Original article. Introduction

Transcription

1 Original article Annals of Oncology 14: , 2003 DOI: /annonc/mdg362 Mature results of a randomized phase II multicenter study of exemestane versus tamoxifen as first-line hormone therapy for postmenopausal women with metastatic breast cancer R. Paridaens 1 *, L. Dirix 2, C. Lohrisch 3, L. Beex 4, M. Nooij 5, D. Cameron 6, L. Biganzoli 3, T. Cufer 7, L. Duchateau 8, A. Hamilton 3, J. P. Lobelle 9 & M. Piccart 10 On behalf of the European Organization for the Research and Treatment of Cancer (EORTC) Investigational Drug Branch for Breast Cancer (IDBBC) 1 Universitair Ziekenhuis, Gasthuisberg, Leuven; 2 Algemeen Ziekenhuis Sint-Augustinus, Wilrijk; 3 IDBBC EORTC, Brussels, Belgium; 4 Universitair Ziekenhuis, Nijmegen; 5 Leids University Medicine Center, Leiden, The Netherlands; 6 Edinburgh University, Western General Hospital, Edinburgh, UK; 7 Institute of Oncology, Ljubljana, Slovenia; 8 EORTC Data Center, Brussels; 9 Pharmacia, Brussels; 10 Institut Jules Bordet, Brussels, Belgium Received 15 July 2002; revised 24 January 2003; accepted 11 March 2003 Background: Women with hormone-responsive metastatic breast cancer (MBC) may respond to or have stable disease with a number of hormone therapies. We explored the efficacy and safety of the steroidal aromatase inactivator exemestane as first-line hormonal therapy in MBC in postmenopausal women. Patients and methods: Patients with measurable disease were eligible if they had received no prior hormone therapy for metastatic disease and had hormone receptor positive disease or hormone receptor unknown disease with a long disease-free interval from adjuvant therapy. They were randomized to tamoxifen 20 mg/day or exemestane 25 mg/day in this open-label study. Results: Blinded independently reviewed response rates for exemestane and tamoxifen were 41% and 17%, respectively. Fifty-seven per cent of exemestane- and 42% of tamoxifen-treated patients experienced clinical benefit, defined as complete or partial response, or disease stabilization lasting at least 6 months. There was a low incidence of severe flushing, sweating, nausea and edema in women who received exemestane. One exemestane-treated patient had a pulmonary embolism with grade 4 dyspnea. Conclusions: Exemestane is well tolerated and active in the first-line treatment of hormone-responsive MBC. An ongoing EORTC phase III trial is comparing the efficacy, measuring time-to-disease progression, of exemestane and tamoxifen. Key words: aromatase inactivators, hormone therapy, metastatic breast cancer Introduction With rare exceptions, metastatic breast carcinoma is incurable, with a median survival time of 18 months from the time of initiation of chemotherapy in patients unselected for estrogen receptor (ER) status, and of 36 months from the time of initiation of hormonal therapy in patients with positive or unknown hormonal receptor status [1 3]. Hormonal therapy is thus a systemic treatment option that can produce meaningful responses and disease stabilization which, in some cases, are of long duration. Predictors of response to hormone therapy include the presence of ER and/or progesterone (PgR) receptors in the tumor, and a long interval from diagnosis to first relapse [4]. *Correspondence to: Dr R. Paridaens, Dienst Gezwelziekten, Universitair Ziekenhuis, Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium. Tel: ; Fax: ; robert.paridaens@uz.kuleuven.ac.be Until recently tamoxifen, which competitively inhibits estrogen binding to the ER, was the standard first-line hormonal therapy for metastatic breast cancer (MBC). The main pathway for estrogen synthesis in postmenopausal women is through conversion of androstenedione to estrone, and testosterone to estradiol by aromatase, an enzyme present in many non-endocrine tissues (muscle, fat, and normal and malignant breast tissue). Inhibition of aromatase represents a non-cross-resistant mechanism to impede the interaction between the estrogen and its specific receptor. The aromatase inhibitor aminoglutethimide was for many years a standard second-line option after tamoxifen failure. However, the new generation of non-steroidal aromatase inhibitors (anastrozole, letrozole) and steroidal aromatase inactivators (formestane, exemestane) are preferred based on greater enzyme specificity and more favorable side-effect profiles. These agents effectively inhibit the peripheral conversion of androgens to estrogens, leading to a >90% reduction in circulating levels of estrogen [5 7]. They represent a major therapeutic advance, in terms of both activity 2003 European Society for Medical Oncology

