| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Diagnostic Radiology, University Hospital, University of
Technology (RWTH), Pauwelsstrasse 30, D-52074 Aachen, Germany.
2 Institute of Medical Statistics, RWTH Aachen, Aachen, Germany.
3 Institute of Medical Physics, Erlangen, Germany.
4 Siemens Medical Solutions, Computed Tomography, Forchheim, Germany.
Received June 5, 2004;
accepted after revision September 9, 2004.
Address correspondence to M. Das
(das{at}rad.rwth-aachen.de).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Two hundred and four patients underwent MDCT of the thorax (Somatom Sensation 16, Siemens). Group 1 was scanned using a standard protocol with 100 mAseffective (mAseff). Group 2 was scanned using a dose modulation template (CareDose). Group 3 was scanned with mAseff = body weight (kg). Group 4 was scanned with a combination of weight-adapted mAseff and dose modulation. All other parameters were kept constant. Signal-to-noise ratio was assessed as an objective measurement for image quality, and subjective image quality was rated by three experienced radiologists on a 4-point scale. Effective dose was calculated using dedicated software.
RESULTS. The mean noise measurement values were 8.31 H for the 100 mAseff protocol for the regression between weight and signal-to-noise (p < 0.0001), 9.08 H for group 2 (p < 0.0001), 9.0 H for group 3 (p = 0.5051), and 9.98 H for group 4 (p = 0.0152). The median image quality was 1 (1 = highest quality) in all subgroups. The mean effective dose was 6.83 mSv, 5.92 mSv, 4.73 mSv, and 3.97 mSv, respectively. The least correlation between weight and image noise was achieved for the individually weight-adapted protocol and in the weight-adapted with CareDose combination.
CONCLUSION. By tube current time product adaptation (kg = mAseff) combined with an online tube current modulation template, a well-balanced examination without significant loss of information was achieved for this specific scanner. Thus, individually adapted protocols for chest 16-MDCT can be recommended.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The Fleischner Society stated recently that "further research into the complex relationship between radiation exposure, image noise, and diagnostic accuracy should be encouraged to establish the minimum radiation doses that provide adequate diagnostic information for standard clinical questions" [9].
Several recommendations for dose reduction have been made for different anatomic regions and with different approaches [10–14].
The purpose of this study was to develop a simple protocol that can be performed easily during routine chest examination for 16-MDCT using individually weight-adapted tube current-time settings.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study, the main focus was to develop a simple and practical guideline for adjusting mAseffective (mAseff) to patient weight; thus, three methods to reduce the radiation dose were compared with routine scanning, with a standard-dose protocol as recommended by the supplier (Somatom Sensation 16 handbook, Siemens Medical Solutions).
Two hundred and four adult patients (121 men, 83 women) were entered into a prospective trial (Table 1). No patient was scanned twice in this study. Informed consent was obtained from all patients and personal data, including patient height and weight, were recorded. Six of the patients received non-contrast-enhanced scans; 198 examinations were performed after IV administration of 90 mL contrast material (iopromide [Ultravist 370], Schering) at a flow rate of 3.5 mL sec followed by a 30 mL saline chaser injection during inspiratory breath-hold on a commercially available 16-MDCT scanner (Somatom Sensation 16). Scanning was performed with a standardized protocol with 16 x 1.5 mm collimation, 0.5 sec rotation time, table feed 30 mm/rotation, and 120 kV. An individually adapted field of view, a matrix size of 512 x 512, and a soft reconstruction kernel (Siemens B30) and a sharp convolution kernel (Siemens B50) were chosen, respectively. Images were reconstructed as 5- and 2-mm thick sections with an increment of 4 mm and 1.5 mm, respectively. The initial 52 patients (group 1) were examined with the recommended "standard" mAseff setting of 100 mAseff. The next 50 patients (group 2) were scanned using commercially available online tube current modulation software (CareDose, Siemens) [15, 16] that is characterized by online monitoring of the attenuation and subsequent tuning of the tube current as a function of the projection angle, with a delay of 360°. For projections with low attenuation, the maximal reduction of the tube current was 90%. For each acquisition, the CT unit calculated the arithmetic mean mAseff throughout the exposure. The next 50 patients (group 3) were examined with patient-specific parameters. The tube current time product was adapted according to the body weight (body weight [kg] = mAseff). Finally, 52 patients (group 4) were scanned using a combined protocol with individually adapted mAseff settings plus additional online tube current modulation.
|
All subgroups were comparable regarding weight, height, and body-mass index (BMI) (BMI = weight in kg divided by the square of height [m]).
The mean weight for all patients was 72.54 ± 14.71 kg, the mean BMI was 24.96 ± 4.54 kg/m2, and the mean height was 1.7 ± 0.09 m. Table 1 gives an overview of parameter distribution in the four groups, including sex.
To measure noise levels, a water syringe was placed ventrally on top of the sternum. Noise levels were measured four times on consecutive axial slices using a region-of-interest methodology. The SD of the attenuation measurements within the water syringe was ascribed to image noise. The mean value of these four measurements was calculated for each patient (Lenzen H, presented at the 1999 annual meeting of the German X-ray Society, Wiesbaden, Germany). To describe dependence between image noise, body weight, and BMI, respectively, noise levels were plotted against both parameters. Univariant and bivariant linear regression analyses were performed to detect correlation between the parameters.
Analysis of variance was performed to evaluate the effect of the methods regarding image noise and mAseff, including posthoc comparisons according to the Bonferroni method.
All examinations were evaluated twice and were first interpreted on a standard PACS (Sectra, Philips Medical Systems) soft-reading workstation during clinical routine. At the second interpretation, three experienced radiologists rated the images for mediastinal and hilar depiction and for image artifacts in soft-tissue window settings (center, 80 H; width, 400 H), and for lung parenchyma in a lung window setting (center, –600 H; width, 1,200 H) on a 4-point scale.
Images with distinct anatomic detail, sharp vessel edges, and clear delineation of small structures were classified as 1 (excellent). Examinations with clear anatomic detail and mild to moderate increase in noise without impairment of diagnostic accuracy were rated as 2 (good); those with further increase of noise without impairment of diagnostic accuracy were rated as 3 (fair); and those with obscured anatomic detail, a distinct increase in noise and extensive blurring, not sufficient for diagnosis were classified as 4 (nondiagnostic). Image quality (ordinal rates scoring 1 to 4) was compared using the Cochrane-Armitage trend test. Level of significance for the statistical analyses was chosen as 0.05. Three factorial analyses of variance were performed to explore the effect of the methods on the effective dose, taking into account weight, BMI, and sex.
The effective dose was calculated for each patient using dedicated software (CTExpo V.1.2, Medical University Hannover).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Objective Image Quality
Noise measurements showed mean values of 8.31 ± 1.61 H for group 1,
9.08 ± 2.1 H for group 2, 9.00 ± 1.16 H for group 3, and 9.98
± 2.32 H for group 4. Variance analysis showed significant influence of
all methods on the image noise (p = 0.0002).
A slight increase of the image noise occurred when group 2 was compared with group 1, as shown with the posthoc comparison test (p < 0.0001). Image noise was significantly higher in group 3 compared with group 1 (p = 0.0498). The highest significant increase of image noise was found in group 4 in comparison with group 1 (p < 0.0001).
The noise levels (SD) were plotted against the estimated body weight and BMI and linear regression analysis was performed (Figs. 1, 2, 3, 4, 5, 6, 7, 8).
|
|
|
|
|
|
|
|
Regression Analysis
Regression analysis for group 1 (Figs.
1 and
2) revealed statistically
significant influence of the patient's weight (pweight
< 0.0001) and BMI (pBMI < 0.0001) on the image
noise. Sex also was a significant influence in group 1
(pfemale = 0.0014; pmale =
0.0357).
