| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Department of Radiology, University of California Davis Medical Center, Research Imaging Center, 4701 X St, Sacramento, CA 95817. Received April 25, 2002; revision requested June 21; final revision received September 24; accepted November 5. Supported in part by grants from the California Breast Cancer Research Program (7EB-0075) and the National Cancer Institute (R01-CA89260). Address correspondence to J.M.B. (e-mail: jmboone@ucdavis.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
MATERIALS AND METHODS: Six phantom cylinders (10–32 cm in diameter) were scanned with a clinical multi–detector row CT scanner. CT noise was statistically characterized for CT technique factors from 80 to 140 kVp and from 10 to 300 mAs. Dose measurements were performed with each phantom. Dilute iodine and tissue contrast were determined with computer calculations validated with measured data. The dose, noise, and contrast data were computer fit, and pediatric CT technique factors (milliampere seconds) necessary to maintain the contrast-to-noise ratio (CNR) were computed.
RESULTS: As compared with that in a reference cylindric adult abdomen of 28 cm in diameter, CNR was maintained at a constant level in pediatric patients of 25, 20, and 15 cm in diameter, respectively, when milliampere second values of 0.557, 0.196, and 0.054 of the adult milliampere second values were used. The corresponding doses were reduced to 0.642, 0.287, and 0.090 of the 28-cm-diameter adult dose, respectively. CT techniques for examination of pediatric heads measuring 15 and 13 cm, respectively, can involve the use of milliampere second values of 0.572 and 0.366 of those used for examination of a standard 17-cm-diameter adult head.
CONCLUSION: CT technique charts for pediatric abdominal and head examinations were produced on the basis of physically measured data; use of these tables will enable pediatric radiation dose to be reduced while CNR is preserved.
© RSNA, 2003
Index terms: Abdomen, CT, 78.1211 • Computed tomography (CT), in infants and children • Computed tomography (CT), radiation exposure • Head, CT, 18.1211, 28.1211 • Radiations, exposure to patients and personnel
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
Chan et al (11) performed CT in children aged 1–12 years with several different milliampere second values, and, by using observer-based subjective image assessment, found that a 40% reduction in milliampere seconds could be used in pediatric cranial CT. Cohnen et al (12) also studied CT dose in pediatric head CT and concluded that a 40% reduction was possible.
Huda et al (13) evaluated abdominal pediatric CT from a physics perspective by using computer computations and reported techniques as a function of patient mass that would deliver constant energy fluence to the CT detectors. Brody (14) has discussed pediatric thoracic CT in general terms and has provided specific recommendations for milliampere second reductions on the basis of patient mass. Whereas many reports address the need for or use of pediatric dose reduction in CT in a clinical manner (15–17), the purposes of this study were to determine patient size–dependent technique factors by using physically measured objective data and to provide guidance on reduction of technique factors on the basis of patient dimensions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
For each patient, the CT image corresponding to the location of the first lumbar vertebra was selected from the CT data set, and the patient’s torso was outlined on this image by one of the authors (E.M.G.), who used mouse-and-cursor software developed for this task. The outline data points were properly scaled by using the known pixel dimensions for each image, and the greatest horizontal dimension (patient width), greatest vertical dimension (patient thickness), and patient circumference were determined from the outline information. An equivalent diameter was computed for each patient, where "patient equivalent diameter" refers to the diameter of a circle that had the same circumference as a patient’s actual elliptic circumference. The equivalence metric was also evaluated by using an equal-area criterion and produced results similar to those of the circumference-based metric.
To verify that the shape of adult patients was similar to that of pediatric patients, we evaluated CT data obtained in 35 pediatric patients by using our PACS according to the same institutional review board–approved protocol. Abdominal CT images were evaluated at the level of L1, and the width and height of the torso were measured as with the adult patients. The mean age of this 35-patient cohort was 9.09 years (SD = 5.08 years), with a minimum age of 0.5 years and a maximum age of 17 years. There were eight patients from the 0–5-year age group, 10 from the 6–10-year age group, 15 from the 11–15-year age group, and two from the 16–17-year age group. The circumference of the pediatric patients was not known; however, the shape (thickness-to-width ratio) of the patients could be compared. The mean patient thickness-to-width ratio was computed and compared statistically with that in the adult cohort.
