Optimum fuzzy filters for phase-contrast magnetic resonance imaging segmentation

PURPOSE

To develop and validate a multidimensional segmentation and filtering methodology for accurate blood flow velocity field reconstruction from phase-contrast magnetic resonance imaging (PC MRI).

MATERIALS AND METHODS

The proposed technique consists of two steps: (1) the boundary of the vessel is automatically segmented using the active contour approach; and (2) the noise embedded within the segmented vector field is selectively removed using a novel fuzzy adaptive vector median filtering (FAVMF) technique. This two-step segmentation process was tested and validated on 111 synthetically generated PC MRI slices and on 10 patients with congenital heart disease.

RESULTS

The active contour technique was effective for segmenting blood vessels having a sensitivity and specificity of 93.1% and 92.1% using manual segmentation as a reference standard. FAVMF was the superior technique in filtering out noise vectors, when compared with other commonly used filters in PC MRI (P < 0.05). The peak wall shear rate calculated from the PC MRI data (248 +/- 39 sec(-1)), was significantly decreased to (146 +/- 26 sec(-1)) after the filtering process.

CONCLUSION

The proposed two-step segmentation and filtering methodology is more accurate compared to a single-step segmentation process for post-processing of PC MRI data.

Hemodynamic performance of stage-2 univentricular reconstruction: Glenn vs. hemi-Fontan templates

Flow structures, hemodynamics and the hydrodynamic surgical pathway resistances of the final stage functional single ventricle reconstruction, namely the total cavopulmonary connection (TCPC) anatomy, have been investigated extensively. However, the second stage surgical anatomy (i.e., bi-directional Glenn or hemi-Fontan template) has received little attention. We thus initiated a multi-faceted study, involving magnetic resonance imaging (MRI), phase contrast MRI, computational and experimental fluid dynamics methodologies, focused on the second stage of the procedure. Twenty three-dimensional computer and rapid prototype models of 2nd stage TCPC anatomies were created, including idealized parametric geometries (n = 6), patient-specific anatomies (n = 7), and their virtual surgery variant (n = 7). Results in patient-specific and idealized models showed that the Glenn connection template is hemodynamically more efficient with (83% p = 0.08 in patient-specific models and 66% in idealized models) lower power losses compared to hemi-Fontan template, respectively, due to its direct end-to-side anastomosis. Among the several secondary surgical geometrical features, stenosis at the SVC anastomosis or in pulmonary branches was found to be the most critical parameter in increasing the power loss. The pouch size and flare shape were found to be less significant. Compared to the third stage surgery the hydrodynamic resistance of the 2nd stage is considerably lower (both in idealized models and in anatomical models at MRI resting conditions) for both hemi- and Glenn templates. These results can impact the surgical design and planning of the staged TCPC reconstruction.

The total cavopulmonary connection resistance: a significant impact on single ventricle hemodynamics at rest and exercise

Little is known about the impact of the total cavopulmonary connection (TCPC) on resting and exercise hemodynamics in a single ventricle (SV) circulation. The aim of this study was to elucidate this mechanism using a lumped parameter model of the SV circulation. Pulmonary vascular resistance (1.96+/-0.80 WU) and systemic vascular resistances (18.4+/-7.2 WU) were obtained from catheterization data on 40 patients with a TCPC. TCPC resistances (0.39+/-0.26 WU) were established using computational fluid dynamic simulations conducted on anatomically accurate three-dimensional models reconstructed from MRI (n=16). These parameters were used in a lumped parameter model of the SV circulation to investigate the impact of TCPC resistance on SV hemodynamics under resting and exercise conditions. A biventricular model was used for comparison. For a biventricular circulation, the cardiac output (CO) dependence on TCPC resistance was negligible (sensitivity=-0.064 l.min(-1).WU(-1)) but not for the SV circulation (sensitivity=-0.88 l.min(-1).WU(-1)). The capacity to increase CO with heart rate was also severely reduced for the SV. At a simulated heart rate of 150 beats/min, the SV patient with the highest resistance (1.08 WU) had a significantly lower increase in CO (20.5%) compared with the SV patient with the lowest resistance (50%) and normal circulation (119%). This was due to the increased afterload (+35%) and decreased preload (-12%) associated with the SV circulation. In conclusion, TCPC resistance has a significant impact on resting hemodynamics and the exercise capacity of patients with a SV physiology.

