Neoadjuvant Anastrozole Permits Breast Conservation

January 1, 2004

SAN ANTONIO-Neoadjuvant anastrozole (Arimidex, A) produced about the same rate of clinical responses as tamoxifen (T) or combined anastrozole/tamoxifen (AT) in estrogen-receptor-positive (ER+) breast cancer and was better than either for shrinking tumors enough to permit breast-conserving surgery, researchers reported at the 26th Annual San Antonio Breast Cancer Symposium (abstract 1).

SAN ANTONIO—Neoadjuvant anastrozole (Arimidex, A) produced about the same rate of clinical responses as tamoxifen (T) or combined anastrozole/tamoxifen (AT) in estrogen-receptor-positive (ER+) breast cancer and was better than either for shrinking tumors enough to permit breast-conserving surgery, researchers reported at the 26th Annual San Antonio Breast Cancer Symposium (abstract 1).

Differences in response to therapy could be predicted after 2 weeks of neoadjuvant therapy by measuring changes in expression of the cell proliferation marker Ki67 (abstract 2).

Ian E. Smith, MD, and Mitch Dowsett, MD, both from the Royal Marsden Hospital, London, reported these conclusions on behalf of investigators from the IMPACT (Immediate Preoperative Arimidex Compared to Tamoxifen) trial of postmenopausal women with ER+ operable breast cancers of 2 cm or greater diameter.

"In the overall population, A and T were similarly effective as neoadjuvant therapy in ER+ operable breast cancer in postmenopausal women. In patients recorded as requiring a mastectomy at baseline, twice as many became eligible for breast-conserving surgery in the A group as in the T group," Dr. Smith said.

This study expanded on work done in the ATAC (Anastrozole, Tamoxifen Alone or in Combination) trial, which showed that anastrozole was more effective than tamoxifen for adjuvant treatment of early hormone-receptor-positive breast cancer, Dr Smith said.

In contrast to a previous randomized neoadjuvant trial (Ann Oncol 12:1527-32, 2001), the IMPACT study enrolled patients who were eligible for breast-conserving surgery as well as those requiring mastectomy. It included 330 postmenopausal women (mean age, 72 years), with core biopsy ER+ invasive, operable breast cancer. Patients were randomized to A (n = 113), T (n = 108), or AT (n = 109) for 3 months’ treatment prior to surgery. Core biopsies were repeated after the second week of treatment. Tissues samples were taken during surgery for biological studies or by core biopsy if surgery was not performed.

The primary study endpoint was objective clinical tumor response (OR)—the product of the two maximum clinical diameters, assessed by calipers and ultrasound. Secondary endpoints included downstaging to permit breast-conserving surgery in patients initially deemed to require mastectomy, identification of surrogate markers for efficacy, and tolerability.

Only 67% of patients (220) had a surgeon’s assessment for feasibility of breast-conserving surgery, of whom 56% were thought to need a mastectomy and 44% were eligible for breast-conserving surgery at baseline.

Study Results

The primary endpoint, OR by caliper, was achieved in 37.2%, 36.1%, and 39.4% of patients in the A, T, and AT groups, respectively (see Table 1). The relative risks (RR) of OR for A and AT vs T are 1.03 and 1.09, respectively.

Results were similar with ultrasound assessment: OR in 23.9%, 20.4%, and 27.5% of patients, respectively. Differences between treatment groups were not significant for either assessment method.

"There was a significant gain for anastrozole in permitting the use of breast-conserving surgery: 46% of patients previously thought to require mastectomy became eligible for breast-conserving surgery, compared with 22% of patients treated with tamoxifen or the combination," Dr. Smith said (see Table 2). In the 124 patients thought to need mastectomy at baseline, 45.7%, 22.2%, and 26.2% in the A, T, and AT groups, respectively, became eligible for breast-conserving surgery after 3 months’ treatment (A vs T P = .03; RR 2.05).

In these 124 patients. overall responses were observed in 39.1% of those treated with anastrozole, 27.8% of those treated with tamoxifen, and 35.7% of those treated with both (see Table 2). Relative risks for A and AT vs T are 1.41 and 1.29. Dr. Smith said that 14.2% of the 239 patients who received planned treatment and had samples accessible for HER-2 showed HER-2 overexpression. In this small subgroup, clinical response was higher with A: 58% vs 22% with T alone and 31% with the combination (P = .09 A vs T; P = .66 AT vs T).

"There does seem to be a strong trend in favor of anastrozole . . . compared to tamoxifen in patients with HER-2-overexpressing tumors," Dr. Smith said.

All treatments were well tolerated, and only 1.8%, 2.8%, and 3.7% of patients in the A, T, and AT groups, respectively, withdrew due to a drug-related adverse event. Total cholesterol was unchanged compared to baseline in the A group and was reduced in the T and AT groups.

Dr. Smith said that vaginal discharge was more common with T (6 of 108) or AT (8 of 109) than with anastrozole (0 of 113). Hot flashes were also more common in the tamoxifen-treated groups (26 of 108 and 28 of 109 vs 18 of 113 with anastrozole). "We saw thromboembolisms only in patients who received tamoxifen, either alone (2 of 108) or in combination (3 of 109)," Dr. Smith said.

Ki67 Marker for Treatment Effects

"Short-term change in Ki67 may predict overall treatment differences between hormonal therapies in the adjuvant setting. We will be using Ki67 as a primary endpoint in our next trial," Dr. Dowsett said. He emphasized that Ki67 is useful only for groups of patients, not for predicting outcomes in individual patients.

The search for a biological marker is being driven by the need for an alternative to the large numbers of patients and lengthy follow-up now required to detect differences between adjuvant treatments. For example, the ATAC trial randomized 9,366 patients and was able to acquire sufficient power to detect differences only after 33 months’ median follow-up.

Dr. Dowsett presented a retrospective analysis asking whether biological response (a change in the proliferation marker Ki67) during neoadjuvant therapy would have predicted the outcome of the IMPACT trial.

Biopsies were performed in 292 patients; data for 259 patients were available for biological analyses. Ki67 expression was assessed in 1,000 cells and expressed as percent positive. A second core biopsy at 2 weeks was done in only 159 patients because it was not a mandatory part of the study.

Reduced Ki67 Expression

Ki67 expression was significantly reduced by all three treatments after 2 weeks, and this reduction continued for at least 12 weeks. Mean reductions in Ki67 were as follows: anastrozole 76% at 2 weeks, 82% at 12 weeks; tamoxifen 59% at 2 weeks, 62% at 12 weeks; AT 64% at 2 weeks, 61% at 12 weeks.

Dr. Dowsett said that the differences in decreased Ki67 expression, which is thought to indicate a decrease in tumor cell proliferation, were significantly greater with anastrozole than with tamoxifen both at 2 weeks (76% vs 59%, P = .004) and at 12 weeks (82% vs 62%, P < .0001).

"There was no significant relationship with clinical response other than a marginal one in the tamoxifen group," Dr. Dowsett said. There was a weak but statistically significant relationship between change in Ki67 at 2 weeks and response to tamoxifen (P = .01) but not to response to anastrozole or AT.

Dr. Dowsett concluded that short-term change in Ki67 expression might predict treatment differences between adjuvant therapies and thus might be useful in designing clinical trials to compare new therapeutic agents in patients with early breast cancer.