Assessment of fetal optic chiasm: an echoanatomic and reproducibility study

ABSTRACT Objectives Our aims were: (1) to perform an echoanatomic correlation study, in order to confirm that the structure identified as the optic chiasm (OC) on ultrasound (US) is indeed this anatomical structure; (2) to assess and compare the reproducibility of two‐ (2D) and three‐(3D) dimensional US in measurement of the OC in normal fetuses; and (3) to assess whether the spatial orientation of the OC changes with increasing gestational age. Methods For the echoanatomic study, the OC was studied in a neonatal specimen, deceased at 29 + 4 weeks, by passing a suture around the OC and visualizing the supposed OC structure on US while pulling gently on the suture. The reproducibility study included 39 women with normal pregnancy at 20–33 weeks undergoing routine obstetric US examination. After the routine exam, the OC was visualized on 2D‐US, and a 2D image and 3D volume dataset were stored for offline measurement. On the 2D images, the diameters of the OC decussation and the optic tract proximal to the transducer were measured. For the 3D volume dataset, multiplanar image correlation with volume contrast imaging (VCI) was used to measure both these diameters and the chiasmocallosal angle (CCA). Two operators each took two sets of measurements of the diameters on 2D‐ and 3D‐US, and intra‐ and interoperator variability were analyzed using Cronbach's alpha intraclass correlation coefficient (ICC), while a single operator took two sets of CCA measurements for assessment of intraoperator variability. Differences in CCA with increasing gestational age were also analyzed by regression, and CCA measurements were divided into three groups according to gestational age and their means compared by one‐way ANOVA. Results During the echoanatomic experiment, when the sling suture was pulled, the hyperechoic X‐shaped structure just below the circle of Willis identified on 2D‐US as the OC was displaced slightly and was eventually cut by the sling, confirming its identity as the OC. Intraoperator variability was low and almost identical for the two operators and the two imaging modalities for measurement of the decussation (ICC for 2D‐US: 0.96 vs 0.95; 3D‐US: 0.95 vs 0.96), but less so for the optic tract (ICC for 2D‐US: 0.95 vs 0.91; 3D‐US: 0.94 vs 0.83). Interoperator variability was low for the decussation (2D‐US: 0.92; 3D‐US: 0.92), but higher for the optic tracts (ICC for 2D‐US: 0.80; 3D‐US: 0.78). The difference between the mean measurement of the two operators was n