| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Ohio State University College of Medicine, Columbus,
OH.
2 Department of Pediatric Cardiology, University Children's Hospital, Im
Neuenheimer Feld 153, 69120 Heidelberg, Germany.
3 Department of Pediatric Radiology, University Children's Hospital, Heidelberg,
Germany.
4 Heart Center, Columbus Children's Hospital, Columbus, OH.
5 Division of Cardiology, Davis Heart and Lung Research Institute, Ohio State
University College of Medicine, Columbus, OH.
6 Children's Radiological Institute, Columbus Children's Hospital, Columbus,
OH.
Received September 21, 2007;
accepted after revision November 12, 2007.
J. G. Eichhorn was supported by a postdoctoral research grant from the Max
Kade Foundation, Inc., New York, NY.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Ten steel stents of varying designs and sizes commonly used in the treatment of congenital heart disease were implanted in polyvinyl chloride (PVC) tubes and three of the 10 stents were partially obstructed with wax by filling 25% (mild) to 60% (moderate) of the lumen with contrast material. On a 64-MDCT scanner, the stents were scanned at tube voltages (kVp) of 80, 100, and 120 and at tube currents (mA) of 40, 80, 120, and 160. CT measurements of inner-stent diameter, strut thickness, and percent lumen (in-stent) stenoses were compared with biplane fluoroscopy of digital angiography.
RESULTS. The stent diameter and percent stenosis on all CT images were consistently smaller than measured on digital angiography but were highly correlated (r = 0.97; p < 0.0001) with improvement as stent diameter increased (93% agreement with digital angiography for 4-mm stent, up to 99% for 25-mm stent; p = 0.001). Moderate stenosis could be assessed better than mild stenosis (99% vs 91% agreement with digital angiography; p = 0.003). Increasing exposure settings improved CT correlation of all measurements for mA up to 120 and kVp up to 100 (98.1% agreement). Higher settings did not improve accuracy (93.9% for 160 mA at 120 kVp; p = 0.03).
CONCLUSION. CT is feasible to assess lumen narrowing of pediatric vascular stents at a wide range of tube settings. The study suggests that it is possible to lower the radiation exposure settings without loss in image quality or accuracy in detecting in-stent stenoses.
Keywords: congenital heart disease • MDCT • pediatrics • phantom • radiation exposure • stents
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MDCT is helpful in planning optimal stent placement by showing the
orientation and extent of stenoses, and it is also helpful in the follow-up
evaluation of stent patency [6,
7]. In contrast to sonography
or MRI, MDCT is surprisingly free of artifacts from the metallic component of
stents [8,
9], which allows the lumen of
the stent to be assessed for in-stent stenosis
[6]. The ability to diagnose
in-stent stenosis on MDCT was evalu ated in a first study of children with
pulmonary artery and aortic stents who also underwent conventional digital
angiography [9]. MDCT
correlated well with digital angiography for all grades of stenosis. Using a
threshold of approximately 
20% stenosis, the sensitivity and specificity
of MDCT exceeded 95% [9].
One of the main limitations of CT is exposure to ionizing radiation. In infants and children, this is of particular concern because of their greater sensitivity to radiation and their longer life span, which increases the risk of developing radiation-induced cancer [10, 11]. The optimal radiation dose settings for performing cardiac CT in infants and pediatric patients are still being established. Preliminary work indicates that diagnostically adequate cardiac CT scans can be obtained at much lower radiation doses than commonly quoted in the literature [9, 12–14].
The purpose of this study was to evaluate the effect of varying tube voltages and currents on the assessment of stent lumen diameter and simulated in-stent stenosis in comparison with digital radiographic angiography in an in vitro model.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Imaging
The phantom was imaged in the helical mode with a 64-MDCT scanner
(Sensation Cardiac 64, Siemens Medical Solutions) with the following settings:
1-mm slice thickness, 319-mm field of view, 0.59 x 0.59 mm pixel
spacing, helical pitch of 1.0, 0.36 s–1 rotation speed, and
B31f (soft) and B70f (sharp) reconstruction kernels in a nearly axial
orientation. Dose settings were 80, 100, and 120 kVp at tube currents of 40,
80, and 120 mA for each kVp. An additional scan at 120 and 160 mA was
performed. Altogether, the phantom was scanned 10 times with 10 varying tube
settings. The phantom was also imaged using angled bi plane fluoroscopy as
would be performed in con ventional digital angiography, with a 1,024 x
1,024 pixel matrix (anteroposterior and 90° lateral projections).
Quantitative Analysis
With software for 3D image display and analysis (VGStudio Max version 1.2,
VolumeGraphics), transverse and longitudinal (perpendicular to each other)
reformations were reconstructed from the CT data at slice thicknesses of 0.5
or 0.6 mm. To improve delineation of the stents, the images were displayed in
a zoom mode with a window width of 1,020 H and window level of 320 H.