To apply the presented techniques to head CT studies, we required knowledge of the typical dimensions of the adult head. Results of head CT examinations in 32 adults (mean age, 47.6 years; SD, 16.6 years) were evaluated by using our PACS, and the anteroposterior length and width of the head were measured at the level of the ocular lens. The measurements were converted to elliptic areas and then to equivalent head diameters by using standard geometric equations.
CT Scanning
A commercial multi–detector row CT scanner (Lightspeed; GE Medical Systems, Waukesha, Wis) was used for the evaluation of image noise, radiation dose, and CT contrast. This scanner has 16 detector rows that can be configured to yield four different section thicknesses: 1.3, 2.5, 3.8, and 5.0 mm. Nonhelical CT techniques were used throughout this investigation, although the principles developed here apply equally to helical CT techniques. For each CT measurement, four simultaneous images with a 5-mm section thickness were acquired by using (nominal) 20-mm collimation. It is generally acknowledged that multi–detector row CT scanners are not as dose efficient as single-detector systems because the x-ray field penumbra in multi–detector row systems is typically positioned to be off the active edge of the detector surfaces (18). The data in this investigation were evaluated in a manner that deemphasized the dose efficiency of a particular scanner (by creating a ratio that cancels this out), and, thus, our results are thought to be applicable to other CT scanner types.
Noise Measurements
Two polymethylmethacrylate (PMMA) phantoms—commercially available American Association of Physicists in Medicine phantoms (Inovision [formerly Nuclear Associates], Cleveland, Ohio) of 16 and 32 cm in diameter—were already available in our laboratory (19). Four other cylindric PMMA phantoms (10, 13, 20, and 25 cm in diameter; 15 cm in length) were designed and machined specifically for this project. Two 12-mm-diameter holes to accommodate a pencil ionization chamber had been drilled into each phantom; one hole was placed at the center of the phantom, and one hole had its center placed 18 mm from the edge of the phantom.
For noise measurements, each phantom was scanned (E.M.G.) at all four tube voltages possible with the CT scanner: 80, 100, 120, and 140 kVp. At each tube voltage, and for each phantom diameter, CT scans were acquired at 10, 20, 30, 50, 70, 100, 150, 200, and 300 mAs. For the three largest phantoms, additional scans were acquired at 440, 600, and 800 mAs. All holes in the phantoms were plugged with PMMA rods or filled with water for this portion of the study. All acquisitions were performed by using a 5-mm section thickness with a 20-mm nominal collimator setting, and each transverse acquisition resulted in the production of four CT images on the quad-detector scanner.
The CT images were reconstructed by using the standard reconstruction filter. The scanning field of view and display field of view for each of the six phantoms were changed to best accommodate the size of the phantom, as is done in routine clinical practice for patient imaging. The display fields of view used were 20, 20, 23, 25, 30, and 38 cm, respectively, for the 10-, 13-, 16-, 20-, 25-, and 32-cm-diameter PMMA phantoms. Neither the phantom nor the table increment was moved between acquisitions, and thus, for each phantom size, the images were nearly perfectly registered for all tube voltage and milliampere second combinations.
The images were transferred from the CT scanner to a Pentium–based (Intel, Santa Clara, Calif) imaging workstation by using the PACS at our institution. Software was custom written (J.M.B. with C/C++ 5.0; Microsoft, Redmond, Wash) to read the CT images, display them, and perform image analyses. The software allowed the placement of square regions of interest (ROIs) on the images obtained with 120 kVp and 300 mAs; these same regions were then analyzed on all the images of each size phantom (ie, on all four images obtained per acquisition performed with each of four tube voltages and each of nine to 12 milliampere second products). Use of the same ROIs was possible because all images for each size phantom were spatially co-registered. Because each of the six phantoms had different hole patterns and diameters, ROIs were manually (J.M.B.) selected to avoid holes. The mean CT number and root-mean-square SD (ie, noise) of the CT numbers was computed for each ROI, and the noise values were averaged across multiple ROIs per image and across the four images obtained with each tube voltage and milliampere second combination. For the evaluation of CT noise per se, measurements were made on ROIs positioned over the PMMA phantom and not in the holes.
