Prediction of polycystic ovarian syndrome based on ultrasound findings and clinical parameters

Objective. To determine the accuracy of sonographic-diagnosed polycystic ovaries and clin- ical parameters in predicting polycystic ovarian syndrome. Methods. Medical records and ultrasounds of 151 women with sonographically diagnosed polycystic ovaries were reviewed. Sonographic criteria for poly- cystic ovaries were based on 2003 Rotterdam Euro- pean Society of Human Reproduction and Embryology/American Society for Reproductive Med- icine guidelines: at least one ovary with 12 or more follicles measuring 2-9 mm and/or increased ovarian volume >10 cm 3 . Clinical variables of age, gravidity, ethnicity, body mass index, and sonographic indica- tion were collected. One hundred thirty-five patients had final outcomes (presence/absence of polycystic ovarian syndrome). Polycystic ovarian syndrome was diagnosed if a patient had at least one other of the fol- lowing two criteria: oligo/chronic anovulation and/or clinical/biochemical hyperandrogenism. A logistic regression model was constructed using stepwise selection to identify variables significantly associated with polycystic ovarian syndrome (p < .05). The valid- ity of the model was assessed using receiver operat- ing characteristics and Hosmer-Lemeshow v 2