The way that data have been presented in several previous studies of mammography results has led to incorrect conclusions about the usefulness of screening women in their 40s, according to a report from the Massachusetts General Hospital (MGH). The article appearing in the May-June 1998 issue of The Breast Journal explains that a year-by-year analysis of the rates of cancer detection in women undergoing screening mammography shows a gradual, steady increase in the number of cancers detected in women ages 40 to 79, with no sudden increase at the age of 50. In contrast, some earlier studies that grouped women by larger age ranges -- for example, comparing all women in their 40s with all women in their 50s -- gave an erroneous impression of an abrupt change in the effectiveness of mammography at age 50.
"Our analysis of five years of screening mammography results supports what we've been saying all along -- there's no big difference between women in their 40s and women in their 50s regarding the usefulness of mammography in detecting breast cancer," says Daniel B. Kopans, MD, MGH director of Breast Imaging and the study's lead author. "We also found that the percentage of screened women who had abnormal mammographic results -- shown by the numbers who were recalled for additional evaluation or referred for biopsy -- was steady thoughout the entire population of women screened, again with nothing changing at 50."
Since it is well known that breast cancer incidence increases with age, the authors write, comparing screening results of large, combined age groups will skew the results toward the extremes. For example, the cancer rate for a group of women in their 30s and 40s would be pulled down by the very low incidence among women in their early 30s. Similarly, the overall rate for a group of women in their 50s and 60s would be raised because of the much greater incidence among women in their late 60s. (Many recent studies have used such large age groupings.)
Kopans and his colleagues reviewed data from more than 72,000 screening mammograms done at the MGH from October 1990 through October 1995. Women screened ranged in age from 40 to 79, and the researchers analyzed the percentages of women who were recalled for additional examination, women who were referred for biopsy, cancers detected by mammography-initiated biopsy and overall cancers detected.
When analyzed on a year-by-year (specific age) basis, results for cancers detected showed a steady increase across the age range studies, with no abrupt increase at or around 50. Percentages of biopsies detecting cancer rose between 0.6 and 0.8 percent each year, from a low of 15 percent at 40 to 43 percent at 79. The percentage of screened women who were called back for additional evaluation decreased steadily from 8 percent down to 5 percent as the women aged; similarly the percentage of women for whom biopsy was recommended showed no age-related change, with slight fluctuations between 0.5 percent and 2 percent. Overall the incidence of mammogram-based additional evaluation and biopsy remained steady throughout the entire group regardless of age, indicating no abrupt change in the rate of false-positive mammograms.
In 1993 the National Cancer Institute withdrew its recommendation for mammography screening of women in their 40s, a decision that generated a continuing controversy. Although the NCI recently reversed the earlier decision and once more supports screening women in their 40s, others disagree. Kopans and his colleagues have consistently advocated for the usefulness of screening mammography in women aged 40 to 49.
The MGH researchers note that several previous studies cited by opponents of screening women in their 40s rely on the sort of broad-age-range grouping that can produce incorrect conclusions. The MGH researchers' own data, if grouped by decades, would show a percentage of cancers detected by mammogram-initiated biopsy of 16.7 percent among women in their 40s and 28.9 percent in women aged 50 to 79. But year-by-year analysis shows cancer detection in 20 percent of biopsies among 49-year-old women and only 20.5 percent of biopsies in 50-year-old women. The very act of chosing 50 as a point of comparison, they write, leads to the impression of a drastic shift at that age.
The authors concluded that the age of 50 has no biological significance in determining when regular mammographic screening should begin and that investigators who publish data on screening in the future should note that age grouping may give misleading and skewed results.
Kopans coauthors are Richard Moore, Kathleen McCarthy, MD; Deborah Hall, MD; Priscilla Slanetz, MD, and Elkan Halpern, PhD, of the MGH Department of Radiology; along with Carol Hulka, MD, and Gary Whitman, MD, of the Departments of Radiology at Beth Israel Deaconess Hospital in Boston and M.D. Anderson Cancer Center in Houston, respectively.