Applied Radiology

Dr. Mulligan is with the University of Chicago Hospital's Breast Imaging Center, Dr. Bailar is with the Department of Health Studies and Harris School of Public Policy, University of Chicago, and Dr. Bieber is with the Department of Reproductive Endo-crinology at the University of Chicago Hospital, Chicago, IL.

Radiologists must be aware of the many dynamic changes in the female population and understand why so many menopausal women are now using hormone replacement therapy (HRT) because it directly affects the interpretation of many mammograms. According to the U.S. Census Bureau, the total number of women in the United States is continually increasing with growth of the general population. Within this growing female population, the proportion of women older than age 55 continues to increase with each census. In 1995, 23 million U.S. women were over age 55, and by 2015 this number is projected to increase to 93 million women. Beginning in 1996, 10,000 Americans, one-half of them women, turn 50 years of age each day. ¹

Life expectancy statistics for women during the last century show interesting changes. In 1901, the life expectancy (at birth) of a woman in North America was 50 years; by 1997 her life expectancy reached 79 years. This increased life expectancy is the result of several factors, most notably improved sanitation, nutrition, housing, medical care, decreased infant mortality, and antibiotic and vaccine discoveries of the 1930s. In 1901, only 6% of women were postmenopausal, past the age of 50, with a life expectancy after menopause of only 20 years. This made menopause a biological marker for the last two decades of a woman's life. This situation changed dramatically in just 95 years. A woman reaching menopause in 1996 can expect to live 30 more years. This is a net-gain in postmenopausal life expectancy of 10 years in one century. ² These facts and projections illustrate how the female population has changed in both age and number.

Since this population is older, they have a different biologic make-up. During a woman's reproductive life, plasma 17B estradiol ranges from 50 to 400 pg/mL, but with menopause estrogen levels drop to <30 pg/mL. This level is lower than the estradiol levels present in a man.

These low estrogen levels in menopausal women cause or contribute to numerous symptoms, ranging from vasomotor hot flashes to major depression, cardiovascular disease, and osteoporosis. In an attempt to mitigate the many problems associated with living longer with less estrogen, increasing numbers of menopausal women are turning to HRT in hope of alleviating these symptoms and gaining important
cardiovascular and skeletal benefits.

Many different estrogen and estrogen/ progestin combinations are prescribed by the millions annually in this country. These include conjugated equine estrogens, which have been widely available for more than 50 years; other oral forms; and transdermal patches. For women with an intact uterus, combinations of estrogen and progestin are generally prescribed, whereas there are no compelling data to use progestins for post-hysterectomy patients.

Radiologists must be aware of these changes in the female population since increased use of HRT can influence the mammogram's appearance and how it is interpreted. ³ Some of the possible effects of HRT on the mammogram include: 1) a diffuse overall increase in the density of the glandular parenchyma; 2) development of a single asymmetric density; and 3) cyst formation.

Increased breast density

The majority (75%) of postmenopausal women initiating HRT will have no appreciable change in the appearance of their mammogram on the 1-year followup exam. In our experience, however, some women (25%) will have changes that can be recognized by the radiologist. The change we see most frequently is a diffuse and generalized increase in the glandular density of both breasts (figure 1). This diffuse and generalized bilateral appearance on the mammogram represents the response of senescent breast tissue to hormone stimulation, producing a benign epithelial hyperplastic change. This mammographic change requires no special diagnostic evaluation as long as the patient has no palpable complaint and there is no evidence of a dominant mass in this changing parenchyma. In the presence of these changes, we recommend the patient remain on HRT.

Frequently, multiple small (2 to 3 mm) round, benign-appearing nodules are visible on the mammogram of a patient taking HRT. If these nodules are nonpalpable and well circumscribed on spot magnification imaging, they usually represent benign glandular proliferative changes. We advise discontinuing HRT for 3 to 4 months and repeating the mammogram to demonstrate their shrinkage or disappearance before resuming HRT. These minimal nodules often increase and decrease in size over the years as some patients start and stop HRT. If these nodules increase in size while the patient is off HRT, biopsy should be considered. The development of an irregularly shaped or spiculated mass while the patient is on or off HRT is a suspicious mammographic finding and a biopsy should be performed promptly.

Benign asymmetric density

A second, but less common, mammographic finding seen following the institution of HRT is the development of a single area of asymmetric density (figure 2). This finding should be evaluated with spot compression views in two projections together with directed ultrasound (US) and directed palpation. If this asymmetric density disperses normally on spot compression, and is non-visualized on US without a palpable finding, it is a benign estrogen effect. We recommend the next mammogram to be a diagnostic exam in 1 year and we encourage the patient to continue HRT.

Cyst formation

Another change occasionally seen on mammogram during HRT is cyst formation. These cysts can be single or multiple and involve one or both breasts. They are usually small, less than 1 to 2 cm, nonpalpable, and well circumscribed on the mammogram. US is used to document their benign nature.

The appearance of several small nonpalpable cysts is of no significance, no interventional action (such as aspiration) is taken, and HRT can continue (figure 3).

