[Prenatal diagnosis of trisomy 18 syndrome with sonogram index scoring system]

To explore the value of sonogram index scoring system in the prenatal diagnosis of trisomy 18 syndrome.

Neonates who had prenatal sonographic screening in our tertiary center were followed up from January 2004 to December 2009. The fetuses who were suspected with abnormalities received karyotype analysis. All fetuses were divided into case group (trisomy 18 group) and the control group (non-trisomy 18 group). The latter group was constituted of fetuses with trisomy 21, trisomy 13, other chromosomal abnormalitis and fetuses with normal karyotype. Logistic regression analysis was done to decide the individual sonographic features of trisomy 18. A score was assigned for ultrasound markers according to their likelihood ratios for trisomy 18 syndrome. A score of 3 was assigned for the sonographic features with likelihood ratio over 200, 2 for those with likelihood ratio between 100 and 200, and 1 for those with likelihood ratio less than 100. The diagnostic efficacy of the ultrasound index scoring system was evaluated by diagnostic test. The optimal cutoff value was determined by receiver operating characteristic (ROC) curve.

The study group included 59 fetuses with trisomy 18. And 26 486 fetuses did not have trisomy 18 syndrome, including 93 fetuses with trismoy 21, 19 fetuses with trisomy 13, 134 fetuses with other chromosomal abnormalities, 3739 fetuses with normal karyotype and 22 501 fetuses with normal appearance after birth. Two or more structural defects were observed in each trisomy 18 fetus. The highest incidence of sonogram abnormalities was extremities abnormalities (85%, 50/59), followed by cardiac defects (83%, 49/59) and central nervous system (CNS) malformations (75%, 44/59). Overlapping fingers, ventricular septal defect and strawberry-shaped skull were the most common abnormalities in extremities abnormalities, cardiac defects and CNS malformations, respectively. Logistic regression identified 16 markers, including choroid plexus cyst, strawberry-shaped skull, enlarged cisterna magna, holoprosencephaly, low-set ears, ventricular septal defect, hypoplastic left heart syndrome, etc. Different scores were assigned according to the likelihood ratios of these markers. In trisomy 18 group, fetuses with the sonographic score of 1, 4, 9, 10 to 16 were 2% (1/59), 9% (5/59), 10% (6/59) and 32% (19/59) respectively, whereas in non-trisomy 18 group they were 2.549% (675/26 486), 0.215% (57/26 486), 0.004% (1/26 486) and zero, respectively. When a score of 4 was used as the cutoff value for diagnosing fetal trisomy 18, the sensitivity and specificity were 0.966 and 0.997, respectively. The area under ROC curve was 0.999.

The ultrasound index scoring system may help to quantify the ultrasound features and has a good diagnostic value for fetal trisomy 18 syndrome. The cutoff value of 4 has the best diagnostic efficacy.