| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 1995;110:704-0714
© 1995 Mosby, Inc.
SURGERY FOR ACQUIRED HEART DISEASE |
London, United Kingdom
Supported by CORDA (Coronary Artery Disease Research Association, London) and by the University of California at Davis.
Received for publication June 13, 1994. Accepted for publication Dec. 8, 1994. Address for reprints: Hugo G. Bogren, MD, Department of Radiology, UC Davis Medical Center, 2516 Stockton Blvd., Sacramento, CA 95817.
Abstract
Magnetic resonance imaging with multidirectional cine velocity mapping was used to study relationships between aortic blood flow patterns and the geometry of thoracic aortic aneurysms and grafts. Ten patients with 13 thoracic aortic aneurysms, single or multiple, or grafts (4) participated in the study. The causes of disease were atherosclerosis (4), Marfan's syndrome (2), trauma (1), and unknown (1), and there were two dissections. Spin-echo imaging and cine velocity mapping in 10 mm thick slices with vertical and horizontal velocity encoding were done. Maps of the two velocity components were processed into multiple computer-generated streaks whose orientation and length corresponded to velocity vectors in the chosen plane. The dynamic arrow maps were compared with previously reported aortic arrow maps from normal subjects. The forward flow occupied the entire lumen in the normal aorta in systole and small vortices were only present in the sinuses of Valsalva. Atherosclerotic aneurysms in the ascending aorta were located at the anterior right and had oblique, eccentric jet flows that created a large secondary vortex in the aneurysm. Patients with Marfan's syndrome had a central jet and two large vortices, one on each side. All other aneurysms, dissections, and grafts had irregular flows and vortices not seen in normal subjects. Magnetic resonance imaging with multidirectional velocity mapping is a powerful noninvasive tool to assess morphologic features and disturbed blood flow in aortic aneurysms and grafts. Recognizably altered flow patterns were found to be associated with altered vessel geometry. The significance of this requires further investigation. (J THORACCARDIOVASCSURG1995;110: 704-14)
The role of hydraulic forces in the formation and propagation of aortic aneurysms is still poorly understood. The initial change in the formation of an aneurysm is structural and results from a degenerative process in the vascular wall. Once the morphologic change has occurred the blood flow pattern is likely to alter, which may cause altered shear stresses and consequent strains and accelerate the propagation of the aneurysm so that a vicious circle is created. The geometry of the aorta is altered in aneurysmal degeneration, and the geometric relationship between the left ventricle and the aortic valve on the one hand and the ascending aorta on the other becomes more and more distorted during the propagation of an ascending aortic aneurysm. Similar morphologic changes occur in aneurysms in the aortic arch or descending aorta that are likely to influence blood flow patterns and change the hydrodynamic forces. The interplay and counterplay between morphologic changes and hydraulic dynamic forces in the development of aortic dissections are also unknown and increased knowledge about the blood flow patterns in dissections might increase our understanding of the pathophysiologic features of the disease process. The final event in the natural history of aneurysms is that they rupture and cause death or that they are surgically replaced with a graft. In this latter case, what are the blood flow patterns in surgical grafts? Do they restore normal blood flow patterns or do they create their own hydrodynamic problems?
A noninvasive tool to visualize blood flow patterns by vector analysis of multidirectional cine magnetic resonance velocity mapping was described by members of our research group and others.
1-6 Mohiaddin and associates 7 described the technique to visualize aortic and pulmonary aneurysms and the reader is referred to that article and the other referenced articles for technical details about the method.
The aim of the present study was to expand the analysis of blood flow patterns in thoracic aortic aneurysms and grafts to try to answer some of the preceding questions, that is, those concerning the interplay or counterplay, or both, between morphologic features and hydraulic hydrodynamic forces in the formation and propagation of aneurysms in the thoracic aorta.
