Abstract
Purpose
To measure respiration-dependent blood flow in the total cavopulmonary connection (TCPC) of patients with Fontan circulation by using free-running, fully self-gated five-dimensional (5D) flow MRI.
Materials and Methods
From July to November 2018, 10 volunteers (six female volunteers, mean age, 25.1 years ± 4.4 [standard deviation]) and six patients with Fontan circulation (two female patients, mean age, 19.7 years ± 7.5) with a TCPC were examined by using a cardiac- and respiration-resolved three-directional and three-dimensional phase-contrast MRI sequence (hereafter, 5D flow MRI). This prospective study was conducted with approval of the local ethics committee, and written informed consent was obtained from all participants and/or their representative. 5D flow data were acquired during free breathing. Data were reconstructed into 15–20 heart phases and four respiratory phases: end-expiration, inspiration, end-inspiration, and expiration. Respiration-dependent stroke volumes (SVs) and particle traces were analyzed from the caval circulation of volunteers and patients with Fontan circulation. Statistical analysis was performed by using parametric tests and scatterplots.
Results
The respiration dependency of caval blood flow was evaluated in all participants and was significantly elevated in patients with Fontan circulation as compared with volunteers. In patients, SV in the inferior vena cava (IVC) showed variations of 120% between inspiration and expiration (P = .002). The flow distribution in the IVC and superior vena cava among the four respiratory phases was differentiated by 20% (range, 9%–30%) and 4% (range, 0%–13%), respectively.
Conclusion
Hemodynamic parameters (volume flow and blood flow distribution) throughout the cardiac and respiratory cycle can be measured using a single scan, potentially providing further insights into the Fontan circulation.
© RSNA, 2019
Summary
A cardiac- and respiration-resolved three-directional and three-dimensional phase-contrast MRI sequence was applied in 10 volunteers and six patients with Fontan circulation, and blood flow rates and flow distribution were shown to be affected by respiration in the total cavopulmonary connection of patients with Fontan circulation.
Key Points
■ Multiple hemodynamic clinical markers can be measured in a single, free-running scan that is simple to set up.
■ The 5D flow MRI technique has a predictable scan time, reduces the risk of failure, and is fully self-gated, which avoids the need of physiologic sensors and increases patient comfort.
■ The 5D flow MRI technique has the potential to improve the understanding of single-ventricle circulation, especially with regard to respiratory-dependent hemodynamics.
Introduction
In patients with Fontan circulation, blood flow dynamics are known to be complex and very heterogeneous among patients (1,2). This is due to the multitude of morphologies and the variations of the procedures themselves (3). Venous return physiology in patients with history of Fontan surgery remains less understood and is variably being attributed to active ventricle function (4,5), to breathing and exercise (5–8), or even to a different cardiovascular model seeing the heart as an impedance pump generating pressure (9). This complicated interplay between geometry and physiology has opened many questions of single-ventricle circulation, and the parameters influencing hemodynamics in patients with Fontan history are not yet fully understood. Fontan geometries can be accurately measured by using MRI or CT. The measurement and simulation of hemodynamics therein has also been shown to be possible, whereas boundary conditions for simulations are uncertain and technical challenges remain to provide an accurate and comprehensive analysis (10). Measuring time-resolved three-directional and three-dimensional (3D) blood velocities using MRI (hereafter, four-dimensional [4D] flow MRI) has initially provided insights into the complicated interplay between anatomy and hemodynamics in these patients (11–16). The quantification of blood flow distributions in the total cavopulmonary connection (TCPC) (14–17) is thought to be a marker for pulmonary arteriovenous malformations (18,19) caused by the absence of hepatic venous blood flow into the pulmonary arteries (20). The measurement of physiology-dependent (breathing and/or exercise) hemodynamics has to date mostly been measured by using two-dimensional flow MRI (6–8,21). Although these studies demonstrate a strong variation of blood flow volumes with breathing processes, they do not allow full 3D coverage nor the quantification of TCPC flow distribution (FD).
