| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Experimental Studies |
1 From the Department of Radiological Sciences, School of Health Sciences (Y.F., M.S.), and Department of Diagnostic Radiology, Graduate School of Medical Sciences (K.A., Y.N., S.S., S.M., Y.Y.), Kumamoto University, 4.24.1 Kuhonji, Kumamoto 862-0976, Japan; Department of Radiology, Kumamoto University Hospital, Kumamoto, Japan (K.K., N.N.); and Philips Medical Systems, Minato-ku, Tokyo, Japan (N.S.). Received September 23, 2004; revision requested November 30; revision received January 23, 2005; accepted February 23. Address correspondence to Y.F.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: The institutional review board approved the participation of the radiologists in the observer performance test, and informed consent was obtained from all participating radiologists. A phantom for measurement of the radiation dose and a phantom containing low-contrast objects were scanned with a 16–detector row CT scanner at 120 kV and 90 kV. For determination of the radiation dose at both 90 kV and 120 kV, the tube current–time product settings were 100–560 mAs, and the doses at the center and periphery of the phantom were measured. To assess low-contrast detectability, we used a 300-mAs setting at 120 kV and 250–560-mAs settings at 90 kV. Five observers participated in the receiver operating characteristic analysis. Area under the receiver operating characteristic curve (Az) values were calculated in each observer. Az values obtained with each of the scanning techniques were recorded, and differences were examined for significance by using the Dunnet method.
RESULTS: The mean Az value was 0.951 at 120 kV and 300 mAs. Az values were 0.927–0.973 at 90 kV and 450–560 mAs, and the differences between those values and values obtained at 120 kV and 300 mAs were not significant (P = .937–.952). A value of 100% was assigned to the radiation dose delivered to the center of the phantom at 120 kV and 300 mAs. The relative dose delivered at 90 kV ranged from 65% at 450 mAs to 79% at 560 mAs.
CONCLUSION: A reduction from 120 kV to 90 kV led to as much as a 35% reduction in the radiation dose, without sacrifice of low-contrast detectability, at CT.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Although the radiation dose can be reduced by decreasing the tube current–time product settings (5–7), this alteration also reduces the contrast-to-noise ratio (CNR). Sigal-Cinqualbre et al (8) reported a weight-adapted contrast-enhanced chest CT technique with low tube voltage that allowed reduction of the radiation dose, as well as reduction of the contrast material dose. The physical density of the lungs is low, however, and low-contrast detectability at chest CT is not as critical as it is at abdominal CT, where it is one of the most important diagnostic factors. Accordingly, efforts to reduce the radiation dose by reducing tube voltage must not result in degradation of low-contrast detectability. The purpose of our study, therefore, was to reduce radiation dose from abdominal CT without degradation of low-contrast detectability by using a technique with low tube voltage (90 kV).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Phantom for Evaluation of Low-Contrast Detectability
The phantom (Catphan 424; Phantom Laboratory, Cambridge, NY) (9,10) contained several cylindrical low-contrast objects of different sizes arranged in a circle; the background was of uniform density. The low-contrast module consisted of low-density phenolic microspheres and high-density polytetrafluoroethylene powder (11). Both the diameter and the length along the z-axis of the phantom were 20.0 cm. The cylinders varied in diameter from 2.0 to 15.0 mm and deviated from nominal contrast levels by 0.3%, 0.5%, and 1.0%. The body annulus, which consisted of acrylic, was set around the phantom to mimic attenuation during body imaging (Fig 1). The diameter and length along the z-axis of the acrylic body annulus were 32.0 cm and 10.0 cm, respectively.
|
CT Scanning
We used a 16–detector row CT scanner (IDT16; Philips Medical Systems). The scanning parameters were as follows: detector configuration, 1.5 mm (detector collimation) x 16 (detectors); section thickness, 5.0 mm; section interval, 5.0 mm; rotation time, 0.75 second; pitch, 0.659; scan field of view, 50.0 cm; and display field of view, 35.0 cm. Scanning was performed at the standard tube voltage of 120 kV (effective energy, 64 keV) and at the low tube voltage of 90 kV (effective energy, 54 keV). In measurements of the radiation dose at 120 and 90 kV, the tube current–time product settings were 100, 150, 200, 250, 300, 350, 400, and 450 mAs and 100, 150, 200, 250, 300, 350, 400, 450, 500, and 560 mAs, respectively. We did not measure the radiation dose at 500 and 560 mAs and 120 kV.
In measurements of the phantom, the tube current–time product settings were 100, 150, 200, 250, and 300 mAs at 120 kV and 250, 300, 350, 400, 450, 500, and 560 mAs at 90 kV. In these scanning techniques, the CNR was measured at 100–300 mAs at 120 kV and 250–560 mAs at 90 kV. Visual evaluation of low-contrast detectability was performed at 300 mAs and 120 kV and at 250–560 mAs and 90 kV.
Radiation Dose Measurements
A 32.0-cm-diameter dose measurement phantom (model 20CT14; RadCal, Monrovia, Calif) was scanned to measure the radiation dose delivered with each scanning technique. The phantom had a 13.0-mm-diameter cavity in the center and four cavities in the periphery. Five glass-rod dosimeters (Asahi Techno Glass, Shizuoka, Japan) (12,13) were inserted by one author (N.N.) in the five cavities, and the radiation dose to air delivered with each scanning technique was measured. The peripheral radiation dose was calculated with averaging of the measurements of the four glass-rod dosimeters inserted in the peripheral cavities of the phantom.
Measurements of CNR
For each scanning technique performed at 100–300 mAs and 120 kV and at 250–560 mAs and 90 kV, we assessed the CNR by using the phantom. We measured the CT numbers of the low-contrast objects in a 15-mm-diameter area and the background of the phantom by using a circular region-of-interest cursor. The region of interest area was kept at 50 mm2. A window width of 100 HU and a window level of 60 HU were set as constants for the images used in this analysis. CNRs were calculated as follows: CNR = (ROIm – ROIb)/SDb, where ROIm and ROIb are the CT numbers of the low-contrast objects in a 15-mm-diameter region of interest and of the background region of interest, respectively, and SDb is the standard deviation of the attenuation values of the background (9).
Visual Evaluation of Low-Contrast Detectability
We used receiver operating characteristic analysis (14,15) to evaluate low-contrast detectability. One author (Y.F.) extracted rectangular areas that were 10 x 30 mm and that included low-contrast objects from the images of the phantom that was scanned with each technique (Fig 2). We used 9.0- and 15.0-mm-diameter low-contrast objects with a contrast difference of 1.0% to evaluate low-contrast detectability. The same author (Y.F.) also extracted rectangular areas that were 10 x 30 mm and that did not include low-contrast objects. Then, we placed the extracted images on black-background images and used them to study observer performance. The observers read the extracted images with a color monitor that had a spatial resolution of 1200 x 1600 (Radiforce R22; Eizo, Ishikawa, Japan) by using a Digital Imaging and Communications in Medicine viewer (Image VINS Pro, version 3.01; Yokogawa Electric, Tokyo, Japan).
|
Five radiologists whose experience ranged from 10 to 20 years (mean, 12 years) participated in the observer study. They were blinded to the scanning technique with which the images were obtained. Our institutional review board approved the participation of these radiologists in this study; informed consent was obtained from all observers. The time for reading the images was not limited, and each observer was allowed to select window levels and window widths on the monitor screen.
A continuous rating scale with a line-marking method was employed to represent each observer's confidence level; marks that represented observer confidence were placed on a line that was 5 cm in length. The observers, who were asked whether an object was present, responded by placing a pencil mark on a bar on the record forms. The words "definitely present" and "definitely not present" were printed at the right and left ends, respectively, of the bar. All images were presented to the observers in random order. Before the test, the radiologists were trained with five representative cases that were not part of the observer study.
Statistical Analysis
The Pearson product moment correlation coefficient (r) was used to investigate the relationship between the CNR and the radiation dose delivered to the center of the phantom at 120 and 90 kV.
A binormal receiver operating characteristic curve was fitted to each radiologist's confidence rating data with quasimaximum likelihood estimation. A computer program (ROCKIT; Charles E. Metz, University of Chicago, Chicago, Ill) was used to obtain binormal receiver operating characteristic curves from the ordinal-scale rating data. The area under the receiver operating characteristic curve (Az) that had the best fit and was plotted in the unit square was calculated for each fitted curve. The Az values obtained with the different scanning techniques were compared by using two-way analysis of variance. When the overall differences were statistically significant, post hoc analysis was performed by using the Dunnet method. We used Az values obtained at 120 kV and 300 mAs as a control in the Dunnet method and compared the results with those obtained at 90 kV and 250, 300, 350, 400, 450, 500, and 560 mAs.
