DOI: 10.1148/radiol.2451060530
(Radiology 2007;245:258-266.)
© RSNA, 2007
Large-Vessel Distensibility Measurement with Electrocardiographically Gated Multidetector CT: Phantom Study and Initial Experience1
Jie Zhang, PhD,
Joel G. Fletcher, MD,
Terri J. Vrtiska, MD,
Armando Manduca, PhD,
Jess L. Thompson, MD,
Madhavan L. Raghavan, PhD,
Robert J. Wentz, BA, and
Cynthia H. McCollough, PhD
1 From the Department of Radiology, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905 (J.Z., J.G.F., T.J.V., A.M., J.L.T., R.J.W., C.H.M.); and Department of Biomedical Engineering, University of Iowa, Iowa City, Iowa (M.L.R.). From the 2005 RSNA Annual Meeting. Received March 24, 2006; revision requested May 24; revision received January 10, 2007; accepted February 20; final version accepted March 16. Supported by the Flight Attendant Medical Research Institute as part of the ECG-gated Multi-detector CT of Aortic Distensibility Project.
Address correspondence to J.G.F. (e-mail: fletcher.joel{at}mayo.edu).
 |
ABSTRACT
|
|---|
The purpose of this study was to prospectively examine vessel distensibility measurements by using electrocardiographically gated multidetector computed tomography (CT) in a phantom compared with measurements by using a digital camera and to examine feasibility in humans. Large-vessel phantoms were constructed, using a pulsatile flow pump, and absolute diameter and percentage diameter changes during pulsation were measured. After institutional review board approval and patient consent were obtained, the abdominal aorta of four patients was scanned with an electrocardiographically gated CT protocol. The mean difference in percentage diameter change between CT and optical measurements for the phantom ranged from –0.47% to 0.14%. The range of area changes in five locations along the abdominal aorta in four patients was 2.97%–37.16%. Findings of this study indicate that electrocardiographically gated CT angiography data reconstructed across cardiac phases permit measurements of large-vessel distensibility in a phantom model and that vessel distensibility measurement in humans may be possible.
© RSNA, 2007
|
INTRODUCTION
|
|---|
Vessel distensibility is an important functional parameter and is an early marker of future atherosclerotic disease. Furthermore, it is associated with hypertension, coronary events, and stroke (1–3). The accurate measurement of vessel distensibility also may permit earlier and more accurate risk stratification for coronary heart disease in patients with hypertension (4) and prediction of abdominal aortic aneurysm rupture (5).
Arterial distensibility is the relative diameter change (or 
D/Ddias, which is the difference in vessel diameter during diastole and systole [D], divided by the vessel diameter in diastole [Ddias]) divided by blood pressure change in the vessel (or P, which is calculated by subtracting diastolic blood pressure from systolic blood pressure). It can be expressed as D/(P · Ddias) (6). Intraarterial pressure can be obtained invasively or estimated noninvasively by using peripheral pressure measurements. Consequently, accurate measurement of the changes in diameter or cross-sectional area of the vessel is essential in vessel distensibility estimation. The morphologic changes of vessel systems have been observed by using radiography, magnetic resonance imaging, and ultrasonography (US) (7–9). However, the use of US is limited to the larger and more accessible arteries, and it is also limited due to interobserver variability, transducer angle, and, in the abdomen, obscuration from bowel gas or lung.
Recent advances in electrocardiographically gated multidetector computed tomographic (CT) technology have made possible four-dimensional imaging of the aorta and its branches (over time) so that multiple temporally resolved CT angiographic data sets can be created during multiple phases of the cardiac cycle. By using a 64-section CT system and partial scan reconstruction techniques, each CT angiographic data set represents vascular motion during 0.165 second of the cardiac cycle. Multiple, temporally overlapping CT angiograms of the vessels obtained during the cardiac cycle allow pulsation to be seen visually and can be examined to obtain measurements of vessels in their distended and nondistended states in diastole and systole. Because of the high attenuation of intraarterial pixels during CT angiography, electrocardiographically gated CT angiography may provide a time-averaged image of maximal vessel excursions in systole. Thus, the purpose of our study was to prospectively examine vessel distensibility measurements obtained by using electrocardiographically gated multidetector CT in a phantom compared with measurements obtained with a digital camera and to examine feasibility in humans.