In group 2 (Figs. 3 and 4), the patient's weight (pweight < 0.0001) and BMI (pBMI < 0.0001) also had a statistically significant influence on the image noise. In contrast with group 1, no statistically significant influence was found regarding sex (pfemale = 0.1961; pmale = 0.0672).
Regression analysis showed a nearly constant relationship between image noise and body weight and image noise and BMI in group 3 (Figs. 5 and 6). This indicated no statistical influence of the patient's weight (pweight = 0.5051) or BMI (pBMI = 0.9636) on the image noise. Sex as a covariant did not have any influence and did not change the given results (pfemale = 0.10867; pmale = 0.0683).
In group 4 (Figs. 7 and 8), body weight and BMI had a statistically significant influence on the image noise (pweight = 0.0152; pBMI = 0.0006). This indicated nearly constant noise related to the body weight and BMI in the two subgroups. Again, sex did not change these results (pfemale = 0.7198; pmale = 0.9706).
Multiple regression analysis for weight and BMI showed that BMI was statistically relevant for group 1 (p = 0.0053) and group 4 (p = 0.0032), while for group 2, both parameters had an influence (pweight = 0.0164; pBMI = 0.0011).
Subjective Image Quality
Median image quality was rated as 1 (Interquartile Distance = 1) for the
soft-tissue window and for the lung-tissue window in all four groups. An
example of excellent image quality from group 4 is shown in Figure
9A,
9B. The Cochrane-Armitage
trend test regarding image quality measured on a rating scale from 1 to 4
showed no influence of any of the methods.
|
|
Dose
In group 1, mAseff was set as 100 (SD = 0) in all patients. In
group 2, mAseff was decreased to 89.51 ± 3.04. In group 3,
mAseff was 71.77 ± 2.68. The lowest mAseff was
found for group 4, with 61.35 ± 12.78.
Significant influence was found regarding effective dose concerning the method (p < 0.001) and sex (p < 0.0001). The calculated effective dose was highest in group 1 with 6.83 ± 0.9 mSv. The effective dose was 5.92 ± 0.92 mSv in group 2, 4.73 mSv in group 3, and the lowest effective dose was found in group 4 with 3.97 ± 0.81 mSv. An overview of the dose is shown in Table 2.
|
Compared with group 1, the mean dose reduction was 13% in group 2, 31% in group 3%, and 42% in group 4.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The aim of our study was to develop an easy-to-use guideline for a 16-MDCT scanner to reduce radiation dose while high image quality is maintained. Different low-dose protocols were used and compared. Online tube current modulation software has been used by other groups previously and a variety of studies [16, 19–21] have proven the potential of attenuation-based online tube current modulation to decrease the amount of radiation dose applied to the patient for single-detector CT.
Thus, we compared an online tube current modulation software protocol, a weight-adapted protocol, and a combination of tube current modulation and weight adaptation with the recommended standard protocol regarding image quality and patient dose.
Wildberger et al. [10] evaluated a similar dose-saving method in which individual mAs-values were derived from body weight [kg] +10, ±0, –10, and –20 mAs. They showed weight adaptation of the mAs –10 would lead to an optimal examination for the explicit scanner type used (Somatom Volume Zoom, Siemens Medical). Tack et al. [11] evaluated an attenuation-based online tube current modulation for thoracic and abdominal studies in MDCT. A mean effective radiation reduction of 16.9% and 20% for the chest and the abdomen, respectively, offers a high amount of dose saving. As the authors stated, it should only be considered to be the second dose-saving strategy after initial adaptation of the mAseff presets to the individual patient. In our study, the online tube current modulation did not achieve as high dose savings as in the study by Tack et al., but it proved to be a valuable additional tool for dose saving.