Differences with P < .05 were considered statistically significant. Calculation of the Pearson product moment correlation coefficient, the two-way analysis of variance, and the post hoc test were performed with a statistical software program (SPSS, version 10.05, SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
CNR Results
There was a direct correlation between the CNR and the radiation dose obtained in the center of the phantom at both 90 and 120 kV (Fig 3). The Pearson correlation coefficient (r) and the corresponding P values were r = 0.99 and P < .001 at 90 kV and r = 0.94 and P < .001 at 120 kV.
|
Low-Contrast Detectability
The mean Az for the five radiologists who performed the visual inspections is shown in Figure 4. By using two-way analysis of variance, the overall difference in the Az values of the eight scanning protocols was statistically significant (P < .001). The mean Az value at 120 kV and 300 mAs was 0.951 (Table 4). At 90 kV, the mean Az value—0.744 at 250 mAs, 0.804 at 300 mAs, 0.837 at 350 mAs, and 0.865 at 400 mAs—was significantly lower than that at 120 kV and 300 mAs (P < .001 at 250 and 300 mAs; P = .003 at 350 mAs; P = .033 at 400 mAs) (Table 4). There was no statistically significant difference between the mean Az value at 120 kV and 300 mAs and the values obtained at 90 kV at any of the other tube current–time product settings that we investigated (Table 4).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
When our observers looked at low-contrast images to assess detectability, the Az values obtained at 90 kV with tube current–time product settings of 450, 500, and 560 mAs did not differ significantly from those obtained at 120 kV and 300 mAs. This suggests that low-contrast detectability at 90 kV with tube current–time product settings higher than 450 is equivalent to that achieved with 120 kV at 300 mAs. Furthermore, the relative dose at 90 kV and 450 mAs was 65% of the dose in the center and 87% of the dose in the periphery of the value (100%) we assigned to the radiation dose delivered to the center of a phantom at 120 kV and 300 mAs. On the basis of these considerations, we postulate that a 35% reduction in the radiation dose can be achieved when scanning is performed at 90 kV rather than at 120 kV without degradation of detectability on low-contrast images. We did not evaluate low-contrast detectability at values higher than 90 kV and 560 mAs because our CT scanner did not allow us to set the tube current–time product at a value higher than 560 mAs.
Low-contrast detectability is one of the most important factors in abdominal, especially hepatic, CT. On CT images, hepatic tumors are recognized by the attenuation differences between the tumor and the hepatic parenchyma, the so-called tumor-to-liver contrast (19). Previously, Awai et al (20) reported that in enhanced hepatic CT, tumor-to-liver contrast was 5–40 HU; however, in some cases, it was only a few Hounsfield units. We used the phantom at a contrast difference of 1.0%; this predicates a 10-HU difference in the CT number between the tumor and the hepatic parenchyma. Most primary liver tumors arise in the chronically damaged liver. The texture of the hepatic parenchyma becomes inhomogeneous as liver damage progresses, and this inhomogeneousness makes it difficult to detect these tumors. Besides frequent follow-up CT studies at short intervals, patients who have confirmed liver tumors or those who are suspected of having them usually undergo multiphase enhanced CT. In this setting, a CT technique with low tube voltage that helps to achieve a reduction in the radiation dose without degradation of low-contrast detectability is of high value.
A disadvantage of the CT technique with low tube voltage is the increase in noise. Boone et al (21) found a relationship between image noise and the tube voltage and the tube current–time product setting in CT. They showed that noise increased at lower tube current–time product settings and lower tube voltage. Huda et al (22) also reported that a decrease in tube voltage increased the noise on abdominal CT images. For CT scanning with low tube voltage, a new adaptive filter must be developed to suppress the increased image noise.
We found that at identical CNR levels, the radiation dose delivered at 90 kV was 0.71 times that at 120 kV. This suggests that, by using 90 kV, the radiation dose can be reduced by 29% without affecting the CNR. On the other hand, on the basis of our analysis of low-contrast detectability, we concluded that scanning with low tube voltage permits a reduction of the radiation dose of 35% compared with scanning at the standard tube voltage (120 kV). Verdun et al (10) reported that there was a statistically significant direct correlation between the mean CNR and the subjective scores of low-contrast detectability in an experiment with a phantom (r = 0.95, P < .05). Those results and our results appear to be inconsistent. We found that the CNR at 120 kV and 300 mAs was almost equivalent to that at 90 kV and 500 mAs. Our Az values for low-contrast detectability obtained at 450 mAs and 90 kV were not significantly different from those obtained at 120 kV and 300 mAs. This suggests that almost the same level of low-contrast detectability is retained at 450 mAs and 90 kV, although the CNR is lower than it is at 300 mAs and 120 kV. This is probably because CNR and low-contrast detectability are not necessarily parallel to each other. We propose that assessment of the possibility of reducing the radiation dose in abdominal CT should be based on the quality of low-contrast detectability because this index directly reflects the diagnostic contribution of abdominal CT.