MATERIALS AND METHODS
Subjects
Ten patients, aged 20 to 79 years (average 56 years), with 13 thoracic aneurysms or grafts (3 patients had two aneurysms) were studied in the supine position at rest with spin-echo imaging and velocity mapping. The causes of the aneurysms were atherosclerosis (4), Marfan's syndrome (2), dissection (2), trauma (1), and unknown (1), and there were four patients with grafts. Five patients had mild or mild to moderate aortic valvular regurgitation (Table I).
|
The data from the patients were compared with those published earlier from 10 normal subjects. 8 These 10 subjects had velocity maps done of the aortic arch and the adjacent parts of the ascending and descending aorta, and three of them had maps done of the proximal ascending aorta as well.
The studies on the patients and the normal subjects were approved by the Ethical Committee of the Royal Brompton Hospital and all subjects signed informed consent forms.
Magnetic resonance imaging (MRI)
A Picker International Vista magnetic resonance machine (Picker International, Inc., Healthcare Products Div., Mayfield Village, Ohio) operating at 0.5 tesla with modified gradient coils and a surface receiver coil was used. Cardiac gated transaxial and coronal multislice spin-echo images (echo time 40 msec) were acquired covering the heart and the thoracic aorta. This was followed by an oblique coronal image equivalent to a left anterior oblique image in the plane of the thoracic aorta located by three points: one at the proximal ascending aorta, one at the top of the aortic arch, and one in the descending aorta. This plane was varied somewhat depending on the location of the examined aneurysm (ascending or descending aorta or the arch). Sometimes a coronal plane was used if such a plane turned out to be the optimal one for visualization of the abnormality in the thoracic aorta. These planes, in most cases a plane aligned with the aortic arch including the aneurysm or graft, was then used for bidirectional cine velocity mapping with a gradient echo sequence with an echo time of 12 msec. Sixteen electrocardiogram-gated frames through the cardiac cycle were acquired and the sequence was run three times, once without velocity encoding and two times with velocity encoded in the direction of the read (vertical in the figures) and phase (horizontal) gradient directions. Slice thickness was 10 mm and the field of view 40 cm. The images were displayed on a 256 x 256 matrix from data acquired from two averages of 128 phase-encoding steps. The acquisition of the bidirectional cine maps took 12 to 15 minutes depending on heart rate, and the whole procedure took approximately 1 hour.
Flow visualization: vector maps
Software had been written 1,2,6 to process the bidirectional velocity maps into computer-generated vector cine maps to display the distribution of flow in the image plane (see figures). The individual vectors are displayed as tapered streaks with the thick end pointing in the direction of the flow and the length of the streak proportional to the relative velocity at the point of origin.
Data analysis
The angle between the long axis of the left ventricle and the ascending aorta was measured on the computer display of the coronal images or from hard copies. The maximum width of the aneurysms, the width of the aortic valvular ostium, and the ratio between these two widths were also measured (Fig. 1 and 
Table 1).
|
RESULTS
The results are summarized in Table II.
|
8 No vortices outside the coronary sinuses were present (Fig. 1).
Ascending aortic aneurysms (Marfan's syndrome)
Spin-echo imaging
The two patients with Marfan's syndrome had the typical finding of a greatly and symmetrically dilated ascending aorta, including the sinuses of Valsalva, ending abruptly at the origin of the innominate artery, giving rise to a shape that has been likened to a chemical retort or an Erlenmeyer flask (Fig. 2, A). The width of the aortic valvular ostium was increased in both cases (25 and 26 mm) giving rise to moderate aortic regurgitation.
|
In one case, small vortices in the coronary sinuses appeared in early systole before the formation of the large vortices. These small vortices were similar to those seen in the normal subjects.
Retrograde flow from the arch joined the vortices more on the left side than on the right. Some of the retrograde flow crossed over from the left to the right sinus in early diastole (Fig. 2, C). Both patients had aortic valvular regurgitation and as the aortic valve opened retrograde flow in the ascending aorta was seen to meet antegrade flow at the aortic opening. Blood was seen to flow retrogradely from the great vessels into the aortic arch and down the ascending aorta as part of the aortic regurgitation.
