A 62-Year-Old Woman With a New Diagnosis of Breast Cancer

DR BURNS: Ms T is a 62-year-old woman with a new diagnosis of left breast cancer. A routine mammogram performed in the fall of 2004 revealed a 9-mm spiculated mass deep within the upper inner quadrant of her left breast. Prior annual mammograms had been normal. No mass was palpable in this area, but ultrasound demonstrated a hypoechoic mass at 10 o’clock a few centimeters from the nipple. A stereotactic core biopsy was performed and disclosed an infiltrating lobular carcinoma.

Her medical history was otherwise noteworthy for recurrent bouts of diverticulitis and osteopenia. She experienced menarche at 12 years of age. She has 2 children and was 21 years old at the birth of her first child. She briefly used oral contraceptives and never used fertility medications. She experienced menopause in her early 50s and took hormone therapy for 1 month. She had previously undergone 2 benign right breast biopsies.

A bone density in October 2003 revealed a T score of –2.0 for the spine and –0.9 for the total hip. She began taking alendronate, which she could not tolerate due to gastrointestinal upset. A repeat bone density in November 2004 revealed a T score of –2.0 for the spine and –1.4 for the total hip. She currently takes calcium and vitamin D.

She works as a banker. She does not smoke and drinks approximately 2 alcoholic beverages per week. There is no family history of breast or ovarian cancer. Her mother had osteoporosis.

Ms T presented to the multidisciplinary breast clinic where she was thought to have a clinical stage I carcinoma of her left breast. She was considered a good candidate for breast-conserving therapy, and in November 2004 she had a wire-localized excisional biopsy under mammographic guidance and sentinel lymph node procedure. Final pathology revealed an infiltrating lobular carcinoma with an invasive component of 9 mm and a histologic grade of 2 (out of 3). The margins of excision were clear. There was a question of lymphovascular invasion within the tumor; the one sentinel node was negative. Immunohistochemical stains for both estrogen and progesterone receptors were positive. Staining for Her-2/neu was indeterminant; fluorescent in situ hybridization showed no amplification of Her-2/neu.

Radiation treatment to the left breast was recommended. Chemotherapy was not recommended given her good prognosis, age, and positive estrogen and progesterone receptors. Adjuvant endocrine therapy was recommended. Exemestane was considered; however, it was not covered by her insurance. She was therefore prescribed tamoxifen (20 mg daily), although she wonders whether an aromatase inhibitor is superior.

I had gone in for a routine examination, and they said that the doctor wants you have to a sonogram because she thinks she's seeing something. She said that it looked very suspicious to her and I needed to have a biopsy done. I went through that procedure in her office, and 4 days later I found out that I had breast cancer. I went to see a surgeon who was recommended by a friend and she said that I needed to have a lumpectomy done, and that I was extremely fortunate that it had been found very, very early. After the surgery I had radiation therapy. And then, I had several long conversations with my medical oncologist about which medication to use. He initially suggested an aromatase inhibitor, but when I called my insurance company, they would not cover it. So then he said considering that, and also considering my bone issues, he wanted me to take tamoxifen.

What is the epidemiology and prognosis for stage I breast cancer in general and for individual patients, for example, Ms T? What is the evidence that adjuvant endocrine therapy prolongs disease-free and overall survival? What are the data supporting the use of aromatase inhibitors vs tamoxifen? What are the adverse effects and costs? What does the future hold? What do you recommend for Ms T?

DR COME: In 2005, an estimated 211 240 new cases of female breast cancer were diagnosed in the United States.¹ The overall incidence rate of breast cancer in US women is 132.9 per 100 000, and this rises with age.¹ For women ages 60 to 64 years like Ms T, the incidence is 394.7 per 100 000. Two thirds of newly diagnosed breast cancers in women older than 50 years are localized to the breast¹ ; Ms T has a stage I breast cancer (T1bN0M0).²

The prognosis for stage I breast cancer is variable, with recurrence-free survival ranging from 70% to greater than 90%.^{3 - 5} Tumor size, histologic grade, proliferative index, and expression of hormone receptors are the most important factors in predicting outcome and in selecting therapy.^{6 - 7} Invasive tumors smaller than 1 cm, histologic grade 1, low proliferative index, and positive for both estrogen (ER+) and progesterone (PR+) receptors represent the most favorable prognosis; conversely, larger tumor size, higher grade and/or proliferative index, and lack of hormone receptor expression are adverse features. The presence of lymphovascular invasion in the tumor specimen,^{8 - 9} overexpression or amplification of Her-2/neu,¹⁰ and very young patient age (<35 years)¹¹ are additional unfavorable factors.