|
MATERIALS AND METHODS
|
|---|
Study Description
To evaluate vessel distensibility measurements by using multidetector CT, a computer-controlled phantom was constructed with a pulsatile flow pump and a distensible tube (Fig 1). During the experiment, a sinusoidal pulse (waveform shown in Fig 1) was set to flow through the tube at a frequency of 1 Hz. For this experiment, the phantom was used, in which a series of four tubes of varying compliance—silicone tube (outer diameter, 7.89 mm), latex tube (outer diameter, 7.38 mm), synthetic compliant translucent tube (10) (outer diameter, 8.63 mm), and balloon tube (outer diameter, 7.24 mm)—mimicked large vessels. The tubes were scanned by using retrospective electrocardiographic gating and a vascular CT angiographic protocol.
To demonstrate clinical feasibility, a pilot study was performed to analyze the cross-sectional area changes during a cardiac cycle in different locations along the abdominal aorta in four patients.
Phantom Study
CT imaging.—The experiments were conducted by using a 64-section CT system (Sensation 64; Siemens Medical Solutions, Forchheim, Germany), which has a spatial resolution of 0.4 mm in the x-, y-, and z-axes; 32 detector rows; and dual z-flying focal spot technology (10). The series of tubes were filled with water, suspended in the air, and placed in the isocenter of the scanner. Data were acquired by using a retrospective electrocardiographically gated vascular CT angiographic protocol. Parameters included a 5-cm field of view, 800 effective mAs (effective tube current–time product is tube current–time product divided by pitch), 0.2 pitch, detector collimation of 32 x 0.6 mm, and a gantry rotation time of 0.33 second. A partial reconstruction algorithm was used, and the temporal resolution at the center of the CT scanner was therefore 0.165 second. An electrocardiographic generator (Lionheart 1; Bio-Tek Instruments, Winooski, Vt) was used to generate an electrocardiographic signal fixed at 60 beats per minute for the CT scanner, and this value corresponded to the pulsatile rate generated by the flow pump.
CT images, which were reconstructed to correspond to every 5% of the R-R interval, were reconstructed by using a B40f kernel for phases from 0% to 95% of the R-R interval and had a section thickness of 0.6 mm. The B40f kernel is a medium-sharp reconstruction kernel routinely used in imaging of the abdomen. With this process, 20 static images of the phantom were created at each scan level.
A radiopaque marker was placed behind the tube during CT scanning to allow the correlation of CT and optical diameter measurements at the same location.
With mathematical software (MATLAB; MathWorks, Natick, Mass), we measured the outer diameter of the tubes at each time by using cross-sectional images at the location of the radiopaque marker. For every phase, the full width at half maximum of the tube was measured (J.Z.) by using the mathematical software–based program in a single section. Multiple measurements were made during the cardiac cycle, and maximum and minimum diameters were chosen for analysis. The optical measurements (J.Z.), as described next, were made in the same location and direction.
Digital camera imaging.—The digital camera used to establish the reference standard for this work was a combination of an actively-cooled charge-coupled–device detector (Orca-ER; Hamamatsu Photonix, Hamamatsu, Japan) and a lens with a 4-mm focal length (TF4DA-8; Fujinon, Saitama City, Saitama, Japan). The camera required manual focus and aperture adjustment. The electronic shutter was controlled by a personal computer with a combination of commercial (Image-Pro Plus; Media Cybernetics, Silver Spring, Md) and custom software. The charge-coupled–device matrix was 1344 x 1024 and produced 12-bit gray-scale images.
The optical measurements and CT measurements were performed within one session. The camera was set up beside the table of the CT scanner. The optical images were acquired as the vascular phantom (pulsating tube) moved out of the imaging plane. Vessel motion (generated by the pulsatile flow pump) stayed the same as it was during CT scan acquisition. The camera control software was configured to provide 12-bit tagged image file format images with spatial resolution of 0.024 mm and temporal resolution of 95 msec (ie, one image per 95 msec).
