EP06.08: the Accuracy and Limitations of Prenatal Diagnosis of Congenital Heart Defects in a Referral Centre
Y. Li,Y. He,Y. Zhang,L. Sun,X. Liu,S. Ge
DOI: https://doi.org/10.1002/uog.16865
2016-01-01
Abstract:In this study, we sought to investigate the accuracy of prenatal echocardiographic screening and diagnosis of congenital heart defect (CHD) by a regional referral centre, i.e. Beijing Anzhen Hospital fetal echocardiography laboratory, compared with a regional registry of birth defect, i.e. Beijing population-based surveillance system on birth defect. A second aim is to assess the limitations and the quality improvement measures of fetal echocardiography for diagnosis of CHD. We studied all 8522 fetal echocardiograms from September 1, 2012 to December 31, 2014 in our institution. Among the studies, data were complete in 2489 cases in the Beijing regional registry. Using the results in the registry of population-based surveillance system in < span style = "line-height:20.8px" > Beijing on birth defect as a reference standard, the diagnostic accuracy variables were determined. In the 2489 cases, 49 cases were true positives, 2415 were true negatives, 29 were false positives (15 VSDs, 4 vascular rings, 2 PSs, 1 TOF, 1 TGA, 1 AVC, etc.), 2 false negatives (2 VSDs), and 4 cases with discordance in diagnosis. Therefore, in this cohort, the sensitivity is 95.6%, specificity 98.8%, false-negative rate 4.4%, false-positive rate 1.2%, accuracy 98.7%, positive-predictive value is 59.7%, negative-predictive value is 99.9% and Kappa value is 0.73, P < 0.001. The ROC analysis showed ventricular septal defect = 2.3 mm as the cut-off value will be the best threshold for diagnosis of VSD. With the limitations of the study in mind, our study suggests fetal echocardiography is highly accurate in diagnosis of CHD at a referral centre. The accuracy can be further improved when diagnostic criteria are optimised for more commonly missed diagnosis, e.g. 2.3 mm cut-off for VSD.