| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, New York University Medical Center, Tisch Hospital,
IRM-236, 560 First Ave., New York, NY 10016.
2 Present address: Department of Medical Imaging, Hunterdon Medical Center, 2100
Wescott Dr., Flemington, NJ 08822.
3 Department of Pediatrics, New York University Medical Center, Pediatric
Cardiology, 530 First Ave., Ste. 9U, New York, NY 10016.
Received December 11, 2001;
accepted after revision October 1, 2003.
Address correspondence to K. J. Roche.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Eleven patients who underwent both MRI and
conventional angiography were retrospectively reviewed. Contiguous 2D cine
gradient-recalled echo images (TR range/TE, 30–80/4.8; flip angle,
20° or 30°) and 3D MR angiographic images (TR range/TE range,
3.8–5.0/1.3–2.0; acquisition time, 13–32 sec) using
gadopentetate meglumine (0.1–0.2 mmol/kg) were obtained. The presence,
size, and course of the pulmonary arteries (main, right, left) and major
aortopulmonary collateral arteries (
5 mm) were determined. Presence of
minor collateral arteries (< 5 mm) was also noted. Results were compared
with findings at conventional angiography.
RESULTS. MRI showed all main (n = 4) and branch (n = 17) pulmonary arteries found at conventional angiography and showed the pulmonary confluence in five of six cases. MRI showed all major aortic collaterals (n = 22) with a highly significant correlation between MRI and conventional angiography measurements (r = 0.84, p < 0.001 [95% confidence interval, –0.35 to 0.40]). One coronary artery collateral was not shown on MRI examination. At MRI, 12 of 14 major and four of seven minor brachiocephalic artery collaterals were shown. MRI showed more minor aortic collaterals than angiography (22 vs 18 vessels, respectively).
CONCLUSION. Combined cine gradient-recalled echo MRI and MR angiography is a reliable method for imaging pulmonary vascular supply in patients with these disorders. Additional prospective studies comparing MRI and conventional angiography may determine whether routine preoperative conventional angiography is required.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The surgical literature has emphasized the need to delineate the pulmonary circulation as completely as possible before operative intervention [4, 5]. The traditional method has been conventional angiography, with MRI examination limited to those patients in whom radiographic data are incomplete. In this study, cine gradient-recalled echo MRI with MR angiography was compared with conventional angiography in a series of patients with this diagnosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI and MR Angiography
All studies were performed on a 1.5-T MR imager (Magnetom Vision, Siemens)
using either a head or phased array body coil. Initial MRI performed for
localization and planning of subsequent series included three-plane scout and
two-plane (axial and coronal) HASTE sequences.
Next, a series of 2D cardiac-gated cine gradient-recalled echo images (TR range/TE, 30–80 [depending on heart rate]/4.8; flip angle, 20° or 30°; number of excitations, 2; mean acquisition time, 30 sec [range, 21–43 sec]) was performed. This series generated seven to 15 images per slice, divided over the cardiac cycle, that could be viewed individually or as a cine loop. A contiguous series of slices (slice thickness, 5 or 6 mm with a corresponding interval) was obtained in the axial plane from the thoracic inlet to the diaphragm and in the coronal plane from the anterior right ventricular wall through the descending thoracic aorta. The time for acquiring the cine gradient-echo images was approximately 20–30 min.
A 3D MR angiography (3D interpolated, spoiled gradient-recalled echo)
sequence (TR range/TE range 3.8–5.0/1.3–2.0; flip angle,
12–50° [mean 27°]; number of acquisitions, 1; mean acquisition
time, 20 sec [range, 9–28 sec]; number of measurements, 2 or 3;
partitions, 40–112 [mean, 68]; effective thickness, 0.9–2.8 mm
[mean, 1.9 mm]; matrix, 90–126 x 256 pixels) was then performed
using IV gadopentetate dimeglumine (Magnevist, Berlex). The dose of contrast
material was 0.2 mmol/kg in nine patients and 0.1 mmol/kg in two patients. The
contrast material was administered through a peripheral IV line using a hand
injection technique (
1–2 mL/sec) in seven patients. Power injection
(Spectris MR Injector, Medrad) was performed in four patients at a rate of 2
mL/sec with a 20-mL saline flush.
For MR angiography, an unenhanced series was first performed. Next, an axial timing run using 1 mL or 10% of the total gadolinium dose (whichever was smaller) was performed. The time to peak enhancement of the descending aorta was determined. The initiation of the angiographic sequence was timed using the following formula: initiation time (sec) = time to peak (sec) + injection time (sec) / 2 – scanning time (sec) / 2. This method ensured that the bolus peak was at or near the center lines of k-space. Two or three contrast-enhanced series were then performed in succession without a time gap. Subtraction images (contrast-enhanced minus unenhanced images) were obtained of the optimally enhanced series. The time for performing the MR angiography was 10–15 min. Off-line processing was performed on a Virtuoso workstation (Siemens). Images were reviewed using source data, multiplanar reconstruction, and 3D volume-rendered image display formats.
Image Analysis
Each MR image was initially reviewed independently by three examiners who
were unaware of the catheterization and clinical data. The examiners then
reviewed the studies together, with differences noted as interobserver
variability and resolved by consensus. For each case, the presence of a main,
right, and left pulmonary artery and the presence or absence of a pulmonary
artery confluence were recorded. The origin (aorta, brachiocephalic arteries
and branches, coronary arteries) of all major collateral arteries ( 5 mm)
was determined. The course of each vessel was traced, and the insertion into
the right or left lung was recorded. Minor collateral arteries (< 5 mm)
usually could not be traced individually to the lungs, although the origin and
total number of minor collateral arteries were also recorded.
At a second review, the size of each pulmonary artery and major aortopulmonary collateral artery was measured independently by the three reviewers. Because the vessel dimensions vary considerably in their course, each reviewer was instructed to measure the maximum lumen diameter.
Conventional Angiography Data
Angiography studies were reviewed after completing all MRI interpretations.
Digitized biplane cardiac angiograms were reviewed using an OptiView Digital
System (OptiMed Technologies). Analog films were reviewed with an SM 3500 Cine
Video system (Sony). A mean of eight injections (range, 4–17) was
performed including right ventriculogram and aortic root injections. Selective
injections were performed in large collaterals, coronary arteries, and
brachiocephalic arteries and branches as needed. One examiner reviewed all the
cardiac catheterization findings. The size of the vessels was measured using
the outer dimensions of the catheter wall as a reference standard.
The results of the MRI consensus interpretation and catheterization data were compared, and discrepancies were recorded. The time between review of MRI and catheterization data was at least 1 month.
Statistical Analysis
Statistical analysis was performed using SPSS version 9.0 (Statistical
Package for the Social Sciences). Interobserver variability for MRI reviewers
was determined using the kappa test. Sensitivity, specificity, and accuracy
were calculated for the MRI consensus interpretation of the main and branch
pulmonary arteries, pulmonary artery confluence, and major aortopulmonary
collateral arteries to each lung using conventional angiography as the gold
standard. Agreement between MRI consensus interpretation and catheterization
was analyzed using the paired Student's t test and the Bland-Altman
plot.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
MRI showed all 22 major aortic collaterals revealed on conventional angiography (sensitivity, 100%; specificity, 100%; accuracy, 100%). MRI was able to show collaterals directly supplying the lung and connecting to the branch pulmonary arteries (Fig. 2). Major aortic collaterals were distributed approximately evenly to the right (n = 10) and left (n = 12) lungs. The origin of 17 (77%) of 22 major aortic collaterals was the descending aorta (Fig. 3). MRI showed one large collateral in the smallest patient (4-kg neonate) arising from the aortic arch that connected to the pulmonary confluence (Fig. 4A, 4B).
|
|
|
|
Fewer major brachiocephalic artery collaterals were found on MRI than at catheterization. On MRI, a total of 12 major collaterals were found versus 14 on angiography (sensitivity, 80%; specificity, 100%; accuracy, 87%). In addition, a single 5-mm right coronary artery collateral to the right lung was not identified on MRI. On conventional angiography, this coronary artery was seen at aortic root injection and confirmed at a selective coronary artery injection.