Although FD in patients with Fontan circulation was recently analyzed by using 4D flow MRI in end-inspiration and end-expiration (22), the method did not allow the measurement of the inspiration and expiration process period (pressure-driven filling and emptying of the lungs), which are known to be the strongest varying drivers of venous flow differentiating preserved from impaired patients with Fontan circulation (23).
In this study, we have proposed a noninvasive imaging method, respiration-resolved three-directional and 3D phase-contrast MRI sequence (hereafter, five-dimensional [5D] flow MRI), that allows the assessment of respiratory-resolved 4D flow MRI, covering the entire respiratory and cardiac cycle (24). As compared with previous application of 4D flow MRI in patients with Fontan circulation, the proposed method is free running (no need for electrocardiography or respiratory sensors); that is, it runs at a predictable scan time (independent of breathing behavior or heart cycle variations), and most importantly allows one to freely reorder the data in various heart cycles and breathing positions and processes.
Materials and Methods
Study Population
From July to November 2018, 10 volunteers (six female volunteers, mean age, 25.1 years ± 4.4 [standard deviation], age range, 21–41 years) without a history of cardiovascular disease (Table E1 [supplement]) and six patients with Fontan circulation (two female patients, mean age, 19.7 years ± 7.5, age range, 10–30 years) without stent material were included (Table E2 [supplement]) in this study. This prospective study was conducted with approval of the local ethics committee, and written informed consent was obtained from all participants before the examination.
MRI Sequence
Acquisitions were performed on a 1.5-T system (Ingenia; Philips, Best, the Netherlands) using a 28-channel surface coil array. A coronal 3D slab was imaged to cover the inferior vena cava (IVC), superior vena cava (SVC), main pulmonary artery (PA), right pulmonary artery (RPA), and left pulmonary artery (LPA) in volunteers and the TCPC (consisting of IVC, SVC, RPA, and LPA) including the conduit in patients. A free-running, fully self-gated, golden-angle stack-of-spirals 5D flow sequence was performed at a fixed scan time of 15:00–16:05 minutes (24). Sequence parameters are summarized in Table 1. Data were acquired without using physiologic signals (electrocardiography or breathing cushion), regardless of respiratory or cardiac frequency changes during the scan. Velocity encoding values were adjusted per participant: 150 cm/sec in volunteers and 50–90 cm/sec in patients with Fontan circulation.
Table 1:
Pulse Sequence Parameter of 5D Flow Sequence
Reconstruction
Data reconstruction was performed offline using tools developed in-house (MATLAB 2016a; MathWorks, Natick, Mass). All data binning was performed retrospectively, based on the extracted self-navigation signal. This signal was acquired at every repetition time between both radiofrequency pulses of the spatial-spectral excitation, leading to a high temporal resolution of about 11 msec. To separate cardiac and respiratory motion patterns, the self-navigation signal was band-pass filtered between 0.1 and 0.5 Hz and 0.6 and 3 Hz, respectively. Details can be found in the study by Bastkowski et al (24). Data were binned into four respiratory phases corresponding to end-expiration, inspiration, end-inspiration, and expiration. The phases end-expiration and end-inspiration were defined as the data acquired in the −15% and +15% periods of the peak breathing position, respectively (7). The phases inspiration and expiration were defined as the periods between the two peak positions (Fig 1). There was no overlap between respiratory phases. In addition, data were binned into 15–20 cardiac phases, resulting in a mean temporal resolution of 48.6 msec ± 8.6 and 49.2 msec ± 7.3 for volunteers and patients, respectively. To account for data undersampling, a regularized conjugate-gradient sensitivity encoding reconstruction was used (25). Image preprocessing included an automated correction of Maxwell terms (26) and background phase correction by using a first-order 3D spatial fit on automatically segmented static tissue for each flow-encoding direction (27). Velocity unwrapping was manually performed after identification of voxels in the areas of interest presenting phase wraps.
Figure 1:
Acquisition and reconstruction scheme of the self-gated golden-angle spiral five-dimensional (5D) flow sequence. Golden-angle spirals are binned into four respiratory and 15–20 cardiac phases using the self-gating signal for respiratory and cardiac motion components. The use of a golden-angle acquisition scheme enables retrospective binning of the data. Respiratory motion and heart rate changes did not influence the acquisition. Four separate four-dimensional flow data sets are reconstructed, one for each respiratory phase. SENSE = sensitivity encoding.