In vitro hemodynamic investigation of the embryonic aortic arch at late gestation

This study focuses on the dynamic flow through the fetal aortic arch driven by the concurrent action of right and left ventricles. We created a parametric pulsatile computational fluid dynamics (CFD) model of the fetal aortic junction with physiologic vessel geometries. To gain a better biophysical understanding, an in vitro experimental fetal flow loop for flow visualization was constructed for identical CFD conditions. CFD and in vitro experimental results were comparable. Swirling flow during the acceleration phase of the cardiac cycle and unidirectional flow following mid-deceleration phase were observed in pulmonary arteries (PA), head-neck vessels, and descending aorta. Right-to-left (oxygenated) blood flowed through the ductus arteriosus (DA) posterior relative to the antegrade left ventricular outflow tract (LVOT) stream and resembled jet flow. LVOT and right ventricular outflow tract flow mixing had not completed until approximately 3.5 descending aorta diameters downstream of the DA insertion into the aortic arch. Normal arch model flow patterns were then compared to flow patterns of four common congenital heart malformations that include aortic arch anomalies. Weak oscillatory reversing flow through the DA junction was observed only for the Tetralogy of Fallot configuration. PA and hypoplastic left heart syndrome configurations demonstrated complex, abnormal flow patterns in the PAs and head-neck vessels. Aortic coarctation resulted in large-scale recirculating flow in the aortic arch proximal to the DA. Intravascular flow patterns spatially correlated with abnormal vascular structures consistent with the paradigm that abnormal intravascular flow patterns associated with congenital heart disease influence vascular growth and function.

Right aortic arch and coarctation: a rare association

OBJECTIVES

Understand anatomical and clinical correlatives to coarctation in right aortic arch.

BACKGROUND

Coarctation of the aorta is rare in patients with a functional right aortic arch. We reviewed a single institutional experience, examining associated diagnoses, diagnostic methodology, and surgical approaches.

METHODS

A retrospective study was performed of our echocardiographic, magnetic resonance imaging, catheterization, and surgical databases from 1988 to 2001.

RESULTS

Of 240 patients with right aortic arch, 10 (4.1%) had coarctation, constituting 1.9% of all native coarctations (n = 524). Nine (90%) had long-segment hypoplasia. Six (60%) had an aberrant left subclavian artery or retroesophageal diverticulum, 3 (30%) had mirror image branching, and 1 (10%) had a double arch with an atretic left arch. Other congenital heart defects were seen in 6 (60%) comprising 3 with ventricular septal defects, and one each with double-outlet right ventricle, cor triatriatum, and pulmonary valve abnormality. No patients with long-segment hypoplasia had bicuspid aortic valve. Six (60%) had vascular rings, and 5 (50%) had other associated syndromes. Magnetic resonance imaging and/or echocardiography successfully diagnosed all of these patients. Although long-segment right aortic arch coarctation courses behind the trachea posteriorly, only 2 needed an extra-anatomic (jump) graft; the remainders were repaired with patch angioplasty.

CONCLUSION

Coarctation with right aortic arch is rare, constituting 4.1% of all patients with right aortic arch, compared with 5-8% of patients with left aortic arch and congenital heart disease. Nearly all had long-segment hypoplasia without bicuspid aortic valve, and half were part of other syndrome complexes. This association can be diagnosed noninvasively and can often be repaired by patch angioplasty.

Deep sedation for cardiac magnetic resonance imaging: a comparison with cardiac anesthesia

OBJECTIVE

To test the hypothesis that safety, efficacy, and image quality in pediatric patients who undergo deep sedation for cardiac magnetic resonance imaging (CMR) for congenital heart disease (CHD) is similar to general anesthesia (GA).

STUDY DESIGN

Retrospective review of all CMR records from 1997-2006. Six hundred sixty patients underwent deep sedation (DS) and 161 underwent GA. Statistics included analysis of variance, chi(2) analysis, and the coefficient of variation.

RESULTS

Diagnoses included a broad spectrum of CHD. No serious adverse events (AE) including mortalities and hospitalizations occurred because of DS. There were 18 (2.8%) self-limited events in the DS group with a success rate of 97.9%. The AE rate for patients undergoing GA was 3.9% (n = 6), including 2 overnight hospital stays with a success rate of 100%. AE and success rates did not differ between the 2 groups. Observers blinded to the patient's group found no difference in image quality. Even in young infants, excellent image quality was obtained.

CONCLUSIONS

Sedation of appropriately screened pediatric patients with CHD undergoing CMR is safe and well tolerated and yields high-quality images similar to GA. GA should be considered for patients with CHD with hemodynamic or airway compromise, in whom sedation has failed, or who have special circumstances.