For stage I breast cancer, primary treatment consists of either mastectomy or the combination of lumpectomy (excision of the tumor with free margins) and radiation therapy to the breast, as was elected by Ms T. These alternatives provide equivalent long-term disease-free, distant disease-free, and overall survival.¹² Lumpectomy without radiation therapy is associated with a higher risk of recurrence in the involved breast. A meta-analysis of 6 trials with greater than 10 years of follow-up in which a total of 4177 women were randomized to undergo breast-conserving surgery with or without radiation found a 68% relative reduction in local recurrence for patients who received radiation (7%) vs those who did not (22%, P<.001).¹³ There is some evidence that the risk of local tumor failure after conservative surgery alone in women 65 years of age and older with small, low-grade, node-negative tumors is low enough that radiation therapy may be omitted in this specific subset.¹⁴ In many women with early breast cancer, medical therapy is added to primary treatment. Systemic treatment with either chemotherapy, tamoxifen, or both has been shown to decrease local recurrence rates in the breast in patients undergoing conservative surgery (lumpectomy) and radiation therapy.¹⁴ However, a decision as to whether, and if so which, adjuvant systemic therapy is added to primary therapy depends principally on an assessment of the risk of subsequent distant tumor recurrence. Furthermore, the impact of these systemic therapies on local control in patients undergoing conservative surgery without radiation therapy is not well studied, and they are not generally considered an alternative to radiation therapy in patients desiring breast conservation.¹⁴

An online evidence-based tool (Adjuvant!) compares the patient's clinical and pathological characteristics with US Surveillance, Epidemiology, and End Results (SEER) tumor registry data to generate 10-year risks of recurrence and mortality.¹⁵ Using results from major adjuvant therapy trials, the Adjuvant! program calculates the proportional and absolute benefits of these various therapies on the patient's risks of recurrence and death. This regularly updated instrument is a helpful aid to clinicians and displays results in a format that can be readily shared with patients. A recent independent validation found a less than 1% variation between the Adjuvant! estimates and actual 10-year outcomes in a large population of women with early breast cancer.¹⁶

Using the Adjuvant! program, Ms T has a 10-year risk of breast cancer recurrence of 18% following lumpectomy and radiation therapy to her breast.¹⁵ Since this overall risk of cancer recurrence includes a 6% 10-year risk of a contralateral second primary breast cancer, the risk of local and distant recurrence attributable to the current cancer is 12% at 10 years.¹⁵ The 10-year mortality from this cancer is only 3%.¹⁵ The discrepancy between recurrence and mortality figures reflects 2 factors. First, women experiencing ipsilateral breast local recurrences following conservative surgery and radiation can be treated by mastectomy. In one recent series, 65% of women treated for ipsilateral breast recurrence were disease-free and 81% were alive 5 years after surgery.¹⁷ Second, there may be long intervals between initial diagnosis and distant recurrence and between recurrence and death for women with hormone receptor–positive breast cancer who will nevertheless ultimately succumb to the disease.¹⁸

Data from numerous studies and from a series of meta-analyses performed every 5 years since 1985 demonstrate that, when added to primary therapy, both endocrine therapy and chemotherapy can reduce recurrence and mortality in women with early breast cancer when added to primary therapy.^{18 - 22} The benefit of endocrine therapy is limited to patients with tumors that express hormone receptors.^{18 - 19} Among postmenopausal women, 79% have tumors that are ER+ and 53% are PR+. The majority of the latter are also ER+ as fewer than 5% of tumors are ER− PR+.²³ Expression of PR is an independent predictor of favorable outcome on endocrine therapy.²⁴

The selective estrogen receptor modulator (SERM) tamoxifen has been the standard of care as adjuvant endocrine therapy for more than 20 years. A recently published meta-analysis, initiated in 2000, reports data from 71 trials of adjuvant tamoxifen involving more than 80 000 women.¹⁸ Fifteen-year observations on the effects of treatment vs control (Table 1) and of the risks of therapy are now available.

