BACKGROUND: Due to advances in 3-dimensional and 4-dimensional ultrasonography, it is possible to detect CF anomalies at 10 weeks of gestation as the facial bones begin to ossify. Rates of prenatally diagnosed craniofacial anomalies vary by region and country partially due to varied screening policies and level of technician expertise. Isolated craniosynostosis is a particular diagnostic challenge due to difficulties visualizing cranial sutures on ultrasound. The purpose of this study was to identify the diagnostic accuracy of ultrasound and magnetic resonance imaging (MRI) for various craniofacial anomalies at our tertiary care center associated with a high volume fetal diagnostic unit.
METHODS: Our institutional fetal imaging database, Fetal Force, was queried to identify patients with suspected craniofacial conditions from January 2002 through August 2019. Parental and demographic data, prenatal imaging, fetal DNA sequencing, postnatal exam findings, and outcomes (delivery, termination, fetal demise, infant demise) were obtained. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of prenatal diagnosis were calculated using postnatal clinical examination as the gold standard. Fetal terminations, demises, and dyads lost to follow-up were excluded from calculations due to lack of gold standard comparison.
RESULTS: Of the 73 parent/fetus dyads identified, 43 fetuses met all inclusion criteria. Thirty dyads were excluded for noncraniofacial anomalies, common facial clefts, or scans obtained due to family history of craniofacial anomalies. The mean maternal age at consultation was 32.8 ± 5 years (range, 22.4–41.3), mean gestational age 26w ± 4w6d (range, 19w–36w5d). 70% (30) of patients were prenatally suspected to have craniosynostosis, 14% (6) micrognathia, 7.0% (3) Binder’s syndrome, and 9.3% (4) a variety of other conditions; microophthalmos, goldenhar, amniotic band syndrome, and Rubenstein-Taybi syndrome. 69.7% (30) of patients received fetal ultrasound and fetal MRI, 23.3% (10) received fetal US only, and 7.0% of patients (3) received fetal MRI alone. Seven fetuses were terminated (Craniosynostosis n = 6, microophthalmos n = 1) and 4 infants with multiple congenital anomalies passed away in infancy. For the diagnosis of any craniofacial anomaly, ultrasound: sensitivity 90%, specificity 43%, PPV 82%, NPV 60%; MRI: sensitivity 86%, specificity 50%, PPV 86%, NPV 50%. For craniosynostosis specifically, ultrasound: sensitivity 100%, specificity 43%, PPV 71%, NPV 100%; MRI: sensitivity 100%; specificity 50%; PPV 82%, NPV 100%. Ultrasound had a sensitivity and PPV of 100% for both micrognathia and Binder’s syndrome. There were 4 false positive diagnoses of isolated craniosynostosis on prenatal ultrasound that were found to be overriding sutures without synostosis or normal head shape variants on postnatal examination. In the setting of syndromic craniosynostosis, careful attention was paid to associated anomalies (ie, hands and feet in Apert Syndrome) to support the diagnosis.
CONCLUSIONS: Although CF anomalies can be detected as early as 10 weeks gestation, most anomalies are diagnosed in the second trimester after the fetal anatomy scan. Finding ways to maximize diagnostic accuracy is paramount given the profound consequences of parental decision-making subsequent to diagnosis. Additionally, it is essential to communicate the degree of doubt associated with each prenatal diagnosis, especially in the setting of isolated anomalies.