Abstract

Poor ovarian response (POR) constitutes a notable clinical challenge within the domain of assisted reproductive technology, primarily attributable to the lack of reliable biomarkers for precise diagnosis and treatment. This study reveals significantly reduced levels of insulin-like growth factor 1 (IGF-1) in the serum, follicular fluid (FF), and granulosa cells (GCs) of patients with POR in comparison to those exhibiting a normal ovarian response (NOR). Notably, FF IGF-1 concentrations demonstrated significant positive correlations with crucial IVF outcomes, including the numbers of metaphase II (MII) oocytes, 2-pronuclear zygotes, and high-quality embryos. To establish causality, we employed complementary in vivo models: systemic insulin-like growth factor binding protein acid labile subunit ( Igfals ) knockout mice and granulosa cell specific IGF-1 receptor ( Igf-1r ) knockout mice. These models collectively demonstrated that disruption of the IGF-1 signaling axis impairs follicle-stimulating hormone (FSH) responsiveness and arrests follicular development at the secondary stage, thereby recapitulating the core POR phenotype. Building on these mechanistic insights, we developed novel clinical prediction tools based on FF IGF-1: a POR risk model [Area under the curve (AUC) = 0.914] and a pregnancy outcome nomogram (AUC = 0.893), both of which significantly outperform traditional ovarian reserve parameters (such as anti-Müllerian hormone and antral follicle count). Decision curve analysis (DCA) further validated a substantial clinical net benefit. This study aids clinicians in the early identification of patients with POR and provides a theoretical foundation for timely intervention and adjustment of treatment strategies.