| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Medicine, Cardiovascular Angiogenesis Research Center, Beth Israel Deaconess Medical Center-East, 330 Brookline Ave, DA 827, Boston, MA 02215. Received October 12, 1998; revision requested December 17; final revision received June 20, 1999; accepted July 19. J.D.P. supported in part by National Institutes of Health (NIH) grant HL55354, R.J.L. supported in part by NIH grant HL MO1 RR01032, and M.S. supported in part by NIH grants HL53793 and HL56993. Address reprint requests to J.D.P. (e-mail: jdp@shogi.bidmc.harvard.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
MATERIALS AND METHODS: Magnetization preparation plus spatial frequency reordering was applied to distinguish new intramyocardial collateral vessels from normal circulation on the basis of geometric differences. A vascular occluder was inserted in 34 pigs, and they were assigned randomly to treatment groups with either placebo or angiogenic growth factor. Collateral extent determined with collateral-sensitive MR imaging was correlated with direct measurements by means of three-dimensional (3D) computed tomography (CT), coronary blood flow distribution determined with microspheres, and findings at histologic examination. Changes in the signal at collateral-sensitive MR imaging before and after treatment were assessed by two observers blinded to treatment.
RESULTS: The collateral extent determined with collateral-sensitive MR imaging correlated well with findings at 3D CT (r = 0.95) and with microspheres (r = 0.86). Furthermore, the collateral extent determined with collateral-sensitive MR imaging increased significantly (P < .001) in response to the administration of an angiogenic growth factor but not to placebo. The correspondence of findings at collateral-sensitive MR imaging to collateral neovascularization was confirmed at histologic examination.
CONCLUSION: The presence of intramyocardial collateral microvessels was accurately determined with collateral-sensitive MR imaging. The technique may be useful in clinical studies of therapeutic angiogenesis.
Index terms: Animals • Coronary angiography, technology, 54.121417, 54.121419 • Coronary vessels, flow dynamics, 54.121417, 54.121419 • Coronary vessels, MR, 54.121417, 54.121419 • Heart, CT, 51.12116, 51.12117 • Myocardium, infarction, 51.121419, 511.121417
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
Although clinically important, identification of collateral circulation has been limited due to insensitivity of the presently available techniques. One available technique is conventional angiography that is capable of identifying only vessels larger than 180 µm in diameter. That limits detection to a subset of epicardial collateral vessels, with complete inability to detect smaller intramyocardial vessels. Another available technique is nuclear perfusion imaging. This approach with either single photon emission tomography (eg, thallium 201 or technetium 99m scanning) or positron emission tomography can be used to indirectly infer the existence of collateral blood supply in the presence of known coronary occlusion by identifying otherwise unexplained preservation of myocardial perfusion (6).
The need for accurate and direct imaging of collateral microvascular development is heightened by recent demonstration of the ability of a number of heparin-binding growth factors to augment the growth and formation of collateral circulation in the setting of chronic ischemia in both myocardial and peripheral tissues (7). The ability to detect and monitor the progress of neovascularization would offer substantial benefits to both animal and human studies of coronary angiogenesis.
In the present study, we investigated the ability of a collateral-sensitive form of magnetic resonance (MR) imaging to depict the presence and quantify the extent of neovascularization in chronically ischemic porcine myocardium. We hypothesized that coronary collateral neovascular development may be identified and quantified with an MR imaging method that uses T2* preparation and spatial-frequency reordering by detecting a dark signal flare due to local heterogeneous susceptibility effects as contrast agent arrives in newly developing small vessels.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
We studied both phantoms and an animal model of chronic ischemia, and we completed four types of validation, including ability to detect treatment response in those receiving vascular endothelial growth factor (VEGF) to stimulate angiogenesis.
Phantom Model of Microvascular Susceptibility Effects
The phantom model was selected to represent sparse collateral microvessels surrounded by tissue fluid in a reproducible and highly controlled environment. This was achieved with a modified dialysis filter (LunDia Alpha 500; Cobe Renal Care, Newport News, Va). The filter is a multicompartmental system with rectilinear adjustable spacing analogous to the human kidney that was designed to contain two spaces in close association, one representing the microvessels, and the other, the surrounding interstitium. Each compartment has separate input and output ports that control the choice of fluid and the rate of its passage.
