Tamoxifen and Breast Cancer Incidence Among Women with Inherited Mutations in BRCA1 and BRCA2
Abstract & Commentary
Synopsis: While increasingly individualized cancer treatment protocols are gaining prominence due to potentially improved efficacy, directed prevention such as that examined in the phase III Breast Cancer Prevention Trial (BCPT) may be effective on a broader scale. In order to identify whether the tamoxifen benefit realized by high-risk women in the BCPT extended to those at even higher risk, women with BRCA1 and BRCA2 mutations were studied. Genomic analysis was performed on the 288 women in the BCPT trial who developed breast cancer, identifying 19 with BRCA1 or BRCA2 mutations. This subgroup was analyzed for tamoxifen vs. placebo use and presence of tumor estrogen receptor (ER) positivity. Although statistical significance was not achieved, preventative tamoxifen appeared to confer a risk reduction of 62% (95% confidence interval, 0.06-1.56) in BRCA2 mutation carriers and appeared to confer no risk reduction in the small number of BRCA1 mutation carriers. Only 14.2% of BRCA1 carriers were ER-positive, while 66.7% of BRCA2 carriers were ER-positive.
Source: King M-C, et al. JAMA. 2001;286:2251-2256.
In 1998, a randomized clinical trial of the National Surgical Adjuvant Breast and Bowel Project, the Breast Cancer Prevention Trial (BCPT) (P-1), revealed that tamoxifen conferred a protective effect for women at high risk for breast cancer.1 Study subjects were selected for age > 60 years, age < 60 years with a 5-year predicted risk for breast cancer of at least 1.66%, or a history of lobular carcinoma in situ. After subjects had ingested placebo or 20 mg tamoxifen daily for 5 years, a risk reduction of 49% overall was identified among the tamoxifen group. Notably, tamoxifen reduced the incidence of estrogen receptor-positive (ER+) tumors by 69%, although there was no reduction in the risk of estrogen receptor-negative (ER-) tumors.
Given these data, King and colleagues hypothesized that women with germline mutations of BRCA1 or BRCA2 tumor suppressor genes could potentially derive even greater benefit from chemical prevention, given their lifetime breast cancer risk of greater than 80%.2 The general BCPT null hypothesis that tamoxifen confers no protective benefit is also applicable to cases (women diagnosed with breast cancer) only. Moreover, the same null hypothesis (no tamoxifen benefit) applies to women with cancer (cases) both with and without BRCA mutations. Therefore, if tamoxifen use yielded no benefit, the number of women with cancer and BRCA1 or BRCA2 mutations who used tamoxifen should equal the number of those who used placebo.
In this case-only study, King et al identified BCPT participants who had been diagnosed with an incident invasive breast cancer between June 1, 1992 (start of the BCPT) and September 30, 1999. Follow-up information was available for 13,195 of the original 13,388 participants. Complete 3-year follow-up was available for 80% of participants; complete 5-year follow-up was available for 65% of participants. Median follow-up was 5.7 years. Among the 13,195 patients with follow-up, 320 were diagnosed with breast cancer. Genomic analysis for BRCA1 or BRCA2 mutations could be performed for 288 (90%) of these 320 women, on tissue prospectively procured from all trial participants.
A total of 19 (6.6%) of the 288 cases screened proved positive for BRCA mutations. Among the 8 women with BRCA1 mutations, 5 were in the tamoxifen group and 3 were in the placebo group (RR, 1.67; 95% CI, 0.32-10.70). Among the 11 women with BRCA2 mutations, 3 were in the tamoxifen group and 8 were in the placebo group (RR, 0.38; 95% CI, 0.06-1.56).
Among the 269 women without BRCA mutations, 87 received tamoxifen and 182 received placebo (RR, 0.48; 95% CI, 0.37-0.61). Among women with ER+ tumors, 132 received placebo and 41 received tamoxifen. Conversely, of those with ER- tumors, 32 received placebo while 36 received tamoxifen. ER status was not available for 28 of the wild-type tumors.
Although the results were not significant, the statistical trend suggested a protective benefit for tamoxifen use among women with BRCA2 mutations. Clinically, this conclusion is believable given the increased frequency of ER+ tumors in this group.3 However, the sample size was simply too small to draw conclusions regarding women with BRCA1 mutations. Even using Fisher’s exact test, the confidence interval was extremely wide for this group, and, furthermore, BRCA1 tumors are associated more frequently with ER- tumors suggesting that tamoxifen would be less relevant as a chemoprotectant.3
Comment by Arden Morris, MD
Tamoxifen has now been shown to play a valuable role in prophylaxis of ER+ breast cancer, reducing risk by 49% among high-risk women, but has not shown a benefit for prevention of ER- malignancy.1 These prophylactic benefits persist for younger women and women at especially high risk, even when weighed against the morbidity risks associated with tamoxifen.4
The subtext in this study by King et al centers on the especially high lifetime risk faced by women with BRCA1 and BRCA2 mutations, prompting some to recommend bilateral mastectomy as the only reliable prophylaxis for women who are likely to develop disease.5 Although tumors associated with BRCA1 mutation tend to be ER-, possibly due to earlier de-differentiation, BRCA2 mutation-associated tumors are more frequently ER+, and therefore are more promising for tamoxifen prophylaxis. King et al have attempted to evaluate this less invasive method of prophylaxis by using data from what may be the last large randomization of tamoxifen vs. placebo in high-risk women (due to ethical implications of randomizing to placebo now).
Although their results did not reach statistical significance, King et al have conducted a cost-efficient, elegant analysis of the available data. The confidence interval for the limited BRCA1 data (95% CI, 0.32-10.70) is too large to draw any conclusions. The confidence interval for the BRCA2 data is substantially narrower (95% CI, 0.06-1.56) but certainly does not reach statistical significance. As King et al point out, the only discernable way to improve confidence intervals would have been to include twice as many subjects in the BCPT initially. Longer follow-up may provide more information; however, median follow-up in this study was a respectable 5.7 years. An alternative and prohibitively expensive study design might have been to enrich the BRCA1 or BRCA2 mutation-carrying cohort selection in the original data collection. On balance, the conclusion that women with BRCA2 mutations should undergo tamoxifen prophylaxis aligns with prior hypotheses and provides an alternative to major surgical intervention.
References
1. Fisher B, et al. J Natl Cancer Inst. 1998;90(18): 1371-1388.
2. Welcsh PL, King MC. Hum Mol Genet. 2001;10(7): 705-713.
3. Phillips KA, et al. J Clin Oncol. 1999;17(11): 3653-3663.
4. Gail MH, et al. J Natl Cancer Inst. 1999;91(21): 1829-1846.
5. Meijers-Heijboer H, et al. N Engl J Med. 2001;345(3): 159-164.