Data Analysis
Data analysis and visualization was performed using GTFlow (Gyrotools, version 3.1.13; Winterthur, Switzerland). Two-dimensional analysis planes were manually drawn and adjusted per cardiac and respiratory phase, perpendicular to the vessel of interest. In volunteers, the vessels IVC, SVC, PA, RPA, and LPA were analyzed. In patients, the vessels IVC, SVC, RPA, and LPA were analyzed. Stroke volumes (SVs) were calculated as time integral over the net flow for each analysis plane and each respiratory phase. To quantify the FD, virtual particles were released into the upstream vessel at every heart phase during one heart cycle, and their arrivals were counted in the downstream vessels (17). To achieve a volumetric quantity, particle counts were weighted by the particle velocities at release time (28). For volunteers, particles were released in the PA and arrivals were counted in the RPA and LPA. For patients, particles were released in the IVC and SVC and arrivals were counted in the RPA and LPA.
FD was then calculated (17,29); for example, FD from IVC to RPA (
):
 |
where PRPA and PLPA are the weighted particle counts arriving in the RPA and LPA, respectively. Differences with regard to the main flow direction were analyzed. The FD of the IVC (FDIVC) with regard to the main flow direction was defined as:
 |
where 
and 
are the mean values of 
and 
over all respiratory phases. Similarly, FDSVC and FDPA were defined.
Internal Consistency Analysis
For data consistency analysis, SV of upstream versus downstream vessels was compared (concept of mass conservation). In volunteers, SV in the PA was compared with the sum of SV in RPA and LPA. In patients, the sum of SV in the IVC and SVC was compared with the sum of SV in the LPA and RPA. To evaluate the particle trace method, SV of the downstream vessels was calculated using the value for FD multiplied by the SV of the upstream vessels, for example:
 |
The result was then compared with the SV of RPA.
Statistical Analysis
Statistical analysis was performed by using RStudio (2018 RStudio, version 1.1.456; Boston, Mass). Normality of the data was tested by using a Shapiro-Wilk test. For normally distributed data, a paired t test was performed to verify if the differences in SV in the IVC, SVC, PA, RPA, and LPA vessels and TCPC between each pair of respiratory phases are statistically significant. The P value for statistical significance was set to P < .05. A power analysis was performed for the paired t test to calculate the sample size at 80% power. A significance level of .05 and an effect size of ±75% (7) was assumed. For these assumptions, a sample size of six patients was calculated. Scatterplots and Pearson correlation coefficients (R) were used for an internal data consistency analysis of SV in the upstream versus downstream vessels. For normally distributed data, a paired t test was used to verify if values used for the internal data consistency analysis are significantly different.
Results
Internal Data Consistency
Scatterplots to compare SVs of upstream versus downstream vessels are shown in Figure 2. Formulae for the compared quantities are shown in each figure. Figure 2, A and C, show the conservation of mass analysis where SV differences amounted to −2.6 mL ± 7.4 (R = 0.93) and 1.9 mL ± 5.5 (R = 0.97) in volunteers and patients, respectively. A comparison of SV calculated using the FD and SV measured in the contours is shown in Figure 2, B, and D. Mean SV differences are −1.3 mL ± 6.3 (R = 0.84) and 1.0 mL ± 5.2 (R = 0.94) in volunteers and patients, respectively.
Figure 2:
Scatterplots of stroke volumes (SVs) in upstream versus downstream vessels for internal data consistency analysis. In, A and C, SV differences of upstream vessels (in volunteers: main PA, in patients with Fontan circulation: IVC + SVC) and downstream vessels (in volunteers and patients: RPA + LPA) are compared. In, B and D, SV of upstream vessels was calculated using the SV multiplied by the flow distribution (FD) to evaluate the particle trace method versus conventional calculation of SV. Results of all respiratory phases are included in these plots. IVC = inferior vena cava, LPA = left pulmonary artery, PA = pulmonary artery, RPA = right pulmonary artery, SVC = superior vena cava.