Saline solution was placed in both the microvascular space and in the interstitial space. The saline solution in the interstitial space was supplemented with manganese chloride to adjust relaxivity to that of the compartment (9). Tubing was connected to the input port of the microvessel space to allow the injected contrast agent to pass through as desired. Imaging with T2*-sensitive sequences was performed during contrast agent arrival to monitor extent of signal dropout around the microvascular tubing. To simulate in vivo conditions, saline solution was infused at a slow rate (analogous to an infusion to keep a vein open). Then during imaging, an 8-mL bolus of gadopentetate dimeglumine diluted 1:4 with saline solution was injected with a power injector (Spectris; Medrad, Indianola, Pa) at 2 mL/sec for 4 seconds followed by an 8-mL bolus of saline solution at 2 mL/sec. Imaging was performed with clinically applicable sequences triggered with an electrocardiographic simulator at 72 beats per minute (one image every 0.83 second). The transit time of contrast agent across the phantom was approximately 2 seconds, which is comparable to that observed through myocardium (10). The phantom was used for preliminary testing with the pulse sequences to observe the presence and conspicuity of the dark flare as gadopentetate dimeglumine arrived in the microvessels. Imaging parameters were adjustedto maximize the contrast-to-noise ratio of the dark flare.
Pig Model of Chronic Myocardial Ischemia
Chronic ischemia was induced in 22 pigs by placing a vascular occluder (Ameroid; Research Instruments, Corvalis, Ore) as previously described (11). In this model, placement of the plastic-encased, size-matched ring of the vascular occluder around the proximal segment of the left circumflex coronary artery leads to progressive coronary occlusion secondary to gradual fluid absorption by the occluder material and the tissue. A magnet-compatible osmotic pump (Alzet Pharmaceuticals, Palo Alto, Calif) was placed in the chest, interior to the left fourth intercostal space, with the pump body outside the pericardium and the flow tip inside the pericardium. The pump delivered saline solution with heparin (50 U/mL) at 4 µL/hr to the territory of the left circumflex coronary artery in all animals as previously described (12). The chest was closed, and all animals recovered. Three weeks later the animals were brought back for baseline studies of coronary circulation with use of colored microspheres and collateral-sensitive MR imaging. After these studies, the animals were randomly assigned to two groups: treatment with growth factor versus placebo (n = 15) or assessment of collateral circulation with ex vivo three-dimensional (3D) computed tomography (CT) (n = 7).
In animals assigned to the growth factor group, osmotic pumps for saline solution were replaced with pumps that contained either 10 µg of recombinant VEGF (rhVEGF165; Genentech, South San Francisco, Calif) supplemented with 50 U of heparin (n = 8) or with only the buffer with the same amount of heparin for sham control (n = 7). The recombinant VEGF stimulates collateral microvessel development. Three weeks later, the animals were brought back for repeated coronary flow studies and collateral-sensitive MR imaging. After acquisition of the final in vivo data, the animals were sacrificed, and the hearts were processed for microsphere blood flow analysis and histologic analysis of neovascularization.
Animals assigned to the 3D CT group were brought back 3 weeks after implantation of the vascular occluder. Collateral-sensitive MR imaging was repeated, the animals were immediately sacrificed, the hearts were rapidly extracted and washed out with saline solution with heparin, and a vascular occluder was placed for ex vivo collateral-sensitive MR imaging and 3D CT.
All animals received humane care in compliance with the institutional and American Physiological Society guidelines. The study was approved by the institutional animal care and use committee.
Coronary Blood Flow Studies
For coronary blood flow studies, 3 x 106 microspheres (selected from sets of six distinct colors, one color per injection) were injected at the time of initial placement of the vascular occluder with the left circumflex artery transiently occluded to allow determination of the extent of territory at risk (12). Three and 6 weeks later, 6 x 106 microspheres were injected to assess the extent of myocardial perfusion. Both microsphere injections and tissue processing were carried out as previously described (12). The extent of coronary blood flow in the myocardium compromised by the vascular occluder was determined by means of selection of tissue slices that corresponded anatomically to the areas of perfusion of the left circumflex artery that contained less than 10% of the microspheres injected at the time of initial placement.
Collateral-Sensitive MR Imaging
Collateral-sensitive MR imaging was initiated with magnetization transfer for background tissue suppression and T2* magnetization preparation with 90° excitation followed by 90° return after a delay adjusted to maximize the return. Then, phase encoding was ordered in alternating pairs beginning at 16-mm wavelength higher and then 16-mm wavelength lower. Thus, the high phase-encoding steps that occurred soonest after the T2* preparation had the strongest T2* sensitivity, whereas those that occurred later had lower T2* sensitivity owing to T1 recovery. Phase ordering in this manner emphasized susceptibility frequencies that corresponded to the heart walls and decreased the large through-space effect from the lower spatial frequencies, which are predominantly affected when the ventricle fills with contrast agent. Acquisition parameters were repetition time msec/echo time msec of 5.6/2.0, flip angle of 8°–12°, field of view of 250 x 250 mm, matrix of 128 x 128, acquisition time for T2* myocardial delineation of 44.8 msec, total imaging time of 717 msec, and electrocardiographic triggering.
