Role of Magnetic Resonance Angiography in the Diagnosis of Major Aortopulmonary Collateral Arteries and Partial Anomalous Pulmonary Venous Drainage
Background— Accurate diagnosis of major aortopulmonary collaterals (MAPCAs) and partial anomalous pulmonary venous drainage (PAPVD) in adult patients with congenital heart disease is important but problematic. Three-dimensional contrast-enhanced magnetic resonance angiography (MRA) provides a minimally invasive technique to allow detailed studies in a single breath-hold.
Methods and Results— We assessed the role of contrast-enhanced 3D MRA in 29 consecutive adult patients with a diagnosis of MAPCAs (n=16) or PAPVD (n=13) made by echocardiogram, cardiac catheterization, or surgical inspection. MRA was performed with a 3D spoiled gradient-echo technique with intravenous gadolinium-DTPA (0.2 mmol/kg). In both types of pathology, there was excellent correlation between MRA and the cardiac catheterization, echocardiogram, or surgical inspection. Additional information was gained for patients with MAPCAs on confluence and size of pulmonary arteries (n=13 had central arteries), pulmonary artery stenosis (n=3), aneurysmal dilatation of pulmonary artery (n=1), and additional anomalous vascular abnormality (n=3). Shunt assessment, where present (9 of 16), showed patency in all cases (100%). For adults with PAPVD, further information was obtained on drainage origin (n=11). There were no complications.
Conclusions— Contrast-enhanced 3D MRA provides a fast, noninvasive, radiation-free method of accurate and comprehensive diagnosis of MAPCAs and PAPVD in adult patients.
Received June 3, 2003; revision received September 19, 2003; accepted September 30, 2003.
Cardiovascular magnetic resonance (CMR) has an important role in congenital heart disease (CHD), but the diagnosis of vascular anomalies in CHD remains problematic, and the clinical presentation can be varied.1 Accurate diagnosis is important for both optimal management and prognostic evaluation.2,3 The utility of CMR in the detection of extracardiac venous and arterial vascular anomalies, such as partial anomalous venous drainage (PAPVD) and major aortopulmonary collaterals (MAPCAs), has been hampered by poor resolution, lack of real-time data, long study times, and lack of 3D reconstruction.4 Recent advances in CMR permit rapid high-resolution imaging of vascular abnormalities.5–9 This improvement in the speed and quality of image acquisition has been facilitated particularly through improved pulse sequence and gradient design, shorter 3D breath-holding sequences, and particularly the use of gadolinium contrast agents for angiography. The introduction of segmented k-space techniques has allowed improved shunt detection by enabling slice-selective dynamic cine imaging within a single breath-hold.10
We evaluated the capability of contrast-enhanced magnetic resonance angiography (MRA) to define the anatomy of the main pulmonary arteries, MAPCA vessels, and pulmonary venous return in adult patients with suspected or known CHD.11 In addition, we sought to determine whether additional information could be gained compared with cardiac catheterization and echocardiographic data.
We prospectively studied 29 consecutive adult patients aged more than 16 years with an established or suspected diagnosis of MAPCAs or PAPVD or an intracardiac shunt who underwent conventional CMR and MRA studies and in whom additional diagnostic data were available for comparison. Patients were referred between March 2000 and July 2002. The age range was 16 to 75 (mean 45) years. These diagnoses were made by echocardiography (transthoracic echocardiography [TTE] or transesophageal echocardiography [TEE]), cardiac catheterization, or surgical confirmation. Ethics approval was obtained for the study, and patients gave informed consent.
CMR and MR Angiography
CMR was undertaken with a Siemens Sonata 1.5-T scanner with 3D fast, low-angle shot (FLASH) gradient-echo sequences with radiofrequency spoiling and a phased-array surface coil. Multislice spin-echo images (half Fourier acquisition single-shot turbo spin-echo [HASTE]) were obtained in 3 orthogonal planes to define the cardiac anatomy and to guide the angiography acquisition. Through-plane cine phase-contrast sequences in the aorta and pulmonary artery were taken to determine the pulmonary:systemic blood flow ratio.8 MRA was acquired in a coronal orientation during a breath-hold in inspiration, before and after intravenous gadolinium-DTPA (Magnevist; Schering). Parameters used were echo time 1.04 ms, repetition time 2.6 ms, flip angle 25°, matrix size 512×256, slice thickness 1.5 mm, receiver bandwidth 700 Hz per pixel, and field of view 30 to 40 cm. Acquisition times were between 14 and 32 seconds (mean 17 seconds) with 1 signal average. The imaging slab of 60 to 80 mm was partitioned into 40 segments. K-space filling was sequential, with the peak gadolinium concentration in left atrium (for PAPVD) or aorta (for MAPCAs) coinciding with sampling at the center of k-space. No patients required sedation or general anesthesia.