2 1392 and tolerability. Anastrozole and letrozole are effective and widely used either as first-line therapy for metastatic disease, or following tamoxifen failure in postmenopausal women, based on favorable and consistent results from several randomized controlled trials [8 14]. Exemestane has the advantage over formestane of being orally bioavailable, and differs from non-steroidal aromatase inhibitors by irreversibly binding to and inactivating aromatase [15]. Exemestane produces a similar magnitude of estrogen suppression as the non-steroidal aromatase inhibitors. Response rates of 7 28% have been reported for exemestane after previous hormonal therapies in phase II trials [16, 17]. A phase III trial reported superiority of exemestane over megestrol acetate after previous tamoxifen in terms of time-to-progression and survival [18]. Interestingly, exemestane also showed modest activity (6% response rate, 25% clinical benefit) in patients pre-treated with non-steroidal aromatase inhibitors [19]. This European Organisation for the Research and Treatment of Cancer (EORTC) Investigational Drug Branch for Breast Cancer (IDBBC) open-label randomized phase II study was designed to explore the activity and tolerability of exemestane as first-line hormonal therapy for metastatic disease in postmenopausal women with hormone-responsive MBC. The mature results are reported herein. Patients and methods Patients In order to be eligible for this study, patients were required to be female, postmenopausal, have a maximum Eastern Cooperative Oncology Group (ECOG) performance status of 2, and have histologically or cytologically confirmed metastatic or locally recurrent inoperable breast cancer with at least one bidimensionally measurable lesion. Postmenopausal status was defined using any of the following: natural menopause or radiotherapy-induced oophorectomy with at least 1 year since last menses; chemotherapy-induced menopause with at least one year from last menses and serum follicle stimulating hormone (FSH), luteinizing hormone (LH) and plasma estradiol levels in the postmenopausal range; surgical castration; and for women younger than 56 years with a hysterectomy but at least one intact ovary, serum FSH, LH and plasma estradiol levels in the postmenopausal range. In order to be considered measurable, lesions had to be measurable in two perpendicular diameters, with at least one diameter 2.5 cm and not pre-irradiated. Exceptions included: liver and extraabdominal lesions that were followed by computed axial tomography (CT) of at least 2.0 cm largest diameter; lesions followed by chest X-ray of 1.5 cm largest diameter; skin lesions that had been photographed of 1 cm largest diameter; and bone lesions that were purely lytic, identified by X-ray or CT scan, surrounded by calcified bone, a minimum of 1 cm diameter, and not previously irradiated unless they had clearly progressed since and the irradiation had occurred at least 3 months before. Patients could not have been treated with hormone therapy, including ovarian ablation, or more than one chemotherapy regimen for metastatic disease. Complete recovery from acute side-effects of chemotherapy and radiotherapy was required. Adjuvant chemotherapy, tamoxifen and ovarian ablation were allowed. Adjuvant tamoxifen had to be discontinued at least 6 months before enrollment in the present trial. Patients primary tumors could be either hormone receptor positive with a disease-free interval (DFI) of at least 6 months, or hormone receptor status unknown with a DFI of at least 2 years from the end of adjuvant therapy (or surgery in the absence of adjuvant therapy). Positive hormone receptors were defined as follows: 10 fmol of H 3 -estrogen or 20 fmol of H 3 -progesterone binding per milligram of cytosol protein for dextran-coated charcoal (DCC) and sucrose density techniques; 0.10 fmol of H 3 -estrogen or 0.20 fmol of H 3 -progesterone per milligram of DNA for the immunofixation and enzyme immunoassay (EIA) detection methods; or a pathology report indicating positive hormone receptors for the immunohistochemistry method. Finally, patients had to have adequate visceral organ function, defined as serum creatinine and bilirubin <1.5 the upper limit of normal (ULN), alanine aminotransferase <2.5 and <5 ULN for patients without and with liver metastases, respectively, and had to provide written or witnessed oral informed consent. Exclusion criteria included: previous history of any other cancer with the exception of non-melanoma skin cancer and curatively treated in situ carcinoma of the cervix; rapidly progressive and/or large volume visceral metastases; central nervous system (CNS) metastases; inflammatory breast cancer; severe co-morbid illness; and prior deep venous thrombosis. Anti-coagulant medication within 2 weeks of registration, and hypolipemic agents started <3 months before study registration were not allowed. Bisphosphonates were allowed provided that the patient in question had at least one non-boney target lesion and continued on bisphosphonates throughout. Bisphosphonates could only be given de novo during the study for the treatment of acute hypercalcemia of malignancy and for a maximum of 7 days. Concomitant pain medications were allowed, as was radiotherapy for painful bone metastases. Study design This was a multi-center study conducted in 13 centers across six countries, and the EORTC Protocol Review Committee and participating institution