| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Division of Pediatric Radiology, Department of Radiology, Box 3808, Duke
University Medical Center, Durham, NC 27710.
2 Radiation Safety Division, Duke University Medical Center, Durham, NC.
3 Department of Radiology, University of Arkansas for Medical Sciences and
Arkansas Children's Hospital, Little Rock, AR.
Received February 6, 2007;
accepted after revision July 22, 2007.
The bismuth material for this study was provided by F & L Medical
Products.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Age-based chest 16-MDCT was performed on an anthropomorphic phantom representing a 5-year-old child. Two scans were obtained in each of four sequences: first, without a shield; second, with a 2-ply bismuth shield; third, using automatic tube current modulation with a scout image obtained after placement of the shield; and fourth, using automatic tube current modulation with a scout image obtained before placement of the shield. Metal oxide semiconductor field effect transistor technology was used to measure the radiation dose in 20 organ locations. Effective dose was estimated using the console dose-length product. Noise was measured by recording the SD of Hounsfield units in identical regions of interest.
RESULTS. The bismuth breast shield reduced the dose to the breast by 26%. Shielding and automatic tube current modulation reduced the breast dose by 52%. Multiple organ doses were lowest when the shield was placed after the scout radiograph had been obtained. When the shield was placed after the scout image was obtained, the mean noise in the range of shielding increased from 11.4 to 13.1 H (superior mediastinum) and from 10.0 to 12.8 H (heart) (p < 0.01). Increased noise, however, was near the target noise index (measured in SD of Hounsfield units) of 12.0 H (SD). Using automatic tube current modulation, the effective dose was reduced by 35% when the shield was placed after the scout and by 20% when the shield was present in the scout.
CONCLUSION. The greatest dose reduction is achieved by placing the shield after obtaining the scout image to avoid Auto mA compensation due to density of shield. With this technique, image noise increased but remained close to the target noise index.
Keywords: CT • CT technology • MDCT • pediatric chest CT • phantoms • radiation dose reduction • shielding
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Two techniques that have independently been shown to lower CT radiation dose are automatic tube current modulation [5-12] and bismuth shielding [13-15]. Automatic tube current modulation using both angular (x,y-axis) and longitudinal (z-axis) techniques has been shown to reduce patient dose [5-12]. Shielding has also been shown to reduce radiation dose, primarily from attenuation of the primary beam.
The combined use of automatic tube current modulation and a shielding device might be expected to reduce radiation dose more than either technique alone. However, it may also be that automatic tube current modulation technology that modulates in the z-axis on the basis of the topogram (scout image) may compensate for the presence of the shield by increasing the tube current and thus increasing the dose during the diagnostic scan when the shield is present while the scout image is obtained. To our knowledge, dosimetry during MDCT using the combination of bismuth shielding and automatic tube current modulation has not been reported. Therefore, the purpose of our study was to determine the effect on radiation dose and to measure image quality (noise) of images obtained using bismuth breast shields with z-axis automatic tube current modulation during pediatric chest MDCT.
|
|
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
Dose-length product (DLP) values were recorded from the console [21] using a conversion factor of 0.021 mSv · mGy-1 · cm-1 for 5-year-old chest MDCT. This factor was based on the fact that the adult chest conversion factor is 0.019 mSv · mGy-1 x cm-1 and on previous MOSFET dosimetry determinations in our laboratory [22]. The same scanning duration was used for each manipulation.
Noise was measured as the SD of Hounsfield units by drawing regions of interest (ROIs) in homogeneous-appearing areas of the scan and recording the SD of Hounsfield units in the ROI. The size of the ROI varied depending on the region measured but was generally a 170 mm2 circular area. Noise was measured in identical regions for each study at a total of 18 different locations in five slice levels from the soft-tissue algorithm display, as shown in Figure 2A, 2B, 2C, 2D, 2E. The mean noise measurement was computed for each of the five slice levels for the four different scanning paradigms in Table 2. The relationship of these mean levels was examined using a two-way analysis of variance.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
With the addition of the bismuth breast shield while keeping all other scanning parameters fixed, the measured average breast dose was reduced by 26% (from 0.38 cGy to 0.28 cGy). Measured doses decreased at all 14 locations under the shield by an average of 18.0% (ranging from 5% in the left lung to 35% in the bone marrow of the spine). Compare these doses with those measured outside the shielded area where doses decreased by an average of 4% at four locations, were unchanged at one location, and increased by 5% at one location.
The changes in tube current that resulted from the different scanning paradigms are illustrated in Figure 3. A fixed tube current of 65 mA was used in the standard age-based chest protocol. By comparison, when automatic tube current modulation was used without the shield present in the scout image, the resultant tube current was lower in the lungs (34-37 mA) and increased through the upper abdomen, eventually reaching the preset tube current ceiling of 65 mA. When automatic tube current modulation was used with the shield present in the scout image, the modulation compensated for the presence of the shield by increasing tube current through the level of the shield. The tube current was 40 mA in the first section that included the shield and increased along the length of the shield, reaching 62 mA in the last section that included the shield (Fig. 3). By comparison, in the examination with automatic tube current modulation but with the scout image obtained before placement of the shield, the tube current was 36 mA in the first section that included the shield and did not increase to more than 36 mA over the entire length of the shield.
With the implementation of tube current modulation with the shield placed after the scout image was obtained, measured doses decreased (average dose decrease, 37.5%) at 19 of the 20 (95%) locations when compared with the fixed tube current scan with the breast shield (Table 1). In the remaining location, the measured dose in the stomach was unchanged (Table 1).
With the implementation of tube current modulation with the shield present in the scout image, measured doses decreased at 19 of the 20 (95%) locations (average dose decrease, 25.5%) when compared with the fixed tube current scan with the breast shield (Table 1). In the remaining location, the measured dose in the left kidney was unchanged (Table 1).
Comparing the examinations with automatic tube current modulation when the shield was placed before the scout was obtained versus when the shield was placed after, measured doses were higher (average dose increase, 48%) at 10 of 14 locations (71.4%) under the shield, unchanged at three locations (21.4%) under the shield, and lower (average dose decrease, 27.8%) at one location (7.1%) under the shield when the shield was present in the scout image (Table 1).
Comparing the overall highest-dose scan (fixed tube current, age-based chest CT protocol, no shield) with the lowest-dose scan (automatic tube current modulation and bismuth shield with shield placed after scout image was obtained), measured doses decreased by an average of 51.3% in the 16 locations under the shield, ranging from a 43.2% decrease in the left lung to a 62.5% decrease in the skin over the sternum. Of note, breast dose decreased by 51.6% when automatic tube current modulation and the bismuth shield were used with the shield placed after the scout as compared with the fixed tube current, age-based chest CT with no shield.
Image noise was measured as described earlier (Table 2). For the CT images in range of the shield, average noise measurements (SD of Hounsfield units) were 8.7 H for the fixed tube current scan, 9.7 H for the fixed tube current scan with breast shield, 10.6 H for automatic tube current modulation and breast shield with shield present in the scout, and 12.9 H for automatic tube current modulation and breast shield with shield placed after the scout was obtained (Table 3). Superior to the shield, average noise measurements were 9.3 H for the fixed tube current scan with no shield, 9.4 H for the fixed tube current scan with the breast shield, 11.7 H for automatic tube current modulation and breast shield with shield present in the scout, and 11.5 H for automatic tube current modulation and breast shield with shield placed after the scout (Table 3). Inferior to the shield, average noise measurements were 11.4 H for the fixed tube current scan with no shield, 11.1 H for the fixed tube current scan with breast shield, 11.2 H for automatic tube current modulation and breast shield with shield present in the scout, and 11.4 H for automatic tube current modulation and breast shield with shield placed after the scout (Table 3). Mean noise measurements were not significantly different with the addition of the breast shield in the fixed tube current scans, except in the heart, where the noise increased from 8.50 to 9.96 H (p = 0.008) (Table 2). When the shield was placed after the scout was obtained with automatic tube current modulation, the mean noise measurement at the level of the heart was significantly greater (from 10.0 to 12.8 H) than when the shield was present in the scout (p = 0.001). Likewise, the mean noise measurement at the level of the superior mediastinum was significantly greater as compared with the scans when the shield was present in the scout (from 11.4 to 13.1 H) (p = 0.01). Average noise at the level of the shield was greater than the target noise index (12.0 H [SD]) when the shield was placed after the scout because automatic tube current modulation did not compensate for the presence of the shield and generated a lower tube current (34-36 mA in the range of the shield) compared with when the shield was present in the scout (tube current of 40-62 mA in the range of the shield). This lower tube current resulted in increased noise.
|
Using the DLP method, the effective doses were 2.0 mSv using the fixed tube current technique with and without the breast shield, as expected because DLP is independent of the scout, unlike with automatic tube current modulation, in which the DLP will depend on the scan modulation. When automatic tube current modulation is used, the projected modulations in tube current are accounted for in the DLP. The total effective dose was 1.6 mSv using automatic tube current modulation with the bismuth shield present in the scout radiograph, yielding a 20% dose reduction as compared with the fixed tube current scans (Table 3). The total effective dose was 1.4 mSv using automatic tube current modulation with the bismuth shield placed after the scout radiograph was obtained, yielding a 30% dose reduction as compared with the fixed tube current scans (Table 3).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Bismuth breast shields have been shown to reduce breast radiation dose by 29% in children [14] and by 27% to 52% in adults, depending on the thickness of the shield [15]. In line with these investigations, we found a 26% dose reduction to the breast in an anthropomorphic phantom representing a 5-year-old using a 2-ply bismuth breast shield. The 4% decrease in dose at four locations not under the shield and the 5% increase in dose at one location not under the shield when the shield is added is not unexpected in that the uncertainty of the MOSFET system is approximately 10% or less for a typical CT dose of 1 cGy.
Three general techniques for automatic tube current modulation are offered by the manufacturer of the scanner used for this investigation: z-axis (longitudinal) modulation, x,y-axis (angular) modulation, and a combination of both. With z-axis modulation, the method used in this study, tube current is modulated along the craniocaudal dimension of the patient. For example, in the setting of chest CT with z-axis modulation, a relatively lower tube current is delivered through the lungs, where there is less attenuation, while a relatively higher tube current is delivered through the upper abdomen, where there is more soft-tissue attenuation. With x,y-axis modulation, a relatively lower tube current is delivered through the thinner portions of the patient (generally the anteroposterior dimension), and a relatively higher tube current is delivered when there is relatively more attenuation (generally the lateral dimension).
The algorithms used for automatic tube current modulation differ depending on the manufacturer [28]. The z-axis modulation used in this investigation (Auto mA) requires the user to select the noise index (SD of Hounsfield units) and minimum and maximum tube current values. The range between these values is where the modulation will take place.
Automatic tube current modulation in the z-axis has been shown to reduce tube current-time product by 18-26% in the setting of adult chest CT, depending on the noise index [7]. Automatic tube current modulation in the x,y-axis has been shown to reduce mAs by 31-39% in the setting of pediatric chest CT [6] and to reduce mean effective mAs by 16.9% in the setting of adult chest CT [5].
Note, however, that tube current reductions are surrogates for the actual patient dose. The current study provides additional insight into the work of prior studies by also assessing the effect on organ doses when automatic tube current modulation is used because dosimetry was performed, rather than recording estimated dose changes based on alterations in tube current. MOSFET dosimetry offers an opportunity for more accurate dose determinations in CT [19]. The 37.5% average organ dose reduction in the current anthropomorphic phantom study with the use of Auto mA tube current modulation with the shield placed after the scout image, then, is slightly higher than previously reported mAs reductions in the setting of adult chest CT with z-axis automatic tube current modulation. This difference may reflect a combination of different automatic tube current modulation technologies, different noise indexes, or automatic tube current modulation when used in a phantom. The 25.5% average dose reduction with Auto mA tube current modulation with the shield present in the scout is similar to previous investigations.
