| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Inje University Ilsanpaik Hospital, Daewha-dong,
Ilsan-seogu, Goyang-si, Gyuanggi-do 411-706, South Korea.
2 Department of Radiology, Inje University Ilsanpaik Hospital, Seoul, South
Korea.
3 Department of Cardiology, Inje University Ilsanpaik Hospital, Goyang-si,
Gyuanggi-do 411-706, South Korea.
Received September 8, 2006;
accepted after revision January 26, 2007.
Address correspondence to G. Hur
(ghurster{at}gmail.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. One hundred patients (59 men and 41 women) who were suspected to have coronary artery disease underwent coronary CTA using a 64-MDCT scanner. In addition to clinical studies, we also performed measurements on water phantoms. Tube current was modulated by the SD of the CT attenuation values measured from the left atrium on unenhanced images scanned at 300 mA. A modulation table was created from data obtained from the studies of water phantoms scanned at various tube currents. Other scanning parameters were identical to those used to obtain unenhanced and contrast-enhanced studies of the 100 patients. The SD values were measured from images scanned at an adjusted tube current, and the images of normal coronary and internal mammary arteries were graded. Radiation doses measured using the volume CT dose index (CTDIvol) were compared between the SD of the CT attenuation values and the modulation parameters suggested by the manufacturer of the scanner.
RESULTS. Image quality was rated as grade 3 (low mottle) on a 4-grade scale by four observers for 92-94 of the 100 patients (average, 92.5%). The mean SD value at an adjusted tube current was 12.1 H with an SD of 0.758 H (target SD = 12 H). A radiation dose reduction of 9-45% was achieved in patients grouped by weight who weighed less than 70 kg, and a reduction of up to 71% was seen in individual cases.
CONCLUSION. Modulating tube current using the SD of CT attenuation values from the left atrium is a highly reliable method of achieving uniform image quality in coronary CTA.
Keywords: cardiac imaging • coronary CT angiography • CT technique • radiation dose • tube current modulation
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Optimization occurs when a diagnostic-quality image is obtained at the lowest radiation dose possible and can be best accomplished in coronary CTA using one of two methods of tube current modulation: The first way is to use attenuation information and the second is to use information from the cardiac cycle (ECG) as modulating parameters. The latter is used mainly for radiation reduction, and the former is used for uniformity of image quality and for radiation reduction in smaller patients (< 65 kg). Ideally, the attenuation modulation parameter should accurately reflect the attenuation value of the imaged body part under actual scanning conditions and parameters. The method chosen also should be practical and easy to use.
Attenuation information can be divided into direct and indirect categories depending on the nature of the information. Indirect information includes body weight, height, circumference, and dimensions that can be obtained from either evaluating the patient or obtaining a topogram. Tube current modulation using indirect information is more effective than using a fixed tube current, but it is less effective than methods that use direct attenuation information [9-15]. Direct attenuation values reflect the tissue type (bone, fat, fluid) and the contents within organs (air, fluid) at a specific level. Angular and z-axis automatic tube current modulation, used by many vendors of CT units, and attenuation data, used by Irie and Inoue [16] for abdominal CT, are examples of modulating tube current on the basis of direct information. The topogram has been the main source of such information, and this type of information is considered to be more accurate than the indirect type [16-19].
|
|
|
Our initial experiences implementing the protocols devised by the manufacturer of the CT units, which use either a fixed tube current of 400 mA or a tube current from a tube current table that lists tube currents on the basis of the patient's body weight, produced images with a wide variation of noise levels because of discrepancies between body size and attenuation values of the scanned region (Figs. 1A, 1B, and 1C) and motivated us to design a new method for more precise modulation of tube current. For our study, automatic tube current modulation was not included in the protocol, and ECG modulation was not available.