Because x-ray noise is Poisson distributed, the (zero frequency) image noise (
) should theoretically be governed by an equation such as = a(mAs)b, where a is a constant and ideally the value of b would be -
, corresponding to the Poisson relationship between image noise and dose in which dose is a linear function of milliampere seconds for a given tube voltage and phantom size. The measured noise values were fit by using this equation, and the fit parameters could be used to accurately interpolate (predict) the CT image noise for phantom thicknesses ranging from 10 to 32 cm in diameter.
Determination of Image Contrast
Contrast on CT images was determined as the difference in CT numbers (in Hounsfield units) between two materials. For the evaluation of contrast, the six PMMA phantoms were rescanned at each tube voltage and at 300 mAs. The middle hole (at the concentric center of the cylindric phantom) was filled with a 0.5% concentration of iodine-based contrast agent (diatrizoate meglumine, 60% Hypaque; Nycomed, New York, NY) in purified water. The hole at the edge of each phantom was filled with purified water. The resulting CT images were evaluated (J.M.B.) as described above for the mean CT numbers of PMMA, water, and dilute iodine contrast agent. These values were measured for all six phantom diameters and all four tube voltages studied. The CT numbers for each material, diameter, and tube voltage were averaged across four CT images.
The physically measured values for the contrast of iodine were calculated as the difference between the CT numbers of the iodinated solution and the CT numbers of water. These data were used to validate computer-simulated estimates of iodine contrast, and the computer-simulated iodine contrast values were used in subsequent CNR computations. PMMA contrast was determined by calculating the difference between the CT numbers of PMMA and those of water. PMMA does not resemble soft tissue well at CT due to its composition, so PMMA contrast values were used here only to validate the accuracy of the computer simulation. The computer simulation values for soft-tissue contrast (ie, the contrast between muscle and adipose tissues) were then used for subsequent evaluation (J.M.B.). The computer simulation methods that were used are described in the Appendix.
Dose Measurements
The CT dose index (CTDI) (20) is typically used for reporting CT scanner dose performance; however, a number of different CTDI values have evolved. In this study, we used CTDI100mm, which refers to the measurement of the CT radiation dose profile integrated along the 100-mm length of the pencil chamber (20,21). Herein we use the term CTDI generally and apply it to measurements performed with PMMA cylinder diameters of any size.
The CTDI100mm was measured at the edge and at the center of each phantom. The weighted CTDI metric, CTDIw, was computed from the edge and center CTDI100mm measurements, and the CTDIw was used for further analysis. The CTDIw is thought to better approximate the mean patient dose than the center or edge measurements. This value was calculated by using the following equation: CTDIw = 
CTDI100mm(center) + 
CTDI100mm(edge), where CTDI100mm(center) is the CTDI100mm value measured in the center of the phantom and CTDI100mm(edge) is the CTDI100mm value measured at the edge of the phantom. Dose measurements were made in PMMA cylinders (J.M.B.) by using a 3-mL pencil ionization chamber with an MDH 1015 exposure meter (Radcal, Monrovia, Calif). After reproducibility of the meter measurements and the CT scanner measurements had been determined by using multiple acquisitions per data point, a single dose measurement was made over a range of parameters. Because dose scales linearly with milliampere seconds for the same-size phantom at the same tube voltage, a setting of 100 mAs was used at all four tube voltages (80, 100, 120, and 140 kVp) and for all six phantom diameters.
The CTDI100mm(edge) and CTDI100mm(center) values were computer fit as a function of patient diameter by using commercially available curve-fitting software (TableCurve 2D; Jandel Scientific, Corte Madera, Calif). The parameterized fits (equations that predict the best curve) were then used to interpolate radiation dose for any patient diameter, and CTDIw was computed from the parameterized values (J.M.B.).