Administration of Contrast
Gadolinium-DTPA (0.2 mmol/kg) was given with a power injector pump at 2 mL/s. An initial test bolus (2 mL) was given to determine the time to peak contrast concentration and to identify late-filling structures. Two sequential MRA acquisitions (early and late contrast scans) were undertaken with a 12-second interval.
Image Analysis and Postacquisition Data Processing
Images were analyzed and processed by a single blinded experienced investigator using 3D reformatting, multiplanar reformation, and maximum-intensity projection techniques (Figures 1 and 2⇓). Processing typically took fewer than 15 minutes. Pulmonary veins were readily distinguished from pulmonary arteries by the more horizontal course of the latter. For MAPCAs, the total number, site of origin and drainage, maximum wall dimensions, and presence of stenoses were determined. The maximum dimension represented the maximal internal diameter of each individual MAPCA, usually at its origin. The anatomy of the pulmonary arteries was determined, including their confluence, absence, and whether or not they were hypoplastic.
We used a scoring system to assess the diagnostic yield of CMR/MRA in each patient compared with alternative diagnostic methods used.12 The following classification was used: 1, confirming the suspected diagnosis plus yielding additional clinically relevant information; 2, confirmation of a previously suspected diagnosis with no additional clinical information gained; and 3, yielding a new diagnosis.
MRA diagnoses were compared with the available diagnostic information, and the agreement (percent) was calculated. Continuous data are reported as mean±SD.
Major Aortopulmonary Collaterals
Of the 29 patients, 16 had evidence of MAPCAs. All scans were undertaken without complications. The mean age range for MAPCA patients was 31.8 (range 20 to 46) years, with a male:female ratio of 10:6. Demographic and anatomic data are presented in Tables 1 and 2⇓. In 15 (94%) of 16 patients with MAPCAs, the diagnosis had been made during the neonatal period, with more recent characterization of vessels angiographically.
Agreement between CMR and MRA and surgical or catheter techniques for detection of MAPCAs was 100% (Table 3). In 36% of cases, CMR/MRA helped both to confirm the diagnosis by catheter study and to provide clinically relevant additional information. The course, origin, and pulmonary connections of the MAPCAs were well demonstrated, with clear delineation of the spatial relationship of the MAPCAs to both adjacent vascular and nonvascular structures. In all cases, MRA confirmed the presence or absence of central pulmonary arteries (13 [81%] of 16 patients had central pulmonary arteries).
MAPCA vessels were classified either as individual vessels (≥2.5 mm diameter) or as multiple smaller vessels that tended to arise in a cluster (each vessel typically <2.5 mm diameter). The majority of MAPCA vessels arose from the descending aorta (63.6% of single MAPCAs and 66.7% of multiple smaller vessels). In most cases, it was possible to tell whether they supplied a single lobe or entire lung (30% drained into the right upper lobe, and 18.1% supplied the right pulmonary artery). Additional findings during the scan included aneurysmal dilatation of a MAPCA to the right pulmonary artery (1), hypoplastic pulmonary artery branch (2), coincident pulmonary venous stenosis (1), anomalous brachiocephalic vein (1), coronary artery to pulmonary artery fistula (1), and evaluation of stenoses of MAPCAs (none seen). In 25.5% of cases, there was a dual supply to parts of the lungs, with flow from both MAPCAs and central pulmonary arteries. Information was also gained on ventricular size, function, and postoperative status (including conduit stenosis and patency of shunts). In patients with MAPCAs, MRA accurately demonstrated the presence and patency of surgically constructed systemic-pulmonary shunts in all patients where present (9 of 16). No stenotic lesions or obstructed shunts were identified.