Respiration-dependent SV
Table E3 (supplement) and Figure 3 show the respiration dependency of SV in volunteers and patients with Fontan circulation in all vessels of interest. The same data are illustrated using spaghetti plots (Figs E1 and E2 [supplement]) to show the subject-specific signal course across respiratory phases. Highest variations are observed in the IVC between inspiration and expiration. In volunteers and patients, the flow in inspiration compared with expiration was elevated by 38.0% and 120.1%, respectively. Lowest variability was measured in the SVC. Table E4 (supplement) lists all P values for the paired t test. In the IVC of patients, the differences between all pairs of respiratory phases were statistically significant. Compared with volunteers, the variability between respiratory phases was approximately three times higher in patients.
Figure 3:
Percentile changes of stroke volumes (SVs) analyzed in the four respiratory phases across partcipants. The SVs in the inferior vena cava (IVC), superior vena cava (SVC), left pulmonary artery (LPA), right pulmonary artery (RPA), and main pulmonary artery (PA) in, A, volunteers and, B, patients with Fontan circulation are shown. Highest flow can be observed in inspiration. Error bars represent the standard deviation.
Respiration-dependent FD
In all volunteers, flow was found to be almost equally distributed between RPA and LPA during all respiratory phases. FDPA in end-expiration, inspiration, end-inspiration, and expiration was 53.0% ± 4.0, 54.8% ± 3.1, 52.9% ± 2.8, and 55.7% ± 6.2, respectively. The mean FDPA over all respiratory phases was 54.1% ± 4.5 in volunteers.
Figure 4 shows velocity color-coded particle traces in the TCPC of one exemplary patient with Fontan circulation. All four respiratory phases are shown. Corresponding Movies 1 and 2 (supplement) visualize flow patterns throughout the cardiac cycle for two patients. Table 2 shows FDIVC and FDSVC for all patients and each individual main flow direction. In each patient, a preferential flow of the IVC and SVC to one PA can be observed, regardless of the respiratory phase. Figure 5 shows particle traces in the TCPC of patient 2, color-coded according to the emitting contour. A high variability in FD between the respiratory phases can also be observed in Movie 3 (supplement) (patient 2). In contrast, little variability in FD can be observed in Movie 4 (supplement) (patient 4). Overall, distribution of flow originating from the IVC showed a more even division into the RPA and LPA compared with the SVC. Mean FDIVC and FDSVC over all respiratory phases were 64.9% ± 16.1 (range, 36.0%–100.0%) and 97.1% ± 6.5 (range, 71.9%–100.0%), respectively. FDIVC over all patients for end-expiration, inspiration, end-inspiration, and expiration was 65.5% ± 16.3, 63.0% ± 13.2, 62.6% ± 9.0, and 68.5% ± 21.1, respectively. FDSVC had a lower variability between respiratory phases and between subjects. In end-expiration, inspiration, end-inspiration, and expiration, FDSVC was 97.1% ± 5.7, 99.3% ± 1.0, 98.2% ± 2.3, and 93.9% ± 10.3, respectively.
Figure 4:
Velocity color-coded particle traces in the total cavopulmonary connection of patient 2 (male patient, aged 17). Particles were released in the inferior vena cava (IVC) and superior vena cava (SVC). From left to right, (A–D) four different respiratory phases during peak systole are shown. Elevated peak-systolic flow is observed during inspiration. LPA = left pulmonary artery, RPA = right pulmonary artery.
Table 2:
Percentile Respiratory-dependent Flow Distribution for Each Individual Patient with Fontan Circulation
Figure 5:
Color-coded particle traces in the total cavopulmonary connection of patient 2 (male patient, aged 17). Particles were released in the inferior vena cava (IVC) and superior vena cava (SVC). From left to right, (A–D) four different respiratory phases in end diastole are shown. In all respiratory phases, an uneven distribution of the SVC flow to the right pulmonary artery (RPA) can be observed. The calculated values for the flow distribution (FD), FDIVCand FDSVC, are shown. Particle traces originating from the IVC show a difference in FD between the respiratory phases. LPA = left pulmonary artery.