The optical images recorded during 10 seconds were analyzed by using the commercial software mentioned previously (Image-Pro Plus; Media Cybernetics) by one author (J.Z.). The maximum and minimum diameters of the tube during pulsation were recorded on each image at the location of the radiopaque marker that was used to allow point-to-point correlation of optical and CT measurements.
Pilot Patient Study
Clinical studies were performed as part of a prospective Health Insurance Portability and Accountability Act–compliant protocol and were approved by the institutional review board of Mayo Clinic College of Medicine, Rochester, Minn. All patients had a body mass index of less than 32 kg/m2, underwent abdominal CT angiography of the aorta for clinical reasons, and consented to the use of additional radiation for research purposes (450 mAs for this study vs 350 effective mAs for routine aortic CT angiography).
After informed patient consent was obtained, electrocardiographically gated CT angiography of the abdominal aorta was performed in four patients (three men and one woman; mean age, 67 years ± 8 [standard deviation]; range, 56–72 years) who were scheduled to undergo CT angiography of the aorta for clinical reasons (abdominal aortic aneurysm, n = 3; for renal donation, n = 1).
All patient imaging was performed by using the 64-section CT system used for the phantom experiments. All patients were scanned with 120 kVp, 450 effective mAs, 0.33-second gantry rotation time, 3.84-mm table feed per gantry rotation, 0.2 pitch, and a detector configuration of 32 x 0.6 mm. A dose of 125 mL of iohexol (Omnipaque 350; Amersham Health, Princeton, NJ) was administered intravenously by using a dual injector at a rate of 4 mL/sec, followed by a 30-mL saline flush (total, 155 mL). Electrocardiographic leads were placed on patients in the standard fashion. CT data sets were reconstructed to correspond to every 5% of the R-R interval for phases between 0% and 40% of the R-R interval and every 10% of the R-R interval for phases between 50% and 90% of the R-R interval. With these reformations, multiple temporally overlapping time-resolved CT angiographic data sets were created. Images were reconstructed with a 1-mm section thickness, 0.8-mm reconstruction interval, and a B20f reconstruction kernel (a smoother reconstruction kernel, given the noise generated from attenuation in the scanning of a patient).
After image acquisition and reconstruction, the aorta was segmented by using locally developed software algorithms that incorporated a segmentation and registration kit (Insight Tool Kit; National Library of Medicine, Bethesda, Md) and a comprehensive set of imaging functions (AVW; Mayo Clinic College of Medicine, Rochester, Minn). Because atherosclerotic plaque can make the aortic lumen asymmetric, we chose to measure aortic distensibility in patients by using cross-sectional area and not luminal diameter, which is one-dimensional. The cross-sectional area of the segmented abdominal aorta was automatically measured (R.J.W., J.G.F.) by counting pixels in the transverse plane at four locations (above the esophageal hiatus, 1 cm above the celiac artery, 1 cm below the renal arteries, and between the inferior mesenteric artery and the aortic bifurcation). The combined cross-sectional area of the common iliac arteries 1 cm distal to the aortic bifurcation was measured across phases in a similar manner (Fig 2).
View larger version (86K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2: An example of the segmented aorta and common iliac arteries in a patient obtained by using an electrocardiographically gated CT angiographic protocol at maximal and minimal cross-sectional areas in the transverse plane. IMA = inferior mesenteric artery.
|
|
Data Analysis
Absolute minimum and maximum diameters were calculated (J.Z.) from CT data and optical images for each tube. The differences in absolute and percentage diameter changes between CT and optical measurements were also computed. Absolute diameter changes in millimeters and percentage diameter changes, calculated as [(Dmax – Dmin)/Dmin] · 100 (where Dmax is maximum diameter and Dmin is minimum diameter) from the CT data, were compared with the measurements obtained with the optical reference standard. For the patient data, the number of pixels in the segmented aorta or iliac arteries was measured by using cross-sectional images (R.J.W. and J.G.F.) at each location. For each location, cross-sectional area in the transverse plane was then normalized to the minimum observed cross-sectional area across the R-R interval.