All adapted protocols showed excellent image quality with a slight increase in image noise. Increased image noise was expected due to technical matters (e.g., pixel noise increased for groups 3 and 4 from 8.3 to 9.0 and 9.8 H, respectively) (Table 2), but this did not result in any deprivation of image quality. All dose-reduced examinations were of diagnostic quality and none had to be repeated. Takahashi et al. [22] showed no differences regarding image quality and diagnostic quality of low-dose protocols in the detection of lung and mediastinal abnormalities for single-slice spiral protocols (50 mAs, 120 kV vs 250 mAs, 120 kV; 10-mm-thick sections each) on additional slices in 50 consecutive patients, despite diminished low-contrast resolution and increased noise levels.
Plotting image noise against body weight and BMI in group 1 graphically showed with a relatively high gradient the heavier the patient (Figs. 1 and 2). The same was true for group 2, but with a gradient not as high as in group 1 (Figs. 3 and 4).
Plotting image noise against body weight and BMI showed nearly zero gradient in linear regression, indicating nearly complete independence of image noise for both parameters in group 3 (Figs. 5 and 6) and a slightly increasing gradient in group 4 (Figs. 7 and 8). Thus, a constant relationship between mAseff and the resulting image quality was achieved in all patients. The slightly increasing gradient in group 4 can be explained by technical limitations of the online tube current modulation software in relatively heavy patients.
Figure 10A, 10B, 10C, 10D gives an example of excellent image quality in group 4 with 41 mAseff compared with an examination with a full dose of 100 mAseff. Even soft-tissue window settings reveal excellent image quality, for example, in the mediastinum as a low-contrast organ.
|
|
|
|
Besides excellent image quality in low-dose protocols, our study consequently showed high dose-saving potential in the majority of the patients. Standard dose scanning as it was performed in group 1 applied an unnecessarily high dose.
Compared with the standard dose scanning, a dose reduction of up to 42% or more could be achieved for our protocols (Table 1). This offers a potentially significant amount of dose saving in the general patient population as adaptation of the mAseff based on the body weight of the patient led to a well-balanced examination, indicating constant noise for all patient weights and BMIs, and reduced radiation without significant loss of information.
Our study showed the potential of dose saving with an easy guideline to adapt the tube current-time setting to weight or to BMI in combination with an online tube current modulation template. Because of easy routine handling, weight-adapted mAseff settings can be proposed for routine chest CT examinations, providing excellent image quality and a high dose-saving potential. Technologists can be easily trained to ask for the patient's weight and adapt the mAseff accordingly, while the on-line tube current modulation is set as default. These settings for routine 16-MDCT chest examinations can be easily used. For other types of CT scanners, analogous protocols may be adapted.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Stolzmann, S. Leschka, H. Scheffel, T. Krauss, L. Desbiolles, A. Plass, M. Genoni, T. G. Flohr, S. Wildermuth, B. Marincek, et al. Dual-Source CT in Step-and-Shoot Mode: Noninvasive Coronary Angiography with Low Radiation Dose Radiology, October 1, 2008; 249(1): 71 - 80. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Herzog, D. M. Mulvihill, S. A. Nguyen, G. Savino, B. Schmidt, P. Costello, T. J. Vogl, and U. J. Schoepf Pediatric Cardiovascular CT Angiography: Radiation Dose Reduction Using Automatic Anatomic Tube Current Modulation Am. J. Roentgenol., May 1, 2008; 190(5): 1232 - 1240. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kubo, P.-J. P. Lin, W. Stiller, M. Takahashi, H.-U. Kauczor, Y. Ohno, and H. Hatabu Radiation Dose Reduction in Chest CT: A Review Am. J. Roentgenol., February 1, 2008; 190(2): 335 - 343. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Leschka, E. Oechslin, L. Husmann, L. Desbiolles, B. Marincek, M. Genoni, R. Pretre, R. Jenni, S. Wildermuth, and H. Alkadhi Pre- and Postoperative Evaluation of Congenital Heart Disease in Children and Adults with 64-Section CT RadioGraphics, May 1, 2007; 27(3): 829 - 846. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