A comparison of vessel measurements on MRI (mean vessel diameter, 0.76 cm;
range, 0.5–1.8 cm) and on conventional angiography is shown in
Figure 5. The mean difference
in vessel measurements between conventional angiography and MRI was 0.02 cm,
and the 95% confidence interval using the Bland-Altman plot was –0.353
to 0.402. Interobserver variability for MRI interpretation was low and evenly
distributed among the three reviewers (mean, 
= 0.90; range,
0.89–0.91).
|
More collaterals from the aorta were seen on MRI than on angiography (22 vs 18 vessels). Most of the minor collateral arteries arose from the descending aorta (20/22 vessels). Two small collaterals were seen from the aortic arch with MRI versus four with angiography. However, the small size and tortuous configuration of these vessels did not permit evaluation of the lung supplied. MRI was less effective at visualizing small collaterals from the right and left brachiocephalic arteries (four vs seven vessels).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Strouse et al. [9], using multiplanar cardiacgated T1-weighted spin-echo images, found that MRI was complementary to conventional angiography in evaluating patients for the size and presence of collaterals and branch pulmonary arteries. In that study, each technique revealed vessels that were missed by the other. The researchers found good correlation between the two techniques for the right (r = 0.82) and left (r = 0.88) pulmonary arteries, although, MR images tended toward smaller measurements compared with angiography.
Cine gradient-recalled echo MRI has also been used to evaluate patients with pulmonary atresia, ventricular septal defect, and major aortopulmonary collateral arteries [12, 13]. Powell et al. [12] studied 13 patients with complex pulmonary artery stenosis or atresia using both cine gradient-recalled echo and T1-weighted spin-echo techniques. The researchers found complete agreement between the MRI and catheterization findings with respect to the delineation of main pulmonary arteries (6/6), branch pulmonary artery hypoplasia or stenosis (13/13), and the origin of major aortic collaterals (18/18). However, they found difficulty in tracing the aortic collaterals along their complete course to the lungs and suggested that MR angiographic techniques using IV gadolinium might be helpful with anatomic delineation.
Ichida et al. [13] compared multiplanar cine gradient-recalled echo MRI with conventional angiography in patients with pulmonary atresia or severe stenosis. The researchers used axial and several oblique planes to evaluate the pulmonary arteries and collaterals and found excellent correlation between MRI and conventional angiography with respect to the main and branch pulmonary arteries (r = 0.97–0.98). MRI also found major aortic collaterals in seven (78%) of nine patients. That study emphasized the importance of oblique imaging planes, but MR angiography was not performed. Acquisition of axial cine gradient-recalled echo images in our study was relatively time-consuming (20–30 min). The sequence used permitted only a single slice per series, typically 25–40 sec. Multisegmented cine gradient-recalled echo sequences, however, in which multiple slice levels are obtained per acquisition, can considerably improve the imaging time. Although this time savings may allow additional cine imaging planes to be obtained, we found that MR angiography provided useful multiplanar information. In this study, we did not compare cine gradient-recalled echo MRI with MR angiography. We found that using two different bright-blood methods allowed complementary evaluation of the vascular anatomy.
Gadolinium-enhanced 3D MR angiography has been used effectively for pulmonary artery evaluation [14]. Improved gradient strengths on newer magnets allow high-resolution (< 1 mm) angiographic sequences to be performed in as little as 8 sec in patients requiring only a small field of view (40 cm). The dose of contrast material is important in obtaining high-quality images. We found that a larger dose of gadopentetate dimeglumine (0.2 vs 0.1 mmol/kg) produced better 3D images. Hany et al. [15] studied both the pulmonary and renal arteries using gadolinium-enhanced MR angiography at doses from 0.05 to 0.3 mmol/kg. The researchers found that 0.2 mmol/kg was required to produce satisfactory images. A study of contrast-enhanced 3D MR angiography techniques in infants and children, including six patients with pulmonary atresia, also recommends a contrast dose of 0.2 mmol/kg for thoracic imaging [16].
A key difference between MR angiography and conventional angiography is in the method of contrast administration. During conventional angiography, contrast material is first administered in one or more injections through a catheter positioned in the aorta. Subsequently, selective injections of the opacified major collateral vessels off the aorta are performed. The coronary, brachiocephalic, and subclavian arteries can also be selectively injected. The use of selective injections is a distinct, though time-consuming, advantage of this technique. The selectively opacified vessel can be easily traced along its entire course through the mediastinum and into the lung. In comparison, MR angiography uses volumetric acquisition of data from 9 to 28 sec while contrast material is being administered through a peripheral IV line. This method produces a series of static images that opacify a variety of vessels such as the aorta, pulmonary arteries, and collaterals in patients with a ventricular septal defect and pulmonary valve atresia almost simultaneously. Using multiplanar and coned 3D techniques, overlapping vessels can be partially eliminated. However, multiple closely located vessels make tracing the course of an individual vessel difficult.
