Breast Cancer Prevention

Breast Cancer Prevention TREVOR J. POWLES Royal Marsden NHS Trust, and Institute of Cancer Research, London, United Kingdom Key Words. Breast cancer Chemoprevention Tamoxifen Raloxifene ABSTRACT Epidemiological, experimental, and clinical data strongly support the possibility that breast cancer will be prevented by using anti-estrogenic interventions in healthy women. Three trials involving over 20,000 women have so far been reported using tamoxifen 20 mg/day or placebo in healthy women to chemoprevent breast cancer. The American National Surgical Adjuvant Breast and Bowel P-1 Project randomized over 13,000 women to take tamoxifen or placebo and showed a 49% reduction in the early incidence of breast cancer. This was associated with a reduction in osteoporotic fractures but increases in the risks of endometrial cancer, cataract, and thromboembolism. The Royal Marsden tamoxifen trial randomized 2,500 women, and the Italian national trial randomized 5,000 women. Interim analyses from these two trials BACKGROUND The promotion of breast cancer by estrogen provides an outstanding opportunity for chemoprevention of this common and frequently fatal disease. Tamoxifen, an orally administered anti-estrogen, is a proven treatment for breast cancer with low symptomatic toxicity. This has encouraged its testing as a chemopreventive agent in healthy women. To this end, a feasibility trial of tamoxifen was begun at the Royal Marsden Hospital (London, United Kingdom) in 1986 to evaluate its acute toxicity, safety, and acceptability, including compliance, in healthy women. Early results from this trial indicated that tamoxifen was a selective anti-estrogen [1] with estrogenic effects on some tissues. With few antiestrogenic effects in the body, acute toxicity was low. Treatment was associated with a significant reduction in serum cholesterol, and no evidence of accelerated bone loss was found. Compliance was high in these healthy women [1]. These encouraging results, together with a report that adjuvant tamoxifen was associated with a reduction in the incidence of contralateral breast cancer in patients treated for primary breast cancer [2], led to the initiation in 1992 of showed no effect on the early incidence of breast cancer. These results, therefore, have not been able to clearly show an overall clinical benefit of giving tamoxifen to healthy women, nor have they shown which women are likely to benefit. Another selective anti-estrogen (SERM), raloxifene, has been used in a clinical trial to prevent osteoporotic fractures in women with low bone mineral density. Annual mammography in this trial has shown an approximate 80% reduction in the early incidence of breast cancer, and further follow-up of this trial continues. New trials in chemoprevention of breast cancer being started or being proposed use luteinizing-hormone-releasing hormone analogues, aromatase inhibitors, and other SERMs. The Oncologist 2002;7:60-64 multicenter trials of tamoxifen versus placebo in healthy women in the United Kingdom/Australasia (International Breast Cancer Intervention Study [IBIS]), in Italy (Italian National Trial [INT]), and in the U.S. (the National Surgical Adjuvant Breast Project [NSABP] P1). In 1994, an osteoporosis trial was started using raloxifene, another selective anti-estrogen that prevented loss of bone mineral density. In order to have breast cancer incidence as a secondary end point, the women in this trial had annual mammography. The first results from these four trials have now been reported and, while promising, they prompt unanswered questions. The follow-up of subjects in all but the NSABP P1 trial continues, and trials evaluating other anti-estrogenic interventions have started or are proposed. It is therefore timely to review the results from these trials and to explore the strategies for future trials. THE NSABP P1 TRIAL (P1 TRIAL) The P1 trial started in June 1992, randomizing healthy women aged >60 years at any risk, or women aged from 35-59 years with a predicted 5-year risk of 1.66% estimated by Correspondence: Trevor J. Powles, M.D., Ph.D., Royal Marsden NHS Trust, Sutton, Surrey SM2 5PT, United Kingdom. Telephone: 020-8661-3361; Fax: 020-8770-7313; e-mail: trevor.powles@rmh.nthames.nhs.uk Received June 15, 2001; accepted for publication October 10, 2001. AlphaMed Press 1083-7159/2002/$5.00/0 The Oncologist 2002;1:60-64 www.theoncologist.com