ethics committees approved the protocol. The design was an open-label phase II trial in which eligible patients were randomized to tamoxifen 20 mg orally or exemestane 25 mg orally once daily. The dose chosen for exemestane was based on phase I and phase II data showing equivalent levels of circulating estrogen suppression but a lower incidence of adverse events and androgen stimulation with 25 mg versus higher doses. Randomization was performed centrally at the EORTC data center using a minimization procedure, and in a 1:1 ratio, after stratification for institution, adjuvant tamoxifen (yes/no), chemotherapy for metastatic disease (yes/no) and dominant disease site (visceral ± other, bone only, bone and soft tissue, soft tissue only). Treatment was continued until progression of disease, unacceptable toxicity, patient refusal or start of any new anti-cancer therapy. Baseline assessments included history and physical examination, performance status assessment, evaluation of all potential tumor sites (bone scan, liver imaging, chest X-ray at a minimum), routine biochemistry and hematology blood analysis. The following parameters were recorded at baseline and during treatment in the context of parallel safety substudies: coagulation parameters (antithrombin 3, prothrombin time, partial thromboplastin time and fibrinogen), lipid parameters (total and high-density lipoprotein cholesterol and triglycerides) and endogenous hormone parameters (thyroid stimulating hormone, free T3, free T4, estrone, estradiol, estrone sulfate and sex globulin binding hormone). The results of these substudies will be reported separately. Patients were assessed for adverse events and tumor response after 8, 16 and 24 weeks of treatment, and thereafter every 12 weeks until week 96, then every 24 weeks in the absence of progression. Adverse events were sought by patient interview at each scheduled visit and recorded using standard National Cancer Institute of Canada Common Toxicity Criteria (NCIC-CTC) toxicity grading scales [20]. Response was assessed using World Health Organization (WHO) criteria [21]. All involved disease sites were re-evaluated and target lesions were re-measured at each tumor evaluation, as were determination of performance status and physical examination. Full restaging was required every 24 weeks. Patients who discontinued therapy in the absence of progression were to be followed every 12 weeks to monitor for eventual progression and for resolution of any recorded adverse events. The requirement for radiotherapy on the only target lesion(s) before 8 weeks of therapy rendered a

3 1393 patient not evaluable for response; the requirement for radiotherapy after 8 weeks was considered to represent progression. Coagulation parameters and routine hematology and biochemistry were drawn at each scheduled follow-up visit. Routine urinalysis and blood samples for lipids and endogenous hormone parameters were collected at weeks 8 and 24, and thereafter every 24 weeks. Lipid parameters were analyzed at a central laboratory. Transvaginal endometrial ultrasound at baseline and every 24 weeks was recommended but optional, with endometrial biopsy left to the discretion of the investigator. End points and statistical methodology A randomized phase II design was chosen not to enable comparison of the efficacy of the two drugs, but to establish a go, no-go rule for exemestane activity and safety before a formal randomized phase III trial. In this way, patients enrolled in the phase II portion of the study could be analyzed within the phase III component if the statistical criteria were met for study extension to phase III. The primary efficacy end point in each arm was response rate (complete plus partial response) assessed by standard WHO criteria. Using a one-stage design and an alpha and beta of 0.10, a response rate of at least 25% for exemestane was considered the minimum acceptable efficacy for consideration of further study of this drug in the first-line hormone therapy setting. This resulted in the requirement of 50 eligible and evaluable patients in each arm. No minimum response criteria were proposed for tamoxifen as it was considered to be the gold standard. Response rates (RRs) were derived based on the intention-to-treat principle together with their 95% confidence interval (CI). A second analysis was planned including only evaluable patients. Given the phase II design and therefore inadequate power, no statistical comparison of efficacy end points between the two treatments was planned or performed. Complete and partial responses (CR and PR, respectively) and disease stabilization [no change (NC)] had to be confirmed by repeat assessment of target lesions no less than 4 weeks after the first documentation. Responses that had not been confirmed were downgraded: for example, an unconfirmed CR preceded by a NC was designated as a NC, an unconfirmed CR preceded by a PR was a PR, and an unconfirmed PR as first response evaluation or an unconfirmed NC followed by disease progression (PD) was