Computer Analysis
The physical CT measurements (E.M.G.) led to the complete computer characterization (J.M.B.) of CT noise and radiation dose (CTDIw). CT contrast (of soft tissue and iodine) was computed by using a computer simulation (J.M.B.), but the computer simulation results were validated by experimental measurements. All data derived from physical measurements and computer simulation were computer fit by using the aforementioned commercially available software. Custom software that enabled numerical least squares solutions for matching CNRs between the different tube voltages and patient diameter combinations was written. With this approach, the milliampere seconds were used as the independent variable and the CNR was matched to reference values on the basis of the minimum least squares criterion. Once the milliampere second value was determined in this manner, it could be used to compute radiation dose for a given phantom diameter and tube voltage.
Other Considerations
Corrections for the differences in composition between PMMA phantoms and human patients were made on the basis of the following analysis (J.M.B.): Patients were included in the model as unit-density water (H2O), and the density of PMMA (C5H8O2) was assumed to be 1.19 g/cm3. The four spectra (at each tube voltage) were simulated (22) to match the measured half-value layers of the CT scanner, and the x-ray attenuation of PMMA was computed for phantom diameters ranging from 8 to 40 cm in 1-cm intervals by using elementally weighted attenuation coefficients (23). For each x-ray spectrum, the diameter of a water-equivalent (mathematic) phantom was computed such that the energy fluence attenuation was identical to that of the PMMA cylinders. The (linear) relationship found between PMMA cylinder diameter and patient diameter was then used to convert the PMMA cylinder diameters to patient diameters.
Statistical Analysis
Linear regression was performed for thickness and width versus equivalent diameter by using Excel (Microsoft). Tests of significance in regard to the slope were performed with the two-sided Student t test, with the t-distribution data provided by the statistical functions in Excel. To demonstrate that the shape of the 87 adult patients was similar to that of the 35 pediatric patients, we computed the thickness-to-width ratio for each patient, and the mean and SD were assessed. The two-tailed t distribution was used as a test for significant differences. P 
.05 was considered to indicate a statistically significant difference.
The CT noise–versus–milliampere seconds analysis was performed by using TableCurve 2D. The CTDI100mm values were also computer fit with TableCurve 2D by using a variety of analytic equations. The experimentally measured contrast values were compared with computer-modeled values by using the Student t test.
The milliampere second and dose reduction data were evaluated by using the simple average and root-mean-square SD computation functions in Excel.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
|
The torso shape—that is, the mean thickness-to-width ratio—for the 87 adult patients was 0.745 (SD = 0.061). The torso thickness-to-width ratio for the 35 pediatric patients was 0.718 (SD = 0.054). Comparison of these values with a two-sided t test indicated that there was no significant difference (P = .657) between the shapes of pediatric and adult patients.
The median adult abdominal diameter was found to be 28.4 cm (range, 20.8–38.3 cm), and 44 (50%) of the 87 adults evaluated had diameters ranging from 26.7 to 31.1 cm. This result led us to consider an adult patient equivalent diameter of 28 cm as the "typical" size. The median head equivalent diameter in the 32 adult patients evaluated was 17.48 cm (mean, 17.42 cm; SD = 0.72 cm; range, 15.9–18.6 cm). This result led us to set 17.0 cm as the typical adult head equivalent diameter. The mean ratio of width to anteroposterior dimension of the adult head was 0.787 (SD = 0.0384). In our analysis, the use of typical abdominal and head diameters that were slightly smaller than the median values allowed us to err in the direction of better image quality (ie, higher CNR) and less substantial dose savings.