Partial Anomalous Pulmonary Venous Drainage
For patients with PAPVD, the mean age was 47.1 (range 16 to 75) years, with a male:female ratio of 6:7. A summary of the clinical characteristics and anatomic findings is shown in Tables 4 and 5⇓. Of the 13 patients, 4 had features of scimitar variant with a hypoplastic right lung. Accordingly, abnormalities of right-sided pulmonary venous drainage were most commonly seen. No patients with complete anomalous pulmonary drainage were seen in the study group. Drainage of the anomalous pulmonary veins was into the inferior vena cava (2), superior vena cava (5), or brachiocephalic vein (1) and via a common pulmonary vein bilaterally in 2 cases, both with atrial isomerism. In 2 patients, the site of drainage was unclear. The most common associated atrial septal defect was sinus venosus (3). In patients with right atrial isomerism (3), CMR/MRA helped to establish the diagnosis of PAPVD and elucidated the relationship of the pulmonary arteries to the bronchi. In 1 patient, MRA identified severe stenosis of a normally draining right upper pulmonary vein, probably after surgery for atrial septal defect closure. The pulmonary to systemic blood flow ratio for PAPVD patients ranged from 1.4 to 3.5 (mean 2.5±0.7).
In 3 patients (23%), MRA established a new diagnosis of PAPVD with clear delineation of pulmonary venous connections. In these cases, previous echocardiographic and catheter studies were either nondiagnostic or the results misinterpreted. Reasons for referral in this population were to exclude constrictive cardiomyopathy or to identify a suspected interatrial shunt. In 5 of 13 patients, the MRA data were considered diagnostic, and subsequent catheter study was deemed to be unnecessary.
This study assessed the value of 3D contrast-enhanced MRA in the detection of 2 common congenital vascular anomalies and uniquely in adult patients only. PAPVD and MAPCAs were chosen as the 2 specific disease states comprising the focus of the present study because our aim was to demonstrate the ability of MRA with conventional CMR to detect both extracardiac venous and arterial anomalies. Furthermore, diagnosis in adult patients by other imaging modalities is problematic. Early detection and treatment are associated with a better outcome but require accurate demonstration of the anatomic distribution of the anomalous vessels.13,14 This is also important for serial evaluation.
The results from this prospective study show that MRA is an accurate, fast, and robust method of diagnosing both pulmonary venous anomalies and MAPCAs in adult patients. There was complete correlation with catheter and/or surgical-based findings. In 3 patients (23%) with PAPVD, this was a new diagnosis, which demonstrates that catheter-based angiography and echocardiography will often not detect these anomalies when there is no clinical suspicion. Overall, MRA provided new, clinically important information in 50% of cases after catheter study in addition to confirming the diagnosis. Adult patients with CHD and particularly with vascular anomalies represent a diagnostic challenge because of limited and suboptimal echocardiographic windows, operator-dependent errors, and poor visualization of vascular structures outside the mediastinum with both TTE and TEE.15 The detection rate for TEE varies between 40% and 85%.16 In both anomalies, cardiac catheterization can be difficult, is invasive, and involves radiation exposure, and serial investigations can be problematic.17 Anomalous vessels may be undetected because of the need for selective and local injection. Sinus venosus defects associated with anomalous pulmonary venous drainage can be missed because of the difficulty in crossing such defects. CMR has a useful role to play in the evaluation of patients with vascular anomalies. Current steady-state free precession gradient cine and parallel processing techniques allow the rapid evaluation of cardiac function, volumes, and mass. In addition, together with fast spin echo with double inversion recovery sequences, anomalous veins, central pulmonary arteries, and MAPCAs can be imaged with good agreement with x-ray angiography. Phase-shift velocity mapping sequences can be used to quantify shunt sizes (eg, pulmonary to systemic flow in patients with anomalous pulmonary venous drainage). Spin-echo techniques can overcome image artifact due to endovascular stents and coils.