a PD. All investigatorreported CRs, PRs and NCs, with the exception of lesions that were only assessed clinically or by ultrasound, were reviewed externally by a panel of two independent radiologists and one medical oncologist who were blinded to treatment assignment. Bone scans were reviewed only to confirm the absence of bone disease when a CR was reported. Other efficacy end points included clinical benefit; (CB defined as the proportion of patients in each arm with an NC of at least 6 months duration, a CR or a PR) and median duration of response and of clinical benefit. Time-toevent end points were measured from the date of treatment start to the date of progression or death, whichever occurred first. Patients who had neither progressed nor died by the date of analysis were censored on the last date of follow-up. Patients who discontinued therapy for any other reason were considered to have progressed on the date they discontinued therapy. The major safety end point was the incidence and severity of adverse events, using the NCIC-CTC grading scales. All patients who started therapy were eligible for safety analysis. The analysis of adverse events was descriptive and summarized as worst grade per patient. Assessment of coagulation parameters, lipid parameters, endogenous hormones and endometrial thickness by transvaginal ultrasound will be analyzed and reported separately. Results A total of 122 patients were randomized between October 1996 and May 1999, 62 to exemestane (E) and 60 to tamoxifen (T). Two patients (one in each arm) were ineligible on the basis of not having MBC, and were excluded from efficacy analyses and patient characteristics. One patient had metastatic renal cell carcinoma and never received any study therapy; the other was found on independent review to have no metastatic sites, and thus was actually receiving adjuvant therapy. Of the remaining 120 patients, an additional seven (five exemestane, two tamoxifen) were not evaluable for response for the following reasons: only target lesion surgically pinned (one); target lesion non-neoplastic (one); target lesions smaller than specified by the protocol (three); and off study for reasons other than PD after only one evaluation (two). These patients were included in the intention-to-treat analysis of response, time-to-event end points and safety. Table 1 summarizes the baseline characteristics for the 120 study patients. The majority of patients had at least two involved sites (80% in both arms) and visceral disease (53% E and 59% T). Bone was the dominant metastatic site in 31% of E- and 32% of T-treated patients. The median times from the end of adjuvant tamoxifen in the 11 (18%) E and 11 (19%) T patients who had been so treated were 42 (range 19 94) and 47 (range ) months, respectively. Among patients randomized to E, 21% and 3% received chemotherapy in the adjuvant and metastatic settings, respectively. Of patients randomized to tamoxifen, 17% had prior adjuvant chemotherapy while 7% had prior metastatic chemotherapy. Response rate and clinical benefit The proportion of patients in each arm with a CR, PR, NC and PD as best response to therapy is shown in Table 2. The overall response rate and clinical benefit rates are also shown. By intention-to-treat, 25 patients (41%; 95% CI 29% to 53%) in the exemestane arm achieved a response (10% CR), while in the tamoxifen arm there were 10 responses (17%; 95% CI 7% to 27%), of which one was a CR (2%). Responses were seen in soft tissue, bone and visceral dominant disease, as outlined in Table 3. Ten and 15 patients in the exemestane and tamoxifen arms, respectively, achieved disease stabilization (NC) of at least 6 months duration, giving overall clinical benefit rates of 57% (95% CI 45% to 70%) and 42% (95% CI 30% to 55%), respectively. When the response and clinical benefit rates were calculated for the subset of evaluable patients, the results were similar (Table 2). Among the 22 patients who had received adjuvant tamoxifen, five out of 11 (45%) treated with exemestane and two out of 11 (18%) treated with tamoxifen responded. A further two (18%) exemestane- and four (36%) tamoxifen-treated patients had a NC of at least 6 months. Exclusion of these 22 patients yielded similar results as for the entire study population: the RR and CB were 40% and 56%, respectively, for exemestane-treated (n = 50), and 17% and 40%, respectively, for tamoxifen-treated (n = 48) patients. Duration of response and clinical benefit At the time of end point analysis, nine exemestane- and five tamoxifen-treated patients were still on treatment without evidence of progression. Table 4 shows reasons for treatment discontinuation, the majority being disease progression. The median

