Fetal Cardiac MRI: Doppler US-gated Cine Imaging in Complex Congenital Heart Disease

See also the article by Vollbrecht et al in this issue.

David M. Biko, MD, is an assistant professor of radiology at the University of Pennsylvania and interim chief of the body imaging division and director of cardiovascular and lymphatic imaging at the Children’s Hospital of Philadelphia. He is also the chief of the pediatric section at the American College of Radiology Institute of Radiologic Pathology. Dr Biko is currently the chair of the Society of Pediatric Radiology Cardiac Committee and a member of the Society of Cardiac Magnetic Resonance Pediatric and Congenital Heart Disease Steering Committee.

Mark A. Fogel, MD, FACC, FAHA, FAAP, is a professor of pediatrics (cardiology) and radiology at the University of Pennsylvania and director of cardiac MR at the Children’s Hospital of Philadelphia. He has published more than 200 peer-reviewed papers on cardiac MR and has held multiple grants from the National Institutes of Health (NIH). He has served on the Board of Scientific Counselors of the NIH and has been a Board of Trustees member of the Society of Cardiac Magnetic Resonance. He published the first functional fetal cardiac MR work in 2005.

Congenital heart disease (CHD) incidence ranges from six to 13.6 per 1000 live births and remains a leading cause of infant morbidity (1). Prenatal fetal echocardiography has enabled improved outcomes in children with high-risk CHD, helping to direct babies to an appropriate level of hospital care. In general, accurate prenatal diagnosis of ductal-dependent CHD necessitates delivery where there is adequate support while directing those with minimal-risk ductal-dependent lesions to deliver safely in other planned locations (2). It has been demonstrated that prenatal diagnosis of hypoplastic left heart syndrome is associated with decreased risk of ductal shock and better preoperative risk factors such as higher preoperative pH and lower lactate levels (1). In one article, prenatal diagnosis of transposition of the great arteries saw substantial improvement in 1-week mortality (3).

Given these data, improving prenatal diagnosis of CHD with the use of other imaging tools such as MRI may improve outcomes as previously demonstrated in prenatal central nervous system imaging (4). Fetal cardiac MRI can serve as a complement to fetal echocardiography, while MRI can simultaneously evaluate the brain and the spine, complementing prenatal US, which is important because neurodevelopment is becoming recognized as a critical component of CHD outcomes (5). The challenges of fetal cardiac MRI include small structures (eg, size of major fetal vasculature at full gestation is 5 to 10 mm in diameter, and measurements of the width of the fetal ventricle at full gestation range from 10 to 30 mm), short duration of the cardiac cycle, use of a large field of view to avoid maternal anatomy wrapping into the image, and fetal motion (6). Finally, fetal cardiac gating is difficult.

Real-time steady-state free precession (SSFP), metric-optimized gating, and self-gating are techniques that have been previously used for fetal cardiac gating. Real-time SSFP allows for dynamic imaging of the heart, which is able to quantify ventricular volumes in utero (7). Unfortunately, this real-time technique has limited temporal resolution (8). Metric-optimized gating is performed using conventional cine images and retrospectively determines the phase of the cardiac cycle in the absence of an electrocardiographic signal. This technique acquires sequential lines of k-space for a predetermined interval, resulting in a matrix of mixed cardiac phases (9). In this technique, k-space needs to be oversampled by sampling over a full range of possible heart frequencies in the fetus. Cardiac motion is then used as a metric for image quality and entropy where hypothetical triggers are adjusted to produce an optimized image. The limitation of this method is that it is susceptible to heart rate variations. Additionally, metric-optimized gating requires postprocessing, which is time-consuming, taking approximately 10 minutes per section (8).

With self-gating, the gating signals are directly extracted from the MRI data, most commonly radial data, and sorted retrospectively, which can be difficult with the small fetal heart and high heart rate (6). Self-gating extracts motion data directly from the image data, using periodic signal variations from the varying blood flow velocities during the cardiac cycle. For instance, in systole, excited protons exit the voxel because of the high velocity of blood flow leading to dephasing of signal and a decrease in transverse magnetization, whereas transverse magnetization increases during diastole. Similar to metric-optimized gating, postprocessing is time-consuming, taking up to 10 minutes per section (8).

In this issue of Radiology: Cardiothoracic Imaging, Vollbrecht and colleagues (10) present a prospective study of 23 female participants with fetuses with CHD. The authors image the cardiovascular system of the fetus by using Doppler US to perform direct interrogation of the blood to synchronize to the cardiac cycle, rather than using metric-optimized gating or self-gating. Using this technique, the authors had comparable findings between fetal cardiac MRI and fetal echocardiography, which was confirmed with postnatal imaging. In a few cases, fetal cardiac MRI depicted additional diagnostic findings that were not seen at fetal echocardiography (10). Previously, the use of Doppler US gating was successful in a small number of fetuses, including four with CHD, showing agreement with findings from echocardiography. Using Doppler US gating, fetal cardiac MRI was able to add additional information to inconclusive findings from fetal echocardiography (8).

The use of this Doppler US technique has the potential to simplify fetal cardiac gating by directly detecting fetal cardiac motion. This enables imagers to use established postnatal cardiac protocols for fetal imaging. As opposed to metric-optimized gating and self-gating, Doppler US gating does not require postprocessing, allowing a real-time assessment of imaging. The disadvantage of this technique is that additional hardware is needed to perform these examinations. Fetal motion can also affect the Doppler signal, which may require repositioning.

Fetal cardiac MRI using Doppler US gating was technically successful and accurate against the reference standard of echocardiography in the cohort of patients reported by Vollbrecht et al (10). Future studies should focus on indications for the evaluation of a fetus with CHD when using fetal cardiac MRI. What structures are best depicted, and at what gestational age, are questions that need to be explored. Although hardware is required for the evaluation of fetal cardiac MRI with Doppler US gating, the hope is that by allowing cardiac imagers to easily apply existing neonatal protocols to fetal cardiac MRI, this technique will become more common among fetal centers. Ultimately though, the test of fetal cardiac MRI is how accurate it is postnatally and if it can predict outcome or direct care better than fetal echocardiography. This complement to fetal echocardiography may lead to a better understanding of CHD. Just as fetal echocardiography leads to improved outcomes in CHD, the hope is that additional data from fetal cardiac MRI will yield similar results.