The goal of this study was to evaluate a new tube current modulation method that uses direct attenuation information from reconstructed images as the modulation parameter by measuring the SD of the CT attenuation values from the left atrium. Unlike the attenuation values obtained from topograms, such as those obtained from automatic tube current modulation, the SD of the CT attenuation values from reconstructed images reflects complex scanning parameters, including ECG gating and body composition. Therefore, a more accurate tube current modulation can be achieved for uniform image quality.
To use the SD values as a parameter, we needed to create a tube current (mA) modulation table using images of phantoms of various sizes obtained with identical scanning parameters before and after contrast administration, including scans obtained with ECG gating and at various tube currents.
This new method of tube current modulation in coronary CTA has not been described elsewhere to our knowledge, and we think that it has advantages over other methods.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ninety-five patients with atherosclerosis (53 men with a mean age of 60.03 years and 42 women with of mean age of 62.23 years) involving the proximal coronary arteries were excluded from visual analysis for two reasons: The first is the difficulty in differentiating between atherosclerosis and noise as the cause of coarseness of vessel walls, and the other is the difficulty in selecting normal vessels of 1.5 mm in diameter in patients with advanced disease. This group was, however, included in the measurement analysis of noise (SD of CT attenuation values at corrected tube current).
Routine preparation of patients for coronary CTA in our institution includes an oral dose of 50-75 mg of atenolol 1 hour before the examination in patients with heart rates of 65 beats per minute (bpm) or greater except for those who have contraindications (e.g., low blood pressure) [20] and a sublingual dose of nitroglycerin (0.6 mg) in all patients before scanning. For each patient, body weight was recorded from a chart but body dimensions were measured using topography (anteroposterior diameter x transverse diameter).
Phantom Study
Water phantoms of four different diameters (180, 245, 300, 350 cm) mounted
on a single frame were scanned using a 64-MDCT scanner (Aquilion 64, Toshiba
Medical Systems) with the same scanning parameters used for CTA before and
after contrast administration except for tube currents, which ranged from 150
to 500 mA in 50-mA increments. An ECG signal of 62 bpm was provided by a
technologist and pitch, which is selected automatically according to the heart
rate, was 0.204. Volume CT dose index (CTDIvol), expressed in
milligrays (mGy), was recorded from the operator console. Images were
reconstructed at a thickness of 3 mm without a gap for scans with unenhanced
parameters and at a thickness of 0.5 mm with 0.3-mm intervals for scans with
enchanced parameters. Four images that did not contain partial volume effects
were selected from the images of each phantom taken at each tube current, and
the SD of the CT attenuation values was measured from the central areas using
a cursor size of 200 mm2. The average values were plotted on charts
(Figs. 2A and
2B). To obtain SD values
between each of the phantom sizes, we calculated median exponential values
using the formula-weighted CT dose index (CTDIw) x
2 
e(µ x
D), which is described in the next section,
Technical Background.
|
|
|
2
w), and body size (D) that can be expressed by
CTDIw x 2 e(µ x
D). This formula was used by Menke
[11], who referred to Nagel
[21], who referred to Brooks
and Di Chiro [22], where
e is the base of natural logarithms (constant) and µ is the mean
attenuation coefficient of the water (phantom) at a defined tube voltage. The
formula indicates that the attenuation value can be calculated by measuring
noise level (SD of CT attenuation values), but this calculation requires
consistent homogeneous fluid of low attenuation at the center of the imaged
body part for accuracy [16,
19,
23].
Tube Current (mA) Modulation Table
To adjust the SD values between the pre- and postcontrast parameters,
300-mA precontrast SD values were subtracted from the postcontrast SD values
and plotted on the precontrast parameter chart
(Fig. 3). The transverse line
with a positive slope (1.5°, line A on
Fig. 3) represents the
increasing differences of SD values between unenhanced and contrast-enhanced
images as the phantom size increases from 180 cm (SD = 0.5 H) to 300 (SD = 1.5
H) cm. The transverse line becomes slightly steeper starting at a phantom
diameter of 300 cm, and the difference of SD values becomes 2 H at a phantom
diameter of 350 cm.
