Interobserver agreement in detailed prenatal diagnosis of congenital heart disease by telemedicine using four‐dimensional ultrasound with spatiotemporal image correlation

• DOI: 10.1002/uog.9059
• Publication Date: 2011-05-24T10:33:32Z
• Authors: B. M. E. Adriaanse; C. H. N. Tromp; J. M. Simpson; T. Van Mieghem; W. J. Kist; D. J. Kuik; D. Oepkes; J. M. G. Van Vugt; M. C. Haak
• Keywords: Heart Defects, Congenital; ACCURACY; Cardiac Volume; 610; Gestational Age; FETAL HEART; spatiotemporal image correlation; Ultrasonography, Prenatal; 03 medical and health sciences; Fetal Heart; 0302 clinical medicine; Pregnancy; 616; Image Interpretation, Computer-Assisted; PREGNANCIES; Humans; TECHNOLOGY; Prospective Studies; Echocardiography, Four-Dimensional; Observer Variation; prenatal diagnosis; DEFECTS; congenital heart disease; Telemedicine; 3. Good health; Pregnancy Trimester, Second; Female; interobserver agreement; telemedicine; NCEBP 14: Cardiovascular diseases; ECHOCARDIOGRAPHY

AbstractObjectiveTo evaluate the clinical accuracy of four‐dimensional (4D) echocardiography in the detailed prenatal diagnosis of congenital heart disease (CHD) in a telemedicine setting.MethodsTen second‐trimester spatiotemporal image correlation (STIC) volumes were sent to three observers in different tertiary care centers with expertise in 4D echocardiography. The 10 volumes were selected based on the type of diagnosis to cover a wide spectrum of CHD anomalies, and also included one normal fetal heart. Observers were asked to provide the diagnosis, the postprocessing modalities used and the time spent on examination, and to give a rating of the confidence for the diagnosis on a 5‐point Likert scale. They were free to consult other colleagues, including pediatric cardiologists, but were blinded to the prenatal diagnosis and the neonatal outcome. A diagnostic scoring system was used to evaluate different aspects of the heart defects. The results were compared with neonatal echocardiography or postmortem findings (‘gold standard’).ResultsIn two cases all observers correctly diagnosed all details of the volume datasets. The observer with the best performance reached perfect agreement in six cases and nearly perfect agreement in three. The volumes were most frequently studied by sectional planes and were analyzed in a median time of 11.0 (range, 2.5–30.0) min. The median confidence score was 4.0 (range, 1.0–5.0).ConclusionsIn a telemedicine setting using STIC volumes, fetal cardiac anomalies can be diagnosed correctly by an expert. However, details required for adequate counseling and planning of postnatal care may be missed. STIC by telemedicine is a promising modality, although not accurate enough for exclusive use in clinical decision making regarding treatment, prognosis or termination of pregnancy. Copyright © 2012 ISUOG. Published by John Wiley & Sons, Ltd.