| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Radiology and Nuclear Medicine (S.A.S., S. Winterhalter, S.S., R.G., K.J.W.) and the Department of Pathology (S.E.C.), Universitätsklinikum Benjamin Franklin, and the Institute of Diagnostic Research (M.K., W.S.), Freie Universität Berlin, Hindenburgdamm 30, 12200 Berlin, Germany; and the Department of Radiology, Charité, Humboldt Universität Berlin, Germany (S. Wagner). Received October 9, 2000; revision requested November 17; revision received May 2, 2001; accepted May 9. Address correspondence to S.A.S. (e-mail: s.schmitz@medizin.fu-berlin.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Stagnation thrombi of different thrombus ages (1, 3, 5, 7, and 9 days) were induced in the external jugular veins of 25 rabbits. Direct MR imaging of thrombi was performed by using a fat-saturated T1-weighted gradient-echo sequence (three-dimensional [3D] magnetization prepared rapid acquisition gradient echo) before and 24 hours after intravenous administration of USPIO (particle size, 25 nm; 200 µmol per kilogram of body weight). Thrombus length on 3D reconstruction images was compared with that depicted on a radiographic venogram and with histologic findings (joint reference standard). In addition, T2*-weighted gradient-echo images were acquired and scored semiquantitatively.
RESULTS: The hyperintensity of the thrombus segment depicted on T1-weighted images (thrombus length determined with 3D reconstruction images divided by true thrombus length) increased significantly after administration of contrast medium at a thrombus age of 3 days (0.6 ± 0.4 [SD] to 0.8 ± 0.4; P = .02), 5 days (0.1 ± 0.1 to 1.0 ± 0.1; P < .001), and 7 days (0 to 0.6 ± 0.4; P = .02), but not at an age of 1 and 9 days. No significant change in the thrombus signal intensity was observed on T2*-weighted images.
CONCLUSION: The animal model showed that direct MR imaging of the thrombus improved 24 hours after USPIO administration with a T1-weighted sequence. No improvement was seen with the T2*-weighted sequence.
Index terms: Animals • Contrast media, experimental studies, 907.129412, 907.12943 • Contrast media, magnetic resonance (MR), 907.129412, 907.12943 • Embolism, experimental studies, 907.77 • Iron • Thrombosis, experimental studies, 907.751 • Thrombosis, MR, 907.751 • Veins, MR, 907.129412, 907.12943
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
If there is USPIO uptake by thrombi, this contrast medium might theoretically affect the signal intensity of the MR image. Depending on their stage of organization, thrombi show a high signal intensity on T1-weighted sequences caused by the formation of methemoglobin (2). This feature enables direct three-dimensional (3D) imaging of the thrombus with a maximum intensity projection reconstruction image (3). In addition, the accumulation of hemoglobin degradation products in macrophages (siderophages), which have been shown to invade the margin of thrombi in large numbers during thrombus organization (4), provides a possible explanation for the signal intensity loss observed in the periphery of thrombi on T2-weighted images (5). In a similar way, USPIO particles produce a strong T1 effect, which can be exploited for MR angiography (6), as well as a strong T2 or T2* effect, which is used diagnostically after phagocytosis by the monocyte-macrophage system and is used in imaging the liver and the lymph node (7).
Assuming that USPIO particles diffuse into the thrombus, one may hypothesize that they, like extracellular methemoglobin, increase the T1 signal of the thrombus (T1 hypothesis). Like blood degradation products, such as hemosiderin, which occur either extracellularly or following phagocytosis by siderophages, they may decrease the T2 or T2* signal (T2 hypothesis). Since the degree of endothelium formation and the presence of siderophages depend on the organizational stage of a thrombus, age-related changes in signal may be observed (age estimation hypothesis). The purpose of the experiments in this study was to test these hypotheses in an animal model.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
On the basis of the method proposed by Wessler (8), stagnation thrombi were produced by inducing blood stasis and a hypercoagulable state. A 20-cm 3-F guiding catheter (Angiomed, Karlsruhe, Germany) was introduced into an ear vein to advance a microcatheter with an outer diameter of 0.61 mm into the external jugular vein and position its tip at the level of the fourth or fifth cervical vertebra. For occlusion of the vessel, an embolizing agent, consisting of a 1:1 mixture of an oily iodinated radiographic lymphographic contrast medium (Lipiodol; Byk Gulden, Konstanz, Germany) and a butylcyanoacrylate tissue glue (Histoacryl; Braun, Melsungen, Germany), was injected through the microcatheter with fluoroscopic control at such a slow rate as to enable its adhesion to the vessel wall. Typically, 0.15 mL of the embolizing agent was required. Following occlusion of the vessel, 100 NIH units of bovine thrombin (Sigma-Aldrich; Chemie, Steinheim, Germany) in 0.1 mL of sodium chloride was injected into the external jugular vein in the direction toward the cranium through a second microcatheter.