Anatomically realistic patient-specific surgical planning of complex congenital heart defects using MRI and CFD

Single ventricle congenital heart defects, which are characterized by cyanotic mixing between the oxygenated and de-oxygenated blood, afflict 2 per every 1000 live births. These defects are surgically treated by connecting the superior and inferior vena cava to the pulmonary arteries. However, such a configuration (also known as the total cavopulmonary connection), results in high energy losses and therefore the optimization of this connection prior to the surgery could significantly improve post-operative performance. In this paper, a surgical planning framework is proposed. It is exemplified on a patient with pre and post surgical MRI data. A pediatric surgeon performed a "virtual surgery" on the reconstruction of the patient's anatomy prior to the actual surgery. Post-operative hemodynamics in the virtually designed post-surgical anatomy and in the actual one are computed using computational fluid dynamics and compared to each other. This framework provides the surgeon to envision numerous scenarios of possible surgical options, and accordingly predict the post operative hemodynamics.

Multi-center transferability of a breath-hold T2 technique for myocardial iron assessment

BACKGROUND

Cardiac iron overload is the leading cause of death in thalassemia major and is usually assessed using myocardial T2* measurements. Recently a cardiovascular magnetic resonance (CMR) breath-hold T2 sequence has been developed as a possible alternative. This cardiac T2 technique has good interstudy reproducibility, but its transferability to different centres has not yet been investigated.

METHODS AND RESULTS

The breath-hold black blood spin echo T2 sequence was installed and validated on 1.5T Siemens MR scanners at 4 different centres across the world. Using this sequence, 5-10 thalassemia patients from each centre were scanned twice locally within a week for local interstudy reproducibility (n = 34) and all were rescanned within one month at the standardization centre in London (intersite reproducibility). The local interstudy reproducibility (coefficient of variance) and mean difference were 4.4% and -0.06 ms. The intersite reproducibility and mean difference between scanners were 5.2% and -0.07 ms.

CONCLUSION

The breath-hold myocardial T2 technique is transferable between Siemens scanners with good intersite and local interstudy reproducibility. This technique may have value in the diagnosis and management of patients with iron overload conditions such as thalassemia.

Nonlinear power loss during exercise in single-ventricle patients after the Fontan: insights from computational fluid dynamics

BACKGROUND

We previously demonstrated that power loss (PL) through the total cavopulmonary connection (TCPC) in single-ventricle patients undergoing Fontan can be calculated by computational fluid dynamic analysis using 3-dimensional MRI anatomic reconstructions. PL through the TCPC may play a role in single-ventricle physiology and is a function of cardiac output. We hypothesized that PL through the TCPC increases significantly under exercise flow conditions.

METHODS AND RESULTS

MRI data of 10 patients with a TCPC were analyzed to obtain 3-dimensional geometry and flow rates through the superior vena cava, inferior vena cava, left pulmonary artery, and right pulmonary artery. Steady computational fluid dynamic simulations were performed at baseline conditions using MRI-derived flows. Simulated exercise conditions of twice (2x) and three times (3x) baseline flow were performed by increasing inferior vena cava flow. PL, head loss, and effective resistance through the TCPC were calculated for each condition. Each condition was repeated at left pulmonary artery/right pulmonary artery ratios of 30/70 and 70/30 to determine the effects of pulmonary flow splits on exercise PL. For each patient, PL increases dramatically in a nonlinear fashion with increasing cardiac output, even when normalized to calculate head loss or resistance. Flow splits had a significant effect on PL at exercise, with most geometries favoring right pulmonary artery flow.

CONCLUSIONS

The relationship between cardiac output and PL is nonlinear and highly dependent on TCPC geometry and pulmonary flow splits. This study demonstrates the importance of studying the TCPC under exercise conditions, because baseline conditions may not adequately characterize TCPC efficiency.

Nonlinear power loss during exercise in single-ventricle patients after the Fontan: insights from computational fluid dynamics

BACKGROUND

We previously demonstrated that power loss (PL) through the total cavopulmonary connection (TCPC) in single-ventricle patients undergoing Fontan can be calculated by computational fluid dynamic analysis using 3-dimensional MRI anatomic reconstructions. PL through the TCPC may play a role in single-ventricle physiology and is a function of cardiac output. We hypothesized that PL through the TCPC increases significantly under exercise flow conditions.