Using electronic calipers, two observers measured independently the cross-sectional stent (single) strut diameters (n = 100 for each observer; 10 stents scanned 10 times with varying tube settings), the inner stent size (n = 100), and the residual lumen diameter in cases of simulated instent stenosis (n = 30; three in-stent stenoses were scanned 10 times). The diameters were measured from cross-sectional images obtained perpendicular to the long axis of the stent (Fig. 2). The measurements of the two observers were averaged for statistical analysis and graphical display of the data.
|
Statistical Analysis
Descriptive statistics were presented with means and SDs. Pearson's
correlation coefficients and simple linear regression models were fitted to
study the linear relationships between the MDCT and digital angiography
variables. In comparing the results of a wide range of stent sizes and tube
settings, the assessed diameters (stent strut, original, and residual stent
lumen) were lined out as the ratio between MDCT and digital angio g raphy
measurements and were given as percent agreement. A two-sided paired Student's
t test was used to compare the stent sizes measured using CT images
(interobserver variability and variability between the varying technical
settings and CT scanners). A value of p < 0.05 was considered
statistically significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The diameters (inner stent size and residual stent lumen) measured by MDCT and digital angiography showed a high correlation with a linear relationship (r = 0.97; p < 0.0001). The inner stent diameters measured on MDCT images were consistently and significantly smaller than those measured on digital angiography (mean agreement, 96.5%; SD, 3.4%). Correspondently, the stent strut was significantly thicker on CT (mean, 200.7%; SD, 11.1%) than measured on digital angiography. There was improvement in agreement between CT and digital angiography measurements as the stent diameter increased (92.6%, SD of 2.1% for the 4-mm stent up to 99.3%, SD of 1.2% for the 25-mm stent; p = 0.001).
The moderate in-stent stenoses (98.7% mean agreement with digital angiography, SD of 2.8%) could be assessed better on CT than mild stenoses (mean, 90.7%; SD, 5.7%; p = 0.003) (Fig. 3). The sharp reconstruction kernel had better agreement for all measurements than the soft kernel: inner stent diameters (soft vs sharp kernel: 93.1% vs 96.5%; p = 0.04) and stent strut thickness (218.3% vs 200.7%; p = 0.01).
|
Comparing the different tube settings (Fig. 4), there was an improvement in agreement with an increase of tube current up to 120 mA and of tube voltage up to 100 kVp (agreement with digital angiography of 98.1%; SD, 2.4% (Fig. 5). Increasing tube current to 160 mA showed no further improvement (93.9% agreement, SD of 1.7% for 160 mA combined with 120 kVp (Fig. 5) but resulted in an increase of artificial lumen narrowing caused by an increase of strut thickness up to 10%. These results correlated with a better assessment of mild stenoses (Fig. 3) using 100 kVp (94.6% agreement; SD, 3.8%) than 80 kVp (89%; SD, 5.6%; p = 0.03) or 120 kVp (mean, 87.5%; SD, 5.3%; p = 0.003) using varying tube currents. Furthermore, the improvement in agreement of 100 kVp (97.7%; SD, 3.1%) versus 80 (96.1%; SD, 2.4%; p = 0.05) and 120 kVp (mean, 94.7%; SD, 3.6%; p = 0.03) could also be shown using a fixed tube current of 80 mA for measuring inner stent and residual lumen diameter.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The assessment of stent lumen diameters and in-stent stenoses using MDCT was highly correlated with digital angiography for all stent diameters (4–25 mm). Both mild and moderate simulated stenoses were identified accurately (with more than 90% for the mild and up to 99% agreement for moderate stenosis). Better differentiation between mild and moderate stenosis has potential therapeutic importance whether there is a subsequent interventional or surgical approach or not.
The difficulty in detecting mild stenoses can be explained by partial
volume averaging artifact from the metallic vascular stent struts that can
partially (or totally) obscure the stented vessel lumen on MDCT
[8,
9]. Previous in vitro and in
vivo studies showed that the stent lumen could be visualized and stent patency
assessed. However, due to artifacts arising from metallic stent struts,
nonocclusive instent stenoses were obscured
[18,
19]. In a study of children
with CHD, MDCT had a low sensitivity for detecting mild (0–30%) in-stent
stenoses but detected all moderate and higher stenoses seen by digital
angiography [9]. The
sensitivity reached more than 98% for stenoses 20% without any
differences between higher grades, which means that 50% of stenoses could be
assessed as well as 70% with a sensitivity of nearly 100% for pediatric
vascular stents with 7-mm median diameter
[9].
