Abstract Background: Estrogen receptor-positive (ER+) breast cancer (BC) has a high affinity for progression in bone, which is often the initial metastatic site. ER+ BC treatment responses may depend on the expression of additional steroid hormone receptors (SHRs), such as progesterone receptors (PR) or androgen receptors (AR). However, the predictive value of these SHRs in determining ER+ BC bone metastases (BMET) outcomes remains largely unexplored. The possible functional impact of PR and AR targeting on bone-specific ER+ BC progression is also not known. To address this knowledge gap, we evaluated the predictive associations of AR and PR expression with clinical outcomes for ER+ BMET. Methods: A retrospective analysis was conducted using BC BMET samples analyzed at Caris Life Sciences, with survival outcomes inferred from insurance claims. HER2-/ER+ BMET (n=1677) specimens were identified using immunohistochemistry (IHC) (HER2-: ≤1+ intensity or ≤10% cells stained, or 2+ 10% with CISH-null reflex test; ER+: ≥1+ AR+: ≥1+ 2) pathogenic gene mutations (mut) known to alter BC survival (ESR1mut or PIK3CAmut); or 3) AR-regulated gene expression. Results: SHR expression was common in ER+ BMET, with 87.1% AR+ and 59.3% PR+. “Triple positive” (AR+/PR+/ER+) BMET were the largest group (53.7%), followed by PR-null AR+/ER+ BMET (33.2%). AR-null ER+ BMET were much less common, with 5.6% or 7.3% PR+ or PR-, respectively. Only 9.4% of PR+/ER+ BMET were AR-, while 38.2% of AR+/ER+ BMET were PR-. AR positivity was associated with the longest OS for ER+ BMET. A similar trend for improved OS was observed in AR-expressing ER- BMET (n=305), though not statistically significant. When stratifying ER+ BMET by the proportion of AR+ cells stained by IHC, having a higher proportion (≥ median) was associated with longer OS only for PR+/ER+ BMET (Cox proportional hazards ratio (HR)=0.70, p=0.003 for high vs low AR % cells stained). There was no significant difference in CDKi TOT across groups. However, OS was significantly and uniquely longer for AR+ (not AR-) ER+ BMET patients (pts) who received CDKi, which was highest when PR was co-expressed (HR=0.62, p0.0001 for pts with vs without CDKi), as compared to PR- (HR=0.73, p=0.012). ESR1mut prevalence was similar in AR+ (17.0%) vs AR- (10.7%) ER+ BMET, with OS reduced in ESR1mut (vs wildtype, wt) AR+/ER+ BMET (HR=1.92, p0.0001). PIK3CAmut were more prevalent in AR+ (51.5%) vs AR- (27.3%) ER+ BMET. However, there was no difference in OS between PIK3CAmut vs wt for any SHR ER+ BMET subtype. While not SHR-regulated, RNA expression of the SHR pioneer factor FOXA1 positively correlated with that of AR, PGR (PR), and ESR1 (ER) in ER+ BMET across all SHR subtypes. Several known AR-regulated genes were among the most highly expressed genes in AR+ vs AR- BMET, consistent with functional AR signaling. Notably, expression of PIP, HMGCS2, and CYP4Z1 was significantly higher in AR+ vs AR- BMET, regardless of ER status. Conclusions: Our findings suggest that AR expression may have utility as a predictive biomarker of improved survival and response to CKD4/6 inhibitors in those with HER2-/ER+ BMET, particularly when co-expressed with PR, with evidence of functional AR signaling in both ER+ and ER- AR-expressing BC BMET. Whether AR in this context is a biomarker or mediator of improved survival cannot be determined. Therapeutic consequences of AR targeting to reduce AR+ BC progression in bone, also an AR-regulated tissue, may hold promise, although whether with an agonist or antagonist remains an intriguing and yet unanswered question. Citation Format: A. M. Rossi, R. N. Plagens, A. Elliott, E. Tapia, G. Sledge, A. Elias, J. Richer, S. Ehsani, J. L. Funk. Evaluation of steroid hormone receptor expression and clinical outcomes in ER-positive breast cancer bone metastases abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS1-11-27.
Rossi et al. (2026) studied this question.