These studies found that 5 years of tamoxifen treatment reduced the annual rates of recurrence by 41% (SE, 0.03%) and mortality by 34% (SE, 0.04%) in women with ER+ disease.¹⁸ The reduction in rate of recurrence is maintained during treatment and for about 5 years after cessation of tamoxifen treatment. Beyond year 10, the recurrence rates in the treatment and control groups are similar, but the gains of the first 10 years are maintained. In addition, the reduction in mortality rate for women treated with tamoxifen compared with controls is still evident at year 15. At 15 years, there is an absolute reduction in recurrence of 11.8% (P<.001) and in mortality of 9.2% (P<.001) for women with ER+ (or ER unknown) tumors receiving 5 years of tamoxifen treatment.¹⁸ Thus, 8.5 women would need to be treated with tamoxifen for 5 years to prevent 1 recurrence at 15 years, while 10.9 women would need to be treated to prevent 1 death at 15 years. The benefits of tamoxifen are largely irrespective of age, menopausal status, nodal status, and use of adjuvant chemotherapy.^{18 - 19} Additionally, for women with ER+ (or ER unknown) tumors, 5 years of treatment with tamoxifen results in a relative reduction in contralateral breast cancer of 39% during the 15 years of observation (incidence rate ratio, 0.61 [95% confidence interval {CI}, 0.50-0.73]).¹⁸

While tamoxifen is a competitive inhibitor of estrogen at the ER, its effects are complex and depend on the local tissue levels of coregulatory proteins.²⁵ In certain environments, tamoxifen has partial agonist effects on the ER. Aromatase inhibitors inhibit or inactivate the enzyme responsible for the conversion of androgen substrates to estrogen.²⁶ Thus, ligand for the ER is depleted; these agents have no potential agonist effect.

Three third-generation aromatase inhibitors are currently available. The nonsteroidal compounds anastrozole and letrozole are reversible inhibitors of aromatase, while the steroidal agent exemestane is an irreversible inactivator of the enzyme. It is not yet clear whether the pharmacological differences between these compounds or potential differences in action at the tissue level are clinically significant. Each aromatase inhibitor lowers plasma estrogen levels by approximately 98% in postmenopausal women.²⁶ The effects of these agents in women with intact ovarian function are not well studied, and at this time they are indicated only in the treatment of postmenopausal women.²⁶ Based on efficacy and tolerability of third-generation aromatase inhibitors in women with metastatic breast cancer,^{27 - 31} these agents have been evaluated in trials in early breast cancer.

Today, most women like Ms T with hormone receptor–positive stage I breast cancer receive primary therapy and adjuvant endocrine therapy.^{32 - 33} For the subset of women with grade 1 tumors less than 1 cm, no adjuvant systemic therapy is an appropriate option.³³ In such patients, the 10-year mortality rate from breast cancer without adjuvant systemic therapy is 1%, and the absolute reduction in mortality attributable to the use of endocrine therapy is less than 0.5%.¹⁵ Conversely, the addition of adjuvant chemotherapy to endocrine therapy in hormone receptor–positive stage I cancer is generally restricted to women with high-risk features such as tumors larger than 1 cm, high tumor grade or proliferative index, weak expression of hormone receptors, overexpression or amplification of Her-2/neu, or very young age (<35 years). In this instance, the 10-year mortality rate from breast cancer is approximately 20% and remains at 13% despite the use of adjuvant endocrine therapy. The addition of chemotherapy results in a further absolute mortality reduction of 5%.¹⁵ For women older than 50 years like Ms T, chemotherapy is less effective.^{18 ,20} Further, women with hormone receptor–positive breast cancer receiving endocrine therapy appear to derive less benefit from chemotherapy than their hormone receptor–negative counterparts.³⁴

More than 40 000 women have been enrolled in trials evaluating aromatase inhibitors in early breast cancer, and several studies using varying designs have reported improvements in disease-free survival when these agents are compared with tamoxifen (Table 2). In each trial, reported follow-up is measured from the onset of the randomized therapy.

The Arimidex, Tamoxifen, Alone or in Combination (ATAC) trial was the first study to report results, is the largest to date, and provides the longest follow-up. A total of 9366 women were randomized to receive 5 years of treatment with anastrozole, tamoxifen, or both.³⁵ Results in the combination group proved no different than tamoxifen alone, so attention has focused on the 6241 women who received either single agent. At 3 years of follow-up, the disease-free survival for women receiving anastrozole was 89.4% vs 87.4% in women treated with tamoxifen (hazard ratio [HR], 0.83 [95% CI, 0.71-0.96]; P = .01).³⁵ With this risk reduction, 50 women would need to be treated with anastrozole to prevent 1 recurrence at 3 years compared with tamoxifen. In the 5-year update, the disease-free survival favored anastrozole, maintaining the HR of 0.83 (95% CI, 0.73-0.94; P = .005) in the subset of women who were confirmed hormone receptor–positive.³⁶ This 17% reduction in recurrence translates to an absolute 3.3% improvement in disease-free survival for women receiving anastrozole. To date, no difference in overall survival has been observed, with 411 deaths in women receiving anastrozole vs 420 in those who received tamoxifen (HR, 0.97 [95% CI, 0.85-1.12]; P = .70).³⁶ Breast cancer deaths were reduced a nonsignificant 12% in women receiving anastrozole (HR, 0.88 [95% CI, 0.74-1.05]; P = .20).³⁶

