Ratio of ovarian stroma and total ovarian area by ultrasound in prediction of hyperandrogenemia in reproductive-aged Thai women with polycystic ovary syndrome: A diagnostic test

Aim To evaluate the performance of ovarian stromal area to total ovarian area (S/A) ratio for the prediction of biochemical hyperandrogenism in Thai women with polycystic ovary syndrome (PCOS). Methods A cross-sectional study was performed in 222 reproductive-aged Thai women with PCOS attending the Gynecologic Endocrinology Unit (GEU), Department of Obstetrics and Gynecology, Faculty of Medicine Siriraj Hospital from May 2007 to January 2009. The patients were interviewed for medical history and examined for anthropometry and clinical hyperandrogenism. Venous blood samples were obtained for androgen profiles. An ovarian ultrasonogram was obtained via transvaginal or transrectal ultrasonography. Results The prevalences of clinical and biochemical hyperandrogenism were 48.6% and 81.1%, respectively. The S/A ratio at a cut-off point of 0.33 had modest predictability for hyperandrogenism, namely, 0.537 area under the receiver-operator curve, 36.6% sensitivity, 72.1% specificity, 83.8% positive predictive value (PPV) and 20.9% negative predictive value (NPV). The combination of clinical hyperandrogenism and S/A ratio improved the predictability for biochemical hyperandrogenism, with sensitivity, specificity, PPV and NPV of 72.1%, 58.1%, 87.8% and 33.3%, respectively. Conclusion The S/A ratio alone is not a good predictor for biochemical hyperandrogenism in Thai PCOS women attending GEU for menstrual dysfunction. The combination of S/A ratio and clinical hyperandrogenism has better performance than the S/A ratio alone to predict biochemical hyperandrogenism.