For 32 section thicknesses (8–40 cm) and four x-ray spectra (128 points in all), the relationship between the equivalent water thickness (Twater) and the PMMA thickness (TPMMA) was evaluated by using linear regression according to the following equation: Twater = m x TPMMA (r2 = 0.9994), where the slope, m, is 1.1207 (m = 0.000914). With this relationship, a 10-cm-diameter PMMA cylinder was found to be equivalent in terms of x-ray attenuation to a 11.2-cm-diameter water cylinder, and a 32.0-cm-diameter PMMA cylinder was found to be equivalent in attenuation to a 35.9-cm-diameter water cylinder. This relationship was used to scale the PMMA diameters to water-equivalent diameters in the reporting of subsequent results, because we believed this correction to be prudent for our presentation of the results in terms of patient-relevant dimensions.
The CT noise measurement results are shown in Figure 2. The symbols correspond to the measured noise values, and the solid lines represent the best fit results according to the following equation: X = aYb, where a is a constant unique to the phantom diameter and tube voltage. This equation produces straight lines with slope b when seen on the log-log axes of the four scatterplots in Figure 2. The regression fit results for the 24 regression lines shown in Figure 2 averaged r2 = 0.9939 (0.0075), with r2 = 0.968 being the worst value. The slope of the lines for the smaller phantom diameters was approximately -, which Poisson statistics for x rays would predict (ie, noise 
dose-1/2). For the larger phantom diameters, however, the slopes became higher in magnitude, probably owing to scattered radiation and other sources of noise in a photon-starved environment.
|
|
Radiation dose, as quantified by the CTDI100(edge) and CTDI100(center) indices, is shown in Figure 4. At constant milliampere seconds, increasing the tube voltage increased both the average energy per x-ray photon and the number of x-ray photons in the beam. Thus, studies with higher tube voltage (at equal milliampere seconds) increased the dose substantially. Smaller-diameter patients experience slightly lower x-ray entrance fluence due to the inverse-square law; however, the dose to smaller-diameter patients is appreciably higher at the same tube voltage and milliampere second settings. This is because there is less self shielding in smaller patients than there is in larger patients. This is analogous to the situation in mammography, where the roentgen-to-rad conversion factors (normalized glandular dose values) are higher for smaller breasts (24,25). The radiation dose curves in the present study were computer fit, with r2 > 0.993 in all cases.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
The section thickness is not automatically reduced for pediatric patients, however. Therefore, where appropriate, the radiologist should consider developing pediatric CT protocols in which the section thickness is reduced in smaller patients to improve spatial resolution along the z axis of the patient. The solid lines (with data points and CIs) shown in Figure 5, A (for relative dose), and Figure 5, B (for milliampere seconds), correspond to use of a 5-mm section thickness, and the dotted lines correspond to use of section thicknesses of 2.5 and 1.3 mm.
When the section thickness is halved (eg, from 5.0 to 2.5 mm), the milliampere second value needs to be doubled to maintain the same CNR, and a four-fold reduction in section thickness (eg, from 5.0 to 1.3 mm) requires that the milliampere second value be quadrupled to maintain constant CNR. Reducing the section thickness by a factor of two and increasing the milliampere second value by a factor of two results in the same energy being imparted for a single acquisition (26,27). However, clinical reality necessitates that a fixed-length region of the patient be scanned, regardless of section thickness. Consequently, halving the section thickness and doubling the milliampere second value requires the acquisition of twice as many contiguous images. With this scenario, both the energy imparted and the mean radiation dose are doubled.
There has been substantial recent interest in the radiation dose received by pediatric patients at CT imaging (10). Because of the exponential relationship between patient thickness and x-ray attenuation, not only is dose reduction possible in pediatric CT, but very large dose reductions are possible for the smallest children. The data in Table 1 in this report demonstrate that children with torso dimensions of 20 cm (equivalent diameter) require only 0.196 of the milliampere second value that the adult (28-cm-diameter) patient requires for the same image CNR to be maintained. This reduction of milliampere seconds results in a radiation dose that is just 0.287 of the adult dose. For the smallest of patients (12-cm equivalent diameter), the radiation dose can be reduced to 0.050 of the adult dose by adjusting the milliampere second value to 0.028 of the adult value.