4 1394 Table 1. Patient characteristics Exemestane (n = 61) Tamoxifen (n = 59) Median age, years (range) 62 (37 78) 63 (46 87) Age group, % <55 years years years ECOG performance status, % Median DFI, years (range) 4.13 (0 29.6) 4.01 (0 27.9) Hormone receptor status, % ER+ and PgR ER+ or PgR (ER and PgR+) 10 9 (ER+ and PgR ) ER and PgR unknown 7 9 Number of involved sites, % > Dominant disease site, % Soft tissue 16 9 Bone Visceral Prior treatment, % Radiotherapy Adjuvant chemotherapy Chemotherapy for MBC 3 7 Adjuvant tamoxifen Median time off adjuvant tamoxifen, 42 (19 94) 47 (14 175) months (range) DFI 0 (stage IV at diagnosis) <24 months months ER, estrogen receptor; DFI, disease-free interval; MBC, metastatic breast cancer; PgR, progesterone receptor. response duration was 16 months (95% CI months) for exemestane and 22 months (95% CI months) for tamoxifen. The median duration of clinical benefit was 14 months (95% CI months) for both exemestane and tamoxifen. Adverse events Overall both drugs were well tolerated. Only one grade 4 adverse event occurred: a deep venous thrombosis and non-fatal pulmonary embolism with grade 4 dyspnea occurring during the first 8 weeks of treatment with exemestane. The patient in question continued exemestane until PD was documented at week 36 (fourth scheduled evaluation). A gastrointestinal bleed of unknown grade but requiring hospitalization occurred in a patient receiving tamoxifen. This patient had a known history of alcohol abuse, which may have contributed to or caused the bleed. She was lost to follow-up immediately afterwards, thus no further details are

5 1395 Table 2. Overall tumor response by treatment assignment Exemestane Patients, n Intention-to-treat (n = 61), % (95% CI) Evaluable patients (n = 56), % (95% CI) Tamoxifen Patients, n CR, complete response; PR, partial response; NC, no change; CB, clinical benefit; PD, progression of disease. Intention-to-treat (n = 59), % (95% CI) Evaluable patients (n = 57), % (95% CI) Overall response rate (29 53) 45 (32 58) (7 27) 18 (8 27) CR PR NC <6 months NC >6 months CB (CR, PR, NC >6 months) (45 70) 63 (50 75) (30 55) 44 (31 57) PD Not evaluable Table 3. Response by dominant disease site and hormone receptor status Exemestane (61 patients) Tamoxifen (59 patients) No. with dominant disease No. of responders No. with dominant disease No. of responders Visceral (34) 35 8 (23) Bone 19 8 (42) 19 2 (11) Soft tissue 10 6 (60) 5 0 (0) ER (45) 49 9 (18) ER-, PgR (0) 5 1 (20) ER and PgR unknown 4 2 (50) 5 0 (0) ER, estrogen receptor; PgR, progesterone receptor. known. Two patients treated with exemestane had grade 3 cardiac arrhythmias (bradycardia in one, atrial fibrillation in the other), which were considered by the investigators to be unrelated to the study drug. Neither of the patients discontinued study medication. Two patients discontinued exemestane in the absence of progression, one due to anxiety about her diagnosis of metastatic disease, the other on the advice of her physician due to a grade 1 rash that resolved with temporary discontinuation and recurred when exemestane was resumed. There were two non-disease-related deaths in the study, both in the tamoxifen arm, one a result of suicide and the other due to an unknown cause. Table 5 shows the incidence of grade 1 4 adverse events for exemestane and tamoxifen (worst grade per patient). Of particular interest is the low incidence of hot flashes, arthralgias, sweating, edema and nausea reported by patients treated with exemestane. Skin toxicity, reported in five (8%) E patients at the grade 2 and 3 level, consisted of one each of grade 2 pruritus, erythema with scaling and vesicular skin rash, and one each of grade 2 and 3 unspecified rash. Discussion Responses, as determined by independent reviewers without knowledge of treatment assignment, occurred in 41% and 17% of Table 4. Treatment discontinuation Exemestane (61 patients) Off study, n Progression Early progression 2 2 Patient request a 2 0 Death b 1 1 SAE c 0 1 Lost to follow-up 0 1 Did not tolerate treatment 0 1 On study, n 9 5 Tamoxifen (59 patients) a Anxiety, n = 1; grade 1 recurrent rash, n = 1. b Suicide, n = 1 (E); unknown cause, n = 1 (T). c SAE, serious adverse event was a gastrointestinal bleed of unknown grade; patient was lost to follow-up immediately afterwards. exemestane- and tamoxifen-treated patients, respectively, in this open-labeled, multi-center, phase II, randomized trial of first-line hormone therapy for postmenopausal women with hormone receptor sensitive, measurable metastatic disease. Clinical benefit

6 1396 Table 5. Grade 1 4 adverse events Exemestane (n = 62) Tamoxifen (n = 59) Grade 1 Grade 2 Grade 3 Grade 4 Grade 1 Grade 2 Grade 3 Anorexia Arthralgias Dyspnea Edema Fatigue Hot flashes Infection Nausea Pain: bone Pain: abdominal/other Sweating Skin Weight gain Weight loss Increased GGT Increased ALP Increased AST or ALT Increased bilirubin GGT, gamma glutamyl transferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase. Grade 4 rates were 57% for exemestane and 42% for tamoxifen, and the duration of response and of clinical benefit was similar in the two arms. Responses with exemestane were seen in visceral, bone and soft tissue dominant disease. Coupled with the low toxicity profile, this data is encouraging. The low incidence of grade 2 and 3 hot flashes, together