Cigarette calculation to predict risk of cancer

SMOKERS and ex-smokers soon will be able to predict their risk of lung cancer

Scientists hope the formula will help people decide if they really want to try a controversial test to detect early stage lung cancer.

The calculation, published in this week's Journal of the US National Cancer Institute, shows a wide

variation in risk.

Consider a 51-year-old woman who smoked 20 a day since she was 14 until stopping nine years ago. The formula puts her chances of getting lung

cancer in the next 10 years at less than one in 100.

Compare a 68-year-old man who smoked 40 a day since he was 18 and hasn't yet quit. He has a one in seven chance of lung cancer by his 78th birthday if he keeps puffing. If he quit smoking today, the risk drops slightly to one in nine.

The formula works only for certain people those older than 50, who smoked at least 10 a day for at least 25 years because it is based on a study that tracked cancer development in just those people.

Researchers from New York's Memorial Sloan-Kettering Cancer Centre created the formula and posted an easy-to-use version for consumers on a web site. Doctors have used a similar model for years that calculates age, family medical history and other factors to predict a woman's risk of getting breast cancer.

But for lung cancer, doctors could give only vague advice: smoking is the chief cause, heavy smokers have the highest risk and that risk drops with each year that passes since kicking the habit.

The new formula will help doctors "be more specific now about who is at greatest risk", said Dr Tom Glynn of the American Cancer Society, who praised the research. That's particularly important as more people consider getting those aggressively advertised, but still unproven, spiral CT scans to hunt early lung cancer, Glynn said. Only 15% of lung cancer sufferers survive five years, mainly because the disease usually is diagnosed very late. There is no proven screening test so far.

The National Cancer Institute is studying whether spiral CT scans, which view the lungs at various

angles, could improve survival by spotting tumours early. There's no answer yet, and the scans have a big problem: up to half detect harmless scar tissue or some other benign lump requiring a risky biopsy or other follow-up testing.

Lung specialists see many patients "wracked by anxiety and concern about what may be in their future" because of ambiguous CT results, said Dr Peter Bach, Sloan-Kettering's lead researcher.

"A lot of chest physicians, I believe, would welcome a way of helping

patients up front decide whether they should have this test in the first place."

First, Bach had to prove there is measurable variation in risk. He turned to the Fred Hutchinson Cancer Research Centre in Seattle, which in the 1990s performed one of the best studies ever to track lung cancer development in 18,000 heavy smokers and ex-smokers. Bach used that data to

determine the effects of age, sex, smoking history and exposure to

cancer-causing asbestos.

He created a model that, while not perfect, largely accurately predicted cancer development among the Hutchinson study participants and among people being screened for lung cancer at the Mayo Clinic.

It's not foolproof, Bach cautioned. Nor does the formula say whether a person should have a CT scan. Instead, people will have a prediction of risk based on data that they can use to make health-care decisions.