Predicting ovarian function loss after chemotherapy and anti-HER2 therapy in young breast cancer patients

Abstract

Background: The ability to predict ovarian function loss after anticancer treatment is important for appropriate oncofertility counseling and to aid in therapy decision-making for young women with early breast cancer (eBC).

Methods: This biomarker analysis of the BETH (NCT00625898) and KAITLIN (NCT01966471) randomized trials investigated anti-Müllerian hormone (AMH) use, alone and combined with follicle stimulating hormone (FSH) and estradiol (E2), for predicting ovarian function loss following currently adopted chemotherapy and anti-HER2 therapy in premenopausal women with HER2-positive eBC. Serum samples were centrally tested measuring AMH, FSH, and E2 using Roche Elecsys assays.

Results: Among 194 included patients (BETH: n = 62; KAITLIN: n = 132), AMH values declined from baseline median 8.44 pmol L-1 to undetectable levels (<0.07 pmol L-1) at the end of therapy, with partial recovery at 36 months (median 0.14 pmol L-1). AMH measured at baseline was predictive of ovarian loss (area under the ROC curve [AUC] = 0.784). Addition of age to AMH slightly improved AUC to 0.800. AMH measured at the end of therapy had AUC 0.741, which increased to 0.785 with addition of age. The combination of AMH at baseline and end of therapy increased prediction to 0.808 and with addition of age to 0.820. Addition of baseline FSH and E2 did not improve prediction in any analysis.

Conclusions: These results support the use of pretreatment measurement of AMH in predicting ovarian function loss in premenopausal women with HER2-positive eBC receiving chemotherapy and anti-HER2 therapy. Measurement of AMH at the end of treatment had reduced accuracy than pretreatment but in combination added slightly to the value of pretreatment sampling.

Figure 1.

Study flow of participants in the BETH and KAITLIN trials to show selection for the present analysis. Abbreviations: AMH = anti-Müllerian hormone; E2 = estradiol; FSH = follicle stimulating hormone.

Figure 2.

Distribution of AMH, FSH, E2 at 3 time points for patients between 18 and 45 years of age. Distribution of AMH stratified by time points using original scale (A) and natural logarithm scale (B), FSH (C) and E2 (D). Abbreviations: AMH = anti-Müllerian hormone; E2 = estradiol; FSH = follicle stimulating hormone.

Figure 3.

Analysis of diagnostic and predictive value of AMH for premature ovarian insufficiency at different time points. (A) AMH levels at 36 months in women with and without POI, and ROC analysis for diagnostic accuracy at that time point. (B) Predictive value of pretreatment AMH for POI at 36 months, AMH levels at pretreatment in women with and without POI at 36 months, and ROC analysis for predictive accuracy at pretreatment. (C) Predictive value of AMH at the end of treatment for POI at 36 months, AMH levels at the end of treatment in women with and without POI at 36 months, and ROC analysis for predictive accuracy at the end of treatment. Abbreviations: AMH = anti-Müllerian hormone; POI = premature ovarian insufficiency; ROC = receiver operating characteristic.

Figure 4.

(A) Distribution of baseline AMH and age for development of premature ovarian insufficiency at 36 months. Blue symbols indicate women who developed POI, red symbols indicate women who did not. (B) Results of model for baseline AMH (log scale) against risk of POI at 36 months for a range of exemplar ages. Abbreviations: AMH = anti-Müllerian hormone; POI = premature ovarian insufficiency .