|
RESULTS
|
|---|
Phantom Study
The minimum diameter (Tables 1, 2) of the four tubes from the optical reference standard ranged from 7.24 to 8.63 mm, whereas the maximum diameter measurements ranged from 7.99 to 8.95 mm. The corresponding minimum and maximum CT diameter measurements ranged from 6.54 to 8.40 mm and from 7.23 to 8.69 mm, respectively. An example of CT and optical images of the tube diameter changes during pulsatile flow is demonstrated in Figure 3. With CT diameter measurements, a systematic mean underestimation of the reference standard measurements of 0.36 mm ± 0.23 was observed for minimum diameter measurements and a mean underestimation of 0.38 mm ± 0.25 was observed for maximum diameter measurements. The difference in the absolute diameter changes between optical and CT measurements ranged from 0.01 to 0.06 mm (mean, 0.020 mm ± 0.034). Optical reference standard measurements for tube diameter changes during pulsatile flow ranged from 0.10 to 0.75 mm, with percentage diameter changes ranging from 1.23% to 10.43%. CT measurements of tube diameter changes ranged from 0.09 to 0.69 mm. Percentage diameter changes measured with CT ranged from 1.28% to 10.45%. The mean difference in percentage diameter change between optical and CT measurements was only 0.10%, with a range from –0.47% to 0.14%.
View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a: (a) Balloon tube diameter change and (b) synthetic compliant translucent tube with compliance modeled after human carotid artery, both at minimum and maximum diameter. Electrocardiographically gated CT angiographic images at left reveal diameter changes of 10.45% for balloon tube and 3.49% for translucent tube, respectively. Optical images obtained with digital camera at right reveal diameter changes of 10.43% for balloon tube and 3.63% for translucent tube, respectively.
|
|
View larger version (50K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b: (a) Balloon tube diameter change and (b) synthetic compliant translucent tube with compliance modeled after human carotid artery, both at minimum and maximum diameter. Electrocardiographically gated CT angiographic images at left reveal diameter changes of 10.45% for balloon tube and 3.49% for translucent tube, respectively. Optical images obtained with digital camera at right reveal diameter changes of 10.43% for balloon tube and 3.63% for translucent tube, respectively.
|
|
Patient Study
The percentage increase in cross-sectional area during systole in the aorta and common iliac arteries during the cardiac cycle for each of the four patients ranged from 2.97% to 37.16% (Table 3, Fig 4). The aorta above the esophageal hiatus consistently demonstrated more distensibility than did the aorta at the other locations.
View this table:
[in this window]
[in a new window]
|
Table 3. Relative Change between Maximum Cross-sectional Area during Systole Compared with Minimum Area during Diastole at Each Location
|
|
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4: Area changes in five locations, along abdominal aorta and common iliac arteries (Fig 2), corresponding to 14 phases of the cardiac cycle for four patients. Normalized cross-sectional areas are the relative area of the aorta or common iliac arteries, compared with the minimum area (as a ratio). The percentage increase in cross-sectional area during systole ranged from 2.97% to 37.16% (Table 3). The aorta above the esophageal hiatus consistently demonstrates more distensibility than the aorta at the other locations. Patient 4 had an abdominal aortic aneurysm, which demonstrated minimal pulsation. IMA = inferior mesenteric artery.
|
|
|
DISCUSSION
|
|---|
Our study findings suggest that measurement of vessel distensibility by using an electrocardiographically gated multidetector CT angiographic protocol is accurate with the ideal conditions employed in our phantom. We observed that CT distensibility measurements agreed highly with the optical reference standard measurements. With the ideal experimental conditions employed in the phantom study, the absolute diameter changes that were used to calculate vessel distensibility have a mean difference of 0.02 mm (range, –0.02 to 0.06 mm), compared with the values observed with the optical reference standard.
