Early second-trimester individualized estimation of trisomy 18 risk by ultrasound.

OBJECTIVE To report two second-trimester ultrasound algorithms for trisomy 18 prediction. METHODS Femur length, gross anomaly, choroid plexus cysts, two-vessel cord, and maternal age were documented in pregnancies undergoing genetic amniocentesis. Stepwise logistic regression was used to identify 1) the significant markers for predicting trisomy 18 when gross anomaly was not considered (algorithm 1) and 2) when gross anomaly was also considered (algorithm 2). Patient-specific risk was calculated based on the significant ultrasound markers plus maternal age. The diagnostic accuracy of each algorithm was determined. RESULTS There were 1167 normal and 47 trisomy 18 cases. The mean gestational ages were 16.5 weeks (standard deviation [SD] 1.5) and 17.9 weeks (SD 1.6), respectively. Algorithm 1 consisted of maternal age, choroid plexus cyst, femur length, and two-vessel cord. The sensitivity and false positive rates were 61.7% and 1.5%, respectively, with an area under the receiver operating characteristics curve of 0.880 (P <.001). Algorithm 2 (age, femur length, gross anomaly, and choroid plexus cyst) had a sensitivity of 72.3% and false positive rate of 0.9% with an area under the curve of 0.956 (P <.001). Comparable detection rates were achieved in early gestation at up to and including 17.5 weeks (72.4% and 82.8%, algorithms 1 and 2, respectively, at a 4.0% false positive rate). CONCLUSION The ultrasound markers were sensitive for trisomy 18 detection in the early second trimester.