Mammography gives no added value in breast cancer screening
9 March 2010
A multicentre study of breast cancer screening has concluded
that mammography is unnecessary in women undergoing MRI and current
guidelines should be revised to reflect this. It found that MRI is
essential for surveillance of women at high risk and also for women at
moderately increased risk of breast cancer.
The multicenter cohort study is published in the Journal of
Clinical Oncology .
Between 2002 and 2007, the EVA trial recruited 687 women who
carried a moderately increased risk of breast cancer (lifetime risk
of 20% and over). Women underwent 1679 screening rounds consisting
of annual MRI, annual digital mammography and half-annual screening
ultrasound examinations. During this time span, 27 women received a
new diagnosis of invasive cancer or DCIS (ductal carcinoma in situ).
Of all imaging methods under investigation (digital mammography,
ultrasound and MRI), MRI offered by far the highest sensitivity: MRI
identified 93% of breast cancers. 37% of cancers were picked up by
ultrasound. The lowest sensitivity was achieved by digital
mammography, which identified only one-third of breast cancers
These results confirm once more that MRI is essential for
surveillance not only of women at high risk, but also for women at
moderately increased risk of breast cancer. Moreover, the results
contradict current guidelines, according to which mammography is
considered indispensable for breast cancer screening.
One aim of the EVA trial was to question this concept and to ask
whether it is still appropriate to require that MRI should only be
used in addition to mammography. The results show that if an MRI is
available, then the added value of mammography is negligible.
Researchers conclude that MRI is necessary as well as sufficient for
screening young women at elevated risk of breast cancer. Since
mammography appears to be unnecessary in women undergoing MRI, its
use is no longer justifiable, and current guidelines should be
revised to reflect this.
Current guidelines for women at high familial risk of breast
cancer recommend annual MRI (with or without ultrasound) and annual
MRI starting at age 25-30.
“These guidelines were set up based on little or no scientific
evidence, and mainly reflect expert opinion”, says Prof. Christiane
Kuhl, radiologist at the University of Bonn and principal
investigator of the EVA trial.
“In the light of the results of the EVA trial, such
recommendations should be revisited. The radiation dose associated
with regular mammographic screening is clearly acceptable and safe”,
underscores Kuhl. “However, regular mammographic screening usually
starts at age 40-50”. The situation is different if systematic
annual mammographic screening is started at age 25-30.
“Not only because these women will undergo more mammograms and
therefore will experience a cumulative lifetime radiation dose that
will be substantially higher, but also because the breast tissue of
young women is more vulnerable to the mutagenic effects of
This appears to be especially true for BRCA mutation carriers.
“Accordingly, we impose more radiation on less radiation-tolerant
breast tissue — for a very limited, if any, diagnostic benefit”.
Therefore, Kuhl advocates a revision of existing guidelines: “It is
no longer justifiable to insist on annual mammographic screening of
women in their thirties if they have access to screening MRI”.
In the past, MRI was used strictly in addition to mammography
only. The allegedly high rate of “false positive” diagnoses and the
allegedly insufficient sensitivity for DCIS were the main reason to
discourage its use as a stand-alone method for breast cancer
“In this multicenter trial, with basic quality assurance
implemented not only for mammography, but also for MRI, we were able
to prove that false positive diagnoses are avoida ble if MRI studies
are interpreted with adequate radiologist expertise,” said Dr Kuhl.
In the EVA cohort, the Positive Predictive Value achieved with
MRI was already even higher than that of mammography or breast
ultrasound. “Moreover, we found that MRI offered the highest
sensitivity especially for DCIS”, adds Dr. Kuhl. “It is simply wrong
to state that we need a mammogram to detect intraductal cancer”
1. Christiane Kuhl, Stefanie Weigel, Simone Schrading, Birke
Arand, Heribert Bieling, Roy König, Bernd Tombach, Claudia Leutner,
Andrea Rieber-Brambs, Dennis Nordhoff, Walter Heindel, Maximilian
Reiser, and Hans H. Schild. Prospective Multicenter Cohort Study to
Refine Management Recommendations for Women at Elevated Familial
Risk of Breast Cancer: The EVA Trial. Journal of Clinical
Oncology. Published online February 2010. doi: