Sept. 3, 2002 — Editor's Note: The controversy over
breast cancer screening rages on, fueled by results from
the Canadian National Breast Screening Study reported in
the Sept. 3 issue of the Annals of Internal Medicine. This study showed
that mammography plus self-examination and physician examination
of women in their 40s did not reduce the death rate from
breast cancer compared to self-examination and physician
examination alone.
In February, the U.S. Preventive Services Task Force
gave B level recommendations for mammographic screening
every one to two years for women in their 40s, although
their review in the same issue of the Annals suggests that
available evidence supporting this position is only "fair."
An accompanying editorial notes that mammograms cause harm
from increased rates of mastectomy and lumpectomy. On the
other hand, an observational study from Florence, Italy,
in the Aug. 24 issue of the British Medical Journal showed that conservative
procedures increased but radical mastectomy rates decreased
with the advent of screening.
Uniquely poised to debate this issue is Michael Baum,
MD, ChM, FRCS, FRCR, emeritus professor of surgery and visiting
professor of medical humanities at University College in
London, U.K. His letter in response to the Annals studies
was published in the same issue of the journal. He has been
a breast cancer surgeon for 30 years, he set up the National
Health Service Breast Screening Programme in the southeast
of England in 1987-1988, and he sat on the government's
advisory board on screening until 1997. He has headed randomized
controlled trials for nearly 30 years, including the CRC
breast cancer trials group, and now chairs the Arimidex/Tamoxifen
Alone or in Combination (ATAC) study, the largest breast
cancer trial worldwide, which plans to enroll 9,000 patients.
Medscape's Laurie Barclay, MD, recently spoke with him about
this issue.
Medscape: Do you have any specific comments or criticisms
concerning the Canadian and Italian studies, in terms of
methodology or conclusions?
Baum: The Canadian trial showed a negative effect
for screening, and because of that it is being subjected
to the most intense scrutiny. In fact, it has the best breast
cancer detection rate of any of the trials, and it is the
only trial to date using individual rather than block randomization.
It is intellectually dishonest for anyone to reject a trial
as rigorous and sound as this because they disagree with
the conclusions. On the other hand, the Italian study, done
by a group for whom I also have the utmost respect, showed
a decrease in mastectomy rate which was offset by a 10%
increase in the rate of all breast surgery. We can't assume
that screening was responsible for the decreased mastectomy
rate, because along with screening come heightened public
awareness and more breast cancer specialist teams. Both
tend to increase the likelihood of smaller cancers at the
time of diagnosis and of breast conserving procedures being
done rather than mastectomies.
Medscape: To what degree are cross-cultural comparisons
valid? Can conclusions based on results from one population
be used to extrapolate to a different population?
Baum: You can't draw conclusions from the Florence
study that would apply to the rest of the world. The U.S.
has the highest mastectomy rate worldwide. Overemphasizing
the results and conclusions from this study and using them
to shape screening policy elsewhere is a cheap trick.
Medscape: Are cost-benefit data available for mortality,
and if so, how do they factor into decisions regarding screening
strategies and their potential outcomes?
Baum: A few years ago, according to the British
Medical Journal, the National Cancer Institute (NCI) appointed
an independent panel which reviewed the available evidence
and advised against screening women under age 50. But they
were threatened that their budget would not be renewed unless
they changed their tune, which is a terrible scandal and
a crass politicization. Based on the available evidence,
the whole question of mammographic screening under age 50
should be disposed of. In terms of women age 50 to 59, we
need to consider the best- and worst-case scenario. At best,
screening decreases the relative risk of breast cancer specific
mortality by 25%. If we assume this to be true, we need
to screen 1,000 women for 10 years to save one life. But
those who set screening policy say we can't tell the women
that because they may not come for screening. I say so be
it. I resent the way the screening program insults the intelligence
of women. The Cochrane review by Gotzche and Olsen [Lancet.
2001;358:1340-1342] describes the worst-case scenario, which
is that screening offers no benefit on breast cancer mortality.