We used tubes of physiologic size (approximately 8 mm), which corresponded in size to iliac arteries. The aorta is even larger. Results of a previous study in which the echo Doppler technique was employed indicated that the minimum and maximum diameters for the transverse aortic arch were 27.7 and 29.2 mm, respectively, for patients with hypertension and 28.0 and 30.2 mm, respectively, for the control group (11). The corresponding diameter changes during the cardiac cycle were 1.5 mm (5.4%) and 2.2 mm (7.9%), respectively, and these diameter changes are larger than the values for absolute changes that we measured in our phantom model. To improve the precision of our CT measurements, we decreased pixel size to 0.1 mm (to permit oversampling) and used a technique with full width at half maximum to permit partial pixel resolution. We previously examined the precision of measurements in a 5-mm static phantom with similar image contrast by using partial scan reconstruction and the same measurement technique with full width at half maximum and found that 10 measurements had a standard deviation of 0.25 mm, with variation decreasing further as image contrast and diameter increased (12).
Our pilot patient study data showed that the cross-sectional percentage area changes ranged from 2.97% to 37.16% along the abdominal aorta (the corresponding diameter changes were from 1.50% to 20.73%). Because we used partial scan reconstruction to improve temporal resolution, we increased effective tube current–time product to 450 mAs from 350 mAs and limited body size to a body mass index of less than 32 kg/m2 to limit noise. Diameter changes of the magnitude corresponding to our clinical observations were measured in our phantom study, and the findings of that study demonstrated accurate measurement of percentage diameter change that ranged from 1.23% to 10.43%; however, the results with the phantom cannot be extended to the human data due to several factors, principally relating to reduced signal-to-noise ratio and lack of control over heart rate. As software is developed to measure human vessel pulsatility, vessel segmentation will be required, with luminal centerlines generated to permit cross-sectional measurements, and coregistration of luminal centerlines will be needed to allow measurements of changes in cross-sectional areas as the aorta moves translationally through the cardiac cycle.
We measured diameter changes or cross-sectional area during the cardiac cycle. For each cardiac cycle simulated by the pulsatile flow in our experiment, the systolic pulse lasts 400 msec. Because we reconstructed images every 50 msec (5% of the R-R interval), approximately eight sampling images were obtained during this period, with each image representing an average of the tube shape over a 165-msec time window (ie, the temporal resolution of the scanner). What are the theoretic limitations of such an approach? Suppose the original cross-sectional area of a tube is S0, the maximum cross-sectional area due to the pulsatile flow is S0 + Sm, the fluence at the systolic period is a sinusoidal curve with a duration of Ts, and the cross-sectional area of the tube changes periodically at a period of T.
In addition, for the retrospectively reconstructed CT images, the sampling interval is t and the temporal resolution is 
< Ts/2. We assume that each sampling point on the curve is an "average" of the curve in a duration of . Suppose the sampling interval t is sufficiently small, it is possible to estimate the accuracy of the measurement on maximum change of the cross-sectional area. The best measurement of the maximum S can be achieved when the sampling position is exactly on the peak of the curve (ie, when t = Ts/2). The worst measurement of the maximum S will occur when t = (Ts – )/2 or t = (Ts + )/2. On the basis of the waveform of the cross-sectional area, one can obtain the ideal form of the diameter change, which can be expressed as follows:
The best and worst measurement of the maximum diameters would be
and
respectively. The maximum change of diameters (Dmax) is
so the best and worst relative accuracy values are
and
respectively.
Therefore, in our experiment, if the diameter is D0 = 8 mm and the diameter excursion is d0 = 0.8 mm, the relative accuracy of diameter change from the optical camera measurement is 2.0%–7.8%. The relative accuracy of diameter change from retrospective reconstructed CT images would be 6.6%–25.1%, which initially appears unacceptable; however, when one considers the fact that the sampling frequency in CT images can be very high, the lower estimation can thus be applied to CT measurement. Therefore, one can approach 7% accuracy by using a CT scanner with a 165-msec temporal resolution, as long as multiple temporally overlapping CT angiographic data sets are employed. If the dual-source CT scanner with an 83-msec temporal resolution is used, a higher accuracy could be expected.
