CHICAGO (AP) -- A new study suggests that screening smokers and former smokers for lung cancer doesn't save lives or prevent advanced disease and may lead to unneeded and harmful treatment. But it's not the final word on CT scans.

Some experts have hoped that the scans, which are a special kind of X-ray that can detect tiny lung abnormalities, will prevent lung cancer deaths by getting people into treatment earlier. But there hasn't been convincing evidence of that.

A large and authoritative scientific study won't be completed for a few years. Without that evidence, the American Cancer Society doesn't recommend the test, which costs $300 to $400, and most insurance companies won't pay for it.

The latest research, appeared in Wednesday's Journal of the American Medical Association, analyzed lung cancer deaths and cases of advanced lung cancer among 3,246 smokers and former smokers who had annual CT scans for about four years.

Researchers compared deaths and advanced cancer cases with rates predicted by a mathematical model. The model — based on a person's age, gender, asbestos exposure and smoking history — has proven valid in previous studies.

The model predicted there would be 33.4 cases of advanced cancer; there were 42. The model predicted 38.8 lung cancer deaths; there were 38.

"We don't see a trace of evidence that a single life was saved, that a single case of advanced cancer was avoided," said study co-author Dr. Peter Bach of Memorial Sloan-Kettering Cancer Center in New York.

CT screening did increase diagnosis and treatment. The people screened were three times more likely to be diagnosed with lung cancer and 10 times more likely to have lung surgery than predicted.

Bach said it's likely the cancers found by scans were so slow-growing they might never have caused death. Meanwhile, scanning missed the fast-growing, fatal cancers, which popped up between the annual scans.

"It's like taking a photo of the night sky when you're looking for a shooting star," Bach said.

Because CT scanning led to more biopsies and surgeries, patients were at risk of complications such as lung punctures, bleeding and infection, Bach said.

"Getting screened for lung cancer with CT scanning is not only unproven, it's potentially a risky endeavor," he said.

Lung cancer is the leading cause of cancer death in the United States and worldwide. About 213,380 Americans will be diagnosed with lung cancer in 2007, and the disease is expected to kill 160,390. Smoking is the main culprit.

Last year, another study was more encouraging about the potential for CT screening. It found that people whose early tumors were detected and promptly removed had an estimated 10-year survival rate of 92 percent.