Theoretically, the artificial lumen narrowing is approximately 60% for coronary stents at 4-MDCT [18], decreases to 25% at 16-MDCT [20], and further with later-generation MDCT scanners [21–23]. Nevertheless, small-diameter stents (< 3 mm) are not assessable very well with 64-MDCT and stent type is very important in the ability to visualize the lumen diameter. For stents with a diameter more than 3 mm, assessment appears to be feasible, and in-stent stenosis was correctly identified with an acceptable sensitivity (42–99%) and good specificity (88–98%) [24, 25].
In previous studies of stent imaging concerning mostly coronary stents in the adult population, radiation exposure and tube settings were not a major issue. Consequently, the commonly used tube voltage was 120 kVp or higher and the tube currents had a wide range of 80–250 mA (effective tube current–time of 680 mAs). The radiation exposures to patients from cardiac CT in adults at these settings are reported to be between 9 and 16 mSv for examinations without dose modulation [26].
The results of our study suggest that higher radiation exposure settings do not correlate automatically with better image quality. In fact, a tube voltage of 120 kVp resulted in an increase in artifacts arising from the stents. This resulted in decreased accuracy compared with digital angiography that we hypothesize is due to greater beam-hardening artifact and greater partial volume averaging. We found that lower kVp of 80 to 100 and mAs of 120 yielded results as accurate as those of higher radiation exposure settings. Of course, higher radiation exposure settings may be necessary in adult patients to penetrate the soft tissues overlying the stent, not just the lungs as in our in vitro model. Because of the reduced X-ray attenuation secondary to lower body mass in children [13, 14], the data from our model would be expected to simulate more closely actual scanning conditions. Recently described cardiac CT protocols for children use a kVp of either 80 or 120 with mA values between 25 and 140 related to the patient's weight [27]. In patients with smaller body habitus (< 50 kg), the studies could be performed at 80 kVp [12]. Such protocols result in a considerable reduction in radiation dose of approximately 30% at 4-MDCT [12] and 60% at 16-MDCT (80 kVp approach compared with the standard 120 kVp protocol) in children [28]. Siegel et al. [13] showed that there is no appreciable difference in image noise in the infant-sized phantoms at the 80- and 120-kVp settings. The image contrast increased with reduction in tube voltage (this effect was greatest in smaller phantoms).
Different Methods of Stent Imaging
At present, conventional digital angiography performed during cardiac
catheterization remains the reference standard for luminal assessment after
stent placement. The disadvantages of digital angiography are that it is a 2D
technique, more invasive, requires longer sedation times, and exposes the
patient to higher doses of ionizing radiation
[9,
29]. The dose range for
pediatric procedures is obviously wide and varies from approximately 5 mSv to
more than 20 mSv for complex interventional approach
[28,
29]. Doppler echocardiography
is limited by field of view, which especially affects visualization of
extracardiac structures and is susceptible to metallic artifacts. The chief
advantage of MDCT over MRI in stent evaluation is the significant signal
dephasing due to inhomogeneities induced in the magnetic field by the metallic
struts [8,
30]. In comparison with
studies using MRI for evaluating in-stent stenoses, MDCT is superior in most
types of stent materials, comparable to MR angiography in nitinol stents, and
inferior only in the new tantalum stents
[8]. Furthermore, the spatial
resolution capabilities of MRI are limited by the strength of gradients.
Current MRI scanners are already operating at the fastest gradients possible,
whereas further improvements with MDCT have been anticipated with the arrival
of more slice detectors.
Study Limitations
Our phantom was designed to simulate pulmonary arteries, where many of the
stents are implanted to treat children with CHD. Nevertheless, there are
limitations to our phantom model. The effects of cardiac and respiratory
motion and the surrounding chest wall soft tissues were not simulated during
the experiments. In addition, the stents were orientated parallel to the
z-axis of the MDCT scanners. This orientation would be expected to
result in less partial volume averaging and improved results than would occur
in vivo in obliquely oriented stents
[22].
In conclusion, this study supports the feasibility of using MDCT to image intravascular stents in pediatric patients because of a remarkable lack of artifacts at a wide range of kVp and mA settings with accuracies up to 99% of digital angiography, the reference standard, in determining stent diameters and, most important, in detecting both mild and moderate in-stent stenoses.
An important finding was that a kVp lower than 120 did not result in increased streak artifacts but actually improved accuracy provided that a minimum tube current of 80 mA was applied. This suggests that it may be possible to lower the radiation exposure settings without loss in stent image quality.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. R. Dillman and R. J. Hernandez Role of CT in the Evaluation of Congenital Cardiovascular Disease in Children Am. J. Roentgenol., May 1, 2009; 192(5): 1219 - 1231. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= 80, 
= 100, and 
= 120 kVp.
For moderate stenoses, there were no significant differences found between kVp
groups (p values between 0.6 and 0.9). For mild stenoses, there was
significant improvement for 100 kVp versus 80 (p = 0.03) and also
versus 120 kVp (p = 0.003). Mean and SD of all tube voltages are
added.
= 120
kVp.