For Immediate Release: August 16, 2000
Contact: Hali Wickner (603) 650-1520 Print Version

Cancer Screening Can Confuse Statistics

Hanover,NH -- Overdiagnosis wreaks havoc with cancer survival statistics and can be more harmful than helpful, a Dartmouth Medical School physician cautions. Writing in an editorial, "Overdiagnosis: an Unrecognized Cause of Confusion and Harm in Cancer Screening," William Black, M.D., professor of radiology and of community and family medicine, comments on a long-term study that showed no mortality reduction in a lung cancer screening trial.

A 20-year follow-up of the Mayo Lung Project (MLP) found no reduction in lung cancer mortality among men who had been offered intense screening compared to those who had not, suggesting that some lung cancers detected through screening have limited clinical relevance. The study results and the editorial are published in the Aug. 16 issue of the Journal of the National Cancer Institute.

"Because overdiagnosis effectively changes a healthy person into a diseased one, it causes overestimations of the sensitivity, specificity and positive predictive value of screening tests and the incidence of disease," writes Black.

The MLP was a randomized, controlled clinical trial of lung cancer screening conducted between 1971 and 1983. No lung cancer mortality benefit was evident in the screening group as of July 1, 1983. Now, Pamela Marcus, Ph.D., of the National Cancer Institute (NCI), and colleagues have extended the follow-up through 1996 -- a median follow-up of 20.5 years.

The MLP, funded by the NCI and conducted by the Mayo Clinic, Rochester, Minn., involved 9,211 male smokers. In the intervention arm 4,618 men were offered (and reminded to receive) a free chest x-ray and sputum cytology every four months for six years. During the six-year screening period, compliance with the scheduled testing averaged 75 percent. Participants in the intervention arm were contacted annually by mail after completing their six years of screening to identify potential lung cancer diagnoses and deaths. The 4,593 participants in the usual-care arm received, at trial entry, the Mayo Clinic1s standard 1970 recommendation to have an annual chest x-ray and sputum cytology. These participants also received the same annual questionnaire as the intervention participants.

A National Death Index (NDI) search and matching algorithm brought the lung cancer death totals to 337 in the intervention arm and 303 in the usual-care arm as of Dec. 31, 1996. The lung cancer mortality rate through 1996 was 4.4 deaths per 1,000 person-years in the intervention arm and 3.9 deaths per 1,000 person-years in the usual-care arm, rates that do not differ significantly statistically. However, participants with lung cancer in the intervention arm had longer survival after diagnosis than those in the usual-care arm. The investigators conclude that the presence of a long-term survival difference in the absence of any lung cancer mortality benefit suggests that some tumors with limited clinical relevance were identified.

Black says the Marcus study provides compelling evidence that a major reduction in lung cancer mortality was not missed in the MLP. He writes that the discrepancy seen between survival and mortality is most likely the result of overdiagnosis, the diagnosis of small lesions that would not have become symptomatic before the individual died of other causes.

Both he and the study researchers point out that these findings add to the controversy surrounding the use of low-dose spiral computed tomography (CT) as a lung cancer screening test. Spiral CT, "far more sensitive than chest radiography," can identify lung cancer lesions before the onset of symptoms, so the potential for overdiagnosis and false-positive results will be even greater, he warns. "It is essential that there be some mechanisms in the screening process to minimize these side effects." To thoroughly evaluate the risks and benefits of this promising technique will require randomized, controlled trials and close monitoring of all causes of mortality.

Furthermore, Black concludes that a balanced presentation of the potential benefits and risks -- including overdiagnosis -- should be made to all prospective screenees to ensure that they can make an informed decision about being screened or enrolled in a randomized trial of screening.

For further information on the study, contact: NCI Press Office, (301) 496-6641.

Hali Wickner

Return to News Archive for July 2000 - Dec. 2000