Movie 1:
Velocity color-coded particle traces in the total cavopulmonary connection (TCPC) of patient 2 (male patient, aged 17). The movie shows a coronal view of the TCPC for all respiratory phases across one heart cycle. Note the difference in flow patterns and velocity distribution between the respiratory phases.
Movie 2:
Velocity color-coded particle traces in the total cavopulmonary connection (TCPC) of patient 4 (male patient, aged 10). The movie shows a coronal view of the TCPC for all respiratory phases across one heart cycle. Compared with Movie 1, flow patterns between the respiratory phases are similar. However, highly accelerated flow can be observed in inspiration compared with expiration.
Movie 3:
Color-coded particle traces in the total cavopulmonary connection (TCPC) of patient 2 (male patient, aged 17). The movie shows a coronal view of the TCPC for all respiratory phases across one heart cycle. Particles were released in the inferior vena cava (IVC) and superior vena cava. Particle traces originating from the IVC show a difference in flow distribution between the respiratory phases.
Movie 4:
Color-coded particle traces in the total cavopulmonary connection (TCPC) of patient 4 (male patient, aged 10). The movie shows a coronal view of the TCPC for all respiratory phases across one heart cycle. Particles were released in the inferior vena cava and superior vena cava. Note that no changes of flow distribution between the respiratory phases are visible in this patient.
Discussion
In this study, we have presented the application of cardiac- and respiration-dependent 5D flow MRI in the Fontan circulation. In spite of being able to measure respiratory-dependent hemodynamics, the presented 5D flow approach has additional advantages as compared with previously presented 4D flow techniques: the acquisition time is similar to conventional 4D flow acquisitions (15–16 minutes), but the scan time is predictable. The scan does not need any external devices or the use of magnetic resonance navigators and is fully free running. Currently, acceleration factors are in the order of two to four. Higher acceleration might be feasible to reduce scan time by using more sophisticated reconstruction techniques such as compressed sensing (30) or k-t–based methods (31,32).
A recent study by Rutkowski et al (22) used a time-averaged streamline analysis to investigate the FD in two respiratory phases: end-expiration and end-inspiration. However, Körperich et al (7) showed what our data confirm (Fig 3, B): highest variability in flow is observed during the process of inspiration and expiration, which remained irrecoverable by the method presented in the study by Rutkowski et al (22). Our approach allows for an arbitrary, freely definable rebinning into cardiac and respiratory states. It is furthermore questionable if streamlines represent the path of actual blood flow (33,12); therefore, we have used time-integrated particle traces.
It is well known that respiration is the main driver of pulmonary circulation in patients with Fontan circulation, and it is also suspected that the analysis of caval blood FD might aid in potentially identifying patients with an impaired Fontan circulation. Real-time two-dimensional flow MRI has been previously used to demonstrate respiration-dependent volume flow in the TCPC (7,23), whereas non–respiratory-resolved 4D flow MRI has been used to quantify FDs in the TCPC (17). We present a technique that allows the retrospective analysis of both of these hemodynamic dependencies. Simultaneously, hemodynamic parameters in any arbitrarily placed two-dimensional analysis plane can be retrospectively measured.
In line with previous results (7,23), caval blood flow was shown to be highly dependent on respiration in patients with Fontan circulation compared with volunteers. SV in the IVC and SVC showed a similar respiratory dependency in volunteers and patients as shown in previous work using real-time two-dimensional phase-contrast MRI (7). Highest differences in SV were observed between inspiration and expiration in the IVC, RPA, and LPA of patients. In contrast, the highest differences in SVC flow was observed between inspiration and end-inspiration.
In contrast to patients, the FD in the pulmonary circulation of volunteers shows low variation between the different respiratory phases. The overall measured blood flow distributions were in line with previous studies that used 4D flow MRI (17,22,12).
