To evaluate the application value of fetal intelligent navigation echocardiography (FINE, 5D Heart) combined with virtual intelligent sonographer assistance (VIS-Assistance^?) in the diagnosis of fetal tetralogy of Fallot (TOF).

From February to January 2020, 57 pregnant women who received echocardiography examination for fetuses at Sir Run Run Shaw Hospital were selected as subjects in this study. Doctor A with rich fetal echocardiography experience collected two to five spatio-temporal image correlation (STIC) volume datasets for each fetal heart. The inexperienced doctor B and the more experienced doctor C screened out the best volume dataset in each fetus, and applied 5D Heart and VIS-Assistance^? for post-processing. According to different spine positions, 57 TOF fetuses were divided into three subgroups: A (spine 2-4 o'clock), B (spine 5-7 o'clock), and C (spine 8-10 o'clock). Five diagnostic views (three-vessel tracheal view, four-chamber view, left ventricular outflow view, right ventricular outflow view, and abdomen/stomach view) and 17 main diagnostic elements of TOF fetuses were scored, and in particular the following parameters were investigated: the display rates of five diagnostic views and 17 diagnostic elements in the three subgroups and the differences between subgroups; comparison of the display rates of 17 diagnostic elements twice by two doctors with different experience; and repeatability and consistency test of two different senior doctors' scoring results in five diagnostic views.

With the application of 5D Heart and VIS-Assistance^?, there was no significant differences in the display rates of five diagnostic views and 17 diagnostic elements between the three subgroups (P＞0.05). The display rates of the diagnostic elements of two different senior doctors are as follows: in the three-vessel tracheal view, the display rates of diagnostic elements (dilated aorta, stenosis of pulmonary artery, trachea, and superior vena cava) were 94.7%-96.5%, 93.0%-98.2%, 70.2%-84.2%, and 86.0%-91.2%, respectively. In the four-chamber view, the display rates of the diagnostic elements (four chambers, atrioventricular valves, mitral valve-tricuspid valve distance, ovale foramen, cardiac cross, ventricular septum, and moderator band) were 100.0%, 100.0%, 93.0%-96.5%, 78.9%-82.5%, 94.7%-96.5%, 100.0%, and 100.0%, respectively. In the left ventricular outflow view, the display rates of diagnostic elements (ventricular septal defect, dilated aorta, and the aorta overriding ventricular septal defect) were 96.5%-100.0%, 100.0%, and 96.5%-98.2%, respectively. In the right ventricular outflow view, the display rates of stenosis of the pulmonary artery was 80.7%-91.2%. In the abdomen/stomach view, the display rate of the diagnostic elements (stomach and descending aorta) was 100.0%. There was no statistical differences for the results between two doctors with different experience, and for scores of the five diagnostic views by the same doctor 1 month before and after (P＞0.05).

5D Heart combined with VIS-Assistance^? has high display rates in various diagnostic views and diagnostic elements in fetal TOF, and has good repeatability and consistency. It has relatively less dependence on the experience of the operating physician, which may has potentials in the prenatal diagnosis of TOF. Using 5D Heart, the fetal spine position within a certain range seems not to affect the TOF image display rate.