Determination of Anatomical Levels in Spina Bifida Fetuses with Ultrasound and MRI.

PURPOSE  The goal of this study was to assess the accuracy of prenatal anatomical level determination by ultrasound (US) and magnetic resonance imaging (MRI) by analyzing the congruence with the "true" anatomical level identified by postnatal MRI. PATIENTS AND METHODS  The first 60 patients undergoing fetal myelomeningocele surgery at The Zurich Center for Fetal Diangosis and Therapy were included in this study. Anatomical levels (i. e., first dysraphic vertebra) determined by prenatal US and MRI were compared to postnatal MRI. The level of agreement between the imaging modalities was evaluated with a Cohen's kappa test. Results > 0.6 were interpreted as good agreement, > 0.8 as excellent. RESULTS  The exact congruence between prenatal US and MRI compared to postnatal MRI was 33 % and 48 %, respectively, for an accuracy within one level difference of 80 % and 90 %, and within two levels difference of 95 % and 98 %, respectively. The level of agreement of prenatal US and MRI compared to postnatal MRI was 0.62 and 0.79, respectively. Most of the prenatally incorrectly assigned levels were assigned too high (worse) than the "true" level (US 88 % vs. MRI 65 %). CONCLUSION  Reliable exact prenatal level determination by US and MRI is not possible. However, the prenatal determination of the anatomical level of the lesion is good within one level margin of error. Prenatal US as well as MRI demonstrate a systematic error towards higher levels. The above considerations must be integrated into prenatal counselling.