Care has to be taken when comparing previous (non–respiratory-resolved) results (12,17) with the presented data, considering the observed differences between respiratory phases in patients with Fontan circulation. The measured respiratory-dependent changes of SV between IVC and SVC in patients with Fontan circulation are expected to lead to a change in flow patterns and therefore to a change in blood flow distribution. Although differences between respiratory phases in FD are small in most patients, even an asymmetric flow in the order of only 20% has been previously considered to identify patients with preferential flow (34). As an example, patient 2 measured using non–respiratory-resolved 4D flow MRI gated on expiration would show a flow distribution of 100% (ie, 100% of the blood from the IVC flowing into the RPA). This would result in the assumption that hepatic blood would never reach the left pulmonary circulation, potentially increasing the risk of a pulmonary arteriovenous malformation. However, our method showed that during end-inspiration, 20% of the blood from the IVC does reach the LPA.
One limitation of this study was the small sample size of six patients, due to which the statistical analysis of subgroups was not feasible. We, however, believe that the extremely large heterogeneity in Fontan geometries and physiologic variations (also shown here in six cases) would limit a statistical analysis of patients with Fontan circulation in general. Therefore, individual hemodynamic parameters were also listed and visualized. A longitudinal study on individual patients with Fontan circulation with follow-up examinations might be needed to allow the understanding of the interplay of hemodynamics with physiology and its implication for patient stratification. The free-running, sensorless approach proposed in this study might facilitate the integration of 5D flow MRI in the clinical realm of patients with Fontan circulation. Another limitation that is not restricted to this study remains the long postprocessing time of 4D and 5D flow MRI. Furthermore, long reconstruction times in the order of 5 hours per respiratory phase remain a hurdle for clinical adoption. An automatic segmentation of the area of interest in combination with a time-resolved vessel contour segmentation is warranted and will facilitate the integration of the proposed method into a larger patient cohort. Finally, the respiration dependency was limited to volume flow and FDs in this study. A larger and longitudinal study is necessary to define potentially more important hemodynamic parameters and their dependence on respiration.
In conclusion, the presented technique allows one to shed light onto the physiologic dependence of blood flow in the TCPC of patients with Fontan circulation. A comprehensive analysis of hemodynamics in patients with Fontan circulation can be performed that covers cardiac and respiratory effects, the main drivers of blood flow in the TCPC. The use of non–respiratory-resolved techniques and analyses might mask out potentially important hemodynamic effects that are important for the understanding of the Fontan circulation.
SUPPLEMENTAL TABLES
SUPPLEMENTAL FIGURES
Disclosures of Conflicts of Interest: R. Bastkowski disclosed no relevant relationships. R. Bindermann disclosed no relevant relationships. K.B. disclosed no relevant relationships. K.W. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: employee of Philips Healthcare since October 2014. Other relationships: disclosed no relevant relationships. D.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: payment from Philips for lectures. Other relationships: disclosed no relevant relationships. D.G. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: employee of Siemens since December 2018 (all work related to this work was performed outside of duties at Siemens; receives royalties from Siemens. Other relationships: disclosed no relevant relationships.