Compared with the optical diameter measurements, with CT measurements we observed a systematic underestimation in both the diameters. We believe this underestimation principally arises from the fact that, with electrocardiographically gated CT, we are averaging vessel excursion over the temporal window (ie, creating a time-rendered maximum intensity projection). We minimized the measurement error by reconstructing images with temporally overlapping data sets (every 50 msec). The partial scan reconstruction, as well as the method with full width at half maximum we used to measure the boundary gradient (ie, vessel edge) instead of the second derivation of the vessel profile (13), may introduce variations in the object size and CT number measurements (12). In our phantom study, we compared CT diameter measurements with measurements obtained with an optical standard. In humans, vessels are often asymmetric due to aneurysm or plaque, and our cross-sectional measurements are more likely to be more meaningful.
Our phantom study had limitations. First, we assumed that the flow pressure changes were consistent during CT scanning and optical image capture.
Second, the phantom study was performed with ideal conditions to minimize noise and maximize spatial resolution: 800 effective mAs, tubes suspended in air (with an increase in the relative contrast between tubes and surroundings to more than 1000 HU), and 5-cm field of view—all unrealistic scanning conditions for the thoracoabdominal aorta. However, 800 effective mAs may be used in cardiac scanning. More important for our experiment, tubes were suspended in air so that comparisons with optical measurements could be made. To minimize noise in our patient studies, we imaged only patients with a body mass index of less than 32 kg/m2 but were able to obtain clinically acceptable images amenable to segmentation by using 450 effective mAs. Because the table speed must be slowed to a pitch of 0.2 to obtain electrocardiographically gated images with our scanner, only the abdominal aorta could be imaged within a single breath hold.
Third, a small diameter change during pulsation may not be measured accurately by using this technique due to artifacts related to partial scan reconstruction algorithms used to achieve the 165-msec temporal resolution. In recent work, researchers showed that, for a phantom tube with an inner diameter of 8.5 mm and an area of contrast enhancement with an attenuation of 400 HU, the error in diameter measurement that arose from partial scan reconstruction artifacts was 3% in water (12).
Fourth, vessels move within the body (eg, in the aortic root) in a translateral direction (in addition to cross-sectional pulsation). Our phantom did not have translateral motion, and our patient measurements were obtained only in the transverse plane. Furthermore, hyperdynamic translateral motion of an artery (eg, in the right-to-left direction) may create motion artifacts that degrade accurate diameter measurements and, thus, require improved temporal resolution to achieve a greater fidelity to vessel motion.
Finally, our phantom experiment was performed at a heart rate of 60 beats per minute. We chose a heart rate of 60 beats per minute, as patients with heart rates greater than 70 beats per minute currently undergo therapy with ß-blockers to lower heart rate for coronary CT angiography. As the heart rate increases, there would certainly be a point at which a higher rate would make such measurements inaccurate. Advances in CT scanner hardware, such as dual-source CT, have led to a decrease in temporal resolution to 83 msec and should permit more accurate vessel distensibility measurements at higher heart rates. Although 90° segmented partial scan reconstruction would theoretically provide the same advantage, our CT system does not permit this reconstruction technique due to image distortion and artifacts created by beat-to-beat variability.
In a pulsating vessel phantom with ideal imaging conditions and an electrocardiographically gated CT angiographic protocol, images of phantom vessel distensibility were generated that reflected distensibility as measured by using our optical reference standard. We used an electrocardiographically gated CT angiographic protocol to retrospectively reconstruct CT angiographic data sets across the cardiac cycle. Particularly, as temporal resolution of CT scanners improves, we believe that electrocardiographically gated CT angiographic data sets may yield additional functional information about vessel wall distensibility. The accuracy of such aortic distensibility measurements in humans remains to be validated. Future studies are needed to investigate aortic distensibility measurements in humans, particularly as they relate to the risk of aortic aneurysm growth and rupture. Disadvantages of this technique are increased radiation dose (owing to the slower table speed, with a dose increase of 25%–50%, compared with non–electrocardiographically gated CT angiographic protocols), the potential for substantial image degradation due to noise (in large patients), and the necessity of a regular heart rate. Potential clinical applications of this technique include risk stratification for the rupture of an abdominal aortic aneurysm or cardiovascular diseases and estimation of left ventricular afterload.
|
ADVANCE IN KNOWLEDGE
|
|---|
- In a phantom model using ideal conditions, electrocardiographically gated multidetector CT angiography can be used to determine large-vessel distensibility.
|
IMPLICATION FOR PATIENT CARE
|
|---|
- It is possible to observe abdominal aortic pulsatility in humans with clinical conditions by using an electrocardiographically gated CT angiographic protocol, but the accuracy of these measurements with physiologic conditions remains unknown.
|
ACKNOWLEDGMENTS
|
|---|
The authors acknowledge the contributions to the manuscript of Lifeng Yu, PhD, who developed the mathematic model relating sampling frequency and temporal resolution to the accuracy of vessel diameter measurements.
|
FOOTNOTES
|
|---|
Authors stated no financial relationship to disclose.
