A nomogram for predicting insulin resistance in polycystic ovary syndrome based on serum amino acid profiles and anti-Müllerian hormone.

Objectives

Insulin resistance (IR) worsens metabolic and reproductive outcomes in polycystic ovary syndrome (PCOS). This study aimed to develop and validate a nomogram incorporating serum amino acid profiles and anti-Müllerian hormone (AMH) for predicting IR in PCOS.

Methods

A retrospective observational analysis was performed on 434 PCOS patients treated at The Fourth Hospital of Shijiazhuang between January 2022 and January 2026. Patients were classified as IR+ (n=184) or IR- (n=250) using a homeostasis model assessment of insulin resistance (HOMA-IR) cutoff of 2.69 and were chronologically assigned to training (n=259) and validation (n=175) cohorts. Anthropometric parameters, hormone levels and serum amino acid concentrations were measured. Multivariable logistic regression identified independent predictors for nomogram development. Receiver operating characteristic (ROC) analysis, calibration plot, bootstrap resampling and decision curve analysis (DCA) were applied to evaluate the model's performance.

Results

Multivariable analysis identified body mass index (odds ratio [OR]=1.357, 95% confidence interval [CI]: 1.180-1.561), waist circumference (OR=1.096, 95% CI: 1.053-1.141), AMH (OR=0.931, 95% CI: 0.879-0.986), leucine (OR=1.028, 95% CI: 1.012-1.044), isoleucine (OR=1.075, 95% CI: 1.045-1.107), valine (OR=1.015, 95% CI: 1.009-1.022) and tyrosine (OR=1.046, 95% CI: 1.004-1.089) as independent predictors. The nomogram demonstrated excellent discrimination in training (area under the curve [AUC]=0.945) and validation (AUC=0.917) cohorts, with good calibration (Hosmer-Lemeshow P=0.554) and clinical utility on DCA.

Conclusion

This nomogram, integrating anthropometric measures, serum AMH and four amino acids, provides an accurate tool for the early identification of IR in PCOS.