Powles 61 the Gail model [3]. Women with a previous history of lobular carcinoma in situ (LCIS) were also eligible. A total of 13,388 women were randomized by September 1993 to receive either tamoxifen 20 mg/day or placebo for 5 years. After a median follow-up of 54.6 months [4], analysis showed a 49% reduction in the incidence of invasive cancers in the tamoxifen arm (p < 0.00001). While reduction in mortality from breast cancer was a part of the primary objective of this trial and no meaningful data had developed for this end point, the trial was stopped and unblinded on the advice of the data monitoring committee. While patients in all categories of risk, including age and family history, showed a reduction in incidence of breast cancer with tamoxifen, the greatest levels of benefit were seen in women with LCIS (relative risk [RR] = 0.44) and atypical ductal hyperplasia (ADH) (RR = 0.14), most of whom were premenopausal. Other premenopausal women defined by the Gail model, including BRCA gene carriers, benefited less. Although the magnitude of this effect in these women has not been reported, women with LCIS or ADH were not included in the Royal Marsden trial. A detailed analysis of the P1 trial results, according to age and considering toxicity, showed that the most benefit accrued to younger women with an elevated risk of breast cancer [5]. However, in this analysis, it is not clear how much of the high risk in these women is attributable to high-risk pathology (ductal carcinoma in situ or ADH) as opposed to Gail risk factors, including a family history, but excluding abnormal pathology. THE ROYAL MARSDEN TRIAL (RMH TRIAL) The RMH trial was started in 1986 as a feasibility trial for the IBIS trial. The eligibility was limited to women with a strong family history of breast cancer whose age-adjusted relative risk of breast cancer was approximately four. Following encouraging toxicity and logistic results from the pilot trial, the IBIS trial was scheduled to start in 1989. However, following reports that tamoxifen was experimentally genotoxic, approval for the start of IBIS was not given until 1992. In 1991, study power calculations for the pilot trial indicated that if accrual to 2,500 was completed, there would be a 90% chance of detecting a 50% reduction in breast cancer incidence by 1998. An interim analysis was therefore planned for this date. The results of this analysis showed an identical incidence of breast cancer for women on tamoxifen or placebo [6]. At the time of this analysis, 2,494 women had been randomized to tamoxifen 20 mg/day or placebo between October 1986 and April 1996. The median age of these women was 47 years with 66% of women pre- or perimenopausal. Women on hormone replacement therapy (HRT) were eligible for entry into the trial and could stop HRT if they wished. Similarly, women not on HRT were allowed to use it while taking tamoxifen/placebo. Analysis of the total time of use of HRT or tamoxifen on a monthly basis throughout the total duration of the trial showed that HRT was used for a total of only 13% of the tamoxifen medication period. There was no evidence of any interaction between HRT and tamoxifen on breast cancer incidence. While the negative result in this trial is statistically compatible with the reduction in incidence reported from the P1 trial (RMH: RR = 1.06, confidence interval [CI] = 0.7-1.7; P1: RR = 0.51, CI = 0.39-0.66), this is unlikely to be the reason for this discrepancy. If statistical power does account for the differences in the results, then pooling of these results will increase the RR for women on tamoxifen in P1 from 0.51 to 0.59 [5]. This level of risk reduction, which would be even less for those women not having LCIS or ADH, would negatively impact the risk benefit ratio, especially for premenopausal women, to a point where clinical benefit would be in doubt. An alternative reason for the negative result in the RMH trial could be that RMH participants included a greater fraction of relatively tamoxifen-resistant individuals. RMH participants were younger than P1 participants, and were more likely to have been carriers of a breast cancer predisposing gene. Individuals with LCIS or atypical hyperplasia, shown to be most sensitive to tamoxifen chemoprevention in the P1 trial, were not included in the entry criteria to the RMH trial. The women in the RMH trial may, therefore, be relatively resistant to tamoxifen chemoprevention compared with those in the P1 trial. It is important to know if these younger women, who are likely to be gene carriers, are resistant to tamoxifen, because they are the women most likely to be attracted to tamoxifen chemoprevention. ITALIAN NATIONAL TRIAL (INT) A total of 5,408 healthy women aged between 35 