Dr. Stephen Feig, director of breast imaging at Mount Sinai Hospital in Manhattan, said that the ultrasound findings were impressive, and that ''anything we can do to increase breast cancer detection we should do if it is feasible.''
But Dr. Feig said he was concerned about the limited nature of the study and the time and expense of screening a wider population with sonograms, which cost about $150, roughly the same as mammograms.
''We need to learn more before screening breast ultrasound becomes standard of care,'' Dr. Feig said, adding that a multicenter trial of such screening was being organized.
It is well known that mammograms, even at their best, miss a lot of cancer, up to 25 percent in women under 50. Young women tend to have dense breasts, which appear white on mammography, making it harder to differentiate small tumors.
Women with very dense breasts, composed primarily of fibrous tissue and milk-producing glands, have four times the risk of developing breast cancer as women of the same age with fatty breasts. Experts suspect that dense tissue means more cells that could potentially become cancerous, but no one is sure.
Research has shown that breast density appears to decrease with age and the number of children a woman bears, and increase with hormone replacement therapy. Because breast size and texture are not reliable predictors of breast density, many women are unaware that they have dense breasts. Mammography is the only accurate test.
In the New England Journal study published today, researchers performed mammograms on 571 identical twins and 380 fraternal twins ages 40 to 70 in Australia, Canada and the United States. They looked for correlations in the degree of breast density within pairs of twins.
Identical twins have identical genes, so that any differences in breast density between them would most likely be because of environmental factors or measurement error. Fraternal twins, on the other hand, on average share only half their genes. The researchers found a much stronger correlation in breast density between identical twins than fraternal twins, suggesting that the contribution of genetic factors to breast density is 60 to 75 percent.
An unsettled question is how women with dense breasts should be screened. Ultrasound has been widely used to evaluate suspicious abnormalities on a mammogram, but whether it should be used as a screening tool for the wider population has been a matter of debate. Several studies in the 1980's showed little benefit of screening breast ultrasounds, and in 1994 the American College of Radiology said that sonograms should not be used to screen for breast cancer.