Validation of the follicular and ovarian thresholds by an 18-MHz ultrasound imaging in polycystic ovary syndrome: a pilot cutoff for North African patients

Abstract

Background: Polycystic ovary syndrome (PCOS) is one of the most prevalent endocrinopathies among young women. Ultrasound evidence of polycystic ovaries is one of its crucial diagnostic criteria. Objectives: Our main objective is to study the contribution of ultrasound data in diagnosing PCOS. In addition, we aim to establish a new cutoff point for the use of ultrasound and to determine its sensitivity as well as its specificity. Design: It was a prospective study, including all patients presenting with clinical hyperandrogenism. Methods: The ultrasound examination of these patients was performed using a novel ultrasound machine (18 MHz) compared to an older ultrasound machine (2 Hz–8 MHz). Inclusion criteria encompassed adult female patients over 18 years presenting symptoms suggestive of PCOS, particularly hyperandrogenism and oligo-anovulation, meeting Rotterdam’s diagnostic criteria. Prior to inclusion, assessments were conducted to eliminate other potential causes explaining hyperandrogenism or menstrual disorders in both groups. Results: We examined 92 patients diagnosed with PCOS. Menstrual disorders were the main symptoms, with amenorrhea being more frequent in the PCOS group (G1) (48.9% vs the control group (G2): 11.1%). The follicle number was significantly lower in the control group, as assessed by both ultrasound machines ( p < 10−3). The accuracy of the new ultrasound device was evaluated compared to the old one using the receiver operating characteristic (ROC) curve, revealing a cutoff of 18 follicles (sensitivity of 68.1%, specificity of 100%) and an area under the curve of 0.955. We found a significant difference between the median values of the number of follicles (NF) by both ultrasound machines (18 vs 12). It was positively correlated with an index of r = 0.916. For the volume, it was distinctively higher in G1 ( p < 10−3). ROC curve analysis revealed an ovarian volume cutoff of 9.25 ml with a sensitivity of 48.9% and a specificity of 100%. Both ultrasound machines were positively correlated with an index of r = 0.979 ( p < 10−3). Conclusion: In conclusion, we were able to establish significant correlations between the new and the old ultrasound devices for both the NF and ovarian volume. Our study is distinctive as it represents the first on the African continent to re-evaluate the ultrasound criterion for PCOS.