Sequencing of Endocrine Therapy in Postmenopausal Women with Advanced Breast Cancer

Endocrine therapy offers the potential for substantial palliation in the majority of postmenopausal women with breast cancer who are candidates based on presence of hormonal receptor [estrogen receptor (ER) and/or progesterone receptor] in the tumor. For the multiple agents available to the clinician (Table 1), the evidence upon which a sequencing strategy can be based will be reviewed.

It is standard practice to use endocrine therapy in most postmenopausal women with resected early breast cancer (1). Tamoxifen was approved by the United States Food and Drug Administration (FDA) for adjuvant therapy in 1986 and was the only agent so approved until September 2002 when the FDA approved the nonsteroidal aromatase inhibitor (NS-AI) anastrozole as an option for adjuvant therapy based on results of the Arimidex, Tamoxifen Alone or in Combination trial (2). Although the position has been taken (3), recently reaffirmed (4) and supported (5), that tamoxifen remains the endocrine agent of choice, anastrozole is recommended in patients with a contraindication or intolerance to tamoxifen. In addition, the American Society of Clinical Oncology Technology Assessment Working Group (3) advised that physicians and patients should review the available data and come to their own decision regarding the choice between tamoxifen and anastrozole. Thus, a discussion of sequencing must consider the population of patients who have a history of prior anastrozole exposure.

Given that there are two endocrine agents (tamoxifen and anastrozole) approved for adjuvant endocrine therapy, the following four categories must be considered when developing a sequencing strategy for the metastatic disease setting.

1. Endocrine therapy naïve. This scenario is becoming increasingly rare given the extensive use of tamoxifen in patients with ductal carcinoma in situ and in the prevention setting in addition to the adjuvant setting.

2. Prior tamoxifen exposure with >1 year between cessation of tamoxifen and identification of metastatic disease. Such patients are considered tamoxifen sensitive.

3. Prior tamoxifen exposure with ≤1 year between cessation of tamoxifen and identification of metastatic disease. Such patients are considered tamoxifen resistant. A similar situation can be considered to exist in the patient who develops breast cancer while receiving raloxifene for osteoporosis based on extrapolation from studies showing low levels of antitumor activity with second- and third generation-selective ER modulators after progression on tamoxifen (6).

4. Prior anastrozole (NS-AI) exposure. Given the recent introduction of anastrozole for adjuvant therapy, it is likely that at least for the foreseeable future, patients will be receiving this agent at the time of disease progression and time since cessation of therapy will not be an issue as is the case with adjuvant tamoxifen.

The endocrine therapy-naïve and tamoxifen-sensitive patients can be viewed as one group in terms of therapy options. Patients will therefore be considered from the standpoint of three categories: tamoxifen sensitive, tamoxifen resistant, and NS-AI resistant (Fig. 1).

Several large randomized clinical trials have produced results upon which a recommendation can be based. Both of the currently available third generation NS-AIs, anastrozole and letrozole, have been compared with tamoxifen in Phase III trials, which provided the basis for approval by FDA (7, 8, 9, 10, 11). The steroidal-AI exemestane is being compared with tamoxifen in a Phase III trial that is an extension of a Phase II trial (12). Accrual to the exemestane/tamoxifen trial has been completed, but no results are available.

A critical review of the pivotal trials comparing the NS-AIs with tamoxifen in advanced breast cancer (13) concluded that there is a reasonable basis for choosing a NS-AI over tamoxifen in this setting. This is based on superior efficacy in terms of time to disease progression and acceptable toxicity, which includes a lower incidence of thromboembolic complications.

In patients who have experienced disease progression on tamoxifen without exposure to a prior AI, the third generation AIs have replaced megestrol acetate. Phase III clinical trials have demonstrated not only improved tolerability but also improved efficacy in some outcome parameters for the three third generation AIs (14, 15, 16). A recent Phase III trial comparing the two NS-AIs, anastrozole and letrozole, found a superior response rate for letrozole for all patients, but the advantage was not seen in the cohort known to be ER positive, and there was no difference in time to progression (17).

In addition to the AIs, fulvestrant represents a fourth option for this population of patients. Fulvestrant, the long-acting formulation of the pure antiestrogen ICI 182,780, has the potential advantage of possessing antagonistic activity against breast cancer without the agonistic activity in other organs such as the endometrium that occurs with the selective ER modulators. It has been studied at the molecular level (reviewed in Ref. 18) and its actions include ER down-regulation. Fulvestrant is thus considered a selective ER down-regulator and is the first and only drug of this class to be approved by the FDA. The FDA labeling for fulvestrant states that it is indicated for treatment of postmenopausal women with hormone receptor-positive breast cancer who have disease progression after antiestrogen therapy. This approval was based on two pivotal Phase III clinical trials in which fulvestrant was compared with anastrozole and found to be comparable in terms of efficacy (19, 20).

