Aromatase Inhibitors: Updates from Major Trials

In 2009, we covered several long-term studies investigating the use of aromatase inhibitors (AIs) alone or sequenced with tamoxifen. This includes the Intergroup Exemestane Study (IES), the TEAM trial, and the BIG 1-98 trial, all of which presented new or updated data at the 2009 SABCS in December.

Update from Intergroup Exemestane Study (IES)

Postmenopausal women with hormone receptor—positive (HR+) breast cancer who switch to the AI exemestane (Aromasin) after taking tamoxifen for 2 to 3 years have superior outcomes compared with women who stay on tamoxifen for 5 years, according to an updated analysis of IES . “These women [who switched to exemestane] have real improvements that persist at least 9 years, which is 11 years from the time of diagnosis,” stated Judith Bliss, MD, The Institute of Cancer Research, Sutton, United Kingdom. “They have a modest but persistent improvement in overall survival [OS ] and a modest reduction in distant recurrence,” she added.

The study included 4724 patients recruited from 366 sites in 37 countries from 1998 to 2003. Slightly less than 50% of patients (44.2%) had nodepositive disease; one-third had received adjuvant chemotherapy. The mean age of the women was 64 years. The analysis presented at SABCS was completed after 91 months of follow-up.

The 4599 women with estrogen receptor—positive and estrogen receptor–unknown status who switched to exemestane had superior OS compared with women who had 5 years of tamoxifen therapy. The absolute difference in OS favoring exemestane was 1.4% at 5 years and 2.4% at 8 years (P = .04). Diseasefree survival (DFS ) was highly statistically significant in favor of exemestane: at 5 years, the absolute difference in DFS was 3%, and at 8 years, it was 4.4% (P = .0009). The annual DFS event rate (including breast cancer recurrence, new primary cancer, or death) was 4% per year. “There is no evidence that the event rate diminishes more than 10 years from diagnosis,” Bliss told the audience.

The risk for local recurrence was .5% per year and the risk of distant recurrence was 2.5% per year. Inter-current deaths are rare, she explained, and as they become more common over time, they can blunt the difference between treatment groups.

“Breast cancer—free survival is a more sensitive estimate of differences,” Bliss said. At 5 years, the absolute difference in breast cancer–free survival favoring exemestane was 2.8%, and at 8 years, it was 4.1% (P = .001). While on treatment, the women who switched to exemestane had a 40% reduction in the risk of a breast cancer event. In the post-treatment follow-up phase, the risk of a breast cancer event was reduced 16% for exemestane users.

All subgroup analyses for the breast cancer—free survival endpoint favored exemestane. Looking at time to distant recurrence, which is strongly associated with breast cancer mortality, exemestane was significantly superior to tamoxifen (P = .01). There were 319 distant recurrences in the group that switched to exemestane versus 370 in the tamoxifen group.

In an exploratory analysis of sites of first reported distant recurrence, exemestane appeared to correlate with a reduction in the risk of bone metastases. Bone-only metastases occurred in 75 patients in the exemestane group compared with 109 in the tamoxifen group. In addition, the exemestane group had 147 distant recurrences that involved bone versus 192 in the tamoxifen group. Bliss said this purported benefit of exemestane was “wildly speculative” and hypothesis generating.

Another surprising finding was an increase in the number of nonbreast primary cancers in the tamoxifen group compared with the exemestane cohort: 159 versus 106, respectively. “Exemestane has an unexpected lower number of reported nonbreast second primary cancers,” Bliss noted. Looking at age at randomization, the evidence for nonbreast primary cancers suggests that these are true second primary cancers and not recurrences, she added. T he findings regarding the metastases involving the bone and the smaller number of nonbreast second primary cancers require further validation, Bliss said.

TEAM trial reports final data

One of the most provocative questions regarding endocrine therapy for HR+ breast cancer is which strategy is superior: a sequential strategy that involves giving tamoxifen up front followed by an AI for a total of 5 years or giving only an AI for 5 years. The final results from the TEAM trial failed to resolve this issue; with both strategies in a dead heat as far as DFS, time to recurrence (TTR), and OS endpoints.

“TEAM was the first trial prospectively powered to detect superiority of 5 years of an AI [in this case, exemestane] versus sequential therapy with tamoxifen followed by AI. This large study showed that either strategy is an appropriate therapeutic option. Translational research may be able to identify subsets of patients who can benefit from each of these strategies,” said Daniel Rea, MD, University of Birmingham, United Kingdom.

The large, open-label TEAM trial randomized 9779 women with invasive, postmenopausal, HR + breast cancer to a total of 5 years of treatment applying either strategy. About half of the breast cancers were node-positive. All women (mean age, 65 y) underwent surgical resection and none had metastases at baseline. Patients who had undergone radiotherapy (68%) and chemotherapy (36%) were permitted to enroll.