Conventional angiography in these patients can take a prolonged period (1–6 hr) to complete, because multiple injections and different catheters may be needed to localize and enter each vessel. An article addressing the issue of radiation exposure in helical body CT applications has focused on potential cancer risks, especially in patients exposed during childhood [17]. Although conventional angiography was not specifically addressed, the use of ionizing radiation in pediatric patients should be minimized or avoided entirely when possible. A potential advantage of MRI examination before conventional angiography may be to provide an anatomic road map that could then decrease fluoroscopic time and radiation dose.
More collateral vessels were shown using MRI than using conventional angiography. During unifocalization, the surgical procedure in which collaterals and existing branch pulmonary arteries are brought together to form augmented pulmonary arteries, these vessels are usually left unaltered. After surgery, they usually regress and are therefore of less importance than major aortic collaterals.
Pulsation of the thoracic aorta and cardiac motion can cause significant artifacts. This occurrence is usually most severe in the aortic root and ascending aorta but can happen throughout the aorta. Although cardiac motion and pulsation artifacts can also affect angiography, the use of selective injections and dynamic cine imaging can be used to largely overcome this problem. A 3D electrocardiographically triggered breath-hold contrast-enhanced MRI sequence has been described that can be used to limit cardiac and pulsatile motion artifacts [18]. Selective coronary angiography showed one major collateral from the right coronary artery that was missed on MRI. MRI of the coronary vessels is especially difficult because of cardiac motion and close location to the cardiac chambers. As already noted, the peripheral injection of contrast material performed during MR angiography does not permit selective opacification of these arteries. MRI sequences that correct for cardiac motion may be especially helpful in delineating coronary artery anatomy.
A limitation of this study was the different methods of contrast
administration, including hand and power injection, rates of injection, and
contrast doses. We used both hand and power injection for contrast material
administration. Our routine protocol for pediatric patients younger than 10
years old is to use hand injection. This method allows manual monitoring of
the IV line pressure in case of infiltration. Also, in sedated patients we
have had fewer patients awaken during the contrast bolus with slightly lower
rates of injection ( 1–2 mL/sec vs 2 mL/sec). Patients older than 10
years can typically undergo contrast injection rates according to our routine
protocol for adults (power injection at 2 mL/sec with a 20-mL saline flush).
One patient (not included in analysis) was uncooperative during MRI. As a
result, selection bias might exist in patients who may be perceived as
uncooperative and therefore not referred for MRI study. Patient size does not
appear to affect MRI referral because our smallest patient was a neonate
weighing 4 kg.
Operative repair in this group of patients is determined by the presence, size, and configuration of the pulmonary vessels [4, 5, 19–22]. A variety of surgical procedures can be performed either in a one-stage operation [21, 23–26] or multistaged series of operations [4, 5]. A trend has developed in some centers toward earlier, single-stage intervention [21, 23–25]. Accurate preoperative anatomic data is critical if a one-stage operation is planned [21].
Even patients with diminutive pulmonary arteries are now considered for
surgical correction [20,
22]. Pagani et al.
[20] suggest that complete
repair of even extremely diminutive pulmonary arteries (
3 mm) is
possible. In the three of 14 patients who underwent MRI, confluence was shown
in two and nonconfluence in one. These findings, not shown on cineangiograms,
influenced the type of procedure (shunt vs conduit placement) performed. Those
authors emphasized the need for accurate preoperative pulmonary artery
assessment and the complementary role MRI can play to angiography.
In conclusion, combined cine gradient-recalled echo MRI and contrast-enhanced MR angiography is a reliable, noninvasive method for delineating the pulmonary blood supply in patients with pulmonary atresia, ventricular septal defect, and major aortopulmonary collateral arteries. At present, MRI can provide a road map for conventional angiography. Future prospective studies comparing MRI and conventional angiography may determine whether routine preoperative conventional angiography is needed.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. I. Boechat, O. Ratib, P. L. Williams, A. S. Gomes, J. S. Child, and V. Allada Cardiac MR Imaging and MR Angiography for Assessment of Complex Tetralogy of Fallot and Pulmonary Atresia RadioGraphics, November 1, 2005; 25(6): 1535 - 1546. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Prince, P. O. Alderson, and H. D. Sostman Chronic Pulmonary Embolism: Combining MR Angiography with Functional Assessment Radiology, August 1, 2004; 232(2): 325 - 326. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