Patient 2 also had severe thoracic scoliosis and a tortuous descending aorta that made two 90-degree bends. Signal loss caused by turbulence was seen at the sharp bends in part of systole and irregular antegrade, retrograde, and transverse flows were seen in diastole.
Atherosclerotic ascending aortic aneurysms
Spin-echo imaging
Patients 3 through 5 (Table I) all had a differently shaped aneurysm compared with those of patients with Marfan's syndrome insofar as the aneurysm was confined to the anterior and right part of the ascending aorta, which had an asymmetric shape (Fig. 3, A). The sinuses of Valsalva were not involved in the aneurysms and the posterior and left wall of the ascending aorta appeared normal.
|
The flow was most rapid along the inner curve of the arch in all three patients. A second vortex was created in the distal arch in patient 3, who also had an aneurysmal proximal innominate artery with large inflow and vortex formation in the proximal arch. Patient 4 had an old dissection in the descending aorta in addition to the ascending aortic aneurysm. He had a vortex in the distal arch just proximal to the old dissection similar to that in patient 3.
In Patient 5, flow reversal was seen in the proximal aorta, as well as in the middescending aorta. Patient 3 had a tortuous descending aorta (Fig. 3) with another vortex.
Ascending aortic aneurysm of unknown cause
The histologic finding in patient 6 was aneurysm, but there was no evidence of medionecrosis or other changes and no evidence of atherosclerosis or aortitis.
The anatomic and blood flow characteristics in this patient were almost exactly as in the patients with atherosclerotic aneurysm. Flow reversal was seen at the transition of the arch and descending aorta as in normal subjects, but there was additional flow reversal with vortex formation in the arch. There was a large amount of retrograde flow in the ascending aorta but no evidence of aortic regurgitation. There was also flow in various directions at various times in diastole in the arch.
Dissections
Three of the patients, 4, 7, and 8, had old dissections in addition to aneurysm or graft. There was disturbed blood flow in all dissections with vortices starting immediately proximal to the intimal tear and continuing in the true lumen of the dissection.
Grafts
There were four patients with grafts, including two with grafts in the ascending aorta after dissection several years earlier. Patient 8, with a Dacron fabric graft, had a thrombosed false lumen in the arch and descending aorta, and Patient 7, with a homograft, has a separate dissection in the descending aorta. Patient 9 had a graft replacement of the aortic arch several years earlier because of an atherosclerotic aneurysm, and patient 10 had a graft in the descending aorta after trauma 4 months earlier. Both of these were Gelseal vascular grafts (Vascutek, Renfrenshire, Scotland).
Spin-echo images
The descending aortic graft was moderately dilated (width 31 mm compared with 18 mm in adjacent descending aorta) whereas the other grafts had normal or slightly increased width.
Velocity maps
Common to all grafts independent of position and age of the graft was that the blood flow was highly disturbed with vortices at both ends of the graft (arch), in the graft (descending aorta) (Fig. 4), and in the graft, as well as at the distal anastomosis (ascending aorta). There was a posterior jet in the dilated descending aortic graft and an anticlockwise vortex started as soon as blood entered the graft in systole (Fig. 4). The flow outside the grafts was also disturbed with other vortices and irregular flow not seen in normal subjects.
|
There was more flow along the inner curve of the aortic arch in all observed cases.
Measurements
The angle between the longitudinal axis of the left ventricle and that of the aorta with the horizontal or transverse plane was measured in the patients and in 10 normal subjects. The angle averaged 30 degrees (25 and 35 degrees) in the patients with Marfan's syndrome, 45 degrees (23, 65, and 47 degrees) in the patients with atherosclerosis, and 21 degrees (15 to 26 degrees) in the 10 normal subjects. It was an average of 30 degrees in all the 10 patients.
