Optimisation of an Algorithm for Fetal Strain by CMR and its Use in Suspected Coarctation of the Aorta

Abstract Prenatal diagnosis of coarctation of the aorta (CoA) remains challenging, with low detection rates and high false positive rates. It is unknown whether fetuses that develop postnatal CoA (true CoA) exhibit impaired ventricular strain prenatally or whether fetal cardiac magnetic resonance (CMR)-derived strain can improve diagnostic accuracy. In addition, fetal strain assessment by CMR requires optimisation and validation. Forty-two third-trimester fetuses underwent fetal CMR at 1.5 T with Doppler ultrasound gating. Left and right ventricular four-chamber strain were assessed using feature tracking and endocardial shortening, respectively. Strain analysis was optimised by adjusting contour stiffness and tracking distance parameters and comparing results with myocardial centreline shortening. Fetal echocardiographic markers associated with true CoA were also evaluated, including the mitral-to-tricuspid valve ratio (MV/TV), aortic isthmus-to-duct ratio (I/D), the combined I/D × MV/TV ratio, and the carotid-subclavian artery index (CSAI). Strain analysis was feasible in 35 fetuses (9 true CoA, 18 false positives, and 8 controls). LV and RV strain values were significantly affected by optimization settings. Although fetuses with true CoA showed a tendency toward lower LV strain, no significant differences in LV or RV strain were observed between groups (all p > 0.05). In contrast, echocardiographic ratios differed between groups (all p < 0.05). Fetal CMR-derived strain is highly dependent on analysis methodology, and values obtained using different techniques are not interchangeable. While echocardiographic markers were associated with outcome, overlap between groups limits their predictive value. In its current form, fetal CMR-derived strain does not improve prenatal CoA diagnosis beyond established echocardiographic markers.