Scientists have found compelling evidence that heredity plays a significant role in determining whether smokers develop lung cancer, a finding that, if confirmed, could explain why some tobacco addicts develop the deadly disease early in life and some lifetime smokers never get cancer.

In two studies published yesterday in the Journal of the National Cancer Institute, researchers report evidence that there is a gene -- found in some people but not in others -- that plays a key role in giving tobacco smoke its carcinogenic potential.

One of the studies also suggests an explanation for why some nonsmokers get lung cancer. It indicates that people who have two copies of the gene -- one from each parent -- but who never smoke have a lung cancer risk 2,200 times higher than that of nonsmokers who lack the gene.

The reports, if true, would explain the long-observed -- but never adequately explained -- phenomenon of lung cancer's high prevalence in some families and utter absence from others. If the gene or genes responsible can be pinpointed, it is possible that a drug to prevent lung cancer and a test to identify smokers most at risk for the disease could be developed.

"There has to be some explanation for why not everyone that smokes gets lung cancer," said University of Minnesota epidemiologist Thomas Sellers, who headed one of the studies published yesterday. "We haven't found it. But we have got some encouragement about how and where to look for a candidate gene."

The researchers stressed, however, that genetic differences account for only part of a smoker's risk of lung cancer, and that all tobacco users -- regardless of their inherited traits -- still run substantial risks of heart disease, stroke, emphysema and other lung diseases caused by smoking.

"If these studies are taken to mean that it is okay for some people to smoke because they are at lower risk for lung cancer, that would be a catastrophic interpretation of this work," said National Cancer Institute researcher Neil Caporaso, who headed the other study.

The suspicion that lung cancer may involve more than cigarette smoke and other environmental factors has been prevalent among epidemiologists for close to 30 years. Until recently, however, scientists lacked the statistical and other tools to analyze genetic and environmental factors simultaneously.

Sellers's study, which involved 337 families in Louisiana in which at least one member had died of lung cancer, is one of the first to use the new technique. By closely analyzing the histories of the families in the study, he and his co-workers were able to determine that lung cancer appeared throughout the families in patterns resembling those in which a single gene was being passed on. The pattern was not consistent with exposure to environmental factors alone.

The study does not identify the gene involved. But Sellers said that it appears to be carried by about 10 percent of the smokers studied and seems to put that group at much greater risk of developing lung disease early in life.

For example, the study estimates that of lung cancers developed by smokers before the age of 50, the gene is solely responsible for about 27 percent and in combination with environmental factors causes another 42 percent. As smokers move into their 60s and 70s, the role of the gene appears to diminish.

The other study, led by Caporaso, implicates one gene as playing a role, through other genes are probably also involved. The study focuses on a gene -- known as CYP2D6 -- which causes cells to make an enzyme needed to break down and process a variety of substances that enter the body, including foods, drugs and various other chemicals.

The enzyme functions more efficiently in some people than others. Caporaso found that people in whom it functioned poorly or not at all -- about 40 percent of those he studied -- had a lung cancer rate one-sixth that of people in whom the enzyme worked at full efficiency. Caporaso speculates that the enzyme is needed to turn the chemicals in cigarette smoke into carcinogens. With little or no enzyme, the smoke is less harmful.

Caporaso said that this by no means proves that CYP2D6 is the lung cancer gene identified by Sellers' study. It is likely, Caporaso said, that more than one gene is needed. He said it is also possible that the key component could be another different gene that is simply inherited with CYP2D6.