False positives plague x-ray lung cancer screening

The lung cancer screening arm of a comprehensive National Cancer Institute trial has found that chest x-rays can detect early lung cancer in asymptomatic people. The downside, however, is that the trial produced many false positives.

Between 1993 and 2001, the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial enrolled 154,942 men and women 55 to 74 years of age. These participants included current and former smokers, as well as individuals who had never smoked.

In the study, which appeared in the Dec. 21 Journal of the National Cancer Institute, Dr. Martin M. Oken and colleagues summarize lung screening data from the interventional arm (those who received chest x-rays) of the PLCO.

Of the 67,038 men and women who received a baseline chest x-ray, 5991 (8.9%) were positive and required additional evaluation. The majority of follow-up tests were chest x-rays, while 23% were CT scans.

Overall, 126 (2.1%) of the 5991 participants with a positive baseline screen were found to have cancer within 12 months.

"The positive predictive value was low," said Dr. Christine Berg, the NCI investigator who leads the PLCO trial. "That means there were a lot of false positives on the initial x-rays. If you get a positive result from a chest x-ray, the message is don't panic."

Berg is not an author of the current study.

Of the cancers detected, 44% were stage I, a higher rate than seen in the general community, according to Berg. She cautioned, however, that this finding does not automatically translate into a mortality benefit.

The five-year survival rate in stage I non-small cell lung cancer is 50% to 70%, which is considerably higher than the rate in stage II NSCLC. These more advanced lung cancers are usually fatal within two years of diagnosis.

"If a beneficial stage shift does occur, it would be reflected in the central analysis in this trial, the comparison of lung cancer mortality among participants in the screening arm with that among participants in the control arm. We will report these data at a later date when sufficiently mature," the authors wrote.

Many lung cancer screening trials have been conducted, and screening has consistently detected more stage I cancers compared with no screening. Screened patients also show longer survival compared with that of patients diagnosed through usual care, but lead time bias and overdiagnosis bias explain, at least in part, these apparent benefits, according to Oken.

It is possible that the small size of previous studies as well as difficulties in the interpretation of the data resulted in the missing of a small but important benefit from the annual chest radiograph. The PLCO trial has enough power to determine if screening offers a benefit over no screening.

An important feature of this study is that it includes 39,115 women. The PLCO is the first major controlled study to evaluate screening for lung cancer in women.

Comparing the results of women and men, researchers found that the incidence of positive screens was lower among women (8.2%) than among men (9.6%). This trend was found in every age group, in current and former smokers, in never smokers, and in smokers with a history of fewer than 30 pack-years or of 30 or more pack-years.

Smokers in the trial (both men and women) showed an increased risk for lung cancer. Among current smokers, 6.3 people per 1000 screened were diagnosed with lung cancer, compared with 4.9 per 1000 for former smokers (those who have quit for fewer than 15 years). Among never smokers, the detection rate was 0.4 per 1000.

For more information from the Diagnostic Imaging archives:

Celebrity death increases awareness of lung cancer

Staging criteria for lung cancer face update

Digital radiography challenges CT-based lung cancer screening

CT exams help smokers kick the habit