CMR without MRA does, however, have significant limitations. Both types of vascular anomalies have a wide region of origin. MAPCAs may arise, as seen in the present study, from the thoracic inlet to the renal arteries. The anatomic area to cover is thus large, resulting in long scan times for complete anatomic coverage and necessitating multiple breath-holds. This is challenging for patients with cardiorespiratory compromise. Furthermore, CMR without MRA has a limited ability to detect small vessels (eg, <2 mm) and those with slow flow or increased tortuosity.4
The role of MRA is complementary to these techniques. MRA has several advantages compared with other imaging modalities. It offers better resolution to detect stenoses and good signal-to-noise ratio, so that vessels down to 0.5 mm with slow flow and intraparenchymal pulmonary vessels can be visualized. It is not dependent on particular windows or the direction of branching.18 It is noninvasive and has no radiation exposure. The large field of view and rapid data acquisition in 1 breath-hold after a single injection into a peripheral vein allow a full delineation of the anatomy with the entire course of the anomalous vessels. In all cases, an initial test bolus was given to assess the optimal scan time. The passage of this bolus was monitored over approximately 40 seconds, during which the patient was free-breathing. This technique was designed to identify and correlate both early- and late-filling vascular structures, including collaterals. Findings were also correlated with baseline gradient-echo and spin-echo images. Offline reformatting allows for evaluation of tortuous blood vessels with clear spatial differentiation of vascular structures through interactive navigation of the 3D volume data, eg, pulmonary arteries from pulmonary veins, as well as definition of infracardiac and retrocardiac pulmonary venous drainage and nonvascular neighboring structures.
Several groups have reported their findings in detecting cardiac vascular anomalies using MRA, but in contrast to the present study, these were either in predominantly pediatric populations or were retrospective studies. Geva et al19 demonstrated the accuracy of MRA in the delineation of all sources of pulmonary blood supply in patients with complex pulmonary stenosis and atresia compared with diagnostic catheterization with x-ray angiography. Although some adult patients were included in their series, unlike the present study, the median age range was 4.7 years. In a retrospective study including 54 patients with anomalous pulmonary venous drainage, MRA was not only accurate but provided additional clinically important information to guide intervention in 74% of patients.12 In contrast to the present study, the age range was 1 day to 60 days. Ferrari et al8 verified the anatomic delineation of partial anomalous pulmonary venous drainage and atrial septal defects seen by MRA with subsequent surgical confirmation in a predominant adult population, which supports the findings of the present study. In our own clinical practice, we have now fully incorporated MRA for assessment of extracardiac structures in these 2 challenging patient groups. Availability of the MRA data guides us on management strategies and limits the duration and radiation exposure of subsequent cardiac catheterization and intervention.
Limitations of Technique
Although the sample size is small, with clear limitations due to patient selection, the confirmation of findings by cardiac catheterization and surgical inspection supports the accuracy of MRA. Not all patients underwent a cardiac catheterization study, because it was deemed unnecessary and unethical in cases in which echocardiographic imaging was considered sufficiently diagnostic in conjunction with the MRA findings. Reflecting the variation in presentation, it is inevitable that some patients will have undergone more baseline investigations or interventions, and this will influence the additional information gained from MRA.
Advances in the use of MRA in the detection of vascular anomalies are likely to come from increased use of real-time sequences in conjunction with fast-gradient cine MR, black-blood spin echo, phase-contrast velocity flow mapping for shunt calculations, and the use of signal averaging or navigator sequences when breathing is problematic. There is also the potential for CMR/MRA-guided intervention.
Contrast-enhanced 3D MRA is a safe, rapid, noninvasive, and robust method to detect vascular anomalies in adult patients with CHD including MAPCAs and PAPVD. It provides complementary information to that gained by echocardiography and catheterization on vascular anatomy and physiology.
This work was supported by the Coronary Artery Disease Research Association (CORDA), The Heart Charity, The Waring Trust, Royal Brompton Hospital, and the British Heart Foundation.
Pilleul F, Merchant N. MRI of the pulmonary veins: comparison between 3D MR angiography and T1-weighted spin echo. J Comput Assist Tomogr. 2000; 4: 683–687.
Reddy VM, McElhinney DB, Amin Z, et al. Early and intermediate outcomes after repair of pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries: experience with 85 patients. Circulation. 2000; 101: 1826–1832.
Redington AN, Somerville J. Stenting of aortopulmonary collaterals in complex pulmonary atresia. Circulation. 1996; 94: 2479–2484.
Pascoe RD, Oh JK, Warnes CA, et al. Diagnosis of sinus venosus atrial septal defect with transesophageal echocardiography. Circulation. 1996; 94: 1049–1055.
Geva T, Greil GF, Marshall AC, et al. Gadolinium-enhanced 3-dimensional magnetic resonance angiography of pulmonary blood supply in patients with complex pulmonary stenosis or atresia: comparison with x-ray angiography. Circulation. 2002; 106: 473–478.