Contact Michael Bernstein (703) 648-8910 Charleta Mason (703) 716-7578[email protected]
For Release September 19, 2000
ACR Dismiss Canadian Screening Study, Urges Women to Continue Having Mammograms
A Canadian study published today in the Journal of the National Cancer Institute concludes that adding mammography for breast cancer screening to physical examinations does not reduce deaths. This study is badly flawed because it is based on poor quality mammograms that were read by radiologists with no specific training in mammography and because of the basic design of the study.
The American College of Radiology (ACR) in leveling this strong criticism at the study today, said the study authors used the same flawed methods in following women 50-59 as they did in a previously published, highly criticized study of women 40-49.
In firmly rejecting the current study's conclusions, the ACR urged women to continue to have mammography screening every year, starting at age 40. For the first time in 40 years, and for the past several years in a row as reported by the National Cancer Institute, breast cancer death rates have declined, in large part due to increased use of screening mammography. "It would be tragic if women were persuaded not to have screening mammography based on the highly questionable results of this latest Canadian National Breast Screening Study-2 (CNBSS-2)", the College said. Other, well conducted, screening trials have documented that mammography does indeed save lives.
In addition to major quality issues, the study design involving volunteers, can not be used to draw conclusions about a general population. The authors correctly point out that the death rate in their control group is far lower than in the average Canadian population.
This was also true in their previously published study of women ages 40-49, where the control group did not have repetitive screening. Their greater than 90% five year survival among women ages 40-49 strongly suggests a randomization imbalance.
In the newly released study of women 50-59, the extremely small percentage of lymph node positive cancers (13%) among women receiving only the Clinical Breast Examination (CBE) is unheard of in any other study raising additional concerns about the allocation of patients in the study groups. The extremely low mortality rate of 105/610 (17%) at 10 years among this same group of women is also unprecedented.
Another major flaw of the study also pertains to the study design. The authors' original bias is evident in the fact that, instead of comparing the benefit of clinical breast examination in addition to mammography to mammography alone, as one would expect with the "gold standard of care", the investigators asked whether mammography added any additional information to clinical breast examination.
The most recent data confirm the poor quality of mammography screening in the Canadian study. In virtually every other comparison of palpable mammography detected cancers, the majority of cancers are detected by mammography alone. Most are non-palpable, even in retrospect.
In NBSS-2, almost as many cancers are detected by clinical examination alone as by mammography and clinical breast examination, although there were more cancers expected from high quality mammography.
What is not understood by many who will read these results is that there was no special training provided for the radiologists reading these mammograms. Mammography screening was not commonly being performed in the early 1980s and radiologists did not have a great deal of experience in interpreting the studies.
Offers by skilled radiologists to train the NBSS radiologists were turned down. On the other hand, individuals who performed the clinical breast examinations in the NBSS were reported to be highly trained and skilled.
In addition to strongly criticizing the NBSS-2 study, the ACR questioned why the Journal of the NCI published such a controversial paper without an accompanying editorial.
###
RSS