| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Radiology, Hospital Santiago Apóstol, Olaguíbel 29, 01004 Vitoria, Spain. e-mail: eanorbe@euskalnet.net
Editor:
As general radiologists involved in screening mammography, we read with great interest the editorial by Dr Jackson in the November 2002 issue of Radiology (1). We would like to comment on certain points.
The author highlights the decrease in mortality from breast cancer, which she attributes mainly to early diagnosis with little influence from improved treatment. This is the reverse view from that of other authors (2).
In the editorial, it is stated that it is difficult to carry out a large clinical experiment, and flaws will always be produced, however small. It seems logical to think that the big studies performed years ago on the usefulness of mammography would also have had some errors. What we general radiologists find hard to understand is that the experts do not come to an agreement on the importance of these biases and their implications for the final result. Thus, studies accepted by some authors are rejected by others (3–5). Curiously, the Malmö mammographic screening trial (6), which is apparently accepted as unbiased (3,5), was strongly rejected in its day by another expert (7).
The mortality rate for breast cancer has been considered to be the correct end point. However, this end point presents various problems that arise from the long-term effects of treatment given to patients with breast cancer (8,9). The "sticky-diagnosis" bias and the "slippery-linkage" bias support the use of all-cause mortality as the end point (10).
We also disagree with the opinion of the author of the editorial that articles that she considers to be biased should not be published in prestigious medical journals because of the potential harm they may cause to the population. Medical journals are precisely where opinions should be given and discussed on medical subjects, independent of the repercussions they may have in the news media. In fact, it has been demonstrated that the news media provide information in an unsuitable and incomplete form, which may also have an influence on patients (11).
Dr Jackson points out the reduction in mortality from breast cancer in the Malmö mammographic screening trial (6), which Miettinen et al establish (in women over 55 years) (5), but she ignores the fact that during the first 6 years, mortality was higher in the group invited to be screened than that in the control group. This means there is an increase in mortality in women who undergo screening when they are younger.
In the editorial, there is no comment on the damage associated with screening described by the authors of the Malmö mammographic screening trial (6), who report the diagnosis of one clinically insignificant cancer for each two deaths from breast cancer in women younger than 50 years (12). Mammographic screening, even if it saves lives, still has negative effects: the diagnosis of clinically insignificant cancers.
When the data from the different randomized controlled studies are analyzed, the risk reduction is emphasized without referring to the absolute values, which are shown by some more modest results: one death avoided for each 1,000 women who take part in a program for 6
years (4,13). The rest of the women would have no advantage, and many of them would even be harmed by the follow-up and false-positive findings.
Perhaps the women who come forward for early detection in breast cancer programs should have all these data explained to them in an easily comprehensible way, thus avoiding the sensationalist and mistaken information published in the news media.
REFERENCES
Department of Radiology, Indiana University School of Medicine, 550 N University Blvd, Room 0663, Indianapolis, IN 46202. e-mail: vjackso@iupui.edu
I thank Drs Añorbe and Aisa for their comments regarding my recent editorial in Radiology (1). This is an area of great emotion and confusion, and the authors point out some frequent areas of misunderstanding.
Drs Añorbe and Aisa state that I attributed the decrease in mortality from breast cancer "mainly to early diagnosis with little influence from improved treatment." My actual statement was that "while some of this mortality reduction has undoubtedly been due to improved treatment of breast cancer, much of the reduction appears to be due to earlier detection of the disease with screening mammography." This is not a statement that little of the reduction is due to treatment, but rather an acknowledgment that while we don’t know the exact contributions of early detection and treatment to changes in mortality, there are data to suggest that screening contributes a large component. Duffy et al (2) examined the contribution of detection and treatment in their recent publication in Cancer, in which they estimated that breast cancer mortality is reduced by 40%–45% with mammographic screening, independent of treatment. In addition, it has already been well established that smaller breast cancers (often detectable only with screening mammography) usually have a better prognosis than that of larger malignancies.
Drs Añorbe and Aisa state that "it seems logical to think that the big studies performed years ago on the usefulness of mammography would also have some errors." I agree completely and stated that in my editorial. It is frustrating to know that it has been, and will continue to be, virtually impossible to perform a flawless randomized control trial of screening mammography. Many people are understandably confused about the lack of agreement by experts on the various trials. When one looks at the preponderance of evidence, however, the trials with the fewest flaws show a statistically significant mortality benefit from screening mammography (3–9).
The authors raise the issue of all-cause mortality as a more appropriate end point than breast cancer mortality for these trials. Most experts believe that breast cancer mortality is the most appropriate end point. This is partly because mortality from breast cancer usually occurs long after the time of diagnosis and treatment. During that period, death may result from many circumstances that have nothing to do with breast cancer or its treatment. For example, if a woman is killed in an automobile accident 1 year after receiving a diagnosis of breast cancer, should that count as mortality related to breast cancer? Obviously not. In almost all studies in medicine, experts agree that it is most appropriate to look at mortality related to the disease being investigated rather than all-cause mortality.
Drs Añorbe and Aisa "disagree with the opinion of the author of the editorial that articles that she considers to be biased should not be published in prestigious medical journals." I had mentioned that this was a common theme in letters to the editor in Lancet following publication of the article by Gotzsche and Olsen (10). I have no problem with publication of controversial articles in the literature, but it is not right to publish flawed or biased articles in highly respected peer-reviewed journals. The purpose of peer review should be to minimize the chance that erroneous information is published in the literature. Many, if not most, experts consider the Gotzsche and Olsen articles (10,11) to be seriously flawed and biased.
Drs Añorbe and Aisa point out that the absolute number of women who will benefit from screening mammography is small. This is partly because breast cancer affects a small percentage of the population. For women whose cancers are detected at an early stage with use of screening, there are benefits over and above the chance of a reduction in mortality. Women with small tumors have more treatment options and are less likely to need highly aggressive therapy. In addition, treatment is more likely to be successful for smaller tumors. It is likely that the patient’s quality of life will be improved, not just her length of life.
I certainly agree with Drs Añorbe and Aisa that women deserve honest information regarding the value of screening mammography or any other medical test or procedure they might undergo. However, this should be done in a calm, balanced, and factual manner. When prominent medical journals publish flawed material that is subsequently portrayed in a sensational manner in the lay press, we all lose.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
