ERAP 2 as a Potential Marker for Ovarian Ageing and Spontaneous Ovulation Recovery in Women With Premature Ovarian Insufficiency

Objective To explore novel biomarkers for clinical prognosis in patients with ovarian ageing, especially in premature ovarian insufficiency (POI). Design Prospective Cohort Study. Setting Reproductive Hospital Affiliated with Shandong University, China. Population Sixty POI patients and 60 control women recruited from 2018 to 2019. Methods Machine learning algorithms were used to screen features of ovarian ageing from public ovarian transcriptome data. The candidate serum biomarker, endoplasmic reticulum aminopeptidase‐2 (ERAP2), was compared between 60 POI patients and 60 control women. A prospective follow‐up of 4 years was conducted on POI patients, and prediction models were established for intermittent recovery of ovarian function and clinical pregnancy based on serum ERAP2 levels. Main Outcome Measures Intermittent recovery of ovarian function and clinical pregnancy. Results Machine learning models identified ERAP2 as a novel biomarker associated with ovarian ageing. POI patients exhibited significantly elevated serum ERAP2 levels compared to controls (5.78 ± 2.29 ng/mL vs. 4.81 ± 2.20 ng/mL, p = 0.018). With a prospective follow‐up of these POI patients, ERAP2 was found to be a new biomarker for predicting intermittent recovery of ovarian function (AUROC = 0.763, 95% CI 0.734–0.792) and clinical pregnancy (AUROC = 0.768, 95% CI 0.749–0.787). Integrating ERAP2 into existing indices significantly improved their prediction accuracy both in predicting intermittent recovery of ovarian function (IDI = 0.166, p = 0.008) and in clinical pregnancy (NRI = 0.442, p = 0.034; IDI = 0.208, p = 0.018). Conclusions Serum ERAP2 can serve as a biomarker for intermittent recovery of ovarian function and clinical pregnancy in patients with POI. Combining ERAP2 with other clinical indicators may facilitate personalised intervention strategies for patients with POI in clinic.