Two hundred and fourteen women were identified as being registered in the compassionate access program for fulvestrant in Saskatchewan between March 2003 and January 2019. Twenty-eight women were excluded (three women did not receive any dose of fulvestrant and 25 did not have adequate follow up information). This study included 186 eligible women with advanced HR+ breast cancer, who had received at least one dose of fulvestrant (Figure 1). The median age at the start of fulvestrant treatment was 63.5 years (IQR: 54.0–74.0). Seventeen (9.1%) women had HR+/HER2+ disease, 81.2% had a WHO performance status of 0–1, and 80.6% of women had a previous diagnosis of early-stage breast cancer, with 89.3% of those having received adjuvant therapy. One hundred and forty-three (76.9%) women had bony metastases, and 57.5% had visceral metastasis. Among the entire cohort, 43% of women had received chemotherapy in the metastatic setting before fulvestrant, and 60.2% of patients received chemotherapy after discontinuing fulvestrant. Overall, 17.7% of women received fulvestrant with a targeted therapy (mostly a CDK 4/6 inhibitor). One hundred and seventy-eight women had endocrine-resistant disease, with 39 (21.9%) having primary resistance and 139 (78.1%) having secondary resistance. Table 1 lists the patient characteristics.
Overall, 102 (54.8%) women had started fulvestrant after ≤ 2 lines of therapy in the metastatic setting, and 84 (45.2%) women received fulvestrant after ≥ 3 lines of therapy. Significant differences were noted between the two groups with respect to age, pathological diagnosis of metastatic breast cancer, median line of therapy, prior chemotherapy, and median time of diagnosis to start of fulvestrant treatment. The median age of women in early-line therapy was 67 years, versus 60 years in women in the later-line therapy group (p = 0.001). The median line of therapy in the group that received fulvestrant after ≤ 2 lines of therapy was 1, compared to 4 in the group who received fulvestrant after ≥ 3 lines of therapy. Women in the ≤ 2 lines of therapy group started fulvestrant at a median of 24.5 months after diagnosis of advanced breast cancer, compared to 44.0 months in women who received fulvestrant after ≥ 3 lines of therapy (p = 0.002). Women in the early-line group were less likely to have received chemotherapy prior to fulvestrant compared to the later-line group (17.6% vs. 73.8%, p = 0.001), and were less likely to have fulvestrant-refractory disease (20.6% vs. 38.1%, p = 0.009).
Overall, 74.7% patients discontinued fulvestrant due to disease progression, 10.2% discontinued due to side effects or patient request and 1.6% due to other reasons. Women in the early-line group were less likely to have fulvestrant-refractory disease (20.6% vs. 38.1%, p = 0.009) compared to the later-line group. The median TTP of the entire cohort was 8 months (95% CI: 5.6–10.4). The median TTP was 12 months (9.4–14.6) in women starting fulvestrant after ≤ 2 lines of therapy, versus 6 months (5.1–6.9) in women with ≥ 3 lines of therapy (p = 0.015) (Figure 2A) [Table 2]. There was no significant difference in median TTP between women who had primary versus secondary endocrine resistance (7 vs. 9 months, p = 0.098). Women who received chemotherapy before fulvestrant had a median TTP of 6 months (4.5–7.4), versus 12 months (8.7–15.3) in those who did not receive chemotherapy (p = 0.039) (Figure 2B).
The OS of all women following commencement of fulvestrant was 21 months (16.0–26.0). The OS was 26 months (16.0–36.0) in women starting fulvestrant after ≤ 2 lines of therapy, versus 16 months (10.5–21.5) in women with ≥ 3 lines of therapy (p = 0.067) (Table 2). There was no significant difference in OS based on primary versus secondary endocrine resistance (p = 0.59). Patients that had visceral metastasis had a lower OS, 18 months (14.1–21.9), versus 32 months (23.0–41.0) in those without (p = 0.029) [Figure 3A]. Patients that received chemotherapy after fulvestrant had a greater OS, 34 months (30.4–37.6), versus 8 months (4.2–11.8) in those that did not (p < 0.001) (Table 2).
Cox proportional Multivariate analysis
Table 3 lists factors that correlate with TTP and provides the hazard ratios for univariate and multivariate analysis of TTP. On univariate analysis, lack of prior chemotherapy and early-line treatment significantly correlated with better TTP. However, on multivariate analysis, only the absence of visceral metastasis was significantly correlated with better TTP, HR: 0.70 (95% CI: 0.50–0.99). Differences in TTP based on the use of chemotherapy prior to fulvestrant or previous lines of therapy when starting fulvestrant were not significant after adjustment for other variables.
Table 4 lists the hazard ratios for univariate and multivariate analysis of OS, after adjustment of important variables. In multivariate analysis, post-fulvestrant chemotherapy, HR: 0.32 (0.23–0.47), clinical benefit from fulvestrant, HR: 0.44 (0.30–0.65), and absence of visceral metastasis, HR: 0.70 (0.50–0.97) were correlated with better OS. The difference in survival based on WHO performance status was not significant in the final multivariate model after adjustment for other variables.