The Breast International Group (BIG) 1-98 4-group trial^{37 - 38} enrolled 8028 women. In 2 of the groups, patients received either letrozole or tamoxifen for 5 years, while women in the other 2 groups crossed from the initial therapy to the other agent after 2 years of treatment and then continued the second agent for the additional 3 years. Thus far, only the results of initial letrozole vs initial tamoxifen in the 8010 eligible women have been reported, as the median follow-up of 26 months is too short to assess the 2 sequential therapy groups (a first report on these groups is anticipated in 2008). Women receiving initial letrozole had a 2.6% absolute improvement in disease-free survival (84% vs 81.4%; HR, 0.81 [95% CI, 0.70-0.93]; P = .003) compared with those receiving tamoxifen,³⁸ such that 38 women would need to be treated with letrozole to prevent 1 recurrence at 5 years compared with tamoxifen. At this time, there is no difference in overall survival.³⁸

Three trials have compared 5 years of tamoxifen treatment with 2 or 3 years of initial tamoxifen followed by 3 or 2 years of an aromatase inhibitor.^{39 ,41 - 42} The women entered into these “switching” trials are not comparable to those enrolled in the ATAC and BIG 1-98 trials since they have already achieved at least 2 recurrence-free years with tamoxifen prior to entering the randomized portion of treatment.

The Intergroup Exemestane Study (IES) randomized 4742 women who had taken 2 to 3 years of tamoxifen therapy to either continue tamoxifen for a total of 5 years or switch to exemestane to complete a total of 5 years of adjuvant endocrine therapy. For the primary end point of disease-free survival, the sequence was superior to continued tamoxifen, with an absolute difference of 4.7% (91.5% vs 86.8%; HR, 0.68 [95% CI, 0.56-0.82]; P<.001) and a median follow-up of 30.6 months.³⁹ Twenty-one women would need to be treated with the sequence of tamoxifen followed by exemestane to prevent 1 recurrence at 30.6 months compared with tamoxifen monotherapy. At a median follow-up of 37.4 months with 339 deaths recorded, there was a nonsignificant trend in overall survival favoring sequential therapy (HR, 0.85 [95% CI not provided]; P = .08).⁴⁰

The Austrian Breast Cancer Study Group 8/Arimidex-Nolvadex 95 (ABCSG8/ARNO95) trial randomized 3223 women who had received 2 years of adjuvant tamoxifen to either continue tamoxifen therapy for a total of 5 years or switch to anastrozole for the remaining 3 years.⁴¹ The patient population in this study is the most “pure” since all of the patients were confirmed hormone receptor–positive and none of the patients received adjuvant chemotherapy. With a median follow-up of 28 months, the sequential therapy group had an absolute advantage in event-free survival of 3.1% (95.8% vs 92.7%; HR, 0.60 [95% CI, 0.44-0.81]; P<.001).⁴¹ Thirty-two women would need to be treated with the sequence to prevent 1 recurrence at 3 years compared with tamoxifen alone. There have been too few deaths to detect a difference in overall survival.

The Italian Tamoxifen Anastrozole (ITA) trial randomized 448 women who had received tamoxifen for 2 to 3 years to continue tamoxifen treatment or switch to anastrozole for a total of 5 years of treatment. All women in this trial had involved axillary lymph nodes at the time of diagnosis. With a median follow-up of 36 months, the recurrence-free survival for women switching to anastrozole was 94.6% compared with 85.8% in women who continued taking tamoxifen. The absolute recurrence rate at 3 years was 5.8% lower in women who switched to anastrozole compared with women who continued tamoxifen (HR, 0.35 [95% CI, 0.18-0.68]; P = .001).⁴² Seventeen women would need to be treated with tamoxifen followed by anastrozole to prevent 1 recurrence at 3 years compared with tamoxifen alone.