MR imaging was performed with a 1.5-T whole-body system (Vision; Siemens Medical Systems, Iselin, NJ) as time series, with use of a four-part phased-array chest coil receiver to enhance the signal-to-noise ratio. First, scout imaging was performed to allow orientation and adjustment; then, a breath-hold baseline set of 12 images triggered with electrocardiography was acquired to establish the effectiveness of breath holding. Pancuronium bromide was administered as needed in the animal model to enforce breath holding when the ventilator was in the pause mode. Then, a series of 24 breath-hold images were obtained with collateral-sensitive MR imaging. During the initial four heart beats, magnetization approached steady state. Gadopentetate dimeglumine was power injected beginning at the fourth heart beat as a 7-mL bolus at 2 mL/sec and followed by 7 mL of saline solution. Twenty additional images were acquired to observe arrival of the contrast agent. The k-space ordering prevented problems relating to left ventricular filling with high-concentration contrast agent. The series was repeated in the four-chamber and short-axis views. Scout imaging was repeated to establish return to baseline prior to subsequent contrast agent injections.
Analysis of Collateral-Sensitive MR Images
At autopsy, the flow distribution at the time of injection was determined by means of microsphere analysis. Collateral-sensitive MR images were segmented to correspond to the tissue samples from the zone compromised by the vascular occluder by selecting areas at the second visit that contained less than 10% of the microspheres injected at the time of initial placement of the vascular occluder. The threshold of 10% was selected by means of discriminant analysis on the basis of observed variance.
On an independent workstation with use of custom software developed by J.D.P., the extent of the collateralized zone depicted at collateral-sensitive MR imaging was defined by means of space-time maps (13,14) coordinated to the time-intensity curves and anatomic location. The border was computed by means of the Fisher discriminant criterion, which was automatically optimized to minimize the total amount of collateralized muscle outside and directly supplied muscle inside. Two technicians blinded to treatment reviewed the space-time maps and confirmed the border recognition with adjustment options, determined the extent and location of collateral dark flash and delayed arrival zones, and reported consensus results. During image data review, the technicians had simultaneous access to (a) a cine format, (b) an instant time-activity curve for the location under the cursor, (c) time-intensity curves for user-defined regions of interest, (d) spatial intensity profiles perpendicular to borders, and (e) space-time maps (11) that combined intensity and time data in a two-dimensional display that facilitates border recognition by allowing direct demonstration of spatial-temporal relations.
Ex Vivo MR Imaging
Explanted hearts were cannulated proximally and imaged with a collateral-sensitive MR imaging protocol similar to that used in vivo but with use of an electrocardiographic simulator set at 72 beats per minute as a trigger to emulate the clinical conditions for the pulse sequence. In addition, the injected contrast agent was diluted 4:1 with saline solution, and the injection rate was reduced to 1.5 mL/sec.
Ex Vivo Elastic-Match CT
After ex vivo MR imaging, 3D CT was performed with different driving pressures so that back pressure did or did not block collateral flow to the territory distal to the occluded left circumflex artery, as previously described (12). Elastic matching (13,14) is a 3D image processing method developed by J.D.P. that compares image sets automatically to determine differences other than local elastic 3D changes of position. Elastic matching of the volume images obtained in these two conditions allowed identification of the zone of myocardium that filled with contrast agent only when back pressure beyond the occluder was released (13). The extent of the collateralized zone at CT was measured as the linear extent in the cardinal axis directions (base to apex, circumferential, and thickness), as total volume, and as volume as a percentage of the total myocardial volume. The data were analyzed blindly and then compared with the Pearson correlation and paired t test corrected for multiple comparisons to an overall confidence of 95%.
