EP27.11: Normal/arcuate vs septate uterus: interobserver reliability/agreement of currently used measurements

To evaluate the inter-observer reliability/agreement of currently used measurements to distinguish between normal/arcuate and septate uterus. We included 100 women evaluated between Jun-2016 and Jul-2016 with suspected uterine anomaly, using a single 3D data-set of each uterus. Two observers using the same initial data-set, independently manipulated the uterus to obtain the coronal plane and performed the following measurements blinded to each other results: indentation depth, indentation angle, and uterine fundal wall thickness; the latter was used to calculated the indentation to wall thickness (I:WT) ratio. Inter-observer reliability and agreement of indentation depth, angle and I: WT ratio were assessed by concordance correlation coefficient (CCC) and limits of agreement (LoA). The inter-rater reliability of indentation depth (CCC=0.99, 95%CI=0.98-0.99) was significantly better than both indentation angle (CCC=0.96, 95%CI=0.94-0.97) and I:WT ratio (CCC=0.92, 95%CI=0.90-0.94). The LoA were: indentation depth = -1.7mm to +2.1mm; indentation angle = -17° to +16°; and I:WT ratio = -75% to + 96%. Among the three measurements frequently used to distinguish between normal/arcuate and septate uterus, the indentation depth was the most reliable; differences up to ± 2mm are expected if the measurement is repeated by another observer.