First-trimester visualization of the fourth ventricle in fetuses with and without spina bifida.

OBJECTIVES

The purpose of this study was to examine the efficacy of nonvisualization of the fourth ventricle for first-trimester detection of spina bifida.

METHODS

A total of 250 digitally stored sonographic examinations at gestational ages of 11 weeks to 13 weeks 6 days (245 normal and 5 randomly interspersed spina bifida cases) were retrospectively analyzed by 4 blinded reviewers for the presence or absence of the fourth ventricle followed by an anteroposterior ventricular dimension measurement. The ventricle size was related to the crown-rump length and gestational age by linear regression analysis and Pearson correlation.

RESULTS

The fourth ventricle was identified in 971 of 1000 image readings (97.1%). False-negative and false-positive readings occurred in 11 of 20 (55.0%) and 20 of 980 (2.0%) cases, respectively (sensitivity, 0.45; specificity, 0.98.). False-negative and false-positive readings were evenly distributed throughout the gestational age range. When the ventricular size was measurable, its mean dimensions increased linearly with gestational age and were below the fifth percentile in 10 of 245 (4.0%) normal and 0 of 4 spina bifida cases, respectively. Intraclass correlation coefficient estimates were calculated based on the 2-way analysis of variance model and found to be 0.30 for a single rater and 0.64 for the mean of 4 raters.

CONCLUSIONS

Nonvisualization of the first-trimester fourth ventricle is a less robust screening parameter for spina bifida than previously published.