Histologic Analysis
Histologic assessment of myocardial tissues was performed with use of 5-µm-thick slices stained with Masson trichrome. The presence of collateral vessels was determined by means of immunocytochemical testing of formalin-fixed tissues. Slides were blocked for 30 minutes with 5% normal goat serum and then incubated with rabbit anti–human von Willebrand factor (1:500) (Sigma, St Louis, Mo) for 30 minutes at room temperature as a marker for endothelial cells and neovascularization. The primary antibody was then detected by using biotinylated antirabbit immunoglobulin G (1:00) (Vector, Burlingame, Calif) followed by streptavidin-biotin (1:00) (Amersham, Arlington Heights, Ill) for 30 minutes at room temperature, respectively. Control slices were stained with the same protein concentration of normal rabbit immunoglobulin G to confirm the specificity of staining.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
The accuracy of this approach is evident from the extent of collateral territory determined with collateral-sensitive MR imaging compared to that determined with the ex vivo images acquired with elastic-match 3D CT after direct injection of intracoronary contrast agent in the setting of pressurized and nonpressurized territory in the left circumflex artery (13). There was a remarkable visual identity between both sets of images, and quantitative analysis of the territory extent also demonstrated close correlation between the two techniques. Similarly, extent of collateral perfusion determined with collateral-sensitive MR imaging compared to that determined with coronary microspheric data also demonstrated close correlation between the techniques. Thus, collateral-sensitive MR imaging provides accurate assessment of newly developed collateral circulation in live animals.
The presence of dark flare signal detected with collateral-sensitive MR imaging was associated with the histologic evidence of intramyocardial collateral microvessels, with the extent of dark flare corresponding to the anatomic extent of intramyocardial neovascularization, which suggests that the latter were responsible for this effect. This association is in accord with our assumption that alterations in geometry associated with development of collateral circulation results in the collateral dark flare. Increase in the area of the dark flare in the VEGF-treated animals compared with that in the control animals further supports the technique as sensitive to the extent of collateral perfusion. The signal at collateral-sensitive MR imaging is not obscured by left ventricular filling with magnetic susceptibility contrast agent, does not obscure myocardial signal, and is not influenced by variations in ventricular wall thickness.
One of the most appealing features of this technique is the ability to identify and monitor developing collateral circulation in the heart. Collateral-sensitive MR imaging is a hybrid of T2*- and T1-weighted MR imaging. Quality imaging of T2* effects requires good shimming. Reordering of the spatial frequency helps minimize filling effects in the left ventricle. Therefore, one limitation of collateral-sensitive MR imaging is that it is not useful in the determination of time of arrival of contrast agent in the left ventricle. Also, collateral-sensitive MR imaging enables identification of new vessels that are much smaller than the image resolution, owing to the signal flare as a result of magnetic susceptibility. In principle, that amplification effect could result in overestimation of the extent of collateral development. However, the good agreement between findings with elastic-match CT, microsphere distributions, and histologic examination indicate that this is not a problem.Practical applications: The technique of collateral-sensitive MR imaging described herein appears promising as a noninvasive quantitative measure of progress of collateral development. Further study of its diagnostic and prognostic accuracy is required to assess its value in clinical practice. The emerging field of therapeutic angiogenesis, as well as other approaches to myocardial revascularization (7), would likely benefit from this technique, which allows serial noninvasive assessment of collateral development. This assessment of neovascularization with MR imaging may also prove useful in the monitoring of neovasculature-suppressive therapy (eg, angiostatin) for treatment of cancer.
|
APPENDIX |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
To emphasize T2* contrast, we added preparation pulses that placed magnetization transversely and then returned it, reduced in length according to T2*. This preparation provided a marker for collateralized myocardium (Fig 2b), but the problem remained of ventricular filling with contrast agent that transiently blotted out myocardium. Thus, a collateralized zone that exhibited delayed arrival (10) was visible, but, if the collateral microvessels were sufficient to provide rapid delivery of nutrient blood supply to heart muscle, the arrival might be masked by the flaring of signal from the ventricle.
The flare of the susceptibility contrast agent needs to be localized to the area immediately around the sparse collateral vessels within the myocardium but not within the left ventricle. We dealt with that problem by taking advantage of the sensitivity of MR imaging to spatial frequency. All MR imaging data are collected as samples of spatial frequency that summarize the state of transverse magnetization as the sum of contributions of waves at various spatial frequencies. The data collected immediately after the preparation pulse report T2* contrast, whereas data collected well after the preparation pulse include T1 recovery and minimal T2* contrast. Therefore, by reordering the acquisition of different spatial frequencies appropriately, differences on a scale of less than twice the thickness of the heart wall are T2* weighted, whereas those on a scale of the size of the left ventricle are T1 weighted and T2* insensitive. This spatial frequency hybrid results in images that are anatomically correct, and when MR susceptibility contrast passes through the collateral microvessels, the myocardium in that vicinity flares darkly. The effect is minimal when the contrast agent fills the left ventricle, however, because the left ventricle is large and its signal dominates very low spatial frequencies that are excluded from the T2* contrast effect. Also, there is no T2* contrast effect when the normal zone fills uniformly with magnetic susceptibility agent because the normal vessels are close enough that there is no appreciable gradient between regions with high and low susceptibility. This approach to MR imaging produced a clearly visible dark flare that corresponded to the location and extent of tissue supplied by collateral microvessels in the heart.