Abbreviations:
- 5D
- five dimensional
- 4D
- four dimensional
- FD
- flow distribution
- IVC
- inferior vena cava
- LPA
- left pulmonary artery
- PA
- pulmonary artery
- RPA
- right pulmonary artery
- SV
- stroke volume
- SVC
- superior vena cava
- 3D
- three dimensional
- TCPC
- total cavopulmonary connection
References
- 1.Kverneland LS, Kramer P, Ovroutski S. Five decades of the Fontan operation: A systematic review of international reports on outcomes after univentricular palliation. Congenit Heart Dis 2018;13(2):181–193. [DOI] [PubMed] [Google Scholar]
- 2.Atz AM, Zak V, Mahony L, et al. Longitudinal outcomes of patients with single ventricle after the Fontan procedure. J Am Coll Cardiol 2017;69(22):2735–2744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.de Leval MR. The Fontan circulation: a challenge to William Harvey? Nat Clin Pract Cardiovasc Med 2005;2(4):202–208. [DOI] [PubMed] [Google Scholar]
- 4.Kamphuis VP, Elbaz MSM, van den Boogaard PJ, et al. Disproportionate intraventricular viscous energy loss in Fontan patients: analysis by 4D flow MRI. Eur Heart J Cardiovasc Imaging 2019;20(3):323–333. [DOI] [PubMed] [Google Scholar]
- 5.Honda T, Itatani K, Takanashi M, et al. Contributions of respiration and heartbeat to the pulmonary blood flow in the Fontan circulation. Ann Thorac Surg 2016;102(5):1596–1606. [DOI] [PubMed] [Google Scholar]
- 6.Hjortdal VE, Emmertsen K, Stenbøg E, et al. Effects of exercise and respiration on blood flow in total cavopulmonary connection: a real-time magnetic resonance flow study. Circulation 2003;108(10):1227–1231. [DOI] [PubMed] [Google Scholar]
- 7.Körperich H, Barth P, Gieseke J, et al. Impact of respiration on stroke volumes in paediatric controls and in patients after Fontan procedure assessed by MR real-time phase-velocity mapping. Eur Heart J Cardiovasc Imaging 2015;16(2):198–209. [DOI] [PubMed] [Google Scholar]
- 8.Wei Z, Whitehead KK, Khiabani RH, et al. Respiratory effects on Fontan circulation during rest and exercise using real-time cardiac magnetic resonance imaging. Ann Thorac Surg 2016;101(5):1818–1825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Furst B. Fontan physiology revisited. Anesth Analg 2016;122(2):578–579. [DOI] [PubMed] [Google Scholar]
- 10.Tree M, Wei ZA, Trusty PM, et al. Using a novel in vitro Fontan model and condition-specific real-time MRI data to examine hemodynamic effects of respiration and exercise. Ann Biomed Eng 2018;46(1):135–147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Markl M, Geiger J, Kilner PJ, et al. Time-resolved three-dimensional magnetic resonance velocity mapping of cardiovascular flow paths in volunteers and patients with Fontan circulation. Eur J Cardiothorac Surg 2011;39(2):206–212. [DOI] [PubMed] [Google Scholar]
- 12.Sundareswaran KS, Haggerty CM, de Zélicourt D, et al. Visualization of flow structures in Fontan patients using 3-dimensional phase contrast magnetic resonance imaging. J Thorac Cardiovasc Surg 2012;143(5):1108–1116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Frydrychowicz A, Arnold R, Harloff A, et al. Images in cardiovascular medicine. In vivo 3-dimensional flow connectivity mapping after extracardiac total cavopulmonary connection. Circulation 2008;118(2):e16–e17. [DOI] [PubMed] [Google Scholar]
- 14.Jarvis K, Schnell S, Barker AJ, et al. Evaluation of blood flow distribution asymmetry and vascular geometry in patients with Fontan circulation using 4-D flow MRI. Pediatr Radiol 2016;46(11):1507–1519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Valverde I, Nordmeyer S, Uribe S, et al. Systemic-to-pulmonary collateral flow in patients with palliated univentricular heart physiology: measurement using cardiovascular magnetic resonance 4D velocity acquisition. J Cardiovasc Magn Reson 2012;14(1):25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Valverde I, Rachel C, Kuehne T, Beerbaum P. Comprehensive four-dimensional phase-contrast flow assessment in hemi-Fontan circulation: systemic-to-pulmonary collateral flow quantification. Cardiol Young 2011;21(1):116–119. [DOI] [PubMed] [Google Scholar]
