Incidence and predictors of ovarian function recovery (OFR) in exemestane patients with breast cancer (BC) with chemotherapy-induced amenorrhea (CIA) using two estradiol assays.

Abstract

521 Background: Aromatase inhibitors (AIs) may promote OFR in patients with CIA. True incidence of OFR is unknown partly because estradiol (E2) assays used in the clinical setting lacking sensitivity. We studied incidence and predictors of OFR measuring E2 by a clinical assay (direct) or by a highly sensitive (indirect) immunoassay, together with other markers of ovarian reserve. Methods: Prospective exploratory study in ER positive BC patients, who had CIA lasting at least two years and postmenopausal E2 levels, while on tamoxifen. Patients received exemestane and underwent assessments of E2 (indirect and direct assay), FSH, antimullerian hormone (AMH), and antral follicule count (AFC) at baseline and at month 1, 3 and 6. We considered OFR if: resumption of menses, or developed (i) both E2 levels inconsistent with postmenopausal women on AI (indirect: >8 pmol/L; direct >165 pmol/L) and FSH (<26 IU/L) levels), or (ii) frankly non-postmenopausal E2 values (indirect >55 pmol/L; direct >220 pmol/L). Results: Fifty-three consecutive patients were enrolled; median age: 48y (41-55). By indirect assay, 15 patients (28.3%; CI95%:16.1-40.4) had OFR, (11 menses, 4 OFR without menses). By direct assay 14 had OFR (11 menses, 3 OFR without menses). Mean time to OFR by indirect assay was 5.7 months (CI95%: 3.1-8.3) and 6.4 months (CI95%: 3.7-9.1) by direct assay. There was no difference in baseline E2 levels among those who recover or not OFR. Mean age was the only baseline marker different between patients who had OFR or not (45.7 vs 49.1 years, t-test p<0.0001). In patients ≤48 years (n=26) rate of OFR was 46%. AMH at baseline was below detection limit in all patients and it did not change during exemestane. Baseline and serial values for AFC were not different among patients who had OFR or not. Conclusions: There is a high incidence (28%) of OFR, in patients with CIA treated with AIs, especially in women ≤48y. The two E2 assays showed similar incidence of OFR and time to OFR, despite having markedly different performance characteristics. AMH was not helpful in defining ovarian reserve and likelihood of OFR.