Why This Study Matters
The phase 3 RxPONDER trial (SWOG S1007) established that women with hormone receptor-positive (HR-positive), HER2-negative, node-positive (1–3 positive nodes) breast cancer and a 21-gene Recurrence Score of 25 or below could often forgo chemotherapy if postmenopausal, while premenopausal women benefited from adding chemotherapy to endocrine therapy. But menopausal status and chronological age are imprecise stand-ins for the underlying biology, and misclassification is common. This sub-study asked a more direct question: can a blood test for ovarian reserve — the biological pool of remaining ovarian follicles — identify who actually benefits from chemotherapy, independent of how a patient is labeled as premenopausal or postmenopausal?
The stakes are overtreatment versus undertreatment. Chemotherapy carries real toxicity and quality-of-life costs; identifying the subset of premenopausal women under 55 with low ovarian reserve who do not appear to benefit could allow many patients to safely avoid it.
Study Design
This is a prespecified biomarker analysis of the phase 3 RxPONDER (SWOG S1007) trial, not a new randomized trial. In the parent trial, 5,083 women with HR-positive, HER2-negative, node-positive breast cancer and Recurrence Score of 25 or below were randomly assigned to chemotherapy plus endocrine therapy or endocrine therapy alone. For this analysis, pretreatment serum estradiol, progesterone, follicle-stimulating hormone, luteinizing hormone, anti-Müllerian hormone (AMH) and inhibin B (INHB) were measured in a blinded fashion using ultrasensitive assays in three sequential cohorts of participants under 55. Associations between each hormone and chemotherapy benefit for invasive disease-free survival (IDFS) and distant relapse-free survival (DRFS) were tested using multivariable Cox regression, adjusting for continuous Recurrence Score, with p-values adjusted for multiplicity.
Patient Population
Of 2,125 RxPONDER participants under 55 who were eligible, 1,556 (73.2%) had baseline serum available for hormone analysis. Of these, 1,221 were classified premenopausal and 335 postmenopausal at trial entry, based on last menstrual period, prior oophorectomy or hysterectomy history, or age when menstrual history was unclear. Participants with serum available were more likely to be premenopausal at entry than those without, but the two groups did not differ in tumor characteristics, per the study.
Chemotherapy Benefit by Ovarian Reserve
Across all women under 55 with evaluable serum, chemotherapy conferred an overall IDFS benefit (hazard ratio 0.65, 95% CI 0.50–0.85) — but that average masked sharply different results depending on ovarian reserve. Using a prespecified AMH cutoff of 10 pg/mL, patients with AMH at least 10 pg/mL (n=998) had a clear IDFS benefit from chemotherapy (hazard ratio 0.46, 95% CI 0.33–0.65), translating to an absolute 5-year IDFS benefit of 8.5%. Patients with AMH below 10 pg/mL (n=558) derived no benefit (hazard ratio 1.27, 95% CI 0.81–1.99). The same pattern held for distant relapse-free survival: AMH at least 10 pg/mL, hazard ratio 0.44 (95% CI 0.28–0.69), an absolute 5-year DRFS benefit of 3.8%; AMH below 10 pg/mL, hazard ratio 1.51 (95% CI 0.85–2.70).
Inhibin B (INHB), another ovarian-reserve marker, showed a similar predictive pattern: patients with INHB at least 12 pg/mL benefited from chemotherapy (IDFS hazard ratio 0.39, 95% CI 0.26–0.59), while those with INHB below 12 pg/mL (56.3% of the cohort) did not (hazard ratio 1.00, 95% CI 0.70–1.43). Combining AMH and INHB in a joint model did not improve prediction beyond either marker alone, due to high correlation between the two (r=0.73).
Overall, we found ovarian reserve was a significantly better indicator of adjuvant chemotherapy benefit than self-reported menopause status, age, or other hormone levels (i.e., estradiol, progesterone, LH, FSH).
— Discussion, Kalinsky et al., Annals of Oncology (2026)
Subgroup Analyses
The authors further stratified results by menopausal status at enrollment and by age band. The source article explicitly flags the postmenopausal and age-banded analyses below as exploratory and, in the case of the age bands, hypothesis-generating given small event numbers.
| Subgroup | IDFS HR (95% CI) | Notes |
|---|---|---|
| Premenopausal, AMH at least 10 pg/mL | 0.46 (0.33–0.65) | Chemotherapy benefit |
| Premenopausal, AMH below 10 pg/mL | 1.21 (0.60–2.43) | No benefit; interaction p=0.035 |
| Postmenopausal, AMH at least 10 pg/mL | 0.62 (0.05–7.33) | Very small subgroup (29 of 335 postmenopausal); underpowered |
| Postmenopausal, AMH below 10 pg/mL | 1.29 (0.71–2.32) | Underpowered |
| Age 40–44 (95.2% high AMH) | 0.36 (0.20–0.66) | Exploratory, hypothesis-generating |
| Age 50–54 | 1.03 (0.68–1.53) | Exploratory, hypothesis-generating |
The source article reports the overall AMH-at-least-10-pg/mL IDFS result (all ages under 55 combined) with a multiplicity-adjusted p-value given as 0.00012 in the abstract and as 0.0001 in the Results section for the same comparison — both figures appear verbatim in the source; the premenopausal-only subgroup above (p=0.0012) is a distinct, separately reported analysis.
Safety Profile
This is a biomarker and predictive-analysis manuscript built on banked serum samples and survival outcomes from the previously reported RxPONDER trial. It does not report new adverse-event or toxicity data, and the parent trial’s safety profile is not restated here. Not specified in source.
Interpretation and Broader Context
The authors conclude that premenopausal women under 55 with low ovarian reserve (AMH below 10 pg/mL) — about one in five premenopausal women under 55 in this cohort — may be able to safely forgo chemotherapy without compromising IDFS or DRFS, potentially sparing a meaningful subset of patients from unnecessary treatment. They note the finding is consistent with the hypothesis that much of the chemotherapy benefit seen in premenopausal women in RxPONDER is mediated by chemotherapy-induced ovarian suppression rather than direct anti-tumor activity, and that whether ovarian function suppression could substitute for chemotherapy in premenopausal women is being tested prospectively in the ongoing NRG-BR009 (OFSET) and OPTIMA-YOUNG trials.
Limitations
The authors flag several limitations of their own analysis. Hormone levels were measured only once, at trial entry, so longitudinal changes in ovarian reserve could not be assessed. The analysis involved multiple comparisons across hormones, cutoffs and subgroups, raising the risk of over-interpretation, and the postmenopausal and age-banded findings are explicitly exploratory and hypothesis-generating given small event numbers in those strata. Gynecologic comorbidities such as polycystic ovary syndrome, which can elevate AMH, were not systematically collected. The authors describe the overall findings as hypothesis-generating pending prospective validation.
In a prespecified biomarker sub-study of the randomized phase 3 RxPONDER trial, serum AMH — a blood test for ovarian reserve — outperformed age and menopausal status at identifying which women under 55 with HR-positive, node-positive breast cancer benefit from adding chemotherapy to endocrine therapy. About one in five premenopausal women under 55 in this cohort had low ovarian reserve and showed no chemotherapy benefit, a potential path to reducing overtreatment. Because the hormone measurements were taken only once, the subgroup findings are exploratory, and the authors call the results hypothesis-generating pending validation in trials now underway, this should be read as evidence supporting further study rather than a change to current treatment guidelines.