METHODS AND RESULTS

MRI data of 10 patients with a TCPC were analyzed to obtain 3-dimensional geometry and flow rates through the superior vena cava, inferior vena cava, left pulmonary artery, and right pulmonary artery. Steady computational fluid dynamic simulations were performed at baseline conditions using MRI-derived flows. Simulated exercise conditions of twice (2x) and three times (3x) baseline flow were performed by increasing inferior vena cava flow. PL, head loss, and effective resistance through the TCPC were calculated for each condition. Each condition was repeated at left pulmonary artery/right pulmonary artery ratios of 30/70 and 70/30 to determine the effects of pulmonary flow splits on exercise PL. For each patient, PL increases dramatically in a nonlinear fashion with increasing cardiac output, even when normalized to calculate head loss or resistance. Flow splits had a significant effect on PL at exercise, with most geometries favoring right pulmonary artery flow.

CONCLUSIONS

The relationship between cardiac output and PL is nonlinear and highly dependent on TCPC geometry and pulmonary flow splits. This study demonstrates the importance of studying the TCPC under exercise conditions, because baseline conditions may not adequately characterize TCPC efficiency.

Delayed-enhancement cardiovascular magnetic resonance identifies fibrous tissue in children after surgery for congenital heart disease

OBJECTIVES

To determine whether delayed-enhancement magnetic resonance imaging can identify fibrous tissue associated with patch reconstructions in postoperative patients with congenital heart disease. Additionally, to determine whether fibrous structures not directly related to the surgical reconstruction exhibited delayed enhancement.

METHODS

Seventy-three patients underwent magnetic resonance imaging. Studies were retrospectively reviewed for the presence of delayed enhancement along the ventricular outflow tracts, cardiac valves, and where available, the ascending aorta. Three groups were identified. Group A patients (n = 34) underwent right ventricular outflow tract reconstruction and ventricular septal defect patch closure. Group B patients (n = 33) had never undergone cardiac surgery. Group C patients (n = 6) had functional single ventricle and underwent Norwood reconstruction.

RESULTS

In group A, 31 of 34 patients had delayed enhancement of the right ventricular outflow tract, and 14 of 34 had delayed enhancement of the ventricular septal defect patch (P < .001). In group B (n = 33), 1 patient with arrhythmogenic right ventricular dysplasia had delayed enhancement limited to the right ventricular outflow tract. The remainder had no delayed enhancement of either outflow tract. Delayed enhancement of the aortic valve and ascending aorta was observed in 13 of 34 (P = .002) and 10 of 26 (P = .05) group A patients, respectively, compared with 2 of 33 and 3 of 24 group B patients. In group C, delayed enhancement of the Norwood reconstruction was observed in 5 of 6 patients (P = .002).

CONCLUSIONS

Delayed-enhancement imaging detects fibrous tissue along regions of reconstruction in patients who have had surgery for congenital heart disease. Furthermore, delayed-enhancement imaging detects fibrous tissue in regions not directly related to the reconstructive surgery, including cardiac valves and the wall of the ascending aorta.

Quantifying regional right ventricular function in tetralogy of Fallot

Right ventricular (RV) function is notoriously difficult to quantify. Patients with tetralogy of Fallot (TOF) have decreased systolic performance. We measure regional RV performance using MRI with 1-dimensional myocardial tissue tagging. By tagging cine-MRI in two views, we measured regional shortening in 12 regions throughout the RV. We image 32 pediatric patients: 21 normal patients and 11 patients with repaired TOF. We establish a normal range for each RV region. TOF patients have decreased shortening on a region-by-region basis. We conclude that regional RV performance can be measured using this technique, and that decreased performance can be demonstrated in TOF patients.

Usefulness of magnetic resonance imaging for the diagnosis of right ventricular dysplasia in children

Cardiac magnetic resonance (CMR) has been helpful in adults in the diagnosis of arrhythmogenic right ventricular dysplasia. Short of direct surgical observation or autopsy, no gold standard exists. CMR diagnostic criteria include right atrial and ventricular dilation, regional right ventricular (RV) wall motion abnormalities, outflow tract ectasia, and myocardial fatty infiltration. To determine whether adult diagnostic criteria are useful in children referred for CMR for this diagnosis, the images and records of 81 patients (aged 11.5 +/- 5.5 years) over an 8-year period were reviewed. Histories included ventricular tachycardia, palpitations, dilated right ventricle, syncope, near sudden death, or family history of RV dysplasia. Four families were studied with parents who had RV dysplasia diagnosed by surgery, explanted heart, or CMR. CMR imaging included T1-weighted imaging, cine, 1-dimensional RV myocardial tagging, and phase-encoded velocity mapping, and 2 patients underwent delayed-enhancement CMR. Only 1 of the 81 patients met 5 of the criteria. None of the others met >2 of the criteria, and only 2 patients met 1 or 2 criteria. For questionable regional wall motion abnormalities, RV myocardial tagging was helpful. In conclusion, CMR of patients with a history suspicious for the diagnosis of RV dysplasia is a low-yield test in children. This may be due to the evolving nature of the disease, which does not manifest itself from a morphologic or ventricular-function standpoint until later in development. Follow-up studies as patients age may be advantageous.