Using the following steps, we created a tube current modulation table from the monogram: First, place a transverse line (line A) to pass through the point of intersection between the 300-mA exponential curve and an SD value equal to 12 H, the desired target noise level; second, place a vertical line (line B) at the point where the 300-mA exponential curve crosses the SD value of 10 H, which is an SD value obtained from unenhanced images of a patient; third, read the crossing point of the vertical and transverse lines that lies between the exponential lines of 200 and 250 mA, closer to 200 mA (Fig. 3). The same method was used to calculate each SD value from 7 to 18 H at three different target noise levels of 11, 12, and 13 (Table 1). The numbers in boldface in Table 1 represent the maximum tube currents and may not generate image quality of the desired noise level. The SD values of the phantoms larger than 280 cm were beyond the correctable range with maximum allowable tube current (500 mA) and were not included for the calculation of the modulation table. In addition, SD values of less than 8 H (phantom size of 197.5 cm) were regarded as 8 H, because the minimum tube current used in the phantom study was 150 mA.
|
Imaging
After topograms were obtained, prospectively ECG-gated sequential
unenhanced scans were obtained with a 3-mm collimation, 0.4-second rotation
time, half exposure, 120 kV, 180-mm field of view, and 300 mA. Two or three
sectional scans from the mid level of the left ventricle were omitted to
reduce radiation. The images were used to determine the scanning range and to
position the heart in the center of the field of view. The SD of the CT
attenuation values of the left atrium at the level of entry for the right
pulmonary veins was measured from four images (single sectional scans) using a
cursor circle of 200 mm2, and the four values were then averaged.
If streak artifacts from vertebral bodies were present, the next image without
such artifacts was used.
For the contrast-enhanced images, a retrospective ECG-gated helical scan was obtained with a 0.5-mm collimated section thickness (64 x 0.5 mm collimation), pitch of 0.204-0.224 (automatically selected depending on the patient's pulse rate), 180-mm field of view, 120-kV tube potential, and 0.4-second rotation time with half reconstruction (200 milliseconds of temporal resolution, 100 milliseconds with two segmentations at higher heart rates).
Tube current was individually modulated using the tube current modulation table (Table 1). CTDIvol, expressed in milligrays (mGy), which was displayed on the operator console, was used to measure radiation dose. One additional slice (3 mm x 4 images) of an unenhanced sequential scan was obtained at the same level in patients who were scanned with a tube current other than 300 mA to measure the SD of the CT attenuation values (Figs. 4A and 4B). The automatic contrast triggering system (SURE Start, Toshiba Medical Systems), which triggered at identification of 140-160 H at the aortic root after injection of 60-80 mL (scaled to body weight) of nonionic contrast material (iohexol [Omnipaque 350, GE Healthcare]), was used in all patients. Contrast material was injected through an 18-gauge IV catheter at a rate of 4.5 mL/s and was followed by a chaser of 30 mL of normal saline. Each patient's body weight was recorded and body dimensions (anteroposterior diameter x transverse diameter) were measured from the topogram at the level of the left atrium.
|
|
For the analysis of uniformity in image quality, a separate set of images of normal coronary arteries and normal internal mammary arteries of 1.5 mm in diameter (1.35-1.65 mm) was created in curved MPR of 1 mm in thickness and color-coded 3D images were created using the same attenuation value threshold setting. The imaging process was performed by a single technologist who used the same display conditions (at 300% zoom: width, 720 H; level, 250 H) for all patients. A total of 100 sets of four images each from 100 patients were stored in separate folders in DICOM format in random order.
Uniformity Analysis
Measurement analysis—The SD of CT attenuation values was
measured from the four unenhanced images obtained with the corrected tube
current using the same size of region-of-interest (ROI) cursor in both groups
of patients (100 patients with normal coronary arteries and 95 patients with
atherosclerosis). Mean values with SDs were calculated for each group. The
data from the patients with atherosclerosis were not included in the
correlation chart of tube currents determined by different parameters to avoid
confusion.