The locations (stomach and left kidney) in which the measured dose did not change with Auto mA were in the upper abdomen, where tube current was the same (65 mA) in the fixed tube current and the automatic tube current modulation scans, likely because an equivalent amount of tube current was necessary to maintain the target noise level. Noise levels in the sections containing the stomach and left kidney dosimeters were similar in the automatic tube current modulation and fixed tube current scans (data not shown).
The use of breast shields in combination with automatic tube current modulation has not been previously addressed. It is conceivable that with the type of z-axis modulation used in this investigation (combined z- and x,y-axis modulation was not available at the time of our study), the increased attenuation present on the scout image due to the shield (and the concomitant increase in tube current in this region) could negate the benefits of both shielding and tube current modulation. We found that Auto mA z-axis modulation compensated for the density of the shield if the shield was present in the scout radiograph by increasing the tube current along the z-axis in the range of the shield (Fig. 3). As seen in Figure 3, the tube current began to increase at the leading edge of the breast shield. The rate of rise increased dramatically after 15 slices, or approximately 7.5 cm beyond the leading edge of the shield. The reason for this unexpected delay in increasing tube current is not known. Although there is a lag before modulation begins, it is relatively short (80 milliseconds). Further analysis of the data with the manufacturer did not provide an explanation (Toth T, GE Healthcare, personal communication).
However, if the modulation began sooner in the upper thorax, we would have expected an even greater dose savings with Auto mA tube current modulation when the shield was placed after the scout. That tube current did not decrease beyond the shield before reaching the abdomen can be explained by the fact that the Auto mA tube current modulation software derives the tube current on the basis of the highest attenuating structure under the detectors at a given time. A portion of the detectors was likely always either over the shield or over the upper abdomen at the end of the scanning. However, even though Auto mA compensated for the shield when the shield was present in the scout, the selected tube currents through the positions that included the shield (40-62 mA) still were lower than the fixed tube current of 65 mA used in the fixed-tube-current, age-based scan. Therefore, some but not all of the tube current benefit was lost when the shield was present in the scout radiograph.
Furthermore, we postulated that placing the shield after the scout image was obtained would maximize dose benefits from both shielding and Auto mA tube current modulation. In fact, placing the shield after the scout was obtained resulted in radiation doses (measured using MOSFET technology) that were lower at 10 locations, unchanged at three locations (right breast, left breast, skin over breasts), and increased at one location in the range of shielding. The effective dose, as expected, decreased by 13%, from 1.6 to 1.4 mSv, when the shield was placed after the scout radiograph with Auto mA tube current modulation.
That breast dose and skin dose superficial to the breasts did not change when the shield was placed after the scout was obtained while almost all the other organ doses in the range of the shield decreased was somewhat unexpected. One explanation is that the breast detectors were located near the leading edge of the shield, where the differences in tube current were not so great (43-50 mA when shield was present in scout image vs 34-35 mA when shield was placed after scout image was obtained). Given the spiral nature of the primary beam, it is also possible that the position of the beam (e.g., posteroanterior vs anteroposterior) relative to the breast detectors was different in the automatic tube current modulation runs with shield placed before and after the scout image was obtained. A posteroanterior beam would be more attenuated by soft tissues before it reaches the breast detectors, whereas an anteroposterior beam would be attenuated by the breast shield before reaching the detectors. This phenomenon has been reported recently with MOSFET and breast dosimetry and is a limitation of this type of surface dosimetry (unlike more central dose locations) in helical as opposed to axial MDCT [29].
It was also unexpected that the average dose in the bone marrow of the right ribs increased when the shield was placed after the scout (Table 1). As noted earlier, all other measured doses in the range of the shield decreased or were unchanged when the shield was placed after the scout with Auto mA. The measured doses at this location for each of the four Auto mA runs (two with the shield present in the scout and two with the shield placed after the scout) were 0.12, 0.14, 0.13, and 0.22 cGy. The 0.22-cGy value appears to be an outlier that may have falsely elevated the average.