The optimal duration of adjuvant endocrine therapy remains uncertain. Although the benefits of 5 years of tamoxifen treatment persist through at least the next 10 years, more than half of recurrences in women with node-negative breast cancer who have received 5 years of adjuvant tamoxifen occur after completion of this therapy.^{18 - 19} The question of whether extending tamoxifen treatment beyond 5 years would improve outcome has been addressed in 3 published studies that randomized a total of 1700 women who were disease-free after 5 years of tamoxifen to either discontinue therapy or extend tamoxifen treatment.^{45 - 47} The National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14 trial with 1172 patients enrolled reported a 7-year disease-free survival of 78% for women who continued tamoxifen beyond 5 years compared with 82% for women who discontinued tamoxifen after 5 years of treatment (HR, 1.3 [95% CI, 1.0-1.7]; P = .03).⁴⁵ Two larger studies, Adjuvant Tamoxifen—Longer against Shorter (ATLAS) and adjuvant Tamoxifen Treatment offer more (aTTom) are ongoing. Follow-up beyond 10 years may be necessary to discern whether extending tamoxifen treatment beyond 5 years is advantageous given the carryover effect from 5 years of therapy.

The pattern of late and continued recurrence risk coupled with evidence to date that extended tamoxifen is not beneficial provided the rationale for the international National Cancer Institute of Canada MA.17 trial.⁴³ In this study, 5187 women who were recurrence-free at the completion of 5 years of tamoxifen treatment were randomized to receive 5 years of letrozole vs placebo. The trial was terminated early when the predefined stopping rule was met. At the most recent update on the entire study population with a median follow-up of 30 months from randomization, absolute disease-free survival (fewer local or distant recurrences or contralateral primary breast tumors) was improved 4.6% with letrozole (94.4% vs 89.8%; HR, 0.58 [95% CI, 0.45-0.76]; P<.001).⁴⁴ Twenty-one women would need to be treated with letrozole to prevent 1 recurrence at 4 years compared with placebo. The 4-year overall survival in the trial was 95.4% for women receiving letrozole vs 95% for those who received placebo (HR, 0.82 [95% CI. 0.57-1.19]; P = .30). A borderline overall survival advantage has emerged in the prespecified subset of women with node-positive cancer (HR, 0.61 [95% CI, 0.38-0.98]; P = .04).⁴⁴

The safety and tolerability of tamoxifen have been well defined during the past 2 decades; a large amount of data from both therapeutic and prevention trials (vs placebo) is available. The experience with aromatase inhibitors is more limited; the longest follow-up available in the adjuvant setting is supplied by the ATAC trial, with a median follow-up of 68 months.³⁶ In this trial, 92% of patients have completed the 5 years of study medication. However, the other major trials have reported adverse effects with median follow-ups of only 2 to 3 years thus far.^{37 - 42}

In the randomized adjuvant trials, aromatase inhibitors are tolerated at least as well as tamoxifen. In the ATAC trial, the withdrawal rate for drug-related adverse events was 5.1% for anastrozole and 7.2% for tamoxifen.³⁵ Menopausal symptoms manifest as hot flashes, night sweats, and mood disturbance are common with both tamoxifen and aromatase inhibitors. The aromatase inhibitor produced fewer hot flashes than tamoxifen in the ATAC trial (P<.001)^{36 ,48} and in the BIG 1-98 trial,^{37 - 38} but this difference was not observed in the 3 trials in which women were randomized to continue taking tamoxifen for 5 years vs switch to an aromatase inhibitor after the initial 2 years of tamoxifen.^{39 - 42} Loss of libido occurred at similar frequency with exemestane and tamoxifen (25% to 28%) in the IES trial and appears to worsen, at least over the first 24 months of follow-up.⁴⁹ During the first 24 months of the ATAC trial, loss of sexual interest was more common in women taking anastrozole than tamoxifen (16% vs 9%, P<.05).⁵⁰ Vaginal discharge and bleeding were more common in women taking tamoxifen in each of the trials, while joint and muscle pain were more common in women receiving aromatase inhibitors.^{35 - 42} Nausea and other gastrointestinal adverse effects have been reported in equal frequency for women taking either aromatase inhibitors or tamoxifen,^{38 - 39 ,41 - 44 ,48} although an excess of diarrhea has been reported for exemestane vs tamoxifen (4.3% vs 2.3%, P<.001).³⁹