Thus, in the patient with tamoxifen-resistant breast cancer, there are four endocrine therapy options: the two NS-AIs anastrozole and letrozole; the steroidal-AI exemestane; and the ER down-regulator fulvestrant.

As Fig. 1 indicates, an AI is a consideration in patients who experience progression on tamoxifen and fulvestrant. Vergote (21) reported at the 2001 San Antonio Breast Cancer Symposium the results of a retrospective collection of outcome data for women treated with endocrine therapy after fulvestrant. Considering patients who had and had not derived clinical benefit from fulvestrant, 15 of 39 (38%) patients and 14 of 34 (41%) patients, respectively, derived clinical benefit from subsequent treatment with an AI (unspecified as to whether steroidal or nonsteroidal). The retrospective nature of these data limits their value but would appear to indicate some potential benefit for AIs after fulvestrant therapy.

A NS-AI is an option after exemestane (Fig. 1). The experience with this sequence is quite limited but outcomes for 10 such patients have been reported with 4 achieving clinical benefit [1 with partial response (PR) and 3 with stable disease (SD); Ref. 22]. Additional information relating to treatment after exemestane or the NS-AIs will be presented below.

The clinical trial evidence upon which to base a sequence strategy is limited in women with disease progression after a NS-AI. The agents for which there are some data include tamoxifen, exemestane, fulvestrant, and estrogens, and these will be considered in turn.

No Phase II trial data are available for patients treated with tamoxifen after failing a NS-AI. Thürliman et al. (23) obtained information through questionnaires for patients treated with tamoxifen after anastrozole in the two pivotal clinical trials with the latter agent. Responses to the questionnaire were obtained for 98 of 137 patients, of whom 11 (11%) were reported to achieve an objective response and 50 (51%) had SD for at least 24 weeks. The retrospective nature of these data limits their value, and the results could be considered a best-case scenario.

A good source of data for the efficacy of tamoxifen after a NS-AI should be the trial of Mouridsen et al. (11), which was double blind and compared tamoxifen with letrozole with an optional crossover upon progression in which the double-blind feature was maintained. It was reported that the patients treated with tamoxifen had a median survival of 19 months from the time of crossover, but neither response rate nor time to progression was reported.

The current understanding of the action of tamoxifen, based on preclinical models of breast cancer (24), is that it has antagonistic (antiestrogenic) effects in an estrogen environment, whereas it has agonistic effects in an estrogen-deprived environment, as might be expected in a patient receiving a potent third generation aromatase inhibitor. Thus, one might expect a relatively low level of antitumor effect with tamoxifen in this setting.

Exemestane is the only agent that has been studied in a formal Phase II trial to determine efficacy in patients with prior NS-AI exposure. Lønning et al. (25) treated 241 women with the currently approved dose of 25 mg/day. The cohort of interest in this trial for today’s practice is that treated with a third generation AI. Taken together, 105 women received anastrozole, letrozole, or vorozole; a total of 4.8% (95% confidence interval, 1.6–10.8) achieved a complete response or PR with a median duration of response of 60.6 weeks, and an additional 15.6% achieved SD for at least 24 weeks. Thus, the clinical benefit rate was 20.4% (95% confidence interval, 13.1–29.5) with a median duration of 37 weeks. Considering all 105 patients, the median time to progression was 12.9 weeks.

The authors of this study (25) examined the effect of 8 weeks of exemestane therapy on serum estradiol (E2), estrone, and estrone sulfate in 20 patients previously treated with a third generation NS-AI. The mean estrogen levels were similar on exemestane to those found before termination of the NS-AI. The authors hypothesized that the beneficial effects of exemestane could be due to the 17-hydroexemestane metabolite, which has androgenic activity (26) and could work through the androgen receptor.

As noted above, the FDA labeling states that fulvestrant is indicated after progression on an antiestrogen. Although a patient who has received treatment with tamoxifen and an AI would be considered a candidate for fulvestrant therapy, the evidence upon which to estimate potential benefit is limited. No Phase II trials of fulvestrant after progression on an AI have been completed, but limited data are available from interim analyses of two ongoing clinical trials.

Perey et al. (27) presented data at the 2002 San Antonio Breast Cancer Symposium on 32 of a projected 93 patients treated with fulvestrant after progression on tamoxifen and an AI. Two (6%) patients achieved a PR and 9 (28%) achieved SD for at least 24 weeks for a clinical benefit rate of 34%.

Steger et al. (28) reported data on 40 patients treated with fulvestrant after tamoxifen and at least one AI. Three (7%) patients achieved a PR, and 17 (43%) patients achieved SD for a clinical benefit rate of 50%.