Author contributions:Guarantor of integrity of entire study, J.Z.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, J.Z., J.G.F., T.J.V., J.L.T., M.L.R.; clinical studies, J.G.F.; experimental studies, J.Z., J.G.F., A.M., J.L.T., M.L.R., C.H.M.; statistical analysis, J.Z., R.J.W.; and manuscript editing, J.Z., J.G.F., T.J.V., A.M., C.H.M.
|
References
|
|---|
- Glasser SP, Arnett DK, McVeigh GE, et al. Vascular compliance and cardiovascular disease: a risk factor or a marker? Am J Hypertens 1997;10:1175–1189. [Medline]
- Laurent S, Boutouyrie P, Asmar R, et al. Aortic stiffness is an independent predictor of all-cause and cardiovascular mortality in hypertensive patients. Hypertension 2001;37:1236–1241. [Abstract/Free Full Text]
- Syeda B, Gottsauner-Wolf M, Denk S, Pichler P, Khorsand A, Glogar D. Arterial compliance: a diagnostic marker for atherosclerotic plaque burden? Am J Hypertens 2003;16:356–362.
- Boutouyrie P, Tropeano AI, Asmar R, et al. Aortic stiffness is an independent predictor of primary coronary events in hypertensive patients: a longitudinal study. Hypertension 2002;39:10–15. [Abstract/Free Full Text]
- Wilson KA, Lee AJ, Hoskins PR, Fowkes FG, Ruckley CV, Bradbury AW. The relationship between aortic wall distensibility and rupture of infrarenal abdominal aortic aneurysm. J Vasc Surg 2003;37:112–117. [CrossRef][Medline]
- O'Rourke MF, Staessen JA, Vlachopoulos C, Duprez D, Plante GE. Clinical applications of arterial stiffness: definitions and reference values. Am J Hypertens 2002;15:426–444.
- Boese JM, Ganten M, Leitermann D. Aortic elasticity measurement using ECG-gated multi-slice computed tomography. Biomed Tech (Berl) 2002;47(suppl 1 pt 1):449–450.
- Krug R, Boese JM, Schad LR. Determination of aortic compliance from magnetic resonance images using an automatic active contour model. Phys Med Biol 2003;48:2391–2404. [CrossRef][Medline]
- Gamble G, Zorn J, Sanders G, MacMahon S, Sharpe N. Estimation of arterial stiffness, compliance, and distensibility from M-mode ultrasound measurements of the common carotid artery. Stroke 1994;25:11–16. [Abstract]
- Flohr TG, Stierstorfer K, Ulzheimer S, Bruder H, Primak AN, McCollough CH. Image reconstruction and image quality evaluation for a 64-slice CT scanner with z-flying focal spot. Med Phys 2005;32:2536–2547. [CrossRef][Medline]
- Gatzka CD, Cameron JD, Kingwell BA, Dart AM. Relation between coronary artery disease, aortic stiffness, and left ventricular structure in a population sample. Hypertension 1998;32:575–578. [Abstract/Free Full Text]
- Zhang J, Fletcher JG, Primak A, McCollough CH. Inaccuracies in the measurement of object size and attenuation resulting from partial scan reconstruction artifacts [abstr]. In: Radiological Society of North America scientific assembly and annual meeting program. Oak Brook, Ill: Radiological Society of North America, 2005; 212.
- Berger P, Perot V, Desbarats P, Tunon-de-Lara JM, Marthan R, Laurent F. Airway wall thickness in cigarette smokers: quantitative thin-section CT assessment. Radiology 2005;235:1055–1064. [Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