Cardiac magnetic resonance imaging for accurate diagnosis of aortic arch anomalies in patients with 22q11.2 deletion

This study demonstrates, in patients with 22q11.2 deletion, the wide variety of complex aortic arch anomalies that can be accurately defined by cardiac magnetic resonance imaging (MRI) compared with echocardiography. From May 1999 to August 2003, 17 of 68 patients (25%) with 22q11.2 deletion referred for cardiac evaluation required cardiac MRI, after echocardiography, to clarify aortic arch anatomy. Images of cardiac anatomy were obtained using steady-state free precession, half-Fourier acquired single turbo spin-echo dark blood sequences and gadolinium for 3-dimensional reconstruction. All MRI findings were abnormal, and in 16 of 17 cases, echocardiography was unable to define aortic arch anomalies correctly compared with MRI.

Total Cavopulmonary Connection Flow With Functional Left Pulmonary Artery Stenosis

BACKGROUND

In our multicenter study of the total cavopulmonary connection (TCPC), a cohort of patients with long-segment left pulmonary artery (LPA) stenosis was observed (35%). The clinically recognized detrimental effects of LPA stenosis motivated a computational fluid dynamic simulation study within 3-dimensional patient-specific and idealized TCPC pathways. The goal of this study was to quantify and evaluate the hemodynamic impact of LPA stenosis and to judge interventional strategies aimed at treating it.

METHODS AND RESULTS

Simulations were conducted at equal vascular lung resistance, modeling both discrete stenosis (DS) and diffuse long-segment hypoplasia with varying degrees of obstruction (0% to 80%). Models having fenestrations of 2 to 6 mm and atrium pressures of 4 to 14 mm Hg were explored. A patient-specific, extracardiac TCPC with 85% DS was studied in its original configuration and after virtual surgery that dilated the LPA to 0% stenosis in the computer medium. Performance indices improved exponentially (R2>0.99) with decreasing obstruction. Diffuse long-segment hypoplasia was approximately 50% more severe with regard to lung perfusion and cardiac energy loss than DS. Virtual angioplasty performed on the 3-dimensional Fontan anatomy exhibiting an 85% DS stenosis produced a 61% increase in left lung perfusion and a 50% decrease in cardiac energy dissipation. After 4-mm fenestration, TCPC baffle pressure dropped by approximately 10% and left lung perfusion decreased by approximately 8% compared with the 80% DS case.

CONCLUSIONS

DS <60% and diffuse long-segment hypoplasia <40% could be considered tolerable because both resulted in only a 12% decrease in left lung perfusion. In contrast to angioplasty, a fenestration (right-to-left shunt) reduced TCPC pressure at the cost of decreased left and right lung perfusion. These results suggest that pre-Fontan computational fluid dynamic simulation may be valuable for determining both the hemodynamic significance of LPA stenosis and the potential benefits of intervention.

Preliminary Investigations into a New Method of Functional Assessment of the Fetal Heart Using a Novel Application of ‘Real-Time’ Cardiac Magnetic Resonance Imaging

OBJECTIVES

Because of quantitative echocardiographic limitations of fetal ventricular volumes as well as poor windows, we sought to determine if real-time magnetic resonance imaging (MRI) could be used.

METHODS

Real-time, functional, true fast imaging with steady-state precession, cardiac MRI was performed on 2 fetuses (one with hypoplastic left heart syndrome and one with ductal constriction). Fetal echocardiography was performed and cardiac index by Doppler was used to validate volume measures by MRI.

RESULTS

This technique was able to visualize the beating heart and assess ventricular volumes. Cardiac index and assessment of right ventricular hypertrophy and dilation by echocardiography were consistent with the ventricular volumes and right ventricular hypertrophy obtained by cardiac MRI.

CONCLUSION

Real-time, functional fetal cardiac MRI is possible and can be used to quantitatively assess ventricular volumes and cardiac index in utero.