Prostate-Specific Antigen, Long a Male Cancer Marker, Emerges as a Window Into Androgen Activity in Women With PCOS

Prostate-specific antigen, or PSA, is one of the most famous molecules in medicine, a protein so tightly associated with the male prostate that its very name has shaped decades of screening programs and public health campaigns. Yet a new study published in Reproductive Sciences suggests that this familiar biomarker may have a second, quieter life in female physiology, one that could reshape how clinicians think about one of the most common hormonal disorders in women. In a prospective case-control study of ninety women with polycystic ovary syndrome, researchers in Ankara, Turkey, found that although serum PSA concentrations did not differ between hirsute and non-hirsute patients when the groups were compared head to head, PSA levels tracked remarkably closely with the severity of hirsutism, with testosterone concentrations, and with markers of insulin resistance across the entire cohort.

The findings arrive at a moment when the scientific community is increasingly aware that polycystic ovary syndrome, or PCOS, is not a single disease but a spectrum of overlapping phenotypes. Affecting an estimated five to ten percent of women of reproductive age, the syndrome is defined by the 2003 Rotterdam ESHRE/ASRM criteria, which require two of three features: oligo-ovulation or anovulation, clinical or biochemical hyperandrogenism, and polycystic ovarian morphology on ultrasound. That broad definition captures women whose experiences of the disorder can differ dramatically. Some struggle primarily with irregular cycles and fertility; others are marked by the visible stigmata of androgen excess, including coarse terminal hair growth on the face and body, a condition known as hirsutism, along with acne and androgenic alopecia. Understanding which biological signals distinguish these phenotypes has become a central goal of reproductive endocrinology.

The rationale for studying PSA in this context rests on a surprising body of evidence accumulated over nearly three decades. Although PSA, encoded by the KLK3 gene, is produced abundantly by prostatic epithelial cells, immunoreactive PSA has also been detected in female tissues, including the breast, the salivary glands, and the endometrium, and measurable amounts circulate in the blood of healthy women at concentrations far below those seen in men. Crucially, early work by Melegos, Diamandis and colleagues in 1997 demonstrated that female serum PSA rises in conditions of androgen excess, and subsequent studies showed that antiandrogen drugs lower PSA levels in hirsute women, providing pharmacological evidence that the protein functions as a downstream readout of androgen action in female tissues. In other words, PSA in women is not a prostate marker at all; it is a marker of what androgens are actually doing to target tissues.

This biological logic made PSA an attractive candidate for the Turkish research team, led by Seyma Donmez Sahip and Batuhan Turgay of the University of Health Sciences at Ankara Bilkent City Hospital, with Turgay also affiliated with Ankara University’s Reproductive Health center. Their hypothesis was straightforward: if PSA reflects androgen activity, and if hirsute women with PCOS experience more intense androgenic stimulation of hair follicles and sebaceous glands than their non-hirsute counterparts, then hirsute patients should display higher serum concentrations of both total and free PSA. To test this, the investigators enrolled ninety women diagnosed with PCOS by Rotterdam criteria between May 2022 and January 2023, stratifying them into hirsute and non-hirsute groups using the modified Ferriman-Gallwey score, the standard clinical instrument for visually grading terminal hair growth across nine androgen-sensitive body areas.

The methodology reflects careful attention to the confounders that plague hormonal research. Venous blood samples were drawn on cycle days two or three, when baseline reproductive hormones are most reliably interpreted, and the panel included total and free PSA alongside total and free testosterone, DHEA-S, luteinizing hormone, and metabolic parameters. Insulin resistance was quantified using the homeostatic model assessment, HOMA-IR, a calculation derived from fasting glucose and insulin that estimates how hard the pancreas must work to maintain normal blood sugar. This detail matters because insulin resistance is not merely a metabolic side effect of PCOS; it is widely regarded as a central driver of the syndrome, amplifying ovarian and adrenal androgen production while suppressing hepatic synthesis of sex hormone-binding globulin, the carrier protein that keeps circulating testosterone biologically inert.

When the two phenotypic groups were compared directly, the results partially confirmed expectations. Hirsute women had significantly higher body mass index, with a p-value of 0.044, higher acne prevalence at p equal to 0.032, and higher free testosterone at p equal to 0.014, painting a coherent picture of a phenotype in which androgenic and metabolic disturbances cluster together. Interestingly, menstrual disturbance ran in the opposite direction: amenorrhea and oligomenorrhea were more frequent among non-hirsute women, a reminder that the severity of ovulatory dysfunction and the intensity of cutaneous androgen effects do not march in lockstep. Ovarian morphology and alopecia showed no significant differences between the groups, and, contrary to the primary hypothesis, neither total nor free PSA concentrations differed significantly between hirsute and non-hirsute patients, with all such comparisons yielding p-values above 0.05.