The most serious complications of tamoxifen are related to its estrogen agonist effects. In an overview of available trials, 5 years of tamoxifen therapy was associated with a risk ratio of 4.2 for endometrial cancer, raising the 10-year risk per 1000 from 3 in the control population to 11 with tamoxifen (P<.001).^{18 - 19} One to 2 years of tamoxifen treatment raised the relative risk to approximately 2.0.¹⁹ Nevertheless, the increase in endometrial cancer with 5 years of tamoxifen was only half as large as the decrease in contralateral breast cancer. Further, the 10-year risk of mortality from endometrial cancer in tamoxifen users was only 2 per 1000.¹⁹ The NSABP tamoxifen prevention trial also reported a risk ratio of 4.01 (95% CI, 1.70-10.90) for tamoxifen use in women older than 50 years; the cumulative incidence of endometrial cancer at 66 months of follow-up was 5.4 per 1000 women for the placebo group vs 13.0 per 1000 women in the tamoxifen group.⁵¹

The second consequence of the estrogen agonist effect of tamoxifen is an increase in venous thrombosis, thromboembolism, and stroke. In the NSABP P-1 prevention trial, which compared tamoxifen and placebo, women older than 50 years using tamoxifen had risk ratios of 1.71 (95% CI, 0.85-3.58) for venous thrombosis, 3.19 (95% CI, 1.12-11.15) for pulmonary embolism, and 1.75 (95% CI, 0.98-3.20) for stroke compared with women receiving placebo.⁵¹ Tamoxifen use had no effect on the incidence of ischemic cardiovascular events in the NSABP P-1 trial and was associated with a nonsignificant reduction in cardiovascular events in the most recent overview.¹⁸

Since aromatase inhibitors have no estrogen agonist effect, their use is not associated with these problems. In the ATAC trial, anastrozole compared with tamoxifen had a relative risk of 0.29 (95% CI, 0.11-0.80; P = .02) for endometrial cancer, 0.64 (95% CI, 0.45-0.93; P = .02) for deep vein thrombosis, 0.61 (95% CI, 0.47-0.80; P <.001) for thromboembolic events, and 0.70 (95% CI, 0.50-0.97; P = .03) for stroke.³⁶

The major concerns regarding the safety of aromatase inhibitors are a consequence of the marked decrease in estrogen levels. After 5 years of follow-up in the ATAC trial, the relative risk of fractures was 1.49 (95% CI, 1.25-1.77; P<.001) for women taking anastrozole with a fracture rate of 11% vs 7.7% for women taking tamoxifen.³⁶ In the ABCSG8/ARNO95 trial, 2.0% of women taking anastrozole and 1.0% of women taking tamoxifen experienced fractures during the median follow-up of 28 months (HR, 2.14 [95% CI, 1.14-4.17]; P = .015).⁴¹ In the BIG 1-98 trial, 5.8% of women taking letrozole vs 4.1% taking tamoxifen had experienced fractures in the first report of this trial with a median follow-up of 26 months.³⁷ The IES trial reported osteoporosis in 7.4% of women taking exemestane vs 5.7% in women continuing tamoxifen at a median follow-up of 30 months (P = .05).³⁹ In this trial, fractures increased nonsignificantly in the exemestane group vs the tamoxifen group (3.1% vs 2.3%; P = .08).³⁹ It is uncertain whether these differences represent a protective effect of tamoxifen on bone, an adverse effect of the aromatase inhibitor, or both. Some protective effect is suggested by the NSABP Prevention trial (tamoxifen vs placebo) in which the relative risk of fracture for women older than 50 years taking tamoxifen was 0.79 (95% CI, 0.60-1.05).⁵¹ In the MA.17 trial, 8.1% of women taking letrozole compared with 6% of women taking placebo (P = .003) reported new onset of osteoporosis during 30 months of observation.⁴⁴ The fracture rates were 5.3% for women receiving letrozole vs 4.6% for the placebo group (P = .25).⁴⁴

A second potential effect of reducing estrogen is an adverse effect on serum lipid levels, although relatively few data are currently available to evaluate this effect. In the ATAC trial after 5 years, ischemic cardiovascular disease increased nonsignificantly for patients taking anastrozole vs tamoxifen (4.1 vs 3.4%; P = .10).³⁶ No difference in hypercholesterolemia or in cardiovascular events has been observed between the letrozole and placebo groups of the MA.17 trial.⁴⁴ In the IES trial, the incidence of acute myocardial infarction was 0.9% (20 cases) in women taking exemestane vs 0.4% (9 cases) in women taking tamoxifen (P = .02), and this is being further investigated.⁴⁰ In the preliminary report from the BIG 1-98 trial, 43.6% of the patients in the letrozole group and 19.2% of the patients in the tamoxifen group had an elevated cholesterol level at least once on semiannual measurements during treatment (no statistics provided). However, the collection methods were not standardized, and no data are provided on the proportion of patients in either group with sustained hypercholesterolemia. More grade 3, 4, or 5 cardiac events occurred in women receiving letrozole than in those receiving tamoxifen (2.1% vs 1.1%, P<.001).³⁸