and 70 years were accrued to this trial between October 1992 and July 1997. Eligibility required participants to have had a hysterectomy (74% had also had bilateral oophorectomy). These women were not at increased risk for breast cancer and 14% were on HRT. An interim analysis, reported in 1998 after a median follow up of 46 months [7], showed no statistical difference between the incidence of breast cancer for women on tamoxifen and placebo. There was, however, a significant reduction in incidence of breast cancer in women on HRT who received tamoxifen. It is possible that the previous oophorectomy in most participants may have compromised any subsequent chemopreventative effect of tamoxifen unless these women also received HRT. THE IBIS TRIAL The IBIS trial started accrual in 1992 and completed accrual of over 7,000 participants in 2000. Eligibility

62 Breast Cancer Prevention depended on risk factors, including a family history of breast cancer that resulted in an estimated increase in the relative risk of breast cancer of two- to threefold. Followup for this trial continues, and the first interim analysis is scheduled in 2002. It is hoped that follow-up will continue for the RMH trial, the INT, and the IBIS trial until 2005 when an overview meta-analysis of these trials could address breast cancer incidence and cause-specific (including breast cancer) mortality. SUMMARY OF CLINICAL BENEFIT FROM TAMOXIFEN CHEMOPREVENTION It is not clear from the results of the P1 trial that the observed reduction in the incidence of breast cancer is of clinical benefit. Because of the uncertainty that the treatment of the cancers in women on placebo would have been as effective and safe, the results of toxicity of all 6,681 women randomized to tamoxifen for 5 years were examined. The 86 tumors in the 6,707 women randomized to placebo were all estrogen receptor positive, mostly <2 cm, and axillary lymph node negative and, therefore, more than 90% curable with primary surgery, radiotherapy and tamoxifen [8]. Furthermore, although the 6,681 women who received tamoxifen had 34 fewer noninvasive breast cancers and 26 fewer osteoporotic fractures, they also had 21 additional endometrial cancers, 14 more strokes, 12 more pulmonary emboli, 13 more deep vein thromboses, and 67 more cataracts compared with women on placebo. A detailed and complex evaluation has been reported to help balance the various effects of tamoxifen in different age and risk groups in the P1 trial [5]. The conclusions are that younger women at higher risk are those most likely to gain benefit, although it is not clear that this remains true if those women with LCIS and ADH are excluded from this analysis. Additionally, it is likely that some groups of atrisk women are resistant to tamoxifen chemoprevention, as suggested by the negative results in the RMH and the INT, and these groups have not yet been clearly identified. As a result, it is impossible to make informed recommendations for individual women. The American Society for Clinical Oncology undertook a technology assessment in 1999 and concluded that, although women at increased risk of breast cancer may be offered tamoxifen to reduce their risk, there are currently insufficient data to determine whether tamoxifen provides an overall health benefit and increased chance of survival [9]. Continuation and follow-up of the tamoxifen placebo-controlled trials are essential in order to provide more clinical data to address these issues. RALOXIFENE (MULTIPLE OUTCOMES OF RALOXIFENE EVALUATION [MORE] AND CONTINUING OUTCOMES RELEVANT TO EVISTA [CORE] TRIALS) In 1994, a double-blind randomized trial of raloxifene, a synthetic selective anti-estrogen (SERM), versus placebo (MORE), was started to evaluate its effect on fractures in healthy postmenopausal women with osteoporosis detected by bone mineral density examination [10]. As a secondary end point in this trial, breast cancer incidence was evaluated. Trial participants had annual mammography. After 30 months of follow-up there was a 76% reduction in the incidence of breast cancer for women on raloxifene, a significant reduction in the incidence of osteoporotic fractures, and no evidence of proliferative effects of raloxifene on the uterus [11]. Because breast cancer incidence was not a primary end point in this trial, the results were considered not acceptable by the regulatory agencies for licensing raloxifene for risk reduction of breast cancer. The CORE trial, which continues the follow-up of participants in the MORE trial, will have breast cancer incidence as the primary end