A limitation of our study is that we did not take into account differences in patient size. In 90-kV scanning compared with 120-kV scanning, higher tube current–time product settings are required to compensate for the lower number of photons. Our CT scanner did not allow us to set the tube current–time product at a value higher than 560 mAs. At constant tube voltage, an increase in patient size can reduce the transmitted energy fluence by two orders of magnitude (22). Therefore, a CT technique with low tube voltage may not be appropriate in overweight patients. Although results of our phantom study suggest that the radiation dose at CT can be reduced without degradation of low-contrast detectability, studies are under way in our institution (Kumamoto University Hospital, Kumamoto, Japan) to determine whether this technique with low tube voltage can be used in a clinical setting (23).
In conclusion, in CT studies with the scanning technique with low tube voltage (90 kV) rather than standard tube voltage (120 kV), the radiation dose at the center and the periphery of the phantom can be reduced by 35% and 13%, respectively, without sacrificing low-contrast detectability. Although studies in patients were not part of this investigation, results of the studies with a phantom suggest that the technique is effective for dose reduction of abdominal CT for patients with relatively light weight (23).
Practical application: The results of this study indicate that radiation dose can be reduced and acceptable low-contrast detectability can be achieved in CT examinations with reduction of tube voltage. The tube current–time product setting at 90 kV should be set so that the volume CT dose index at 90 kV is 70% of that at 120 kV. With respect to patient weight, a standard deviation at 20 HU was considered the threshold for acceptable noise (8). The noise increased with patient weight, and according to our clinical data reported in another study, the noise value was higher than 20 HU in patients who weighed more than 80 kg (23). Therefore, the technique with low tube voltage should be used only in patients whose body weight is less than 80 kg.
|
FOOTNOTES |
|---|
Abbreviations: Az = area under the receiver operating characteristic curve • CNR =contrast-to-noise ratio
See also Science to Practice and the article by Nakayama et al in this issue.
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantors of integrity of entire study, Y.F., K.A.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, Y.F., K.A., Y.N., S.S., S.M.; experimental studies, Y.F., K.A., Y.N., K.K., N.N., N.S.; statistical analysis, Y.F., M.S.; and manuscript editing, Y.F., K.A., Y.N., Y.Y.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. H. McCollough, L. Guimaraes, and J. G. Fletcher In Defense of Body CT Am. J. Roentgenol., July 1, 2009; 193(1): 28 - 39. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. T. Schindera, R. C. Nelson, S. Mukundan Jr, E. K. Paulson, T. A. Jaffe, C. M. Miller, D. M. DeLong, K. Kawaji, T. T. Yoshizumi, and E. Samei Hypervascular Liver Tumors: Low Tube Voltage, High Tube Current Multi Detector Row CT for Enhanced Detection Phantom Study Radiology, December 1, 2007; 246(1): 125 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Yanaga, K. Awai, Y. Nakayama, T. Nakaura, Y. Tamura, M. Hatemura, and Y. Yamashita Pancreas: Patient Body Weight tailored Contrast Material Injection Protocol versus Fixed Dose Protocol at Dynamic CT Radiology, November 1, 2007; 245(2): 475 - 482. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Nakayama, K. Awai, Y. Funama, D. Liu, T. Nakaura, Y. Tamura, and Y. Yamashita Lower tube voltage reduces contrast material and radiation doses on 16-MDCT aortography. Am. J. Roentgenol., November 1, 2006; 187(5): W490 - W497. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. McCollough Automatic Exposure Control in CT: Are We Done Yet? Radiology, December 1, 2005; 237(3): 755 - 756. [Full Text] [PDF]
|
|
|
|
|
|
Y. Nakayama, K. Awai, Y. Funama, M. Hatemura, M. Imuta, T. Nakaura, D. Ryu, S. Morishita, S. Sultana, N. Sato, et al. Abdominal CT with Low Tube Voltage: Preliminary Observations about Radiation Dose, Contrast Enhancement, Image Quality, and Noise Radiology, December 1, 2005; 237(3): 945 - 951. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