In addition to efficacy and adverse effects, cost is a consideration in comparing aromatase inhibitors and tamoxifen. While the impact of medication cost is a complex topic that cannot be addressed independent of other factors such as efficacy and patient preference, many postmenopausal women have no coverage for oral prescription medication and also are on fixed or low incomes. In this situation, the expense of adjuvant endocrine therapy increases from approximately $100 for a 90-day supply of generic tamoxifen to between $633 and $655 for a 90-day supply of any of the aromatase inhibitors.⁵²

Over the next several years, additional follow-up from the existing trials will provide data on overall survival, allow a comparison of sequential tamoxifen and an aromatase inhibitor with treatment with an aromatase inhibitor alone, and afford further observation of the safety and tolerability of these agents. Understanding of resistance to endocrine therapy is rapidly progressing and already translating into clinical efforts to abrogate these mechanisms. Combinations of endocrine agents (eg, aromatase inhibitors plus the pure estrogen antagonist fulvestrant) are of interest,⁵³ as is the simultaneous use of endocrine therapy and agents directed at other biological targets, in particular the epidermal growth factor pathway.⁵⁴ New SERMS with properties different from tamoxifen and additional pure estrogen antagonists are likely to be developed.⁵⁵ Finally, the molecular profiling of tumors is beginning to provide not only prognostic but also predictive information on who will benefit most from which treatments.⁵⁶ Analysis of gene expression patterns performed on the initial diagnostic pathology may ultimately inform the selection of available endocrine agents and combinations for women like Ms T.

For Ms T, the first question is whether to recommend any adjuvant systemic therapy. Her projected overall 10-year recurrence risk of local and distant recurrence is 12%.¹⁵ Since the Adjuvant! program does not consider lymphovascular invasion and combines tumors less than 5 mm with those between 5 mm and 9 mm (T1a vs T1b), Ms T's risk could be underestimated by the online tool.

Adjuvant endocrine therapy would reduce her 10-year risk of local and distant recurrence from 12% to 6% using regimens containing an aromatase inhibitor or to 8% if tamoxifen alone is given.¹⁵ In addition to reducing the risks of local and distant recurrence for her current tumor, adjuvant endocrine therapy offers a prevention effect, lowering the risk of a new primary breast cancer by 50% or possibly more.^{35 ,51}

Each of the phase 3 adjuvant endocrine trials evaluating third-generation aromatase inhibitors report a statistically significant advantage in disease-free survival favoring these agents (Table 2). While only the subset of node-positive women in the MA.17 trial have demonstrated an overall survival advantage to date,⁴⁴ a broader survival advantage may emerge in other trials with further follow-up and additional events. Based on these data, the American Society of Clinical Oncology recommends that adjuvant therapy for postmenopausal women with hormone receptor–positive breast cancer include an aromatase inhibitor to lower the risk of tumor recurrence but does not specify when or for how long an aromatase inhibitor should be used.^{57 - 58} Further, it is unclear whether aromatase inhibitors are slightly superior in all patients or whether specific subset(s) of patients account for most or all of the improved outcome when these drugs are compared with tamoxifen.

Monotherapy with an aromatase inhibitor as an alternative to tamoxifen results in a 3% absolute improvement in disease-free survival at 5 years.³⁶ The studies in which women received tamoxifen for 2 to 3 years and then switched to an aromatase inhibitor also appear promising (Table 2),^{39 - 42} but currently these 2 strategies have not been compared directly. At 3 years of follow-up in the ATAC trial, the difference in disease-free survival between anastrozole and tamoxifen is 1.6% for women with hormone receptor–positive tumors, and the curves show no separation in the first 12 months.³⁵ Thus, there is a small therapeutic advantage in beginning treatment with an aromatase inhibitor during the first 2 to 3 years; however, it is uncertain whether the longer-term outcome will favor monotherapy with an aromatase inhibitor vs some sequence of tamoxifen and aromatase inhibitor. The BIG 1-98 trial compares 5 years of letrozole with the sequences of tamoxifen followed by letrozole and vice versa, but the results will not be available for approximately 3 years.^{37 - 38} To address the question of whether an aromatase inhibitor alone or the sequence of tamoxifen followed by an aromatase inhibitor is superior, statistical models that adjust and then compare data from the ATAC, BIG 1-98, IES, and ABCSG8/ARNO95 trials have been created. Three recent studies produced conflicting results.^{59 - 61} The outcome is dependent on the assumptions used in building the model; in each case, the advantage for the preferred approach was quite small. Some preliminary data suggest that the effectiveness of aromatase inhibitors relative to tamoxifen is much greater in the subgroups of women who have tumors that are ER+ PgR−⁶² and/or overexpress Her-2/neu.⁶³ If confirmed, this would support the initial use of an aromatase inhibitor in such patients.