point. FUTURE STRATEGIES Following the results of the P1 and the MORE trials, the NSABP started a new trial in postmenopausal women at moderate risk determined by the Gail model. Participants were randomized to receive either tamoxifen 20 mg/day or raloxifene 60 mg/day. The primary objectives of this trial are to compare breast cancer incidence and toxicity, especially any effects on the uterus, in women on tamoxifen and raloxifene. It is planned to enter 22,000 women, 8,000 of whom have already been randomized. There are concerns that this trial does not have a placebo control arm, and that an excess of endometrial events in the tamoxifen arm may result in closure of the trial before the incidence of breast cancer on tamoxifen and raloxifene has been established. Additionally, mortality is again not an end point of this trial. However, this trial is important and will substantially contribute to our knowledge of the role of SERMS in breast cancer chemoprevention. The results will need to be linked to the continuing follow-up in the placebo-controlled trials (RMH, IBIS, INT, CORE) in order to establish the overall clinical benefit of tamoxifen and raloxifene chemoprevention. In considering trials of new anti-estrogenic interventions, the possibility that different groups of women at risk for breast cancer may be variably sensitive to different chemoprevention measures has influenced the trial designs. In highrisk healthy women identified as BRCA1 or 2 gene carriers (or likely to be BRCA carriers based on pedigree risk assessment), three trials are proposed by the National Cancer Research Institute in the United Kingdom and IBIS group. The intervention is based on ovarian suppression using the

Powles 63 luteinizing-hormone-releasing hormone analogue Zoladex in women under the age of 45. In order to reduce the risk of expected accelerated bone loss and osteoporosis, three different pilot studies are proposed. The RAZOR study would add raloxifene to Zoladex, the TIZER study would add the synthetic HRT tibolone, and the GISS study would use the bisphosphonate ibandronate. These pilot studies have started accrual and will involve 150 women and look at recruitment, compliance, bone markers, quality of life, and lipid profile. Any of these trials that have an attractive profile of activity will be extended into definitive multicenter, multinational trials. Women in these trials will have identified BRCA1/ BRCA2 mutations and/or pedigree risks of such gene mutations. It will be essential that the ER status and other phenotypic features of the cancers that occur are established. For women at about a two- to fourfold greater agematched relative risk of breast cancer because of a family history, a new IBIS trial is proposed (IBIS II) that will randomize 10,000 women to receive either tamoxifen 20 mg/day, or the aromatase inhibitor anastrazole 1 mg/day, or placebo. There is evidence now accumulating that the new third generation aromatase inhibitors, such as anastrozole or letrozole, are as effective or more effective than tamoxifen for treatment of established breast cancers [12]. This appears to be achieved by causing very low or unrecordable levels of estrogen in the plasma. Adjuvant trials using these agents are under way, but the results of long-term tissue toxicity from these very low levels of estrogen in healthy women have not been established, especially in the brain, the bones, and the genital tract. It has been shown that low plasma levels of estradiol are associated with a reduced risk of breast cancer and are also associated with significantly lower levels of bone mineral density [10, 13]. However, the cut-off for loss of bone density is substantially higher than that for reduced breast cancer incidence, and the levels achieved by aromatase inhibitors are much lower than both of these. It is therefore imperative that the results of trials of these very powerful aromatase inhibitors are predicted by rigorous toxicity studies. The very low levels of estrogen achieved for 5 years in otherwise healthy women could result in long-term unacceptable toxicity to normal tissues, which may not become evident for many years. Because of the potential long-term risks in healthy women of any endocrine intervention, trials in low-risk women must be considered with great caution. Dietary intervention studies are possible in these low-risk women. There is evidence that a high intake of isoflavonoid phytoestrogens in the diet is associated with a reduced incidence of breast cancer [14]. Supplementation of the Western diet with isoflavonoids may reduce