Additional follow-up of these trials and results from a North Central Cancer Treatment Group trial (N0032), which to date has accrued 57 patients, should clarify the potential benefit of fulvestrant after progression on an AI. The results thus far suggest a low objective response rate but a higher rate of SD, the latter likely reflecting to some extent the natural history of disease in the patients enrolled.

Before the introduction of tamoxifen in the 1970s, high-dose estrogen was the endocrine treatment of choice in postmenopausal women with advanced breast cancer. Tamoxifen was considered to be preferable to estrogen therapy not because of a superior efficacy but rather improved tolerability as demonstrated in Phase III trials (29, 30). One of these trials (29) involving diethylstilbestrol (DES) versus tamoxifen has been updated. This trial was small in size by today’s standards but has mature survival data in that 95% of the 143 eligible patients have died (31). There was no significant difference between the two agents in terms of response rates and time to progression. However, survival was modestly and significantly better for women initially treated with DES (adjusted P = 0.039) with median survival of 3.0 versus 2.4 years and 5-year survival rates of 35 and 16%, respectively.

Lønning et al. (32) performed a Phase II trial of high-dose DES in women with prior endocrine exposure. Eligibility criteria included evidence of endocrine sensitivity as indicated either by a prior response or by 6 months of disease stability on an endocrine agent. Thirty-two patients were entered with a median of four prior endocrine regimens (range, 2–10) and can be considered heavily pretreated. All of the patients had received prior treatment with AIs, including 19 patients with exposure to a third generation agent (anastrozole and/or exemestane), 5 patients to a second generation agent (formestane), and 8 patients to the first generation agent aminoglutethimide. In terms of hormonal receptor status, one-quarter had either ER- and progesterone receptor-negative tumors (2 patients), or the receptors were unknown (6 patients). Considering these patient characteristics, it is remarkable that 10 (31%) achieved a complete response (4 patients, 12%) or PR (6 patients, 19%), and half of these responses lasted for >1 year.

High-dose estrogens are not tolerated by all patients as illustrated by 6 patients (19%) in this trial stopping therapy because of side effects. This proportion is on the same order as our experience (29) in the first-line treatment setting where 9 of 74 patients (12%) discontinued DES because of side effects. However, it is this author’s experience that a stepwise escalation of the estrogen dose over 1–2 weeks will ameliorate the toxicity and make the treatment tolerable for some patients.

Insights into potential mechanisms of resistance to estrogen deprivation therapy, as with the aromatase inhibitors, come from the work of Masamura et al. (33) who hypothesized that response observed to subsequent endocrine therapy could be related to increased sensitivity to E2 because of adaptation to E2 deprivation by tumor cells. These investigators grew MCF-7 cells in serum-free medium to eliminate E2 for extended periods of time and designated them long-term estrogen-deprived (LTED) cells. When the E2 dose response curve was generated, a typical bell-shaped curve was seen with an initial increase in stimulation at lower doses, a peak stimulation, and a progressive diminution in stimulation at higher E2 concentrations. The LTED cells showed maximal stimulation at 10^-14 m E2 compared with 10^-10 m for the wild-type MCF-7 cells. Thus, the LTED cells had maximal stimulation at a level of E2, which was 1/10,000^th of that for the wild-type cells.

Song et al. (34) studied LTED and wild-type MCF-7 cells with regard to the effect of E2 concentrations on apoptosis. Apoptosis was induced by high concentrations (≥0.1 nm) E2 in LTED cells, with a 7-fold increase over controls and a concomitant 60% decrease in growth but not in wild-type MCF-7 cells. The authors presented data showing that only LTED cells expressed Fas protein and suggested that high-dose estrogen may induce tumor regression in postmenopausal women through activation of Fas-mediated apoptosis. These laboratory findings combined with clinical data from Lønning et al. (32) are provocative and indicate that further study of estrogens in the metastatic setting are warranted (35).

The progestins, particularly megestrol acetate and less so the androgen fluoxymesterone, are commonly used as salvage endocrine therapy. Although fluoxymesterone has been relegated to a low standing in the endocrine therapy hierarchy for years, this has been the case for the progestins only since the introduction of the third generation AIs and fulvestrant. Despite the use of these agents in clinical practice, there are no prospective data on their value to the patient with prior AI exposure.

The recent introduction of anastrozole as an option for adjuvant therapy in postmenopausal women has resulted in a new category of patients who can present with metastatic disease. In the foreseeable future, most patients treated with NS-AIs will likely present as NS-AI resistant, but there are no Phase III trials of agents upon which to base sequencing strategies in these patients. Phase II data are available for exemestane and high-dose estrogens and retrospective data for tamoxifen and fulvestrant.

