Predictive capability of slice-to-volume 3D black blood imaging-augmented 4D flow CMR for aortic coarctation in third-trimester fetuses (RF_FR_314)

Coarctation of the aorta (CoA) can be challenging to diagnose in utero with fetal echocardiography, even in expert hands. A prior fetal 4D flow cardiovascular MR (CMR) study found that decreased flow at the distal arch was associated with CoA¹. However, the spatial resolution of fetal 4D flow is limited as rapid scanning is needed to minimize fetal motion, making detection and flow quantification in small vessels, such as a hypoplastic arch, challenging. Slice-to-volume reconstruction for black blood (BB) imaging has been proposed to quickly capture small fetal vascular structures^2,3. Thus, this study aims to investigate the capability of a combination of BB and 4D flow (BB+4D) to detect CoA and to compare flow quantification performance versus conventional 4D flow CMR (4D-alone) for healthy and suspected CoA fetuses. We hypothesize that BB+4D better captures flow at the arch than 4D-alone.

Methods:

CMR data including 4D flow (Doppler ultrasound gated, gradient recalled echo) and multi-planar BB (T2-weighted single shot turbo spin echo) were prospectively acquired with IRB approval on a Philips 3T Ingenia for normal and suspected CoA fetuses. BB data were motion corrected and reconstructed in 3D using SVRTK². Post-processing for 4D flow was conducted using a custom MATLAB tool⁴, and flow was quantified by either BB+4D or 4D-alone: based on segmentation of the BB data or from a phase contrast MR angiography created from the 4D flow data alone. The main pulmonary artery, ductus arteriosus (ductus) and aorta were segmented using 3D Slicer, and BB segmentation was aligned to 4D flow using a rigid registration function in FLIRT⁵ (Fig. 1). Net flow at the ductus, distal arch, and descending aorta was quantified using both segmentations. Mass conservation (ductus+arch=descending aorta) was evaluated using orthogonal regression and Pearson correlation. Predictive capability of arch flow for postnatal CoA was evaluated using simple logistic regression and receiver-operating characteristic curve. P< 0.05 was considered statistically significant.

Results:

Thirteen normal and 14 suspected CoA fetuses were included (Table 1). Seven suspected CoA fetuses were confirmed as CoA postnatally (true positives). BB+4D-based net flow was lower than 4D-alone at the ductus and descending aorta (Fig. 2A). Regression models found mass conservation was better in 4D-alone (Fig. 2B). Logistic regression found the coefficient for arch flow was significant for BB+4D (-0.053; P=0.048), but not for 4D-alone (-0.035; P=0.06) although the area under the curve was not significantly different (P=0.51; Fig. 2C). Figure 2D shows a streamline visualization for two true positive case where 4D-alone was not able to detect the distal arch or an arch branch, leading to incorrect flow measurement locations.

Conclusion: Although flow quantification performance was not comparable with 4D-alone, BB+4D shows potential to better capture arch flow for CoA cases and can be helpful in diagnosing CoA prenatally using 4D flow.

Fig. 1 Data analysis workflow. The aorta (red), pulmonary artery (PA), ductus arteriosus (ductus) and descending aorta (DAo) were segmented from black blood images and phase-contrast MR angiography (PCMRA). The black blood segmentation was aligned to 4D flow using a transformation matrix obtained by rigid registration of the PA-ductus-DAo part (blue). Net flow at the ductus arteriosus, distal arch, and descending aorta was quantified using 4D flow-based (4D-alone) and black blood-based (BB+4D) segmentation masks.

Fig. 2 Comparison between conventional 4D flow analysis (4D-alone) and a combination of black blood and 4D flow (BB-4D). (A) Net flow at the ductus arteriosus (ductus), distal arch, and descending aorta. (B) Orthogonal regression and Pearson correlation to evaluate mass conservation. The gray dashed lines indicate the lines of identity and black lines are the regression lines. (C) Receiver operating characteristic curves for distal arch flow. The gray dashed line indicates where the area under the curve (AUC) is 0.5. (D) Systolic streamlines of two true positive case. Flow measurement locations for the distal arch are shown by black rectangles and the features detected only by BB-4D are highlighted by red arrows. MPA, main pulmonary artery.