with the absence of grade 2 and 3 sweating, may be related to the steroidal nature of the compound and to the fact that a major metabolite of exemestane, 17-OH exemestane, has weak androgenic activity [7]. There was also a low observed incidence of arthralgias in the exemestane-treated women, a sideeffect experienced by 15% of patients receiving non-steroidal aromatase inhibitors. One exemestane-treated patient had a non-fatal pulmonary embolism, and there were no deep venous thromboses or pulmonary emboli in tamoxifen-treated patients. The incidence of thromboembolic events in the Pharmacia safety database (Pharmacia, unpublished data) of >1000 patients with metastatic disease treated with exemestane in clinical trials is 1.3% (14 events, including four myocardial infarctions). In the randomized trial comparing megestrol acetate with exemestane as second-line hormonal treatment for metastatic disease, the incidence of reported dyspnea (grade not given) was 0.3% among 366 exemestanetreated patients, and there were no venous thromboembolic events [12]. These data are reassuring and suggest that the risk of thromboembolic complications secondary to exemestane is low; however, this drug has not been studied extensively in women with a history of venous thromboembolic disease. The severe event observed in our limited series may be fortuitous and not necessarily treatment, but rather disease related. The numerous large-scale randomized adjuvant studies presently running will generate accurate data in the near future, acquired in an appropriate clinical setting, allowing an unbiased comparison of the safety profiles of tamoxifen and exemestane. The response rate, clinical benefit rate and duration of response observed for exemestane in this trial compare favorably with rates observed for first-line non-steroidal aromatase inhibitors (Table 6). Notwithstanding the absolute numbers, it must be pointed out that the literature regarding the activity of first-line non-steroidal aromatase inhibitors is based on large, blinded, randomized, controlled trials, whereas these data on exemestane are on fewer patients and is not blinded. The findings in this study are being explored further in an ongoing randomized phase III trial, which is powered to compare the efficacy of tamoxifen and exemestane. While exemestane results in irreversible inactivation of aromatase, the non-steroidal aromatase inhibitors anastrozole and letrozole lead to reversible inhibition [23]. In theory, exemestane could therefore result in more effective global or intra-tumoral aromatase inhibition. Another difference between these agents is that steroidal aromatase inactivators can cause profound and prolonged reduction in aromatase levels, while anastrozole and letrozole can lead to aromatase up-regulation, thus possibly leading to the emergence of a resistant cell population in a shorter time period with the non-steroidal aromatase inhibitors [24, 25]. One

7 1397 Table 6. Comparison of characteristics and outcome for patients treated with tamoxifen or with a non-steroidal aromatase inhibitor, within the frame of a randomized phase III study conducted in first-line metastatic breast cancer patients Characteristics Bonneterre et al. [8] Nabholtz et al. [9] Mouridsen et al. [10] T A T A T L Total sample size, n % response rate % Clinical benefit % ER unknown % visceral dominant % >2 disease sites Not specified Not specified % metastatic at diagnosis % adjuvant anti-estrogens Minimum DFI from adjuvant T (months) Measurable criteria Uni- or bidimensional (UICC criteria) Uni- or bidimensional (UICC criteria) Independent response rate review No No Yes Uni- or bidimensional (UICC criteria) A, anastrozole; ER, estrogen receptor; L, letrozole; T, tamoxifen; UICC, International Union against Cancer; WHO, World Health Organization. study, in which all patients (n = 241) who had received prior nonsteroidal aromatase inhibitors (anastrozole, letrozole or vorozole in 44%) were treated with exemestane, reported a 6% RR and a 25% clinical benefit rate, suggesting that this may in fact be the case [19]. Although the magnitude of circulating estrogen suppression is similar with the non-steroidal inhibitors and exemestane, down-regulation and irreversible aromatase inactivation by exemestane may provide enhanced inhibition of the interaction between estrogen and the estrogen receptor at the tumor level. Limited series have also reported that responses and clinical benefits can occur the other way, with non-steroidal aromatase inhibitors administered after demonstrated disease progression under exemestane, although this has not been examined in a prospective randomized trial setting. Only a head-to-head comparison of exemestane and anastrozole (which is underway in first-line metastatic patients with visceral disease), including a cross-over upon progression, will allow quantification of non-cross-resistance in both sequences, and will enable determination of whether one class is superior to the other. In this trial, the RR to tamoxifen