Breast doses of 0.1-0.7 Gy have been shown to be associated with an increased risk of breast cancer [25, 26]. Although the breast dose in our study was low (0.4 cGy without the shield), appropriately minimizing any radiation dose, especially for breast parenchyma in young individuals, is prudent. In keeping with this ALARA principle, we found that breast dose was reduced by 26% with the breast shield alone and by 52% when using the breast shield and automatic tube current modulation in an anthropomorphic phantom representing a 5-year-old child.
When measured doses decreased, image noise increased. The highest-dose scan (fixed tube current chest CT without shielding) was also the least noisy. The lower noise measured in the nonmodulated examination was due, we think, to tube current levels that were higher than needed (such as in the lungs). For nonmodulated examinations, it is likely that protocols have evolved to maintain a minimum noise level for diagnostic quality through intrinsically higher attenuating areas, meaning other regions will have less noise than is needed. For example, in the nonmodulated examinations, average noise measurements cranial to the shield were 9.3 and 9.4 H (SD) and caudal to the shield were 11.4 and 11.1 H (SD) in the runs with and without the breast shield, respectively (Table 3).
The lowest-dose scan (automatic tube current modulation and shield, with shield placed after the scout image was obtained) was the noisiest scan. The increase in image noise at the level of the heart and superior mediastinum (in the range of shielding) when the shield was placed after the scout was significantly (p < 0.01) higher compared with when the shield was present in the scout radiograph with automatic tube current modulation. It is unclear why the heart and not the superior mediastinum had a significantly higher amount of noise with nonmodulated scanning with and without the shield. This may represent a statistical error.
When the shield was placed after the scout image was obtained, the measured noise index (12.9 H [SD]) in the range of shielding was slightly higher than the selected target noise index of 12.0 H (SD). Although this increase in noise is statistically significant, it does not necessarily equate to a reduction in diagnostic quality. For example, a prior investigation (Darsie J et al., presented at the 2005 annual meeting of the Radiological Society of North America) supports a noise index of 12.9 H (SD) to be in the acceptable range for pediatric chest CT. In addition, the target noise index was 12.0 H (SD), and the breast shields increased noise above this level by 0.9 H (SD), a difference that is arguably below threshold for visual detection. In the automatic tube current modulation runs, not unexpectedly, noise measurements were similar both cranial to and caudal to the shield. The automatic tube current modulation algorithm modulates tube current to maintain a constant noise level throughout the scanning.
Limitations of our study include that only one type of CT scanner and therefore only one automatic tube current modulation technology was investigated. For example, not all automatic tube current modulation products described by McCollough et al. [28] modulate on the basis of the scout radiograph. Therefore, the effect of placing the shield before or after the scout radiograph cannot be predicted in these settings. The use of an anthropomorphic phantom representing a single age is also a limitation of this study. However, it is expected that similar dose reductions would be seen with anthropomorphic phantoms representing different ages. Also, because there are no respiratory or cardiac motion artifacts in a phantom study, the doses and image noise changes we found may not completely translate into a patient study. The use of the SD of Hounsfield units in a region of interest as a proxy for image quality is also a limitation of this study. The effect of the combined use of bismuth shields and automatic tube current modulation on diagnostic quality of images needs to be assessed. Finally, our conversion factor used to compute effective dose was an estimate of a chest CT protocol for a 5-year-old because standards are not available for pediatric chest 16-MDCT. Although the effective dose would differ if another conversion factor were used, the change in the effective dose we determined (13% decrease) when the shield was placed after the scout compared with when the shield was present in the scout is independent of the factor used.
In conclusion, the use of a bismuth breast shield in addition to one technique for z-axis automatic tube current modulation further reduces detected radiation dose as determined in a pediatric anthropomorphic phantom. Dose reduction is greatest when the breast shield is placed after the scout radiograph is obtained. Although placing the breast shield after the scout radiograph is obtained does result in increased image noise in the range of the shield because the tube current is calculated without factoring in the attenuation of the breast shield, the increased image noise is still close to the noise target level.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
