Lung-cancer screening in the United States is built largely around smoking history. That remains sensible: tobacco is the dominant preventable cause of the disease.

But a study published September 17 in Science identifies an inherited genetic variant that may be a powerful exception to that framework. People carrying the variant, known as EGFR T790M, had substantially higher odds of developing lung cancer, with the strongest association among people who had never smoked.

The finding is important. It is also easy to overstate.

The study does not show that every carrier will develop cancer. It does not establish that people without a smoking history should immediately seek repeated CT scans. And it does not yet justify population-wide genetic testing or lung-cancer screening.

What it does show is that a rare inherited mutation can confer a strong susceptibility to lung cancer—strong enough to challenge a screening system organized primarily around behavior.

A mutation with two biological roles

The EGFR gene helps regulate how cells grow and divide. In many lung cancers, changes in EGFR arise during a person’s lifetime inside tumor cells. These are somatic mutations: they are present in the cancer but are not necessarily inherited by the patient’s children.

EGFR T790M can also occur in the germline, meaning it is present in the body’s inherited DNA from birth. That distinction matters. A tumor may acquire a mutation during the biological history of one cancer. A germline mutation can create a predisposition that exists before any tumor appears.

Researchers have known since 2005 that inherited EGFR T790M occurred in families with unusual clusters of lung cancer. Earlier family studies showed that the variant appeared to matter, but they could not reliably measure the size of the risk.

The new study attempted to address that limitation through scale. Jaclyn LoPiccolo of Dana-Farber Cancer Institute, Pasi Jänne and colleagues analyzed genetic and health information from more than 3.3 million research-consented participants whose data were available through 23andMe.

The researchers identified the variant in approximately one person in 15,850 in the study population. Carriers had approximately 25 times the odds of lung cancer compared with noncarriers. Among never-smokers, the estimated association was roughly 62-fold. The mutation was not significantly associated with 17 other cancers examined in the study.

Those figures are striking, but they require translation.

A relative-risk or odds estimate does not tell an individual carrier the probability of developing cancer. Absolute risk depends on age, sex, family history, other genetic factors, environmental exposures and the background rate of disease in the relevant population.

Nor does “62 times” mean that the mutation is more dangerous than smoking in every practical sense. The comparison involves different groups and different baseline risks. Smokers who carry the mutation still face elevated risk; the mutation does not make smoking safe or irrelevant.

The narrower and better-supported conclusion is this: in the study dataset, inherited EGFR T790M was a very strong marker of lung-cancer susceptibility, particularly among people who had never smoked.

Why Southern Appalachia matters

The study also found that the variant was more common in the United States than in British- and Irish-descended populations outside the country. The researchers traced most copies to a shared ancestral lineage associated with British and Irish settlers and estimated that a founder event in Southern Appalachia occurred roughly 200 to 225 years ago.

A founder event occurs when a relatively small population carries a genetic variant into a new or isolated community. If descendants remain connected through generations, a variant that is rare in the broader population can become more common locally.

Dana-Farber reported that the mutation may occur in approximately one in 2,000 people in parts of the Southeastern United States, compared with approximately one in 15,000 nationally. The study identified concentrations in Southern Appalachian populations, including areas associated with Tennessee and Alabama.

That does not mean the mutation is confined to Appalachia, or that ancestry alone can determine whether someone carries it. Populations move, family histories overlap and commercial genetic databases do not represent every community equally.

The geographic pattern may nevertheless be useful. It suggests that a variant too rare to detect easily in ordinary clinical populations can become more visible when researchers combine large-scale genetic data with family history, ancestry and regional distribution.

For people in Alabama and other parts of the Southeast, the result is potentially relevant without being immediately actionable for the general public. The study does not establish that everyone with Southern Appalachian ancestry should be tested. It does suggest that clinicians may need to ask more precise questions when a family has multiple cases of lung cancer, particularly among people who never smoked.

Large datasets still have limits

The study’s principal strength is its size. A rare variant cannot be studied reliably through a handful of families alone. More than 3.3 million participants provided a much larger comparison group than earlier investigations.

The paper was published in Science, and its main results are consistent with earlier reports that inherited EGFR T790M appears in families with unusual lung-cancer patterns. The researchers also drew on additional data from the INHERIT study, a research effort focused on inherited lung-cancer risk.