was relatively low compared with other recent randomized trials with tamoxifen as the gold standard first-line hormone therapy (Table 6). Considering that the population studied was small and the confidence intervals are wide, the true response rate may be as high as 27%. Factors which may in part account for the lower response to tamoxifen in this trial include the independent review of responses, the fact that adjuvant tamoxifen was allowed, and the burden of disease (visceral dominant disease 59%; more than two involved sites 49%) compared with the other trials. Among tamoxifen-treated patients, the independent response review led to a downstaging from a response to a non-response in two cases, and an upgrade from a non-response to a response in one case. The first-line trials that compared anastrozole and tamoxifen did not independently review responses, which may in part account for the higher response rates they reported for tamoxifen (32.6% and 17%) [8, 9]. The letrozole versus tamoxifen trial did independently review responses and reported a RR for tamoxifen of 20%, closer to the rate we observed [10]. Compared with the first-line trial of letrozole versus tamoxifen, the proportion of women in this trial with more than two disease sites was 49% (versus 11%), visceral dominant disease 59% (versus 45%), hormone receptor unknown disease 8.5% (versus 33%), and adjuvant anti-estrogens 17% (versus 18%), all factors that could decrease the probability of responding to hormonal interventions [10]. Although cross-trial comparisons are problematic, it should be noted that the 95% CI for the tamoxifen RR in the two trials are overlapping: in ours, the RR was 17% with a 95% CI of 7% to 27%, and in the Mouridsen trial, the RR was 20% with a 95% CI of 17% to 24%. The 95% CI of 12% to 23% for the RR to tamoxifen in the Nabholtz et al. anastrozole versus tamoxifen trial [9] also overlaps with ours [7]. It must be pointed out, however, that the confidence intervals in our study are wide due to the small sample size, and that a more accurate assessment of tamoxifen efficacy in this heavily diseaseburdened population will emerge with the phase III data. Importantly, it has been recognized that long-term disease stabilization can provide meaningful palliation and is readily achievable with hormone therapy [26, 27]. In this study, clinical benefit was achieved in 42% of patients treated with tamoxifen, which compares favorably with clinical benefit rates reported for tamoxifen in other first-line studies (55.5%, 46% and 38% in the Bonneterre, Nabholtz and Mouridsen trials, respectively). Thus, although the RR was somewhat lower, the overall benefit observed for tamoxifen was well within the range observed with other first-line trials. Globally, this study reports a promising overall RR, clinical benefit rate and response duration for exemestane as first-line therapy for MBC among postmenopausal women. The incidence of serious toxicity is low and exemestane is well tolerated. Although these data are encouraging, they must be interpreted in light of several design weaknesses in the trial. These include that it

8 1398 was not blinded to either the patient or investigator, and that the trial size was small, which does not allow for an accurate assessment of the true efficacy of exemestane, or for a statistical comparison of end points between exemestane and tamoxifen. Although it is no substitution for up-front double blinding, the response and clinical benefit rates reported were adjudicated by two independent radiologists who had no knowledge of treatment assignment. This phase II trial was never designed to compare the two hormone therapies directly; however, a phase III extension is now being conducted by the EORTC. Acknowledgements We would like to thank the women who participated in this trial. We also thank Ms C. Yague, Ms E. Ngalula Kabanga and Ms V. Batter for their assistance with data management, and Ms M. Delval for secretarial help. We are grateful to the Fonds Heuson for having supported the fellowships of Drs C. Lohrisch and A. Hamilton. The following investigators and their institutions also participated in the study: investigators not listed as coauthors are Dr K. Roozendaal (Onze Lieve Vrouw Gasthuis, Amsterdam Oost, The Netherlands), Dr L. Mauriac (Fondation Bergonié, Bordeaux, France), Dr R. Coleman (Weston Park Hospital NHS Trust, Sheffield, UK), Dr J. Nortier (Diakonessenhuis, Utrecht, The Netherlands), Dr J. Jassem (Medical University of Gdansk, Gdansk, Poland), and Dr C. Seynaeve (Daniel den Hoed Kliniek, Rotterdam, The Netherlands). References 1. Chan S, Friedrichs K, Noel D et al. Prospective randomized trial of docetaxel versus doxorubicin in patients with metastatic breast cancer. J Clin Oncol 1999; 17: Mouridsen H, Sun Y, Gershanovich M et al. Final survival analysis of the double-blind, randomized, multinational phase III trial of letrozole (Femara ) compared to tamoxifen as first-line hormonal therapy for advanced breast cancer. Breast Cancer Res Treat 2001; 69: 211 (Abstr 9). 