Rare inherited EGFR mutation raises lung cancer risk 25-fold, study finds

Lung cancer has long been framed as a disease of smoking, and for good reason. Tobacco use remains the single largest driver of the disease, and public health campaigns, screening guidelines and social stigma have all been built around that assumption.

Yet somewhere between 10 and 20 percent of lung cancer cases diagnosed in the United States each year occur in people who never smoked, a population that has historically had few answers about why they developed the disease. For those patients, the standard explanation has often been an unsatisfying shrug.

A study published in Science on September 17, 2026 by investigators at Dana-Farber Cancer Institute and 23andMe Research Institute now offers a concrete answer for a small slice of that group. The researchers report that carriers of a rare inherited mutation in the Epidermal Growth Factor Receptor (EGFR) gene, known as T790M, had a 25-fold higher risk of lung cancer.

The work drew on de-identified, aggregated genetic data from more than 3.3 million 23andMe research-consented participants, making it one of the largest efforts yet to tie a specific inherited variant to lung cancer risk.

Key Facts

Dana-Farber Cancer Institute announced on September 17, 2026, that carriers of the inherited EGFR T790M mutation had a 25 times higher risk of lung cancer than noncarriers. Among people who never smoked, carriers were more than 60 times as likely to develop the disease. Among people who smoked, the risk was about 10 times higher than in noncarriers who smoked. The researchers identified EGFR T790M as one of the strongest inherited risk factors for lung cancer discovered to date.

To find carriers, the team turned to the 23andMe research cohort. The Boston Globe reported on September 17, 2026, that investigators ultimately identified 641 T790M carriers out of more than 10 million consented participants, a rate of roughly 1 in more than 15,000 people in the United States. When they narrowed the analysis to the more than 3.3 million people who had indicated whether they had had lung cancer, they calculated the 25-fold and 60-fold risk figures.

Prevalence is not uniform across the country. Jaclyn LoPiccolo, an attending physician at Dana-Farber and the lead author, noted that the mutation is found at about 1 in 15,000 people nationally, but in parts of the Southeastern United States it can be as frequent as 1 in 2,000 people. Across Alabama, Mississippi and Tennessee, the rate is about 1 in 2,078. Medical Xpress reported on September 17, 2026, that the variant was not linked to any of 17 other common cancers studied, a result the team called a surprise. Additional data came from the INHERIT Study (NCT05587439), led at Dana-Farber by LoPiccolo and Pasi A. Janne in partnership with GO2 for Lung Cancer and the Addario Lung Cancer Medical Institute.

The variant's history is unusual. Using 23andMe data, the researchers traced the mutation to Europe, probably Britain or Ireland, and said the vast majority of carriers inherited it from the same ancestral lineage. "We could trace that lineage to British and Irish settlers in the United States and show that the mutation became enriched after a founder event and genetic bottleneck in Southern Appalachia about 200 years ago," LoPiccolo said. The Boston Globe reported on September 17, 2026, that the variant became established in the South roughly 200 to 225 years ago.

The findings were published in Science. Alexander Gusev, a co-senior author, said, "One of the remarkable findings here is just how strong an effect a single mutation can have." Co-senior author Janne added, "By studying more than three million people, we were able to demonstrate just how strongly this inherited mutation is associated with lung cancer."

Analysis

The striking element of this study is not that an inherited mutation raises lung cancer risk, but how much it raises it. Gusev said, "To my knowledge it's one of the strongest cancer risk increasing mutations that has ever been found." What this really means is that inherited susceptibility, rather than behavior, can be the dominant driver of lung cancer for a small but identifiable group of people, and that group may be findable long before a tumor appears.

TIME reported on September 17, 2026, that the T790M variant was first discovered in a European family in 2005 whose members had not smoked but had developed lung cancer. Two decades later, the scale of the 23andMe dataset has allowed researchers to quantify what was previously a family curiosity. "While we knew that T790M was associated with lung cancer, we didn't have a population large enough to determine how common the variant was, how strong its effect, and how the risk varies in different groups," LoPiccolo said.

Context matters here. Smoking alone confers roughly a 4-fold increase in lung cancer risk, according to the study's numbers, which means the inherited variant's effect is several times larger than the behavioral exposure it is so often compared with. Gusev framed the two as additive rather than interchangeable: "Smoking is bad for lung cancer. This mutation is bad for lung cancer. When you do both, your risk is the sum of those two risks."

Independent voices are cautious about scope. The Boston Globe reported on September 17, 2026, that Deborah Caswell, a lung cancer biologist at University College London who was not involved in the new study, said the variant likely plays a role in only a small portion of never-smoker cases, while describing the amount of increased risk as huge. The bigger picture here is that lung cancer in people who never smoked is not one disease but a collection of causes, and one of those causes has now been given a name, a number and a map.

Why It Matters

Lung cancer screening in the United States is driven almost entirely by smoking history, and the study's authors say that could change. "Today, lung cancer screening is driven almost entirely by smoking history," LoPiccolo said. "Our findings raise the possibility that, in the future, screening could also be dictated by inherited genetic risk." If that shift happens, it would be the first time a germline variant, rather than a behavioral exposure, brought a person into a lung screening program.

The immediate human consequence is already visible. TIME reported on September 17, 2026, that patient Frank McKenna, a personal trainer in Virginia Beach, was diagnosed with Stage IV lung cancer in 2016 despite never smoking. A genetic test found he carried EGFR T790M, and he began a targeted therapy that he continues today. His daughter, now 33, later tested positive for T790M, but there are no evidence-based monitoring recommendations for her yet. That gap, between knowing a person carries a high-risk variant and knowing what to do about it, is the central practical problem the study leaves open.

The paper did not quantify absolute risk, and the team is now investigating that number. Nadia Litterman, executive director of the Susan Wojcicki Foundation, which funded the Lung Cancer Genetics Study that provided data, said, "That's the world we are trying to work toward, and this is a major step toward that."

Next Up

The researchers' next step is absolute risk. A relative risk of 25-fold sounds enormous, but without knowing how many carriers actually develop lung cancer over a lifetime, clinicians cannot write a screening guideline or decide who should receive annual imaging. LoPiccolo's team is now working on that figure, and the INHERIT Study, run in partnership with GO2 for Lung Cancer and the Addario Lung Cancer Medical Institute, is positioned to follow carriers over time.

Whether the findings reach the communities where the variant is most common is a separate question. Dana-Farber Cancer Institute reported on September 17, 2026, that the mutation was not associated with any of 17 other cancers, which narrows the clinical target but also means that any screening program built around T790M would need to focus specifically on the lung. For families like McKenna's, the science has arrived before the guidance.