The ratio between the maximal aneurysmal width and the width of the aortic valve was 2.55 (2.5 and 2.6) in the patients with Marfan's syndrome and 3.4 (3.2 to 3.7) in the three patients with atherosclerotic aneurysms. The difference was because of widening of the aortic valve in the patients with Marfan's syndrome (25.5 versus 21 mm average) inasmuch as the width of the aneurysm was not significantly different (66 versus 69 mm).
DISCUSSION
Morphologic features of ascending aortic aneurysms
The two Marfan's syndrome aneurysms had the classic appearance of a pear-shaped symmetric dilation confined to the ascending aorta and including the sinuses of Valsalva, which is well known from x-ray angiography, echocardiography, 9 computed tomography, 10 and earlier MRI studies. 11-16 We also reviewed the spin-echo images in two other patients with Marfan's syndrome who did not have vector mapping. The shape of the ascending aorta was identical to that of the others.
The three atherosclerotic ascending aneurysms and two others that were not subjected to velocity mapping had also a characteristics shape. In all five patients, the aneurysms were confined to the anterior right part of the ascending aorta and the posterior left part of the sinuses appeared normal (see Fig. 3, A). This is in agreement with our and others' experience from x-ray aortograms, in which the aneurysms are more prominent on the right, the outside, or the convex side of the ascending aorta on left anterior oblique projections. However, we are not aware that this shape, seen in Fig. 3, A, has ever been described in detail in the radiologic, surgical, or pathologic literature. 17-19 This is probably because MRI can image morphologic features in more detail than other techniques, including autopsy and specimen studies, so that the aneurysmal lumen, its walls, and all the adjacent structures are seen simultaneously. The morphology of these five atherosclerotic ascending aortic aneurysms is quite diagnostic and sets them apart from the Marfan's syndrome aneurysms (Fig. 2, A) although probably not from inflammatory aneurysms (see following section).
Atherosclerotic aneurysms are rare in the ascending aorta and much more common distally. 18 The most common cause of aneurysms in the ascending aorta is degenerative disease of the media (medionecrosis and mucoid degeneration) as seen in Marfan's syndrome or forme fruste of Marfan. Inflammatory diseases predominate in the arch and descending aorta in our time, whereas they were the most common, and most often appeared in the ascending aorta, when syphilis was the dominant cause of aneurysms. 17 We have no experience with MRI of luetic or other inflammatory aneurysms, but they are described in the pathologic literature as being anterior, 17 probably similar in location to atherosclerotic aneurysms.
The ascending aortic aneurysm with normal or uninterpretable histologic findings (case 6, Table I) looked exactly like the atherosclerotic aneurysms. The patient was only 39 years old, which makes atherosclerosis unlikely. We assume by exclusion that this aneurysm was inflammatory.
Blood flow patterns in ascending aortic aneurysms
The main difference between the atherosclerotic and the Marfan's syndrome aneurysms was that the jet was obliquely directed toward the right anterior wall in the atherosclerotic aneurysms and was central in Marfan's syndrome and that one large vortex to the left of the jet formed in the former whereas there were two large vortices, one on each side of the jet, in the latter (see Results section and Figs. 2 and 3). This is probably to be expected from the morphologic features of the aneurysms, but it could also be the other way around, that is, that the morphologic features are a result of hydraulic forces exerted by the jet or the vortices, or both forces. A possible cause of an eccentric jet would be misalignment between the left ventricle and the ascending aorta; in other words, that a more horizontally aligned ventricle would eject blood to the right into a more vertically directed aorta. We therefore measured the inclination of the left ventricular long axis and that of the ascending aorta with the transverse or horizontal plane in the patients and compared the results with those in 10 normal subjects. The left ventricle was more horizontally oriented than the aorta in all cases, but there was no significant difference between the results in the normal subjects and those in the patients with Marfan's syndrome or atherosclerotic aneurysm in this small series. The angle was quite large in two patients with atherosclerotic aneurysms (65 and 47 degrees) but normal in one (23 degrees), and it was very small (9 degrees) in the patient with an unclear aneurysmal cause but with an aneurysm of the same morphologic type as the atherosclerotic aneurysms. It seems therefore unlikely that the apparent misalignment would exert much effect on the morphologic features of ascending aortic aneurysms. It may seem difficult to explain this apparent misalignment between the left ventricle and the aorta, which is also seen in normal subjects, but one must keep in mind that the records are two-dimensional and that a twisting motion of the left ventricle may align the left ventricle and aorta, at least in normal subjects.