3. Klijn J, Beex L, Mauriac L et al. Combined treatment with buserelin and tamoxifen in premenopausal metastatic breast cancer: a randomized study. J Natl Cancer Inst 2000; 92: Paridaens R, Sylvester RJ, Ferrazzi E et al. Clinical significance of the quantitative assessment of estrogen receptors in advanced breast cancer. Cancer 1980; 46: Demers LM. Effects of fadrozole and letrozole on the inhibition of aromatase activity in breast cancer patients. Breast Cancer Res Treat 1994; 30: Geisler J, King N, Dowsett M et al. Influence of anastrozole (Arimidex), a selective, non-steroidal aromatase inhibitor, on in vivo aromatisation and plasma oestrogen levels in postmenopausal women with breast cancer. Br J Cancer 1996; 74: Paridaens R, Thomas J, Wildiers J et al. Safety, activity and estrogen inhibition by exemestane in postmenopausal women with advanced breast cancer: a phase I study. Anticancer Drugs 1998; 9: Bonneterre J, Thürlimann B, Robertson JFR et al. Anastrozole versus tamoxifen as first-line therapy for advanced breast cancer in 668 postmenopausal women: results of the Tamoxifen or Arimidex Randomized Group Efficacy and Tolerability Study. J Clin Oncol 2000; 18: Nabhlotz JM, Buzdar A, Pollak M et al. Anastrozole is superior to tamoxifen as first-line therapy for advanced breast cancer in postmenopausal women: results of a North American multicenter randomized trial. J Clin Oncol 2000; 18: Mouridsen H, Gershanovich M, Sun Y et al. Superior efficacy of letrozole versus tamoxifen as first-line therapy for postmenopausal women with advanced breast cancer: Results of a phase III study of the International Letrozole Breast Cancer Group. J Clin Oncol 2001; 19: Dombernowsky P, Smith I, Falkson G et al. Letrozole, a new oral aromatase inhibitor for advanced breast cancer: double-blind randomized trial showing a dose effect and improved efficacy and tolerability compared with megestrol acetate. J Clin Oncol 1998; 16: Hamilton A, Piccart MJ. The third-generation non-steroidal aromatase inhibitors: a review of their clinical benefits in the second-line hormonal treatment of advanced breast cancer. Ann Oncol 1999; 10: Buzdar AU, Jonat W, Howell A et al. Anastrozole versus megestrol acetate in the treatment of postmenopausal women with advanced breast carcinoma: results of a survival update based on a combined analysis of data from two mature phase III trials. Cancer 1998; 83: Gershanovich M, Chaudri HA, Campos D et al. Letrozole, a new oral aromatase inhibitor: randomised trial comparing 2.5mg daily, 0.5mg daily and aminoglutethimide in postmenopausal women with advanced breast cancer. Ann Oncol 1998; 9: Buzdar AU. Role of aromatase inhibitors in advanced breast cancer. Endocr Relat Cancer 1999; 6: Jones S, Vogel C, Arkhipov A et al. Multicenter, phase II trial of exemestane as third-line hormonal therapy of postmenopausal women with metastatic breast cancer. J Clin Oncol 1999; 17: Thürlimann B, Paridaens R, Serin D et al. Third-line hormonal treatment with exemestane in postmenopausal patients with advanced breast cancer progressing under aminoglutethimide: a phase II multicentre multinational study. Eur J Cancer 1997; 33: Kaufmann M, Bajetta E, Dirix LY et al. Exemestane is superior to megestrol acetate following tamoxifen failure in postmenopausal women with advanced breast cancer: results of a phase III randomized double-blind trial. J Clin Oncol 2000; 18: Lonning PE, Bajetta E, Murray R et al. Activity of exemestane in metastatic breast cancer after failure of nonsteroidal aromatase inhibitors: a phase II trial. J Clin Oncol 2000; 18: MacDonald J, Haller D, Mayer R. Grading of toxicity. In MacDonald J, Haller D, Mayer R (eds): Manual of Oncologic Therapeutics. Philadelphia, PA: Lippincott 1995; Miller AB, Hoogstraten B, Staquet M et al. Reporting results of cancer treatment. Cancer 1987; 47: Paridaens R, Dirix L, Beex L et al. Exemestane (Aromasin ) is active and well tolerated as first-line hormonal therapy (HT) of metastatic breast cancer (MBC) patients (pts): results of a randomised phase II trial. Proc Am Soc Clin Oncol 2000; 19: 83a (Abstr 316). 23. Di Salle E, Ornati G, Paridaens R et al. Preclinical and clinical pharmacology of the aromatase inhibitor exemestane (FCE 24304). In: Motta M, Serio M (eds): Sex hormones and Antihormones in Endocrine Dependent Pathology: Basic and Clinical Aspects. Amsterdam: Elsevier Science 1994; Miller WR, Dixon MJ. Anti-aromatase agents: preclinical data and neoadjuvant therapy. Clin Breast Cancer 2000; 1 (Suppl 1): Miller WR. Biology of aromatase inhibitors: pharmacology/endocrinology within the breast. Endocr Relat Cancer 1999; 6: Howell A, Mackintosh J, Jones M et al. The definition of the no change category in patients treated with endocrine therapy and chemotherapy for advanced carcinoma of the breast. Eur J Cancer 1988; 24: Bernhard J, Thürlimann B, Hsu Schmitz SF et al. Defining clinical benefit in postmenopausal patients with breast cancer under second-line endocrine treatment: does quality of life matter? J Clin Oncol 1999; 17: