| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Full Paper |
Department of Radiology, College of Medicine, The Catholic University of Korea, Seoul, Korea
Correspondence: Dr K J Ahn, Department of Radiology, St. Mary's Hospital, #62 Youido-Dong, Youngdeungpo-Gu, Seoul 150-713, Korea
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsDiscussionReferences |
|---|
2 analysis). Re-circulation artefact demonstrated the preservation of the posterior wall contour (19/21), ill-defined margin (19/21), less dark defect (18/21), and no involvement of the common carotid artery (19/21). On the contrary true stenotic defect demonstrated focal loss of posterior wall contour (8/8), sharp margin (8/8), dark defect (8/8), and involvement of common carotid artery (4/8). No significant difference was noted in the size of the defect between the two groups (p>0.05). The specificity of 2D-FT TOF MRA for carotid stenosis was markedly increased after application of above signs. These distinguishing signs are very helpful in differentiating re-circulation artefact from true stenotic defect. |
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
However, various kinds of artefact secondary to patient motion, local field inhomogeneity, slow blood flow, and higher order motion are obstacles to its more widespread use [4, 5]. Of these multiple causes of artefacts, artefactual loss of signal intensity in the posterior bulb of the internal carotid artery is particularly problematic [6, 7]. This artefact is known to occur secondary to vortex flow in the bulb of the internal carotid artery and is referred to as re-circulation artefact, flow separation artefact, or non-laminar flow artefact [8].
Atherosclerotic plaques tend to be located precisely in this region of predicted vortex flow, thus making it very difficult to distinguish this artefact from true stenotic defect, and this re-circulation artefact is often misdiagnosed as true stenotic defect.
The purpose of our study is to find appropriate distinguishing features of re-circulation artefact that would allow its differentiation from true stenotic defect on 2D-FT TOF MRA.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
We used a standard 1.5 T Magnetom Vision Plus MR system (Siemens, Erlangen, Germany) with 15 mT m–1 gradients using radiofrequency transmit receive head/neck coil. The 2D-FT TOF MRA was acquired using 50–60 thin (2–3 mm), contiguous or overcontiguous, axial slices, and a FLASH sequence. Imaging parameters were as follows: repetition time (TR), ms/echo time (TE), ms/number of excitations, 25/9/1; flip angle (FA), 35°; field of view (FOV), 200 mm x 150 mm; slice thickness, 3–0.33 mm (overlapping); matrix size, 134 x 256; imaging time, 3 min and 36 s. To suppress the venous flow, a pre-saturation slab, 40 mm thick, was applied at a 2.5 mm offset cephalad to the slice being imaged. Projection MRA was created from the stack of 2D slices using a maximum intensity projection algorithm.
Contrast angiography was performed via femoral artery catheterization using film–screen technique (Siregraph D2; Siemens, Forchheim, Germany) and digital subtraction angiography (Neurostar plus; Siemens, Forchheim, Germany). Anterior–posterior oblique and lateral projections of the extracranial carotid arteries were obtained for each vessel.
For CE-MRA of the carotid artery, initial 2D FLASH phase-contrast MRA was obtained to guide the position of the 3D volume with following parameters: TR/TE, 80 ms/8 ms; FA, 30°; FOV, 280 mm x 280 mm; matrix size, 192 x 256; slice number, 1; slice thickness, 50 mm; scan time, 17 s.
The imaging slabs were positioned to cover the carotid bifurcation identified in the sagittal FLASH 2D phase-contrast MRA sequence. We injected 0.2 mmol kg–1 MR contrast medium at a rate of 2–2.5 ml s–1, either by hand or using pump injector. After injecting the contrast medium, we injected the same amount of saline to push the remaining contrast medium in the line and venous access system. We used a bolus tracking method to trigger the imaging sequence. A 3D FLASH sequence with a 10 s scan time had the following parameters: TR/TE, 3.2 ms/1.2 ms; FA, 30°; FOV, 260 mm x 160 mm; matrix size, 110 x 256; an effective thickness, 1.67 mm. A coronal 60-mm thick 3D volume was obtained before and after contrast injection and a subtraction image was generated.
Distinguishing features between re-circulation artefact and true stenotic defect
Re-circulation artefact was defined as a filling defect at the carotid bulb on 2D-FT TOF MRA; negative on contrast medium based angiography. True stenotic defect was defined as a definite filling defect on both 2D-FT TOF MRA and contrast medium based angiography. The vessels without filling defect on 2D-FT TOF MRA were excluded from this study to identify distinguishing features between re-circulation artefact and true stenotic defect.
On 2D-FT TOF MRA, we first evaluated whether the posterior wall contour, representing the posterior margin of arterial opacification along the carotid bulb was preserved, and categorized this as "preserved" or "not preserved". The character of the margin of the filling defect was also evaluated, and was categorized as "ill-defined" or "sharp". Next, we evaluated the darkness of the filling defect. We categorized the darkness of the filling defect into "dark" or "less dark". In this categorization, "dark" indicated that the darkness of filling defect was equal to the background darkness, and "less dark" indicated an ambiguous filling defect. Common carotid artery involvement by the filling defect was also evaluated, and categorized as "involved" or "not involved". Finally, the size of filling defect was measured by the longest diameter, and classified as "less than l cm" or "more than l cm".
The imaging findings of contrast medium based angiography and 2D-FT TOF MRA were analysed by two experienced radiologists to compare the findings of filling defects between the re-circulation group and the true stenotic defect group, and the results were adapted by consensus by two radiologists. We used chi-squared (2) analysis to test the statistical difference of each distinguishing feature between the two groups, and regarded them as significantly different when the p value was less than 0.05.
Sensitivity and specificity of 2D-FT TOF MRA
We also evaluated changes in specificity and sensitivity in diagnosing stenosis at the carotid bulb on 2D-FT TOF MRA, before and after employing the distinguishing signs from the previous study. First, diagnostic sensitivity and specificity of carotid artery stenosis were obtained from the analyses of 45 vessels each examined by both 2D-FT TOF MRA and contrast medium based angiography, by two radiologists not involved with and unacquainted with the results of the previous study.
After an interval of 15 days a new diagnostic sensitivity and specificity were obtained when the same radiologists re-analysed the 45 vessels using the signs that distinguished re-circulation artefact from true stenotic defect. The diagnosis was made by consensus of the two radiologists as in the previous study.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
In cases of re-circulation artefact, the posterior wall contour was well preserved in 19 of 21 vessels on 2D-FT TOF MRA (Figures 1, 2
). The remaining two vessels showed an interrupted contour. The vessels with true stenotic defect showed complete discontinuity of the posterior wall contour in all eight vessels (p<0.001, Figure 3).
|
|
|
). The remaining two vessels showed a sharp margin between filling defect and normal blood flow. All eight true stenotic defects showed a sharp margins (p<0.001, Figure 3).
A significant difference was also found in the darkness of the filling defect on 2D-FT TOF MRA in the re-circulation artefact group and the true stenotic defect group (p<0.001). In the case of the re-circulation artefact, 18 of 21 vessels showed ambiguous and less dark filling defects than that observed in true stenotic defects (Figures 1, 2). The remaining three vessels demonstrated dark filling defects. True stenotic defects showed definitely dark defects in all eight vessels (Figure 3).
In five of eight vessels with true stenotic defects, the common carotid arteries were involved by an extension of filling defects. However, in the case of re-circulation artefacts only two of 21 vessels showed common carotid artery involvement (p<0.01). The average longest diameters of the filling defects of the two groups were 0.9 cm and 1.0 cm, respectively (p>0.05).
Sensitivity and specificity of 2D-FT TOF MRA
Diagnostic sensitivity and specificity before employing the distinguishing signs were 64.3% and 45.2%, respectively. When using the diagnostic criteria of the previous study the sensitivity remained unchanged as 64.3%, but the specificity improved to 80.6% (Table 1).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Flow separation and reversal is known to occur in the normal carotid bulb. Some degree of flow reversal is present in nearly all healthy individuals. By making vessels from animals and human post-mortem transparent, Motomiya et al [9] obtained the exact flow patterns and distributions of fluid velocities in the human carotid artery bifurcation. Their results convincingly demonstrated the phenomena of flow separation and the formation of paired spiral secondary flows.
Flow reversal in the normal carotid bifurcation was also well documented by Doppler colour flow imaging: Middleton et al [10] identified areas of flow reversal in all but one bifurcation among 50 healthy volunteers. According to this study, flow reversal was typically located peripherally, just proximal to the bifurcation, and opposite the origin of the external carotid artery. On average, flow reversal occupied 33% of the lumen of the carotid bulb and extended for a distance of 14 mm. The average duration of flow reversal was 22% of the total cardiac cycle, with reversal typically occurring at peak systole.
Because atherosclerotic plaques tend to be located precisely in the region of predicted flow reversal, this area of the carotid bifurcation is one of the most frequent sites of atherosclerotic stenosis. Consequently, this re-circulation artefact may cause false positive results on 2D-FT TOF MRA.
Heiserman et al [6] suggested that cases of artefactual loss of signal intensity in the posterior bulb of the internal carotid artery were characterized by a relatively mild decrease in the signal intensity and smooth borders whereas, atherosclerotic lesions in this location usually demonstrate a more marked loss of signal intensity and possess less regular margins.