We also looked into the possibility that the relationship between the width of the aortic valvular ostium and the aneurysm might be different and possibly play a role in the formation or hemodynamics of the two types of ascending aortic aneurysms (see Table I and Results section). The ratios were smaller in the patients with Marfan's syndrome, most likely because of a stretched aortic ring, which increased the numerator. Both patients had aortic insufficiency. One patient with atherosclerotic aneurysm had also a wide aortic valvular ostium with a 25 mm diameter, like the patients with Marfan's syndrome, but had no aortic regurgitation.
Basic fluid mechanical principles would suggest the jet to be central in aneurysms, 20 as it indeed is in Marfan's syndrome. Computer simulation of blood flow patterns on the basis of solving continuity and Navier-Stokes equations in a fusiform aneurysm also suggested that the jet should be central. 21 The severely asymmetric jet in atherosclerotic aneurysms is therefore difficult to explain.
The fact that the jet hits the wall of the atherosclerotic aneurysm may suggest that it exerts an influence on the further expansion of the aneurysm. On the other hand, such a jet is missing in the symmetric Marfan aneurysms, which attain the same width without a jet hitting any wall. The vortices in the expanding parts of both types of aneurysms cannot be excluded to play a role in their further growth, although the hydrodynamic force is likely to be much smaller than the tension force. According to the law of La Place, the wall tension (for a given blood pressure) increases in proportion to the square of the radius of the curvature. 22 Where diameter increase is large, this has major consequences for tension in a wall that is (presumably) already structurally abnormal. The more the vessel dilates, the more it is subjected to stresses that predispose to further dilation. For a given flow rate, mean forward velocity decreases in inverse proportion to the square of the diameter. When flow moves from a region of normal diameter into a region of increased diameter, momentum is conserved, and the stream maintains its initial velocity, but separates from the diverging wall or walls. 20,21 Flow separation is associated with recirculation, that is, local reversal of flow, in the separation zone. The recirculating movement can take the form of a vortex. The pulsatility of flow and any asymmetry or dilation can enhance the tendency to vortex formation.
It is therefore likely that aneurysm formation is primary and dominant and has fluid dynamic consequences that are secondary and do not contribute significantly to the propagation of the aneurysms.
The vortices sometimes moved distally in systole, there were sometimes multiple vortices, and there was, in general, irregular flow, not only in the aneurysms, but also in the arch with recirculations not seen in normal subjects. 8,23 The blood flow was therefore quite disturbed, which may influence flow to the whole body, perhaps especially to the coronary arteries and brain, which depend on a fine-tuned interplay between antegrade and retrograde flow. 23-26
Blood flow in dissections
The three dissections that were unrelated to the aneurysms or grafts all had disturbed blood flow with vortices. Patient 4 had an atherosclerotic ascending aortic aneurysm, a relatively normal arch without aneurysm, and a separate dissection in the descending aorta. The hemodynamic trauma appeared to be in the right anterior wall of the ascending aorta, which was hit by the jet, and one would expect the dissection to start there, if there was much force in a jet. The dissection in the descending aorta seems to support our view that morphologic changes are primary and not the result of hydraulic, hemodynamic forces.