Well-conducted Phase III trials provide evidence for choosing first-line therapy for advanced disease in patients who have not received a NS-AI. In tamoxifen-sensitive patients, one can choose either of the NS-AIs, anastrozole or letrozole. In tamoxifen-resistant patients, one can choose either of the NS-AIs anastrozole and letrozole, the steroidal-AI exemestane, or the ER down-regulator fulvestrant.

Additional clinical trials are needed to determine an optimal sequencing strategy. This is particularly the case in patients with progression on aromatase inhibitor therapy.

Dr. Steven Come: In most of these studies in resistant disease, the response is almost entirely just stable disease.

Dr. Ingle: There is something to be said for looking at objective response rates, especially when these are highly selected patients. If you are giving a third- and fourth-line endocrine therapy, then these patients had very indolent or very favorable disease.

Dr. Matthew Ellis: One of the major reasons why clinicians don’t use high-dose estrogen is concern about one, thromboembolism and two, hypocalcemia. A third safety issue is fluid accumulation and heart failure. Perhaps you could give us some guidance from your experience.

Dr. Ingle: When we used DES the major problem was fluid retention, and incontinence was the most disabling problem for several women. In using estradiol, I have not seen the problems that we used to see with DES. Dr. Santen and I have been talking about doing a clinical trial. We have treated 5 patients with high-dose estradiol and the first 4 had an excellent response. They tolerated the drug much better than I remembered with DES. You escalate slowly over a week and a half to limit nausea and vomiting; you have them watched for fluid retention, then you act quickly if they have it.

Dr. Richard Santen: Dr. Ingle published a study in 1981 comparing tamoxifen with DES (N. Engl. J. Med., 304: 16–21, 1981). Four years ago he did a follow-up on those patients and it turned out that those who got DES survived statistically significantly longer than the patients treated with tamoxifen. In other studies the premenopausal patients did not respond to DES. The patients who were 5 years postmenopausal had about a 20% response and those who were to 20–30 years postmenopausal had approximately a 40% response rate. My hypothesis would be that long-term estradiol deprivation is a factor and that perhaps these patients have up-regulation of Fas receptor and other changes that would enable estrogen to stimulate apoptosis. This is a new concept that needs to be tested.

Dr. Ingle: The other thing to keep in mind about estrogens is individual differences in response and tolerability. Twenty-five years ago, Carter’s group did a randomized study showing responses from 1.5 mg/day DES all of the way up to 1500 mg (J. Am. Med. Assoc., 237: 2079–2080, 1977). I once put a woman on DES 5 mg/day who became extremely nauseated; on 1 mg/day she had a beautiful response. Our ability to do dose response testing is not to be applauded in many instances. Here it is still very empirical.

Dr. Mina Bissell: There is no reason why the combination therapies that you and Dr. Nicholson discussed should not include estrogen because these pathways all talk to each other. The ER, EGFR and PI3 kinase receptor all talk to each other. So when you add the estrogen, you are doing combination therapy, and that is the very important concept that people should think about.

Dr. Rachel Schiff: If we keep our HER2-expressed cells on estrogen, the tumor grows, but under the microscope the cells look nonaggressive. The moment you do any endocrine therapy you get a much more aggressive phenotype. At least in high growth factor signaling, it may be that classical genomic activity of ER sustains proliferation but inhibits the metastatic phenotype.

Dr. Paul Goss: When you withdraw an aromatase inhibitor in a postmenopausal woman, you get a measurable return of the estrogen. If Dr. Santen’s theory is translatable to human breast cancer in the patients, you should get a withdrawal response, even at those levels of estrogen. That again is another concept that has not really been tested, but I am not aware of any anecdotal reports of withdrawal responses to aromatase inhibitors.

Dr. Ingle: The problem with trying to come up with an evidence-based sequencing strategy is that there is no evidence. The pharmaceutical firms are to be congratulated for doing these large pivotal trials. But sequencing of therapies is an important issue, and where are we going get the studies?

Dr. Ellis: Of all of the strategies that have been discussed here, I think there is rising enthusiasm for the high-dose estrogen.

Dr. Goss: The other strategy we’ve discussed is decremental estrogen reduction. Is it really necessary to estrogen-ablate all of the patients? That again is not a question that will be answered easily by the pharmaceutical industry, but it is equally important to the add-on therapy with estrogen.

Dr. Ingle: How you choose the proper dose for many medications is not clear. We did a tamoxifen pharmacokinetic study that found over a log difference in the levels of 4-hydroxy-tamoxifen and N-desmethyl-tamoxifen (Clin. Cancer Res., 5: 1642–1649, 1999). These differences are related to genetic polymorphisms, probably, and in years to come that will be more important to understand.