Abstract
OBJECTIVE: To contrast the diagnostic accuracy of 3D ultrasonographic markers of ovarian morphology to 2D ultrasonography and serum anti-Müllerian hormone (AMH) levels for detection of polycystic ovary syndrome (PCOS). Any gains in diagnostic accuracy when combining ovarian markers was also explored. METHODS: 30 women with PCOS based on the National Institutes of Health (NIH) criteria and 30 age- and body mass index (BMI)-matched controls were evaluated. Transvaginal ultrasound scans (GE Voluson E or S Series; 5-12 MHz) and blood draws were conducted during the early follicular phase for women with regular cycles or at a random time for women with irregular cycles. Ovarian images were analyzed for the number of 2-9mm antral follicles in the entire ovary (FNPO), number of 2-5mm antral follicles (FNPO2-5mm) and 6-9mm antral follicles (FNPO6-9mm), and ovarian volume (OV) using 3D and 2D ultrasonography by one of six trained raters. 3D versus 2D ovarian markers were compared using Bland-Altman agreement statistics for matched-pair analysis. Diagnostic accuracy of ovarian markers for PCOS was determined by Receiver Operating Characteristic (ROC) curve analysis. Stepwise logistic regression analysis determined if combinations of ultrasonographic markers or AMH yielded greater diagnostic potential than single parameters. RESULTS: 3D measurements of FNPO, FNPO2-5mm, and OV had greater diagnostic accuracy for PCOS compared to their 2D counterparts. The most accurate markers for PCOS were 3D FNPO (AUC=0.938) and 3D FNPO2-5mm (AUC=0.933), followed by 2D FNPO (AUC=0.922) and 2D FNPO2-5mm (AUC=0.906). Despite overall agreement between 3D and 2D FNPO measures, 3D ultrasonography overcounted FNPO2-5mm (2.00 ± 3.68, P=0.006) and undercounted FNPO6-9mm (-3.00 ± 3.49, P<0.001) in women with PCOS. AMH (AUC=0.898) had greater diagnostic accuracy than 3D and 2D OV (AUC=0.854 and 0.849, respectively), but remained inferior to both 3D and 2D follicle counts. The addition of OV or AMH did not improve diagnostic accuracy for follicle counts. However, the addition of AMH did substantially improve the diagnostic accuracy of 3D OV (AUC=0.918) and 2D OV (AUC=0.911). CONCLUSIONS: 3D ultrasonographic markers of ovarian morphology outperformed 2D markers in the diagnosis of PCOS, with 3D markers of follicle excess being the most accurate measure overall. AMH had inferior diagnostic accuracy compared to FNPO across both modalities but was superior to OV. Addition of AMH improved the predictive power of OV on 2D and 3D ultrasonography thereby offering a robust alternative when image quality prevents an accurate assessment of follicle counts. (This study was funded by the National Institutes of Health Grant [R01HD09374, R56HD089962, UL1 TR000457-06, T32DK007158-41] and the Canadian Institute of Health Research [FRN 146182].)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call