Improved detection of coarctation of the aorta using speckle-tracking analysis of fetal heart on last examination prior to delivery.

OBJECTIVE

The false-positive rate for prenatal diagnosis of coarctation of the aorta (FP-CoA) commonly exceeds 50%, with an accurate detection rate of < 50%. This study was conducted to determine if the sensitivity for prenatal detection of true CoA and the FP-CoA rate could be improved by evaluating the fetal epicardial size and shape in the four-chamber view (4CV) and the endocardial right (RV) and left (LV) ventricular size, shape and contractility.

METHODS

We analyzed retrospectively Digital Imaging and Communications in Medicine (DICOM) clips of the 4CV from the last examination prior to delivery in a series of 108 fetuses with CoA suspected prenatally by pediatric cardiologists using traditional diagnostic criteria. Postnatal evaluation distinguished those fetuses which subsequently required CoA surgery (true positives; true CoA) from those that were FP-CoA. Postnatal cardiac abnormalities were identified for each group. For the prenatal evaluation, we measured the 4CV end-diastolic epicardial area, circumference, length, width and global sphericity index. Speckle-tracking analysis was used to compute the endocardial RV and LV end-diastolic area, length, 24-segment sphericity index, 24-segment transverse width and the following functional parameters: fractional area change; global longitudinal, free-wall and septal-wall strain; basal-apical-length, basal free-wall and basal septal-wall fractional shortening; septal-wall annular plane systolic excursion; 24-segment transverse-width fractional shortening; and LV end-diastolic and end-systolic volumes, stroke volume, cardiac output and ejection fraction. In addition, the RV/LV end-diastolic area ratio was computed. Using a control group of 200 normal fetuses, the mean and SD for each of the above cardiac measurements was used to compute the Z-scores for each measurement in each of the 108 study fetuses. Logistic regression analysis was then performed on the Z-score values to identify variables that separated the true CoA group from the FP-CoA group.

RESULTS

Of the 108 study fetuses, 54 were confirmed postnatally to have true CoA and 54 were FP-CoA. Right/left area disproportion > 90  centile was present in 80% (n = 43) of the true-CoA fetuses and 76% (n = 41) of the FP-CoA fetuses. Fetuses with true CoA had a significantly greater number of associated cardiac abnormalities (93%, n = 50) compared with the FP-CoA fetuses (61%, n = 33) (P < 0.001). The most common associated malformations were bicuspid aortic valve (true CoA, 46% (n = 25) vs FP-CoA, 22% (n = 12); P < 0.01), aortic arch hypoplasia (true CoA, 31% (n = 17) vs FP-CoA, 11% (n = 6); P < 0.01), ventricular septal defect (true CoA, 33% (n = 18) vs FP-CoA, 11% (n = 6); P < 0.05) and mitral valve abnormality (true CoA, 30% (n = 16) vs FP-CoA, 4% (n = 2); P < 0.01). Logistic regression analysis identified 28 variables that correctly identified 96% (52/54) of the fetuses with true CoA, with a false-positive rate of 4% (2/54) and a false-negative rate of 4% (2/54). These variables included the epicardial size in the 4CV, size and shape of RV and LV, and abnormal contractility of RV and LV. The area under the receiver-operating-characteristics curve was 0.98 (SE, 0.023; 95% CI, 0.84-1). There was no significant difference in the percent of fetuses with RV/LV area disproportion between those with CoA and those that were FP-CoA.

CONCLUSIONS

Speckle-tracking analysis of multiple ventricular measurements may be helpful to refine the diagnosis in fetuses that are suspected to have CoA prenatally. © 2020 International Society of Ultrasound in Obstetrics and Gynecology.