Blood flow in grafts
Common to all grafts was a rather severely disturbed blood flow with formation of single or multiple vortices. This was the case also in the graft that was only 3.5 months old (see Results section and Fig. 4). There were vortices in and around the grafts and also disturbed flow elsewhere, suggesting a nonideal situation. Approximately 200,000 prosthetic vascular grafts are implanted per year in the United States, and there has been and is much ongoing research to find the ideal graft, which has been defined as follows: "For a synthetic vascular graft to function as a normal blood vessel, it must contain some or all of the activities and properties which make natural blood vessels function normally. These include prostacyclin (PGI2), endothelium-derived relaxing factor (EDRF), tissue plasminogen activator (tPA), heparin and other glycosaminoglycans, thrombomodulin, ADPase, compliance, and undoubtedly other as yet unknown factors." 27 Emphasis has been placed on thefactors mentioned, 28 but we have not had a tool until now to examine how the grafts carry blood flow. This study of blood flow patterns in grafts is small, but suggests that further research be done regarding blood flow in grafts of varying age and composition, anastomosis technique, and so on. Magnetic resonance velocity mapping is a powerful noninvasive tool to be used in research regarding the ideal prosthetic vascular graft, which, with due respect for all the aforementioned activities and properties, first and foremost must carry blood flow in a reasonably normal fashion.
Miscellaneous blood flow observations
Blood was seen to enter the left coronary sinus from above in diastole and also to cross over from the left to the right just above the sinuses (Fig. 2, C) as predicted from earlier indirect studies that used velocity mapping in the mid ascending aorta. 24
Blood was also seen to flow in and out of the great vessels in the arch in various ways as described in detail in the Results section, which was likewise predicted in earlier studies that used through-plane one-directional velocity mapping. 23,26 Multidirectional MRI velocity imaging with computer-generated vector maps is an ideal tool to study flow patterns in the aorta and its major branches.
Limitations of the study
The main limitation of the flow pattern analysis is that it was confined to a 10 mm slice of the aorta and that only two directions were recorded. It is possible to make a complete "seven-dimensional" analysis of the flow, 29 but it would be enormously time consuming and not practical at the present speed of MRI studies. The anterior-posterior direction was not analyzed in the patients and it is quite possible, even likely, that more vortices are present in that direction.
The study is also limited in the numbers of the different categories of aneurysms and grafts, but shows features of disturbed flow common to all categories.
We did not analyze the consequences of the disturbed flow, but because it involves energy losses it must increase cardiac work. Neither did we have the opportunity to analyze flow during exercise.
Conclusions
MRI spin-echo imaging and multidirectional cine velocity mapping with computer-generated velocity vectors are suitable noninvasive techniques to reveal detailed morphologic features and blood flow patterns in thoracic aortic aneurysms and grafts. All aneurysms, grafts, and dissections had severely disturbed blood flow, which entails energy losses and must increase cardiac work. The technique may be applied to test efficiency of grafts.
Footnotes
*On sabbatical leave from the Department of Radiology, University of California, Davis, Medical Center, Sacramento, Calif. 
References
This article has been cited by other articles:
|
|
 |
|
  Z. Touat, L. Lepage, V. Ollivier, P. Nataf, U. Hvass, J. Labreuche, M. Jandrot-Perrus, J.-B. Michel, and G. Jondeau Dilation-Dependent Activation of Platelets and Prothrombin in Human Thoracic Ascending Aortic Aneurysm Arterioscler. Thromb. Vasc. Biol., May 1, 2008; 28(5): 940 - 946. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Marui, T. Mochizuki, T. Koyama, and N. Mitsui Degree of fusiform dilatation of the proximal descending aorta in type B acute aortic dissection can predict late aortic events. J. Thorac. Cardiovasc. Surg., November 1, 2007; 134(5): 1163 - 1170. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Houston, S. J. Gandy, W. Milne, J. B. C. Dick, J. J. F. Belch, and P. A. Stonebridge Spiral laminar flow in the abdominal aorta: a predictor of renal impairment deterioration in patients with renal artery stenosis? Nephrol. Dial. Transplant., July 1, 2004; 19(7): 1786 - 1791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Robicsek Myocardial infarction with angiographically normal coronary arteries J R Soc Med, January 10, 2002; 95(10): 528 - 528. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