In practice, a dose reduction of 95% can be traded off for better spatial resolution. For example, the 5-mm section thickness can be adjusted to 1.3 mm, providing a four-fold increase in z-axis resolution (and a four-fold increase in milliampere seconds) while increasing the relative dose from 0.050 to only 0.199 (ie, still providing a dose reduction of 80% compared with the dose in the adult patient). The notion of maintaining constant CNR presumes that the detection task is equivalent between adults and pediatric patients. If there are clinical reasons to suspect that the contrast of a structure is intrinsically lower in the pediatric patient, then one could lower the amount of dose reduction suggested here.
The milliampere second and dose reduction factors presented in this report were determined by combining both iodine and soft-tissue CNR curves, as well as combining the results from CT performed with 100, 120, and 140 kVp. When these individual data are normalized to unity at 28 cm (as in Fig 5), the curves almost completely overlie each other. The very small gray regions (representing ±2 SD) in Figure 5 demonstrate this. Combining the iodine and soft-tissue curves is consistent with the notion that contrast material–enhanced CT images still must demonstrate good CNR in soft-tissue areas. Although our reduction factors resulted from combining data from CT performed with three different tube voltages, it is our experience that the majority of CT studies are performed at 120 kVp. The milliampere second and dose reduction curves presented assume that the milliampere seconds–reduced CT study of a pediatric (or small adult) patient is performed at the same tube voltage as that used in the nominal adult protocol for that type of examination.
Brenner et al (10) estimated that 500 individuals younger than 15 years who are examined with adult CT techniques in the United States will eventually die from cancers attributable to the radiation received. The data given in Table 1, coupled with diameter-versus-age estimates (28) and year 2000 population-versus-age data from the U.S. Census Bureau (29), allowed us to estimate the relative mortality from pediatric CT if the milliampere second values are reduced as suggested in Table 1. The population distribution for the age range of 0–14 years (inclusive) was used to weight the dose reduction factors given a section thickness of 5 mm that are listed in Table 1. Prior to this weighting, age was converted to patient diameter by using reference data (28). If the dose reduction factors in Table 1 are used, and assuming that CT examination frequency is evenly distributed across ages 0–14, population doses would be reduced to about 23% of current levels. Given the estimate of Brenner et al (10) that 500 fatal cancers would occur with use of current CT techniques, the use of the technique factors in Table 1 would reduce the number of fatal cancers in the 0–14 age range to 116, reducing the number of fatal cancers per year by 384.
The technique and dose reduction computations reported in this study were derived from use of a single type of CT scanner (a quad-detector GE Lightspeed). The dose efficiency (as expressed by the CTDI per 100 mAs) of other scanners may vary substantially. However, given that dose efficiency is generally a multiplicative factor, the data shown in Figure 5 and given in Table 1 were normalized to cancel out scanner-specific dose efficiency factors. Thus, the data should be generally applicable to other types of scanners. Nevertheless, for those interested in developing their own dose reduction tables for other types of CT scanners, the methods described here may be useful toward that end. The conversion factors in Table 1 used to adjust milliampere seconds and dose to the 2.5-mm and 1.3-mm section thicknesses were applied with the assumption that dose efficiency does not change as a function of section thickness, but most scanners are in fact less dose efficient with thinner sections. The CTDI100mm values for different section thicknesses published by scanner manufacturers (or measured by the user) can be used to tailor the data in Table 1 to a specific scanner model.
We also note that for larger patients, an increase in milliampere seconds is required to achieve a constant CNR— for example, a 35-cm-diameter patient would require a 4.3-fold increase in milliampere second values. Alternatively, for larger patients, the radiologist may consider developing protocols in which section thickness rather than the milliampere second value is increased. Doubling the section thickness from 5 to 10 mm achieves the same effect (in terms of CNR) as doubling the milliampere second value at a constant section thickness: In both cases, twice as many photons reach the detector. With this in mind, and with the 35-cm-diameter patient as an example, the section thickness could be doubled (from 5 to 10 mm) and the milliampere second value could be increased by half of the recommended value give in Table 1 (for a milliampere second value increase of 2.16) to achieve the same CNR levels. Increasing the pitch also reduces dose by a factor of p/p', where p is the previous pitch and p' is the newer, higher pitch value.
