|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2008.17.9341 on December 15 2008 © 2009 American Society of Clinical Oncology.
Flawed Inferences About Screening Mammography's Benefit Based on Observational DataDivision of Quantitative Sciences, The University of Texas M. D. Anderson Cancer Center, Houston, TX
Department of Public Health Sciences, University of Toronto, Toronto, Ontario, Canada
Departments of Surgery and Medical Humanities, University College, London, United Kingdom
Department of Epidemiology; and US Cochrane Center, Johns Hopkins University, Baltimore, MD
Department of Ambulatory Care and Prevention, Harvard Medical School, Boston, MA
Nordic Cochrane Centre, Copenhagen, Denmark
Department of Epidemiology and Clinical Medicine, Ecole des Hautes Etudes en Sante Publique (EHESP School of Public Health), Rennes, France
Department of Pathology, Ullevål University Hospital, Oslo, Norway
The Dartmouth Institute for Health Policy & Clinical Practice, Hanover, NH; and the VA Outcomes Group, White River Junction VA Medical Center, White River Junction, VT
Department of Health Sciences, University of Leicester, Leicester, United Kingdom
Norwegian Institute of Public Health, Oslo, Norway To the Editor: A recent article by Badgwell et al1 finds that among breast cancer patients at least 80 years of age, the 5-year breast cancer–specific survival was 82% among women who were nonusers of screening mammography, as opposed to 88% among those who were irregular users of screening and 94% among regular users. The investigators are cautious about the implications of this observation: "Our findings add to the accumulating evidence that the use of regular mammography may be beneficial for older women."1(p6) However, they are not cautious enough. Observations such as theirs are expected with any screening program and cannot address benefit. The authors point out that their study was subject to a healthy-person bias. While true, the real culprits are lead-time and length biases, which they do not mention. These biases are elementary and fundamental in cancer epidemiology. But they seem to have been unknown also to the Journal's editors who handled the manuscript—had they understood them, they could not have accepted the manuscript for publication. Lead-time bias is the easier of the two to understand. A woman whose cancer is detected n years early through breast cancer screening will live n years longer after her tumor is discovered. The pure bias of n years adds to the survival time of all women whose tumors were detected by screening. Because of the heterogeneity of the disease, the value of n is highly variable and is unknown for any particular woman. The average of n is also unknown; it is commonly estimated to be 3 to 5 years and is almost certainly greater in older women who generally present with low-grade hormone receptor–positive disease. "Length" in length bias refers to the tumor's presymptomatic period when it is mammographically detectable, called the sojourn time. Aggressive tumors have shorter sojourn times because they grow faster. Indolent tumors have longer sojourn times. Screening finds tumors in proportion to the lengths of their sojourn times. (Here is an analogy: when you look into the sky and see a shooting star, it is more likely to be one with a longer arc. Or, when you select a potato chip from a newly opened bag, it is more likely to be a bigger one.) Screening preferentially selects tumors with longer sojourn times; therefore, tumors detected through screening are slower growing and less lethal. This is length bias. A special case of length bias is overdiagnosis, when screening finds a tumor that grows sufficiently slowly that it would not have caused symptoms or death in the woman's lifetime. Overdiagnosis and its sequelae are among the harms attributable to screening, which the authors do not consider at all. Overdiagnosis occurs at all ages, but it is pronounced in older women for two reasons. One is competing risks: older women are likely to die as a result of other causes. The other reason is that tumors arising in older women tend to be less aggressive (especially when detected mammographically) than those found in younger women. Lead-time and length biases give rise to the stage shift associated with screening mammography that is always observed in studies such as that of Badgwell et al.1 Quantifying these biases precisely is difficult, if not impossible. This is why investigators in several countries have conducted randomized screening trials. None of the trials included women in their 80s. The oldest women included in the screening trials (in Sweden) were in their 70s. Because the sample size was small, the results were far from conclusive. But women in this subset who were not screened actually had lower breast cancer mortality than those who were screened. In sum, credible evidence of a beneficial effect of breast cancer screening for older women does not exist, but there is solid evidence of harms. The article by Badgwell et al1 might have had minimal impact on women were it not for the news media picking up the story. Coverage was predictable and considerable. The media painted stronger conclusions than the authors may have intended. The American Society of Clinical Oncology fanned the flames with its totally inaccurate and misleading press release entitled, "Women 80 and Older Benefit from Mammography, but Few Are Screened."2 The Journal of Clinical Oncology erred in publishing this article. It was a disservice to women, young as well as old. Such publication reveals a seriously defective editorial process at the journal. AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES
1. Badgwell BD, Giordano SH, Duan ZZ, et al: Mammography before diagnosis among women age 80 years and older with breast cancer. J Clin Oncol 26:1–8, 2008. 2. American Society of Clinical Oncology. Women 80 and older benefit from mammography, but few are screened. http://www.asco.org/.
Related Articles
Related Editorial
Related Reply
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2009 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
