numbers.”

For more information:

Pasi A. Jänne, MD, PhD, can be reached at Pasi_Janne@dfci.harvard.edu.

Jaclyn LoPiccolo, MD, PhD, can be reached at Jaclyn_LoPiccolo@dfci.harvard.edu.