Study Finds No Abrupt Change at Age 50 Of Mammography Yield on Breast Biopsy

BOSTON--Previous study results showing that screening mammographyis beneficial only in women over age 50 may stem from inappropriategrouping of age-based data rather than an actual change in benefitat age 50, Daniel B. Kopans, MD, and his colleagues assert ina new study.

"Our study, and careful review of other studies, shows thatthere is no abrupt change in the diagnosis of breast cancer detectedby mammography at any age, but rather a steady increase in detectionas one would expect, since the incidence of breast cancer increaseswith age," Dr. Kopans says.

Based on an analysis of data from the Breast Cancer ScreeningProgram at the University of California at San Francisco, Kerli-kowskeet al recommended screening mammography for all women aged 50or older but only for high-risk women aged 40 to 49 (JAMA 270:2444-2450,1993). However, Dr. Kopans points out that the authors reachedthis conclusion by comparing data from women aged 30 to 49 withthose from women aged 50 to 70 or more.

This inappropriate grouping of extremes, he says, skewed the results,causing the appearance of an abrupt change at age 50 instead ofa continuum.

Such inappropriate groupings are generally done because thereare not enough women in the study to analyze the data by smallerage increments. In the San Francisco study, women aged 30 to 39who have a very low prevalence of cancer were grouped with thoseaged 40 to 49, while women aged 70 and older who have a very highprevalence were grouped with 50 and 60 year olds.

But even grouping women into ages by decades is too broad to showthe actual gradual increase in the ability of mammography to findcancer with increasing age, Dr. Kopans argues.

The Mass General Study

In Dr. Kopans' study from the Massachusetts General Hospital andHarvard Medical School, where he is associate professor of radiology,the researchers reviewed the results of 4,778 biopsies performedbetween 1978 and 1995 as a result of abnormalities found onlyon mam-mography examination.

The mammograms were a mix of screening and diagnostic tests, butall lesions were clinically occult and detected only at mammography.

The women ranged in age from 40 to 79, with those aged 40 to 49accounting for 26% of the total; those aged 50 to 59 for 30%;those 69 to 69 for 28%; and those 70 to 79 for 16%.

The analysis showed that the positive predictive value for thebiopsies increased steadily with increasing patient age, startingat approximately 12% at age 40, and rising 0.8% to 1% each year,to reach a value of 46% at age 79.

Although the data for women older than 50 years differed significantlyfrom the data for those younger than 50, Dr. Kopans says, theanalysis showed no evidence of any abrupt change at age 50 yearsthat is independent of a steady change (Radiology 200:357-360,1996).

He concludes that researchers should be cautious in grouping datawhen analyzing phenomena that change gradually. "They shouldcease using age 50 years as a point of analysis, since it mayresult in a serious misrepresentation of the facts," he says.

Dr. Kopans colleagues from the Department of Radiology were RichardMoore, Kathleen McCarthy, MD, Deborah Hall, MD, Carol Hulka, MD,Gary Whitman, MD, Priscilla Slanetz, MD, and Elkan Halpern, PhD.