As they elegantly demonstrate, the trials suggesting any
benefit are seriously flawed because of both methodological
and statistical issues. The HIP trial suggesting a positive
benefit used obsolete mammographic equipment. It was a brave
attempt in its time, but it would never even be published
today. These criticisms cannot be ignored. I suspect that
the truth lies somewhere between the best- and worst-case
scenario, that the relative risk reduction in breast cancer-specific
mortality by mammographic screening is somewhere between
0% and 25% -- not high enough to justify it.
Medscape: What is the likely impact of breast screening
programs on other outcomes like mastectomy rates and rates
of conservative procedures?
Baum: The impact of screening on the mastectomy
rate is difficult to determine, but it seems clear that
screening brings about more surgery overall. In the U.K.,
one third of detected breast cancers are subjected to mastectomy,
but there is extreme variation, up to 50% in some regions.
One concern is that screening is detecting many more cases
of ductal carcinoma in situ (DCIS), which is not necessarily
a good thing. Outside a screening program, about 1.0% of
breast cancers are DCIS, but this jumps to 20.0% with a
screening program. Although the natural history of DCIS
is unknown, a conservative estimate from autopsy series
is that 50% progress to invasive disease, and some series
suggest that the actual figure is only 1 in 7. So many of
those cancers detected by the Florence series probably represent
cases of DCIS which would not have progressed if left to
their own devices.
Medscape: Based on these studies, your own experience,
and other pertinent information, what would you recommend
concerning optimal strategies for breast cancer screening?
Baum: No one in their right mind, outside the U.S.,
would offer mammography screening to women under age 50.
America has to ask herself why she is out of step with the
rest of the world. It is difficult to convince the lay public
why screening may not be a good thing, to explain concepts
like ascertainment bias. On the other hand, it seems that
any fool believes the mantra that "earlier is better," and
those who convince women to take part in mass screening
exploit this. There's no good evidence that catching a slow-growing
cancer on mammography one doubling time before it would
be picked up on physical examination translates into any
survival benefit.
I think it's time women wised up to the fact that they're
being duped. I'm not saying that no woman should have a
screening mammogram, but when she does, it should be based
on a rational decision and on informed consent. She should
weigh the benefit of one life saved per 1,000 over 10 years
of screening against the psychological risk of false alarms
and the physical risk of unnecessary surgery. The worst
lie of all is that she should be screened just for reassurance.
A negative mammogram has virtually no reassuring value,
because interval cancers that crop up between mammograms
are the ones most likely to kill you. Then there's the issue
of radiation exposure, which is probably nil for an occasional
mammogram. On the other hand, the policy of beginning annual
screening five years before the earliest age breast cancer
appeared in the family -- say at age 30 -- translates into
a lot of radiation exposure to a young breast by the time
a woman reaches age 50. Another absurd policy is to get
a "baseline" mammogram at age 35, so that there's something
to compare at age 40. The problem with this is that the
baseline study doesn't just sit in a drawer for five years;
it gets interpreted right away and the woman may end up
with unnecessary surgery at age 35 based on an equivocal
finding.
Unfortunately, in the U.S., screening is a huge commercial
industry, and in the U.K., it's a politically charged issue
guaranteed to win votes. Everyone wants to fight cancer,
so screening seems like a quick fix, but the long-term fix
will involve spending more money on treatment and research.
It's time for those in charge of health policy to do what's
best for the patient, and to put aside these other considerations.
Medscape: How receptive are U.S. and U.K. physicians
and policy makers to your views?
Baum: Often people shoot the messenger when they
don't like the message. I can speak with authority even
though Americans don't like what I say. I know about numbers
and can translate relative risk reductions into absolute
benefits. The screening zealots don't like that! So they
suggest that because I recommend against screening, that
I hate women. In fact, I love women and I'm on the side
of womankind. I have a very strong family history of breast
cancer, and I've committed my whole professional life to
trying to help women overcome this disease. Interestingly,
after I resigned in disgust from the government's advisory
board on screening in 1997 because evidence was being buried
that demonstrated that screening wasn't all that it was
cracked up to be, I was appointed to chair the government's
committee on PSA screening. So I guess that tells you something.
Reviewed by Charlotte E. Grayson, MD
.gif){kind=spacer}