An alternative to measuring the patient’s width with calipers is to measure the patient’s circumference with a cloth measuring tape—this is probably a more accurate measurement and would take only a few seconds to perform. The act of a technologist measuring the pediatric patient’s circumference is an overt demonstration of an imaging center dose reduction program, which would be appreciated by parents attuned to the pediatric CT dose issue. We have found that application of standard distance measurement software available with the CT scanner to the scout view provides a good determination of patient width (or thickness), and this software is much faster for the CT technologist to use.
The following example demonstrates the approach to reducing milliampere seconds discussed in this study:
The standard protocol for an adult abdominal CT examination (in a patient of approximately 28 cm in diameter) at a certain institution is 120 kVp, 5-mm section thickness, a helical pitch of 1.5, and 320 mAs. A pediatric patient arrives for an abdominal CT examination, and the technologist measures the circumference of the patient (in the region of the torso to be scanned) by using a cloth measuring tape and finds a circumference of 69 cm.
The 69-cm patient circumference corresponds to an equivalent diameter of 22 cm (Table 1, 69/
= 22). The milliampere second reduction factor with a 5-mm section thickness listed in Table 1 is 0.304, and, thus, the appropriate milliampere second value at CT is 97.3 mAs (ie, 320 mAs x 0.304). The technologist rounds up to 100 mAs on the scanner (milliampere second selection on CT scanners is discrete), and the patient is scanned with 120 kVp, 100 mAs, a 5-mm section thickness, and a pitch of 1.5. The dose to the patient has been reduced by a factor of 0.42 ([0.408 x 100]/97.3), and, thus, a 58% (100 x [1-0.42]) reduction in dose was achieved while the same CNR as in the standard adult CT study was maintained.
The technique factors provided allow the development of pediatric CT protocols in which constant image quality (ie, CNR) is maintained but the radiation dose to pediatric patients is dramatically reduced. A spreadsheet of Tables 1 and 2 will be provided to any interested party to use as their CT scanner "technique chart," upon e-mail request to the first author. The spreadsheet can be easily updated to include the consequences of changing pitch and the dose inefficiency that occurs with use of thinner CT sections. Furthermore, the milliampere second reduction factors can be multiplied by the typical milliampere second values used in adult CT at an institution, saving the technologist the need to compute milliampere second reduction factors on a case-by-case basis. We believe that use of such a chart would result in the production of relatively constant image quality across a range of patient sizes.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
(E), a thickness t of a substance (iodine, muscle, adipose tissue, or PMMA) with linear attenuation coefficient µ(E), and a PMMA cylinder with attenuation coefficient 
(E) and diameter D, the effective linear attenuation coefficient, µeff, was computed for the evaluated substance as follows:
|
The effective linear attenuation coefficient of a substance, µeff, and that of water, µw, were computed as above, and the CT number was computed by using the following equation (18): CT number = 1,000(µeff - µw)/µw. The elemental composition of PMMA was assumed to be C5H8O2, and the muscle and adipose tissue compositions used were those reported by Johns and Cunningham (30). Dilute iodine contrast (Ciodine) was computed as Ciodine = CTiodine+water - CTwater, where CT represents the CT number in Hounsfield units. The modeled iodine concentration was best fit by using the least-squares technique, and the resulting value was consistent with the known iodine concentration. Soft-tissue contrast was computed as Csoft tissue = CTmuscle - CTadipose.