Following embolizations, 16 of the 25 animals developed swelling of the ear on the embolized side. None of the animals showed abnormal drinking or eating behavior.
Experimental Design
Twenty-five experimental animals were used in which thrombi were induced in a first session at day 0. The animals were assigned to five groups of five animals each and were examined in a second session with radiographic venography and MR imaging (precontrast imaging study) on different days (day 1, 3, 5, 7, or 9) (Fig 1). The groups were designated as D1, D3, D5, D7, and D9, according to thrombus age. Precontrast imaging was followed by injection of the contrast medium, and, in a third session, all animals underwent a second MR imaging examination with the same technique. Postcontrast MR imaging was performed 24 hours after contrast administration on day 2, 4, 6, 8, or 10. The second MR imaging study was followed by histologic examination.
|
Radiographic Venography
Radiography was performed on the cervical veins on the left and right side in two planes following injection of contrast medium into the ear vein through an intravenous indwelling cannula. This procedure was performed for in vivo documentation of the thrombus in the second session according to the previously defined schedule (days 1–9).
MR Imaging
While still anesthetized, the animals were subsequently imaged in a 1.5-T magnet machine (Vision; Siemens, Erlangen, Germany) with a standard knee coil. The anesthetized animals were positioned supine on a U-shaped expanded plastic support. A plastic bag containing water was placed on the ventral side of the neck to improve fat saturation.
T1-weighted images in the coronal plane were obtained by using a 3D magnetization-prepared rapid gradient-echo sequence (9). A variation of this sequence was used to suppress the fat signal by an inherent selective water excitation pulse and the blood signal resulting from the black-blood effect of the sequence. We used a parameter setting for thrombus imaging as suggested by Moody et al (3). The magnetization preparation consisted of an inversion pulse and a 20-msec inversion time followed by a rapid gradient echo for collecting the signal with a repetition time msec/echo time msec of 10.3/4.0, a flip angle of 15°, and 100 partitions. The sequence includes a delay time after the rapid gradient echo, which was set to 1,000 msec to allow for magnetization recovery.
The other parameters were slab thickness, 80 mm; section thickness, 0.8 mm; field of view, 100 x 200 mm; matrix, 128 x 256; pixel size, 0.78 x 0.78 x 0.8 mm; acquisition time, 5 minutes 30 seconds. The two-dimensional source data sets were used to generate 3D maximum intensity projection reconstruction images, with a thickness of 1.5 cm representing a subvolume of the total volume acquired.
T2*-weighted transverse images were generated with a 3D gradient-echo sequence (fast low-angle shot [FLASH]) (10). The inherent bright-blood effect of this sequence makes the venous lumen appear hyperintense relative to fat and muscle. Moderate T2* weighting is used to depict intracellularly accumulated blood degradation products or USPIO particles with low signal intensity. The following imaging parameters were used: 54/18; flip angle, 15°; slab thickness, 80 mm; partitions, 40; section thickness, 2 mm; field of view, 80 x 80 mm; matrix, 256 x 256; acquisition time, 22 minutes 8 seconds; pixel size, 0.31 x 0.31 x 0.2 mm.
Specimen Preparation
After postcontrast MR imaging, the animals were killed with an overdose of xylazine. The xylazine and 20 mL of saline were injected into an ear vein of the thrombus-bearing side to prevent postmortem clot formation in the veins of interest. The external jugular vein and facial vein were prepared, fixed in 10% formalin for 24 hours, and cut into vascular cylinders 3 mm long. After embedding the specimens in paraffin, a 3-µm slice was cut from the caudal end of each cylinder and stained with hematoxylin-eosin.
