Diagnosing Septate Uterus Using Three-Dimensional Ultrasound Using Three Different Classifications: An Interobserver and Intraobserver Agreement Study

Objective Currently, there are up to three different classifications for diagnosing septate uterus. The interobserver agreement among them has been poorly assessed. Objective To assess the interobserver agreement of nonexpert sonographers for classifying septate uterus using the European Society of Human Reproduction and Embryology/European Society for Gynaecological Endoscopy (ESHRE/ESGE), American Society for Reproductive Medicine (ASRM), and Congenital Uterine Malformations by Experts (CUME) classifications. Methods A total of 50 three-dimensional (3D) volumes of a nonconsecutive series of women with suspected uterine malformation were used. Two nonexpert examiners evaluated a single 3D volume of the uterus of each woman, blinded to each other. The following measurements were performed: indentation depth, indentation angle, uterine fundal wall thickness, external fundal indentation, and indentation-to-wall-thickness (I:WT) ratio. Each observer had to assign a diagnosis in each case, according to the three classification systems (ESHRE/ESGE, ASRM, and CUME). The interobserver agreement regarding the ESHRE/ESGE, ASRM, and CUME classifications was assessed using the Cohen weighted kappa index (k). Agreement regarding the three classifications (ASRM versus ESHRE/ESGE, ASRM versus CUME, ESHRE/ESGE versus CUME) was also assessed. Results The interobserver agreement between the 2 nonexpert examiners was good for the ESHRE/ESGE (k = 0.74; 95% confidence interval [CI]: 0.55-0.92) and very good for the ASRM and CUME classification systems (k = 0.95; 95%CI: 0.86-1.00; and k = 0.91; 95%CI: 0.79-1.00, respectively). Agreement between the ESHRE/ESGE and ASRM classifications was moderate for both examiners. Agreement between the ESHRE/ESGE and CUME classifications was moderate for examiner 1 and good for examiner 2. Agreement between the ASRM and CUME classifications was good for both examiners. Conclusion The three classifications have good (ESHRE/ESGE) or very good (ASRM and CUME) interobserver agreement. Agreement between the ASRM and CUME classifications was higher than that for the ESHRE/ESGE and ASRM and ESHRE/ESGE and CUME classifications.