Visual analysis—Using images obtained from a database of images collected before the use of the new tube current modulation method, we created a 4-grade scale to evaluate image quality for visualization of the coronary and internal mammary arteries that were 1.5 mm in diameter: 1 = high mottle, 2 = medium mottle, 3 = low mottle, 4 = minimal or no mottle. Sample images displayed in color-coded volume rendering and curved MPR for each grade of image noise level were agreed on by four observers (Figs. 5A, 5B, and 5C). In a blind review, four radiologists, all of whom are experienced in CTA, rated all of the image sets by importing them onto a DICOM image viewer with two monitors: one for the images being evaluated and the other for the sample images displayed for comparison during the analysis (Table 2).
|
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The correlation chart between tube current levels determined by SD values of unenhanced images, body weight, and a fixed tube current of 400 mA (manufacturer's guideline) shows wide differences in tube currents used in all sizes of patients. Either these patients would have received radiation in excess of the dose needed to obtain diagnostic quality images or image quality would have been poor due to inadequate dosing (Fig. 6).
Using SD values to modulate tube current, we found that patients who weighed 69 kg or less received 9-30% less radiation (CTDIvol) than if body weight modulation had been used and 18-45% less than if a fixed tube current of 400 mA had been used. There were individual patients who received 71% less radiation. Patients who weighed 70 kg or more received 1-11% more radiation (Table 3) to achieve the desired noise level. Body dimensions showed a better correlation with the SD of the CT attenuation values than did body weight (Figs. 7A and 7B).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Automatic tube current modulation is known to be an effective method of reducing radiation dose to the abdomen and pelvis [18, 24, 25], but it is expected to be limited in coronary CTA because of the relatively smaller angular or z-axis fluctuation of attenuation at the level of the heart. In addition, z-axis automatic tube current modulation cannot be used with ECG modulation when both techniques are implemented at the same time [1]. Our method will not affect the use of ECG modulation or angular automatic tube current modulation because the tube current is selected before scanning starts. In addition, an SD value-based tube current modulation table can be used without cost by users of the same scanner model (Aquilion 64), and no installation of software is required. Users of other models of scanner made by Toshiba or of scanners made by other manufacturers can create a tube current modulation table through a series of phantom tests by referring to our method.
The new modulation method can be automated by creating an ROI cursor of a predetermined size to be placed at the center of left atrium and by calculating average SD values to match with the tube current modulation reference table. A system with automated software based on this principle will enhance the use of tube current modulation for the uniformity of images and radiation dose reduction in smaller patients. Although the method is designed specifically for coronary CTA, it also can be used for imaging other body parts with low z-axis fluctuation of attenuation and central homogeneous tissue (fluid), such as for CT of the organs surrounding the bladder.
The left atrium was chosen in this study for the measurement of SD of CT attenuation values because of its central location and thin myocardium; thick myocardium can falsely increase SD value. The SD values from four images were averaged to minimize measurement error. To evaluate the reproducibility of SD measurements, SD measurements on four images obtained from each of 20 randomly selected patients were performed by four observers. The intraclass correlation coefficients were 0.993, 0.991, 0.990, and 0.994, respectively, using SPSS software (version 10.0, SPSS) for Windows (Microsoft), suggesting a high level of reproducibility.
For evaluation of uniformity of image quality, directly measuring the SD of CT attenuation values from the left atrium on contrast-enhanced images may seem to be the most accurate and simple method. However, we found that this method is not accurate because of nonuniform contrast enhancement. Instead, we measured the SD of CT attenuation values from unenhanced images obtained at an adjusted tube current before obtaining the contrast-enhanced images.
There is a bias to perform the measurement analysis of noise (SD of CT attenuation values) from unenhanced images to evaluate the uniformity of contrast-enhanced images because there is a significant difference in scanning parameters (radiation dose) between the two images. Although we adjusted for this difference in the process of creating the tube current modulation reference table, we needed additional assurances, so we visually analyzed images of both the coronary and internal mammary arteries.