At a median follow-up of 5.1 years, 60% of participants had completed all 5 years of treatment. At that time, 127 local recurrences had occurred and 820 distant metastases had been reported. More than 85% of patients in both groups were disease-free at 5 years, with no difference in outcomes observed between the two strategies (HR, 0.97). Similarly, more than 90% of patients were alive with or without disease at 5 years, with no difference in OS between the two strategies (HR , 1.00). “We hardly ever see an HR of 1,” Rea said.

Analysis of secondary endpoints failed to detect superiority for continuous AI therapy versus sequential tamoxifen followed by AI. No difference between strategies was seen for either TTR or nodal status at baseline.

Toxicities were as expected with both tamoxifen and exemestane. “Tamoxifen-treated patients had more tamoxifen-like adverse events, including gynecologic effects, while exemestane was associated with typical AI-like effects, such as musculoskeletal complaints,” said Rea.

Vaginal toxicity was significantly greater for those initially treated with tamoxifen: 17 endometrial cancers occurred in the tamoxifen-treated patients versus 7 in the group on continuous exemestane. In addition, 187 women in the tamoxifen group experienced endometrial abnormalities compared with 20 patients in the AI group. Vaginal discharge and bleeding were also significantly greater in patients who took tamoxifen (P <.001 for all comparisons). The incidences of cardiovascular disorders were similar between the groups, with the exception of a higher rate of venous thromboses in the tamoxifen users (P <.001). “Surprisingly, we saw more hyperlipidemia in the exemestane group,” Rea said.

Musculoskeletal events were significantly higher with continuous exemestane, which was not unexpected. Fractures occurred in 5% of the exemestane group versus 3.5% of those initially randomized to tamoxifen (P <.001). The incidence of osteoporosis was double: 9.9% versus 5.4% for those originally randomized to tamoxifen (P <.001).

BIG 1-98: Adjusting for selective crossover

Meredith Regan, ScD, assistant professor, Department of Medicine, Harvard Medical School, Cambridge, Massachusetts, presented new data from BIG 1-98 that show even after taking into account the effect of patients crossing over from tamoxifen to letrozole (Femara), 5 years of treatment with letrozole was still associated with significantly better OS than 5 years of tamoxifen therapy (P <.05). Explaining the rationale for conducting the analysis, Regan, who is group statistician for the International Breast Cancer Study Group conducting the trial, said, “We have an obligation to present updated results accounting for women who crossed over to the other arm following report of positive results. In trials with selective crossover, an intent-to-treat (ITT) analysis is no longer relevant for patient care.”

BIG 1-98, a phase III randomized double-blind, controlled trial, included 4922 postmenopausal women with HR + early breast cancer. They were randomly assigned to treatment with either the AI letrozole or tamoxifen.

When BIG 1-98 results were first presented in 2005, they demonstrated superior DFS and reduced disease recurrence risk for women who received letrozole for an average of 18 months. At that time, the tamoxifen-only treatment arm was unblinded, and 25% of women in that group selectively crossed over to the letrozole arm.

After a median follow-up of 76 months, an ITT analysis revealed significant benefits for letrozole regarding DFS (HR, 0.88; P = .03) and time to distant recurrence (TDR; HR, 0.85; P = .05), but the OS benefit did not reach the level of statistical significance (HR, 0.87; P = .08).

While selective crossover is in the best interest of patients, it does complicate further trial analyses, Regan said. “The trial becomes a hybrid of a randomized trial and an observational study,” she commented, adding that accurate assessment of outcomes requiring longer follow-up, such as OS, becomes difficult. With the ITT analysis, the effects of letrozole were attenuated by the crossovers, which moved the randomized trial more toward an observational study. As a result, Regan said, statistical methods needed to be modified.

The solution lies in alternative modeling strategies, Regan explained. In the case of BIG 1-98, they used inverse probability of censoring weighted analysis (IPCW)—an established modeling method for addressing bias due to treatment selection—to weight the follow-up for the women who stayed on tamoxifen so that they accounted not only for themselves but also for the censored follow-up of matched patients who crossed over. The weighting adjusts for factors associated with OS and selective crossover, including baseline factors like age, nodal status, tumor grade, and time-varying performance status. It estimates outcomes that would have been observed had there been no selective crossover, she said.

With IPCW and censoring follow-up at selective crossover, the OS HR for letrozole was 0.83 (P <.05), a statistically significant benefit. “The ITT analysis on updated data from BIG 1-98 is inaccurate and no longer relevant for patient care,” Regan concluded. “Our best estimate is that 5 years of letrozole is significantly better than tamoxifen for DFS, OS, and TDR.”

In an interview, Regan emphasized, “We owe it to clinicians and to patients to learn as much as we can from trials. This method deals with selective crossover and gives an estimate of the important longer-term outcome of OS. It gives clinicians good information for making decisions for patients.”