We observed and analysed the marginal character of filling defects from the viewpoint of sharpness (ill-defined or sharp) rather than marginal irregularity. We thought that the categorization of filling defects into ill defined or sharp was more objective than categorization by marginal irregularity in depicting the character of filling defect margin. The results of Heiserman et al are in keeping with our results. In the present study, the vessels with true atherosclerotic stenotic defects showed a definitely dark defect and sharp margin, whereas re-circulation artefact showed ambiguous, less dark defect and ill-defined margin. We believe that these differential points are very useful for discriminating re-circulation artefact from true stenotic defect.
We also examined several additional differential points: (1) preservation of the posterior wall contour adjacent to the filling defect; (2) involvement of the common carotid artery; and (3) size of filling defect. The first and second items showed statistically significant differences between re-circulation artefact and atherosclerotic true stenotic defect. However, there was no significant difference in the size of filling defect.
While the posterior wall contour of the carotid bulb was maintained without break or loss in the case of re-circulation artefact, it was not preserved in the presence of a true stenotic defect. In atherosclerotic stenosis the filling defects on 2D-FT TOF MRA develop secondary to atheromatous plaques, and atheromatous plaques are firmly attached to the inner wall of the internal carotid artery. Consequently, the filling defects caused by atheromatous plaques are not able to preserve the posterior wall contour. There should be a break or a loss of the posterior wall contour at the attachment site of the atheromatous plaque.
However, in case of re-circulation artefact, although the vortex flow in the carotid bulb can reduce the blood flow signal intensity, it can not eliminate it completely in the presence of blood flow through the carotid bulb. Thus, the inner wall contour adjacent to the filling defect is preserved, and the signal intensity of the filling defect caused by re-circulation artefact may be less dark than that of a true stenotic defect.
Moreover, the margins of filling defects in re-circulation artefact appeared to be ill defined, whereas those of true stenotic defect appeared to be sharp. If one considers that vortex flow is a whirling phenomenon, the cause of an ill-defined margin associated with re-circulation artefact on 2D-FT TOF MRA can be explained.
In cases of atherosclerotic stenotic defects, the filling defect showed a tendency to extend to the common carotid artery. However, filling defects secondary to re-circulation artefact were confined to the carotid bulb of the internal carotid artery. As previously described, re-circulation artefact developed by vortex flow in the carotid bulb is due to the specific anatomical shape of the carotid bulb. Thus, the filling defects caused by re-circulation artefact have to stay in the bulbous portion of carotid bulb.
After educating the two radiologists on the four signs that distinguished re-circulation artefact from true stenotic defect, diagnostic sensitivity remained unchanged while specificity markedly improved. In other words, by identifying re-circulation artefacts, which are often misdiagnosed as true stenotic defect, we were able to accurately identify false positive cases, resulting in our increased specificity.
Although the diagnostic specificity of 2D-FT TOF MRA is improved by using our distinguishing signs, its low sensitivity in actually detecting stenosis makes it an inadequate replacement for CE-MRA as a screening technique. CE-MRA has many advantages over 2D-FT TOF MRA: substantially reduced patient motion due to short scanning time, increased coverage of a large volume with excellent contrast-to-noise properties, and the availability of very short TEs. Thus, 2D-FT TOF MRA may be used in limited situations where CE-MRA can not be used, for example: using MR equipment that can not perform CE-MRA, patients that are allergic to MR contrast media, patients with poor intravenous access, pregnant women, and patients with severe renal insufficiency.
We used CE-MRA as a reference study, like digital subtraction angiography, for evaluating the presence of stenosis at carotid arteries. According to previously reported studies on CE-MRA, the sensitivity and specificity of maximum intensity projection images are over 93% and 85%, respectively, compared with conventional angiography as the reference standard and by using a 70% threshold for internal carotid arterial diameter stenosis [11]. The sensitivity is very high, and although the specificity appears somewhat low, this is due to overestimation of the stenosis by spin dephasing at the stenotic and post-stenotic sites. We used CE-MRA only to determine the presence of a stenosis rather than to grade the stenosis, enabling us to use CE-MRA as a reference study to determine the presence of stenosis.
In conclusion, the presence of re-circulation artefact in the bulb of the internal carotid artery on 2D-FT TOF MRA is indicated by the preservation of the posterior wall contour, an ill-defined margin, a less dark filling defect, and no involvement of the common carotid artery. Theses findings can help discriminate re-circulation artefact from a true stenotic defect on 2D-FT TOF MRA.
Received for publication March 7, 2003. Revision received February 9, 2004. Accepted for publication March 30, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