Reference Standard
The histologic sections were evaluated by a pathologist (S.E.C.), who reported the presence of thrombus in each histologic cut without knowledge of the MR imaging findings. Thrombus makeup was summarized for each group by using established criteria for thrombus age estimation (4). Hard copies of the venograms were evaluated by a radiologist (S.A.S.), who measured the length of the thrombus in the external jugular vein without knowledge of the MR imaging data. The venograms and findings of histologic examinations jointly served as the reference standard. By consensus, a decision was made by the pathologist and radiologist if the thrombus length as determined by means of histologic examination and venography differed by 3 mm or more. In cases of smaller length differences, the result obtained with the venogram was considered to represent the true length, since the histologic result is affected by shrinkage of the specimens during preparation.
MR Image Analysis
The effect of USPIO on thrombus depiction with the T1-weighted sequence was determined by measuring thrombus length on 3D reconstruction images with the T1-weighted magnetization-prepared rapid acquisition gradient-echo sequence. This analysis was performed by the same radiologist at least 1 month after evaluation of the venograms. Only those segments of the thrombus were measured that were definitely hyperintense and could be clearly demarcated from their surroundings. When there was incomplete visualization, only the visible portions were measured. The ratio of the thrombus length with the magnetization-prepared rapid acquisition gradient-echo sequence to the true thrombus length as determined by the reference standard was calculated as a measure of agreement between the reference standard and the magnetization-prepared rapid acquisition gradient-echo sequence and is referred to as the length ratio. A length ratio of 1 expresses complete depiction of the thrombus with the magnetization-prepared rapid acquisition gradient-echo sequence, while a ratio of 0.4, for instance, means that depiction is incomplete and only 40% of the length of the thrombus is visible on the magnetization-prepared rapid acquisition gradient-echo sequence.
The T2*-weighted gradient-echo images were analyzed semiquantitatively by using a representative section characteristic of the individual thrombus and showing an area that could be definitely assessed by the reference standard. When a thrombus could not be distinguished from surrounding blood, the reference standard was used for localizing it. In cases of heterogeneous thrombi, the signal intensity accounting for most of the area or that of the central zone when such thrombi showed a concentric composition was used. Thrombus signal intensity was determined on a five-point scale: 1, no signal, same signal intensity as compact bone; 2, low signal intensity, between 1 and 3; 3, moderate signal intensity, isointense with muscle; 4, high signal intensity, between 3 and 5; and 5, very high signal intensity, higher than the signal intensity of cerebrospinal fluid or, if no cerebrospinal fluid was depicted, higher than the signal intensity of lymph nodes.
Statistical Analysis
An analysis of variance was applied to compare thrombus lengths determined by the reference standard between days. The results of the T1-weighted technique were summarized for graphic representation by calculating mean values with standard deviations for each thrombus age (days). Precontrast and postcontrast image length ratios determined by T1 measurements were compared by using the analysis of variance for repeat measurements. The results of the T2*-weighted measurements were expressed as the median, minimum, and maximum for each thrombus age (days). Since the signal intensity scale was assumed to be nonlinear, a Wilcoxon test was applied to compare pre- and postcontrast image signal intensity values of the overall population (n = 25). No statistical comparison was performed between the individual groups owing to the small number of animals per group (n = 5). Significance was assumed at a level of P < .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
MR Imaging: T1 Effects
The 3D reconstruction images from the T1-weighted magnetization-prepared rapid acquisition gradient-echo sequence already showed thrombi in animals of groups D1, D3, and D5 before contrast medium administration. The thrombi were in part seen as worm-shaped hyperintense structures (Fig 2). No portions of the thrombi were seen on the nonenhanced images in groups D7 and D9 (Fig 3). The mean length ratio calculated for all 25 animals together was as low as 0.1 ± 0.3 (SD) for the nonenhanced T1-weighted sequence.
The images acquired 24 hours after contrast medium administration showed a significant increase in the length ratio from 0.1 ± 0.3 to 0.5 ± 0.5 in all groups taken together (n = 25, P < .001) (Figs 2, 3, 5). Analysis of the animal groups imaged on different days showed that the proportion of the thrombus depicted by the magnetization-prepared rapid acquisition gradient-echo sequence was dependent on thrombus age (Fig 5). A significant increase in the length ratio from the precontrast to the postcontrast MR imaging study value was seen in the following groups: group D3, from 0.6 ± 0.4 to 0.8 ± 0.4 (P = .02); group D5, from 0.1 ± 0.1 to 1.0 ± 0.1 (P < .001); and group D7, from 0 to 0.6 ± 0.4 (P = .02). No significant difference between nonenhanced and enhanced images was found for groups D1 (P = .3) and D9 (P = .4).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Animal Model
Human thrombosis is difficult to simulate in animals (11) because it is multifactorial in origin. The factors involved in its occurrence are described by the so-called Virchow triad: blood stasis, endothelial lesion, and hypercoagulable state. The sequential evolution, the heterogeneous makeup, and the dynamic balance of thrombus apposition, on the one hand, and lysis and organization, on the other hand, pose considerable problems for experimental thrombus induction. Surgical methods, such as the one proposed by Wessler (8) that combines the induction of stasis by means of ligation with the creation of a hypercoagulable state by means of thrombin injection, have been used for experimental investigation of the potential of use of MR imaging in the diagnosis of thrombosis (8,12,13). A drawback of this method is that the healing process occurring in response to the intervention is associated with inflammation of the vessel wall and its environment, which makes differentiation of the thrombus from the wall difficult (13).