breast cancer incidence. We have now started a pilot double-blind randomized trial of an extract of red clover (Promensil ) at the Royal Marsden Hospital and Christie Hospital (Manchester, UK). Women aged 35 to 65 years with one first-degree relative with breast cancer are randomized to receive Promensil, a quality-controlled (good clinical practice standards) extract of 80 mg of isoflavones or placebo once per day. In the pilot phase of this trial, we have randomized over 300 healthy women. Monitoring of these women involves clinical and biological assessment including mammographic breast density, endometrial ultrasound, endocrine studies, bone mineral density measurements, and lipid profile. If Promensil is shown to have a favorable biological profile, the trial will extend into a national multicenter trial aimed at evaluation of breast cancer incidence and other outcomes. CONCLUSIONS Chemoprevention trials of breast cancer in healthy participants offer the attractive potential for prevention of this major disease. However, because of the low number of beneficial events that are likely to occur in these clinical trials, it is essential that any proposed long-term medication is of the highest established safety, particularly because toxicity may remain latent for many years. With tamoxifen we have been lucky. It is an agent that has an established safety profile in many millions of women with breast cancer, and there is now a clear-cut indication that it will substantially reduce the early incidence of breast cancer. We now need to prove that this reduction in early incidence will translate into real clinical benefit that outweighs any long-term toxicity, however rare. We can then build on to this by identifying agents that work better with less toxicity in healthy women. It is also essential that we identify which women benefit from such interventions, so that women who will not benefit will not be subjected to unnecessary risk. The strategy is clear. It needs to be carefully integrated and orchestrated in order to establish that breast cancer can be prevented safely. REFERENCES 1 Powles TJ, Hardy JR, Ashley SE et al. A pilot trial to evaluate the acute toxicity and feasibility of tamoxifen for prevention of breast cancer. Br J Cancer 1989;60:126-131. 2 Cuzick J, Baum M. Tamoxifen and contralateral breast cancer. Lancet 1985;2:282. 3 Gail MH, Brinton LA, Byar DP et al. Projecting individualized probabilities of developing breast cancer for white females who are examined annually. J Natl Cancer Inst 1989;81:1879-1886. 4 Fisher B, Costantino J, Wickerham D et al. Tamoxifen for prevention of breast cancer: report of the National Surgical

64 Breast Cancer Prevention Adjuvant Breast and Bowel Project P-1 Study. J Natl Cancer Inst 1998;90:1371-1388. 5 Gail MH, Costantino JP, Bryant J et al. Weighing the risks and benefits of tamoxifen treatment for preventing breast cancer. J Natl Cancer Inst 1999;91:1829-1846. 6 Powles T, Eeles R, Ashley S et al. Interim analysis of the incidence of breast cancer in the Royal Marsden Hospital tamoxifen randomised chemoprevention trial. Lancet 1998;352:98-101. 7 Veronesi U, Maisonneuve P, Costa A et al. Prevention of breast cancer with tamoxifen: preliminary findings from the Italian randomised trial among hysterectomised women. Italian Tamoxifen Prevention Study. Lancet 1998;352:93-97. 8 Early Breast Cancer Trialists Collaborative Group. Tamoxifen for early breast cancer: an overview of the randomised trials. Lancet 1998;351:1451-1467. 9 Chlebowski R, Collyar DE, Somerfield M et al. American Society of Clinical Oncology technology assessment on breast cancer risk reduction strategies: tamoxifen and raloxifene. J Clin Oncol 1999;17:1939-1955. 10 Ettinger B, Black D, Mitlak B et al. Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene: results from a 3-year randomized clinical trial. Multiple Outcomes of Raloxifene Evaluation (MORE) Investigators. JAMA 1999;282:637-645. 11 Cummings SR, Eckert S, Krueger KA et al. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA 1999;281:2189-2197. 12 Cohen MH, Hirschfeld S, Flamm Honig S et al. Drug approval summaries: arsenic trioxide, tamoxifen citrate, anastrazole, paclitaxel, bexarotene. The Oncologist 2001;6:4-11. 13 Cauley JA, Lucas FL, Kuller LH et al. Bone mineral density and the risk of breast cancer in older women: the study of osteoporotic fractures. Study of Osteoporotic Fractures Research Group. JAMA 1996;276:1404-1408. 14 Adlercreutz CH, Goldin BR, Gorbach SL et al. Soybean phytoestrogen intake and cancer risk. J Nutr 1995;125(suppl 3):757S- 770S.