I believe that sequential therapy would be a good choice for Ms T. Initiating treatment with tamoxifen for a few years would likely have a beneficial effect on her bone density. This is particularly important for Ms T since she has underlying osteopenia, a family history of osteoporosis, and a demonstrated intolerance to oral bisphosphonates. After 2 to 3 years of tamoxifen, she would switch to an aromatase inhibitor. While each of the available aromatase inhibitors is effective, additional safety data will be available by that point, and one agent may emerge as preferred overall or in specific situations, such as underlying osteopenia. Further data as to the use of intravenous bisphosphonates in this setting will also be available.⁶⁴

While, in my opinion, the contribution of adjuvant endocrine therapy for Ms T is worthwhile, the benefit of adjuvant chemotherapy in a 62-year-old woman with a 9-mm stage I cancer that expresses both estrogen and progesterone receptors is too small to justify the adverse effects and risks. A standard, anthracycline-based chemotherapy regimen would reduce her 10-year recurrence risk by 2.8% in the absence of endocrine therapy, and by less than 2% (1.8% if tamoxifen, 1.3% if regimen contains an aromatase inhibitor) if Ms T receives adjuvant endocrine therapy.¹⁵

QUESTION: Can you make an argument for treating for a month, then getting off for a month, and then going back?

DR COME: Rotating on and off therapy could not be considered standard care at this point and could reduce the efficacy of treatment. Nevertheless, the concept of cycling low (such as with an aromatase inhibitor) estrogen levels with periods of exposure to moderate or even high doses of estrogen has been a subject of investigation.⁶⁵ The hypothesis behind such an approach would be that altering estrogen levels might prevent the adaptation of the tumor to a specific estrogen “setting,” thereby delaying the emergence of resistance and prolonging the overall time that a patient might remain on endocrine treatment. Traditionally, therapeutically induced changes in estrogen levels, for example, switching from tamoxifen with its partial estrogen agonist effect to an aromatase inhibitor that dramatically lowers estrogen, have been reserved for the time of tumor progression rather than being used on a fixed schedule. Perhaps the observed benefit of the “switching studies” discussed above—a sequence I have recommended for Ms T—provide some evidence that altering the hormonal milieu before resistance emerges could be productive.

QUESTION: Initially Ms T was going to start her therapy with an aromatase inhibitor, but her insurance company declined it. Where are the insurance companies on the curve of approving the aromatase inhibitors?

DR COME: I have not experienced rejections by insurance carriers for an aromatase inhibitor. Positive results involving each of the 3 third-generation aromatase inhibitors in early breast cancer have now appeared in peer-reviewed publications, and I believe these agents are broadly accepted even though there are unanswered questions about how best to use them. The American Society of Clinical Oncology has supported their use.⁵⁸ Since there is no evidence, as yet, that there are important differences among them in terms of efficacy or safety, it is possible that an insurer might not have all 3 drugs in its formulary. The greater barrier to access is cost.

QUESTION: You made a good point that about 20% of women with early stage breast cancer are going to show up with a recurrence and are ultimately going to die. And yet, you dismissed chemotherapy as an option. What about chemotherapy to improve the survival advantage in that 20%?

DR COME: In the case of Ms T, her residual risk of distant recurrence with endocrine therapy is about 6%, possibly even less. Chemotherapy could perhaps lower that by about 25%, or an absolute reduction of an additional 1.5%, taking her from a risk of 6% to 4.5%. Balanced against the adverse effects and risks, the therapeutic ratio for chemotherapy would thus be pretty low, although admittedly still elected by some patients. For women with early breast cancer and a higher residual risk despite endocrine therapy, chemotherapy is indeed indicated. The latest overview analysis emphasizes the independent value of both modalities.¹⁸ If Ms T had a residual risk of recurrence of 20% after factoring in the benefits of endocrine therapy, I would certainly recommend adjuvant chemotherapy in an otherwise healthy 62-year-old.