But the story changed when the researchers abandoned group comparisons in favor of correlation analysis across the whole cohort. Using Spearman rank correlation, a non-parametric method appropriate for biological variables that do not follow normal distributions, the team uncovered strikingly strong positive associations. PSA correlated with the Ferriman-Gallwey hirsutism score at a rank correlation coefficient of 0.896, with total testosterone at 0.779, with HOMA-IR at 0.838, with DHEA-S at 0.807, and with luteinizing hormone at 0.804, all with p-values below 0.001. In clinical research, correlation coefficients of this magnitude are rare; values approaching 0.9 indicate that two variables rise and fall together almost in lockstep across the study population. The implication is that PSA behaves as a continuous integrator of androgenic and metabolic load, even if the binary split into hirsute and non-hirsute categories is too coarse to reveal differences in group means.

This distinction between categorical and continuous analysis carries real methodological weight. Phenotypic classification by Ferriman-Gallwey score necessarily dichotomizes a graded biological variable, and thresholds for hirsutism vary by ethnicity and remain imperfectly standardized. A molecule whose expression is driven by tissue-level androgen action may respond smoothly to the full range of androgen exposure, so that averaging PSA values across a threshold-defined group obscures the very dose-response relationship that correlation analysis exposes. The strong PSA-HOMA-IR correlation adds a further layer, suggesting that PSA may capture the intertwined androgen-insulin axis that many investigators consider the engine of PCOS pathophysiology, in which hyperinsulinemia potentiates luteinizing hormone-driven ovarian theca cell androgen synthesis and reduces binding-protein-mediated sequestration of free testosterone.

The new findings also sit within a contested literature. Prior studies of PSA in PCOS have produced mixed results: some reported elevated PSA in affected women compared with controls, a 2019 systematic review and meta-analysis in Hormone and Metabolic Research found overall differences supporting PSA as a potential androgen-excess marker, and work in adolescent girls explored its diagnostic value, while other investigations failed to replicate clear group-level differences. The Ankara study’s design differs from much of this literature in that it compared PSA across PCOS phenotypes rather than between PCOS and healthy controls, asking a subtler question about within-syndrome heterogeneity. The answer, that PSA does not separate hirsute from non-hirsute patients but correlates powerfully with hirsutism severity, androgen concentrations, and insulin resistance, suggests that PSA’s utility may lie not as a diagnostic classifier but as a quantitative index of androgenic-metabolic activity.

For patients and clinicians, the practical implications remain preliminary. PSA is not proposed as a replacement for testosterone measurement or for careful clinical assessment of hirsutism, acne, and cycle regularity, and the study’s authors emphasize that their results underscore the heterogeneity of PCOS and the need for comprehensive phenotypic evaluation rather than a single biomarker shortcut. Yet the research adds to a growing recognition that molecules once considered sexually dimorphic in absolute terms may carry meaningful physiological information in both sexes, and that the androgen signaling pathway, so central to PCOS, leaves measurable fingerprints throughout the body. If future longitudinal studies confirm that serum PSA tracks with long-term metabolic risk in women with PCOS, a cheap and widely available immunoassay, already deployed in laboratories worldwide, could find an unexpected second career in women’s health, transforming a prostate cancer biomarker into a lens on the hormonal and metabolic complexity of one of medicine’s most heterogeneous syndromes.

Subject of Research: Serum prostate-specific antigen as a marker of androgenic and metabolic activity in hirsute and non-hirsute phenotypes of polycystic ovary syndrome

Article Title: Serum Prostate-Specific Antigen Concentrations in Polycystic Ovary Syndrome: A Comparative Analysis of Hirsute and Non-Hirsute Phenotypes

Article References: Sahıp, S. D., Eyol, M. M., Ersoy, E., Erden Deveci, E., Bahadır, I. B., & Turgay, B. (2026). Serum Prostate-Specific Antigen Concentrations in Polycystic Ovary Syndrome: A Comparative Analysis of Hirsute and Non-Hirsute Phenotypes. Reproductive Sciences. https://doi.org/10.1007/s43032-026-02187-4

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02187-4

Keywords: polycystic ovary syndrome, prostate-specific antigen, hirsutism, hyperandrogenism, PCOS phenotypes, insulin resistance, free testosterone, biomarkers, reproductive endocrinology, Ferriman-Gallwey score, HOMA-IR, Rotterdam criteria