An alternative procedure suggested by Erdman et al (12) that avoids inflammation-induced artifacts is the endovascular induction of thrombus by application of a current. However, in pilot experiments performed by our group, the thrombus was destroyed when the electrode was withdrawn. We, therefore, developed a model of endovascular thrombus production by combining the induction of stasis and a hypercoagulable state through catheter embolization and thrombin injection, respectively. In this model, a largely organized thrombus, as shown by nearly complete invasion of the thrombus by mononuclear cells (day 9), develops from static blood interspersed with fibrin threads (day 1) within 9 days.
Pilot experiments with this model have shown that only wall thickening and netlike thrombus residues are present after 11 and 13 days (Schmitz SA, et al, unpublished data, 1999). These thrombi are very similar to human stagnation thrombi in their makeup, but they organize much faster (4,14–16). We attribute the faster organization to species-specific factors, the animals’ young age, the small diameter of the thrombi, the absence of wall lesions associated with phlebosclerosis or valve insufficiency, and in particular to the absence of underlying disease that promotes the development of thrombosis in the animal model. The two latter factors cause and maintain thrombosis in humans.
The stagnation thrombus model has advantages that make it seem to be particularly suitable for MR imaging. By far, the majority of deep leg and pelvic vein thrombi in humans are stagnation thrombi (17). Stagnation thrombi are of particular clinical relevance because they are brittle, and early stages often show no wall adhesion, which is why they often cause embolization (18). Another advantage of this model is that the external jugular vein of rabbits is large enough to be depicted with clinical MR imagers. Occlusion of the vein is tolerated by the animals because there is sufficient venous drainage through the contralateral side. The embolisate provides a reliable point of orientation, which is identified by all examination procedures used (venography, MR imaging, and histologic analysis) and which, together with the influx of the facial vein that is regularly present, allows for the precise anatomic localization of the thrombus in the external jugular vein.
T1 Effects
Application of superparamagnetic iron oxide particles resulted in a significant increase in the thrombus portion depicted on the 3D reconstruction image of the T1-weighted magnetization-prepared rapid acquisition gradient-echo sequence. This effect was statistically significant in all animals taken together (n = 25) and in groups D3, D5, and D7 (n = 5 each). These results demonstrate that USPIO particles are indeed taken up by thrombi. It is highly unlikely that this effect is due to maturation processes, since the thrombus segments of thrombi that are 5, 7, and 9 days old, which have a high signal intensity on nonenhanced images, are so small that they are negligible. It is, therefore, unlikely that intraindividual follow-up by using nonenhanced MR imaging, for instance from day 5 to 6, will show an increase in signal intensity.
The absent T1 effect after USPIO administration in thrombi of group D9 is assumed by the authors to be attributable to endothelialization of the thrombi and to adverse T2* effects. The absent T1 effect after USPIO administration in group D1 may be speculated to be due to a still low concentration of methemoglobin in the thrombus. Furthermore, fresh thrombi adhere to the vessel wall over long stretches, and the contrast medium may thus have more difficulty in reaching them. It is also conceivable that a net outflow of water from the thrombus due to shrinkage prevents diffusion of the contrast agent into the thrombus.
T2* Effects
The T2*-weighted images showed changes of thrombus signal intensity in individual animals after administration of USPIO. However, there was no consistent trend within any particular group (eg, a decrease in signal intensity in all animals of one group). A comparison of thrombus signal intensities on pre- and postcontrast images obtained in all animals taken together did not indicate a statistically significant difference. Thus, the course of T2* signal intensities does not support the hypothesis that there is USPIO uptake by thrombi.