- 17.Bächler P, Valverde I, Pinochet N, et al. Caval blood flow distribution in patients with Fontan circulation: quantification by using particle traces from 4D flow MR imaging. Radiology 2013;267(1):67–75. [DOI] [PubMed] [Google Scholar]
- 18.Dasi LP, Whitehead K, Pekkan K, et al. Pulmonary hepatic flow distribution in total cavopulmonary connections: extracardiac versus intracardiac. J Thorac Cardiovasc Surg 2011;141(1):207–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.McElhinney DB, Marx GR, Marshall AC, Mayer JE, Del Nido PJ. Cavopulmonary pathway modification in patients with heterotaxy and newly diagnosed or persistent pulmonary arteriovenous malformations after a modified Fontan operation. J Thorac Cardiovasc Surg 2011;141(6):1362–70.e1, e1361. [DOI] [PubMed] [Google Scholar]
- 20.Vettukattil JJ. Is the hepatic factor a miRNA that maintains the integrity of pulmonary microvasculature by inhibiting the vascular endothelial growth factor? Curr Cardiol Rev 2017;13(3):244–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Claessen G, Claus P, Delcroix M, Bogaert J, La Gerche A, Heidbuchel H. Interaction between respiration and right versus left ventricular volumes at rest and during exercise: a real-time cardiac magnetic resonance study. Am J Physiol Heart Circ Physiol 2014;306(6):H816–H824. [DOI] [PubMed] [Google Scholar]
- 22.Rutkowski DR, Barton G, François CJ, Bartlett HL, Anagnostopoulos PV, Roldán-Alzate A. Analysis of cavopulmonary and cardiac flow characteristics in fontan Patients: Comparison with healthy volunteers. J Magn Reson Imaging 2019;49(6):1786–1799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Körperich H, Müller K, Barth P, et al. Differentiation of impaired from preserved hemodynamics in patients with Fontan circulation using real-time phase-velocity cardiovascular magnetic resonance. J Thorac Imaging 2017;32(3):159–168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bastkowski R, Weiss K, Maintz D, Giese D. Self-gated golden-angle spiral 4D flow MRI. Magn Reson Med 2018;80(3):904–913. [DOI] [PubMed] [Google Scholar]
- 25.Pruessmann KP, Weiger M, Börnert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med 2001;46(4):638–651. [DOI] [PubMed] [Google Scholar]
- 26.Bernstein MA, Zhou XJ, Polzin JA, et al. Concomitant gradient terms in phase contrast MR: analysis and correction. Magn Reson Med 1998;39(2):300–308. [DOI] [PubMed] [Google Scholar]
- 27.Walker PG, Cranney GB, Scheidegger MB, Waseleski G, Pohost GM, Yoganathan AP. Semiautomated method for noise reduction and background phase error correction in MR phase velocity data. J Magn Reson Imaging 1993;3(3):521–530. [DOI] [PubMed] [Google Scholar]
- 28.Gaeta S, Dyverfeldt P, Eriksson J, Carlhäll CJ, Ebbers T, Bolger AF. Fixed volume particle trace emission for the analysis of left atrial blood flow using 4D Flow MRI. Magn Reson Imaging 2018;47:83–88. [DOI] [PubMed] [Google Scholar]
- 29.Fogel MA, Weinberg PM, Rychik J, et al. Caval contribution to flow in the branch pulmonary arteries of Fontan patients with a novel application of magnetic resonance presaturation pulse. Circulation 1999;99(9):1215–1221. [DOI] [PubMed] [Google Scholar]
- 30.Lustig M, Donoho D, Pauly JM. Sparse MRI: The application of compressed sensing for rapid MR imaging. Magn Reson Med 2007;58(6):1182–1195. [DOI] [PubMed] [Google Scholar]
- 31.Knobloch V, Boesiger P, Kozerke S. Sparsity transform k-t principal component analysis for accelerating cine three-dimensional flow measurements. Magn Reson Med 2013;70(1):53–63. [DOI] [PubMed] [Google Scholar]
- 32.Jung B, Stalder AF, Bauer S, Markl M. On the undersampling strategies to accelerate time-resolved 3D imaging using k-t-GRAPPA. Magn Reson Med 2011;66(4):966–975. [DOI] [PubMed] [Google Scholar]
- 33.Bogren HG, Buonocore MH. Helical-shaped streamlines do not represent helical flow. Radiology 2010;257(3):895–896; author reply 896. [DOI] [PubMed] [Google Scholar]
- 34.Jarvis K, Schnell S, Barker AJ, et al. Caval to pulmonary 3D flow distribution in patients with Fontan circulation and impact of potential 4D flow MRI error sources. Magn Reson Med 2019;81(2):1205–1218. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.