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, J.D.P., M.S., R.J.L.; study concepts, J.D.P.; study design, J.D.P., M.S.; definition of intellectual content, J.D.P., M.S.; literature research, J.D.P., M.S.; experimental studies, J.D.P., R.J.L., M.S.; data acquisition, J.D.P., R.J.L.; data analysis, J.D.P., R.J.L., M.S.; statistical analysis, J.D.P., R.J.L., M.S.; manuscript preparation and editing, J.D.P., M.S.; manuscript review, J.D.P., M.S., R.J.L.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIXReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Saeed, A. Martin, P. Ursell, L. Do, M. Bucknor, C. B. Higgins, and D. Saloner MR Assessment of Myocardial Perfusion, Viability, and Function after Intramyocardial Transfer of VM202, a New Plasmid Human Hepatocyte Growth Factor in Ischemic Swine Myocardium Radiology, October 1, 2008; 249(1): 107 - 118. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Voisine, A. Rosinberg, J. J. Wykrzykowska, Y. Shamis, G. F. Wu, E. Appelbaum, J. Li, F. W. Sellke, D. Pinto, C. M. Gibson, et al. Skin-derived microorgan autotransplantation as a novel approach for therapeutic angiogenesis Am J Physiol Heart Circ Physiol, January 1, 2008; 294(1): H213 - H219. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Saeed, D. Saloner, A. Martin, L. Do, O. Weber, P. C. Ursell, A. Jacquier, R. Lee, and C. B. Higgins Adeno-associated Viral Vector-Encoding Vascular Endothelial Growth Factor Gene: Effect on Cardiovascular MR Perfusion and Infarct Resorption Measurements in Swine Radiology, May 1, 2007; 243(2): 451 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Buehler, M. A.M.J. van Zandvoort, B. J. Stelt, T. M. Hackeng, B. H.G.J. Schrans-Stassen, A. Bennaghmouch, L. Hofstra, J. P.M. Cleutjens, A. Duijvestijn, M. B. Smeets, et al. cNGR: A Novel Homing Sequence for CD13/APN Targeted Molecular Imaging of Murine Cardiac Angiogenesis In Vivo Arterioscler. Thromb. Vasc. Biol., December 1, 2006; 26(12): 2681 - 2687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Tio, E. S. Tan, G. A.J. Jessurun, N. Veeger, P. L. Jager, R. H.J.A. Slart, R. M. de Jong, J. Pruim, G. A.P. Hospers, A. T.M. Willemsen, et al. PET for Evaluation of Differential Myocardial Perfusion Dynamics After VEGF Gene Therapy and Laser Therapy in End-Stage Coronary Artery Disease J. Nucl. Med., September 1, 2004; 45(9): 1437 - 1443. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.H. Tayebjee, G.Y.H. Lip, and R.J. MacFadyen Collateralization and the response to obstruction of epicardial coronary arteries QJM, May 1, 2004; 97(5): 259 - 272. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Yang Imaging of Vascular Gene Therapy Radiology, July 1, 2003; 228(1): 36 - 49. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Laham, M. Simons, J. D. Pearlman, K. K. L. Ho, and D. S. Baim Magnetic resonance imaging demonstrates improved regional systolic wall motion and thickening and myocardial perfusion of myocardial territories treated by laser myocardial revascularization J. Am. Coll. Cardiol., January 2, 2002; 39(1): 1 - 8. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. B. Freedman and J. M. Isner Therapeutic Angiogenesis for Coronary Artery Disease Ann Intern Med, January 1, 2002; 136(1): 54 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Simons, R. O. Bonow, N. A. Chronos, D. J. Cohen, F. J. Giordano, H. K. Hammond, R. J. Laham, W. Li, M. Pike, F. W. Sellke, et al. Clinical Trials in Coronary Angiogenesis: Issues, Problems, Consensus : An Expert Panel Summary Circulation, September 12, 2000; 102 (11): e73 - e86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Simons, B. H. Annex, R. J. Laham, N. Kleiman, T. Henry, H. Dauerman, J. E. Udelson, E. V. Gervino, M. Pike, M.J. Whitehouse, et al. Pharmacological Treatment of Coronary Artery Disease With Recombinant Fibroblast Growth Factor-2: Double-Blind, Randomized, Controlled Clinical Trial Circulation, February 19, 2002; 105(7): 788 - 793. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