But scale does not remove every source of uncertainty.

The 23andMe research population is not a random sample of the United States. The main analysis focused on more than 3.3 million people of European ancestry, limiting how confidently the results can be extended to populations that were less represented.

Health information collected through a consumer database may also differ from medical-record data. Some participants may have incomplete diagnoses, inaccurate smoking histories or uneven access to clinical testing. The researchers used statistical methods to address such issues, but no observational dataset guarantees perfect classification.

Most importantly, the study demonstrates an association. It does not by itself explain why some carriers develop lung cancer while others do not.

Other inherited variants, environmental exposures, age, sex, immune response and additional changes in lung cells may influence whether cancer develops. Dana-Farber says further work is needed to understand why the mutation appears to affect lung cancer without showing the same association with the other cancers examined.

A genetic marker can therefore be highly informative while still being insufficient to determine an individual’s fate.

Does the finding change lung-cancer screening?

Possibly—but not yet in the way some headlines may imply.

Current U.S. screening recommendations are designed primarily around age and smoking exposure. The researchers argue that inherited genetic risk could eventually become another basis for identifying people who might benefit from screening. Their proposed direction is targeted genetic testing followed, where appropriate, by personalized low-dose CT screening.

That is a research possibility, not a settled clinical recommendation.

CT screening can detect lung cancer earlier, but it also carries costs and risks. Scans expose patients to radiation, although at low doses. They can identify small nodules that are harmless but require repeat imaging, invasive procedures or prolonged anxiety.

A screening program must show not merely that it finds more tumors, but that it improves outcomes enough to justify those harms. That evidence does not yet exist for EGFR T790M carriers at the population level.

The next studies would need to determine:

  • how many carriers develop lung cancer over time;
  • the ages at which cancers appear;
  • whether periodic low-dose CT detects tumors early enough to improve survival;
  • how often screening produces false alarms;
  • whether benefits differ by sex, smoking history or ancestry; and
  • whether testing is practical and equitable outside specialized cancer centers.

A screening strategy built from a striking association can still fail if it is applied too broadly, too early or without adequate follow-up care.

The access and interpretation problem

The finding raises a difficult question about genetic information: who should be offered testing, and who should pay for it?

The study authors suggest that people with several close relatives who developed lung cancer, multiple lung nodules or multifocal lung cancers may wish to consult a genetic counselor. They also point to the potential relevance of ancestry in parts of the Southeast where the variant appears more common.

A genetic counselor can help interpret family history and explain the limits of testing. That matters because a result may affect not only the person tested but biological relatives who may share the variant.

Access to counseling and screening is uneven, however. A discovery that benefits only people who can reach specialized clinics, pay for testing or navigate complicated insurance rules could widen an existing health-care gap.

Regional genetic findings can also be mishandled socially. A population-level association may be turned into an identity claim, a stereotype or a reason to treat ancestry as a substitute for clinical evidence. The study supports more careful investigation of risk. It does not justify genetic suspicion directed at entire communities.

What the study establishes—and what it does not

We know that inherited EGFR T790M is rare.

We know that, in a study of more than 3.3 million participants, carriers had much higher odds of lung cancer than noncarriers, particularly among people who had never smoked. We know the variant appears to be more common in parts of the Southern Appalachian region than in the broader population. And we know the finding is consistent with earlier reports of familial lung-cancer clusters.

We do not yet know the absolute lifetime risk for every carrier. We do not know why some carriers remain cancer-free. We do not know whether genetic testing followed by regular CT screening will reduce mortality enough to justify its costs and harms. And we do not know whether the same risk estimates apply equally across populations that were not well represented in the dataset.

The study is therefore neither a reason for complacency nor a warrant for panic.

It identifies a genuine inherited risk factor and gives researchers a better estimate of its strength than they previously possessed. The next task is not to convert that estimate into a slogan. It is to determine how the information can be used—carefully, selectively and fairly—to find disease earlier without turning probability into destiny.

Sources

  • LoPiccolo et al., “Germline EGFR T790M mutation and lung cancer risk,” Science, September 17, 2026. PubMed
  • Dana-Farber Cancer Institute, study summary and researcher statements, September 17, 2026. Dana-Farber
  • American Association for the Advancement of Science, EurekAlert summary of the Science study. EurekAlert