Thrombus Age Estimation
Assessment of thrombus age might yield important clinical information on whether a thrombus will respond to therapy. A potential for identifying different maturation stages of thrombi is suggested by the results of our T1-weighted MR imaging studies, which demonstrated that USPIO uptake occurred in thrombi in groups D3, D5, and D7, but not in very fresh (group D1) and in more organized (group D9) thrombi. Thrombi of groups D1 and D9 might be differentiated by means of the signal intensities on postcontrast T2*-weighted images, on which thrombi of group D9 tended to show a lower signal intensity.
Contrast Medium
Our hypothesis assumes that the high concentration gradient of USPIO between blood and thrombus enables uptake of USPIO particles by thrombi that lack an endothelial layer. This hypothesis is corroborated by the histologic examination findings, which showed that endothelium formation does not begin until day 7. However, complete endothelialization is difficult to demonstrate due to the complex surface of thrombi. The marked increase in the thrombus portions depicted with the magnetization-prepared rapid acquisition gradient-echo sequence in groups D3, D5, and D7 suggests a higher affinity of the contrast medium for the thrombus or parts of it. One may speculate that this affinity is caused by the unspecific binding of the particles to platelets since dextran, which is chemically closely related to carboxydextran, shows such an increased affinity for platelets and induces platelet aggregation following intravenous administration (19).
A problem is posed by the high dose of 200 µmol Fe per kilogram of USPIO, which is by a factor of 2–4 greater than the dosages of other USPIO preparations tested in humans. Also, the long time between contrast medium administration and MR imaging may be an obstacle to using this contrast medium in a routine clinical setting. However, the aim of the present study was merely to demonstrate that there is a contrast effect with a diagnostic potential. The contrast medium preparation, the coating of the particles, the dosage, and the imaging delay after administration may be optimized further.
Practical applications: USPIO-enhanced direct MR imaging of the thrombus combines the advantages of high spatial resolution with isotropic voxels less than 1 mm with those of a high signal from the thrombus. These features and the effective suppression of fat, blood, and flow signals even enable 3D reconstruction imaging with a quality achieved for few other pathologic conditions. Theoretically, direct imaging of the thrombus has more intrinsic advantages than indirect angiographic procedures, such as radiographic venography and flow-dependent MR imaging techniques like time-of-flight and phase-contrast imaging. Direct MR imaging of the thrombus, for instance, demonstrated thrombi in occluded vessels that did not fill at radiographic venography and that are presumably not even depicted by using flow-sensitive MR imaging techniques (Fig 3).
These advantages make USPIO-enhanced direct MR imaging of the thrombus an interesting candidate for diagnosis of problematic cases, such as pelvic vein thrombosis. Sinus or cerebral vein thromboses might be another interesting area for application of this technique, especially for identifying thrombosis of small veins. Since the magnetization-prepared rapid acquisition gradient-echo sequence additionally offers the possibility of cardiac and respiratory triggering, it may also be used for assessing pulmonary artery embolism and coronary thrombosis. Moreover, the observation that fresh thrombi of group D1 showed no uptake of USPIO particles in contrast to those of groups D3, D5, and D7 indicates that absent USPIO accumulation in combination with a second MR technique for thrombus detection might help to identify fresh thrombus portions. Such information may be of interest in establishing the indication for lysis therapy or surgery.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.A.S.; study concepts, S.A.S., S.Wagner, M.K.; study design, S.A.S.; literature research, R.G.; experimental studies, S.Winterhalter, R.G., S.S.; data acquisition, S.W.; data analysis/interpretation, S.A.S., S.E.C.; statistical analysis, S.A.S.; manuscript preparation, definition of intellectual content, and editing, S.A.S.; manuscript revision/review, K.J.W., W.S.; manuscript final version approval, S.A.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Li, J. Salanitri, S. Tutton, E. E. Dunkle, J. R. Schneider, J. A. Caprini, L. N. Pierchala, P. M. Jacobs, and R. R. Edelman Lower Extremity Deep Venous Thrombosis: Evaluation with Ferumoxytol-enhanced MR Imaging and Dual-Contrast Mechanism--Preliminary Experience Radiology, March 1, 2007; 242(3): 873 - 881. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sirol, V. Fuster, J. J. Badimon, J. T. Fallon, J.-F. Toussaint, and Z. A. Fayad Chronic Thrombus Detection With In Vivo Magnetic Resonance Imaging and a Fibrin-Targeted Contrast Agent Circulation, September 13, 2005; 112(11): 1594 - 1600. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
