| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Harvard Medical School Massachusetts General Hospital Boston, MA 02114
As someone who has spend the past 25 years learning about and understanding lead-time bias, length bias sampling, selection bias, publication bias, and all of the biases and issues that have caused controversies in breast cancer screening [1], I have been astonished and amused that the concepts involved in determining the efficacy of other imaging screening tests (such as those for lung cancer or colorectal cancer) are just being "discovered." To those of us involved in breast cancer screening it is "déjà vu all over again"!
First came the article by Obuchowski et al. [2] in which the authors dismissed the experience with mammographic screening as not being a good model of a screening test. Now Robert Stanley, in his 2001 ARRS presidential address [3], mentions mammographic screening, only in passing, as an example of the risks of finding pseudodisease. As the saying goes, "What are we? Chopped liver?"
Stanley was absolutely correct in warning radiologists that screening is different from diagnosis. What he neglected to mention was that Moskowitz [4] explained this more than 20 years ago. The issues involved with tests that screen healthy individuals are, fundamentally, different from those of tests used to evaluate individuals who are ill. It is well established that merely finding cancer earlier does not mean that the test actually saves lives or even alters the course of the disease. All of the factors involved in validating screening tests are so well understood in the breast imaging community that they are thoroughly described in textbooks [5]. Although many of us would like to use surrogate end points to validate screening tests, the only way to prove the efficacy of these interventions is through carefully designed and executed randomized, controlled trials that have sufficient power to answer the questions being asked [6]. Rather than "reinventing the wheel," those who are interested in screening tests might want to familiarize themselves with the breast imaging literature. All of the issues have been thoroughly explored. In the words of the philosopher George Santayana, "Those who cannot remember the past are condemned to repeat it."
References
Cleveland Clinic Foundation Cleveland, OH 44195-2980
With 14 references on lead-time bias, length bias, pseudodisease, and the like, dating back to the 1960s, it is difficult to construe our article [1] as "reinventing the wheel." Those involved in breast cancer screening know well the scientific importance of these biases and related issues. Our article simply applies these well-established ideas to multidetector CT for pulmonary and colorectal screening, using mammography as a point of reference, in hope that the science will precede the lust for profit from multidetector CT screening.
References
University of Alabama at Birmingham Birmingham, AL 35249-6830
Kopans' letter regarding my presidential address [1] serves to emphasize my point that screening is indeed different from diagnosis. Experts in mammography have long been aware of the differences and have cautioned their colleagues in breast imaging accordingly. However, the concepts of potential biases inherent in screening seem to have been forgotten or overlooked by some, particularly those purveyors of whole-body CT of healthy, asymptomatic adults.
My exposure to some of the issues in screening is partly derived from the works of Cole and Morrison [2], clinical epidemiologists who were developing these concepts in the 1960s and 1970s. Cole is a professor emeritus in the School of Public Health at the University of Alabama at Birmingham, my home base. Additional useful insights were provided by Black et al. and Welch et al. [3,4,5,6], who have presented many of the bias concepts in a format I could understand. Because I was familiar and comfortable with these works, they were cited in the article. The omission of works by experts in mammography wasn't intended to diminish their contributions.
My presentation was an effort to remind the audience of lessons learned over the years—that is (to borrow from Kopans), that the issues involved with tests that screen healthy individuals are fundamentally different from those of tests used to evaluate individuals who are ill. Likewise, merely finding cancer earlier does not mean that the test actually saves lives or even alters the course of disease. The proliferation of whole-body CT screening centers proves that these messages bear repeating.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Berlin Breast Cancer, Mammography, and Malpractice Litigation: The Controversies Continue Am. J. Roentgenol., May 1, 2003; 180(5): 1229 - 1237. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