The internal mammary artery was included in the visual analysis of images to avoid bias from motion artifact and possible unrecognized atherosclerosis that might degrade image quality. It is coincidental that the internal mammary artery, which is affected little by cardiac pulsation, is consistently included in the imaging field, and it has a very low incidence of atherosclerosis because of the low number of vasa vasorum [26-28].
Contrast enhancement will increase the attenuation coefficiency of contrast-enhanced images; however, this could not be factored into the phantom test or the calculation of the tube current modulation table because of technical limitations.
Because there is no universally accepted method of grading image quality in coronary CTA, we used a grading scale similar to the one that Irie and Inoue [16] and Rizzo et al. [25] used for abdominal CT. Although the optimal image noise level has been suggested at an SD value of 10-11 H for the chest and 11-12 H for the abdomen by vendors of CT units and by some authors [16, 18], the optimal diagnostic image quality has not been scientifically supported in coronary CTA. The optimal desirable noise level may be determined on the basis of the study purposes (e.g., significant stenosis, quantification of stenosis, or soft plaque analysis). However, we have been using 12 H (equivalent to 11 H on contrast-enhanced images) as the desired target noise level in more than 1,000 patients. Although these data have not been published in the literature, this image quality has been well received among referring cardiologists.
Another limitation to this study is that the normalcy of coronary arteries was not confirmed with conventional angiography. Recent reports of the accuracy of coronary CTA using 64-MDCT show a high negative predictive value (93-99%), and the results were from using the criterion of more than 50% stenosis as an abnormal segment [29-31]. Because we used "no visible atherosclerosis" as the criterion for normal coronary arteries, the negative predictive values in our cases are even higher.
At some institutions, radiologists prefer to perform coronary CTA without obtaining a precontrast calcium scoring scan to reduce radiation dose [32, 33]; however, this practice may not necessarily reduce the radiation dose because a larger-than-necessary scan range is often selected to avoid "failed scan range determination" when the topogram is used alone. A 1-cm excess of scan range (average scan range = 11.4 cm) will increase the total dose-length product (DLP) by 8-9% in our institution (postcontrast, 85% of total dose; topogram, 1%; bolus tracking, 7%; calcium scoring, 7% in dose analysis of 20 patients in our institution), and this is about the same radiation dose as the entire unenhanced scan at 300 mA.
In summary, the SD of CT attenuation values from the left atrium is an accurate parameter with which to modulate tube current for coronary CTA. This method has not been described elsewhere to our knowledge, and we think that this method has advantages over other methods, as we discussed earlier. The method, either manual or automated (when it become available), will give operators an additional choice for tube current modulation in coronary CTA.
Acknowledgments
We thank Ki Su Park, RT, a chief CT technologist, for his invaluable
assistance in CT imaging and Edward Y. Hur for his outstanding editing. We
also thank Hyun Hoi Her, a chief service engineer from Daihan Metra, a local
distributor of Toshiba Medical Systems, for phantom studies.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. B. Lim, G. Hur, S. Y. Kim, Y. H. Kim, S. U. Kwon, W. R. Lee, and S. J. Cha Coronary Stent Fracture: Detection with 64-Section Multidetector CT Angiography in Patients and in Vitro Radiology, December 1, 2008; 249(3): 810 - 819. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Niinuma, R. T. George, A. Arbab-Zadeh, J. A.C. Lima, and C. A. Henrikson Imaging of pulmonary veins during catheter ablation for atrial fibrillation: the role of multi-slice computed tomography Europace, November 1, 2008; 10(suppl_3): iii14 - iii21. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. P. Shuman, K. R. Branch, J. M. May, L. M. Mitsumori, D. W. Lockhart, T. J. Dubinsky, B. H. Warren, and J. H. Caldwell Prospective versus Retrospective ECG Gating for 64-Detector CT of the Coronary Arteries: Comparison of Image Quality and Patient Radiation Dose Radiology, August 1, 2008; 248(2): 431 - 437. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