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, J.M.B.; study concepts, J.M.B., E.M.G., S.W.G.; study design, J.M.B., E.M.G., J.A.S.; literature research, E.M.G., J.M.B.; clinical studies, J.M.B., J.A.S., E.M.G.; experimental studies, J.M.B., E.M.G.; data acquisition, J.M.B., E.M.G., J.A.S.; data analysis/interpretation, J.M.B., E.M.G.; statistical analysis and manuscript preparation, J.M.B.; manuscript definition of intellectual content, J.M.B., S.W.G.; manuscript editing, J.M.B.; manuscript revision/review, J.M.B., J.A.S.; manuscript final version approval, J.M.B., S.W.G.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONAPPENDIXREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J BRANDBERG, L LONN, E BERGELIN, L SJOSTROM, E FORSSELL-ARONSSON, and G STARCK Accurate tissue area measurements with considerably reduced radiation dose achieved by patient-specific CT scan parameters Br. J. Radiol., October 1, 2008; 81(970): 801 - 808. [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]
|
|
|
|
|
|
J. G. Eichhorn, C. Jourdan, S. L. Hill, S. V. Raman, J. P. Cheatham, and F. R. Long CT of Pediatric Vascular Stents Used to Treat Congenital Heart Disease Am. J. Roentgenol., May 1, 2008; 190(5): 1241 - 1246. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Masciari, A. D. Van den Abbeele, L. R. Diller, I. Rastarhuyeva, J. Yap, K. Schneider, L. Digianni, F. P. Li, J. F. Fraumeni Jr, S. Syngal, et al. F18-Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography Screening in Li-Fraumeni Syndrome JAMA, March 19, 2008; 299(11): 1315 - 1319. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Tubiana, S. Nagataki, L. E. Feinendegen, D. A. Dimitroyannis, D. P. Frush, M. J. Goske, M. Hernanz-Schulman, P. Soyer, H. Varnholt, D. J. Brenner, et al. Computed Tomography and Radiation Exposure N. Engl. J. Med., February 21, 2008; 358(8): 850 - 853. [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]
|
|
|
|
|
|
P. Bohy, V. de Maertelaer, A. Roquigny, C. Keyzer, D. Tack, and P. A. Gevenois Multidetector CT in Patients Suspected of Having Lumbar Disk Herniation: Comparison of Standard-Dose and Simulated Low-Dose Techniques Radiology, August 1, 2007; 244(2): 524 - 531. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A J Van Der Molen, W J H Veldkamp, and J Geleijns 16-slice CT: achievable effective doses of common protocols in comparison with recent CT dose surveys Br. J. Radiol., April 1, 2007; 80(952): 248 - 255. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P C Shrimpton, M C Hillier, M A Lewis, and M Dunn National survey of doses from CT in the UK: 2003 Br. J. Radiol., December 1, 2006; 79(948): 968 - 980. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. C. Chan, B. N. Joe, F. V. Coakley, E. L. Prien Jr, R. G. Gould, S. Prevrhal, W. C. Barber, K. S. Kirkwood, A. Qayyum, and B. M. Yeh Gallstone Detection at CT in Vitro: Effect of Peak Voltage Setting. Radiology, November 1, 2006; 241(2): 546 - 553. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. H. McCollough, M. R. Bruesewitz, and J. M. Kofler Jr CT Dose Reduction and Dose Management Tools: Overview of Available Options. RadioGraphics, March 1, 2006; 26(2): 503 - 512. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Funama, K. Awai, Y. Nakayama, K. Kakei, N. Nagasue, M. Shimamura, N. Sato, S. Sultana, S. Morishita, and Y. Yamashita Radiation Dose Reduction without Degradation of Low-Contrast Detectability at Abdominal Multisection CT with a Low-Tube Voltage Technique: Phantom Study Radiology, December 1, 2005; 237(3): 905 - 910. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. H. Mulkens, P. Bellinck, M. Baeyaert, D. Ghysen, X. Van Dijck, E. Mussen, C. Venstermans, and J.-L. Termote Use of an Automatic Exposure Control Mechanism for Dose Optimization in Multi-Detector Row CT Examinations: Clinical Evaluation Radiology, October 1, 2005; 237(1): 213 - 223. [Abstract] [Full Text] [PDF] |
) and thickness (T, 
) as a function of patient diameter. (W and T are defined geometrically in the inset image, along with the equivalent diameter, Diaeq.) These data demonstrate that the relative proportions of width and thickness remain approximately constant among patients of different size. Identity = line of identity (where x = y).
= 120 kVp, 