If a new cyst is palpable, or if there is uncertainty about its cystic versus solid nature on US exam, an US-guided aspiration should be performed and its disappearance should be documented with a stat postaspiration mammogram so that HRT can continue uninterrupted (figure 4).

Possible causal relationship with breast cancer

A potential problem with HRT is the possibility that it could be associated with the development of breast cancer (figure 5). This is a matter of great concern to many patients before and during HRT. The radiologist should understand the basic statistical risks of developing cancer while on HRT and be able to describe these risks to the patient.

Scientists and statisticians, in an attempt to find a definitive answer to this question have done meta-analyses of hundreds of published studies in the scientific literature that reported data suggesting either increased or decreased incidence of breast cancer with HRT (figure 6). In the final assessment, statisticians observed that the risk of developing cancer varied widely from study to study. Because of this observation, and despite analysis of numerous epidemiologic studies, there is no solid consensus regarding postmenopausal estrogen use and risk of breast cancer. The "bottom-line" summary of many studies displayed in figure 6 shows minimal, almost statistically insignificant, increased relative risk (RR 1.10) of breast cancer associated with estrogen use.

During analysis of these multiple studies, scientists gave particular attention and scrutiny to the type, duration, and dosage of estrogen used in HRT. These analyses resulted in several causal inferences that we will briefly explain, but we must first define relative risk.

RR is the incidence of a condition in an exposed group divided by the incidence in a non-exposed group. A RR of 1.0 means no increased risk for the exposed group compared with the non-exposed. An example of increased RR would be women with first-degree relatives with breast cancer having an increased risk (RR 2.3) of getting breast cancer as compared with women without such a family history. ⁴ Another example is seen in figure 6, in which there is a reported increased RR of 1.1 associated with HRT and breast cancer. For comparison, this risk is even smaller than the reported slightly increased RR of 1.3 of developing breast cancer in women who have merely had an early menarche, at <12 years of age. ⁵

It is generally accepted that the type of exogenous estrogen used in HRT can be important. For example, one large study of postmenopausal women using HRT therapy found only a very slight increase in overall risk (RR 1.10) of developing breast cancer. ⁶ The portion of women in this study group taking estradiol preparations had a slightly higher risk (RR 1.20) and those women in the study using the injectable forms of estradiol had a four-fold (RR 4.0) increased risk. It is important to note that those women in this study using conjugated estrogen had no increased risk (RR 1.0) of developing breast cancer.

The relationship between the duration of HRT and breast cancer risk can also be important. Some authors find a modest increase in relative risk with increasing duration (>5 years) of estrogen use, although other authors found no increased risk with time. ⁷ Therefore, the literature is not conclusive on this issue.

A third causal factor observed by scientists is the daily dosage of estrogen, which may influence risk. There is evidence that dosage of 1.25 mg of conjugated estrogen may increase risk, while a lower dosage of 0.625 mg does not appreciably increase risk (figure 7).

In view of these three generally accepted findings about the risk of breast cancer while using HRT, we recommend that the radiologist interacting with postmenopausal women undergoing HRT: 1) encourage the postmenopausal patient to make an informed decision about HRT therapy by sharing what you have learned of this topic; 2) obtain baseline mammogram and breast palpation before HRT begins; 3) encourage the use of estrogens in the lowest appropriate dose; 4) advocate annual mammography and breast palpation; and 5)
evaluate new-onset mammographic densities, masses, or palpable complaints detected on physical exam.

Conclusion

Radiologists interpreting screening and diagnostic mammograms can now appreciate the changes in the female population, including increasing numbers of menopausal women electing to use HRT to mitigate the symptoms of menopause and improve cardiovascular and skeletal health.

This widespread hormone use can affect the appearance of the mammogram in a portion of women by producing increased densities, benign cysts, and nodules. Radiologists must be familiar with these mammographic changes and be prepared to perform any additional diagnostic imaging necessary to confirm their benign nature.

Finally, radiologists should be prepared to discuss with the patient the small but possible increased risk of breast cancer with HRT. This discussion should emphasize annual mammographic screening and physical exams for the early detection of breast cancer, no matter what its etiology. AR

REFERENCES

1. U.S. Census Bureau

2. National Center for Health Statistics: Division of Vital Statistics, Washington, DC, personal communication, 1999.

3. Laya MB, Gallagher JC, Schreiman JS, et al: Effect of postmenopausal hormone replacement therapy on mammographic density and parenchymal pattern. Rad 196:433-437, 1995.

4. Sattin RW, Rubin GI, Webster LA, et al: Family history and the risk of breast cancer. JAMA 253:1908-1913, 1985.

5. Dupont WD, Page DL: Breast cancer risk associated with proliferative disease, age at first birth, and a family history of breast cancer. Am J Epidemiol 125:769-779, 1987.

6. Bergkvist L, Adami HO, Persson I, et al: The risk of breast cancer after estrogen and estrogen-progestin replacement. N Engl J Med 321:293-297, 1989.

7. Dupont WD, Page DL: Menopausal estrogen replacement therapy and breast cancer. Arch Int Med 151:67-72, 1991.