Original ResearchFree Access

Diagnostic Reference Ranges for Pediatric Abdominal CT

Published Online:https://doi.org/10.1148/radiol.13120730

Abstract

Purpose

To develop diagnostic reference ranges (DRRs) and a method for an individual practice to calculate site-specific reference doses for computed tomographic (CT) scans of the abdomen or abdomen and pelvis in children on the basis of body width (BW).

Materials and Methods

This HIPAA-compliant multicenter retrospective study was approved by institutional review boards of participating institutions; informed consent was waived. In 939 pediatric patients, CT doses were reviewed in 499 (53%) male and 440 (47%) female patients (mean age, 10 years). Doses were from 954 scans obtained from September 1 to December 1, 2009, through Quality Improvement Registry for CT Scans in Children within the National Radiology Data Registry, American College of Radiology. Size-specific dose estimate (SSDE), a dose estimate based on BW, CT dose index, dose-length product, and effective dose were analyzed. BW measurement was obtained with electronic calipers from the axial image at the splenic vein level after completion of the CT scan. An adult-sized patient was defined as a patient with BW of 34 cm. An appropriate dose range for each DRR was developed by reviewing image quality on a subset of CT scans through comparison with a five-point visual reference scale with increments of added simulated quantum mottle and by determining DRR to establish lower and upper bounds for each range.

Results

For 954 scans, DRRs (SSDEs) were 5.8–12.0, 7.3–12.2, 7.6–13.4, 9.8–16.4, and 13.1–19.0 mGy for BWs less than 15, 15–19, 20–24, 25–29, and 30 cm or greater, respectively. The fractions of adult doses, adult SSDEs, used within the consortium for patients with BWs of 10, 14, 18, 22, 26, and 30 cm were 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9, respectively.

Conclusion

The concept of DRRs addresses the balance between the patient’s risk (radiation dose) and benefit (diagnostic image quality). Calculation of reference doses as a function of BW for an individual practice provides a tool to help develop site-specific CT protocols that help manage pediatric patient radiation doses.

© RSNA, 2013

Supplemental material: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.13120730/-/DC1

Introduction

There is limited information describing estimates of radiation doses used to obtain computed tomographic (CT) scans in patients in the United States (1). Data in children are particularly lacking (2,3). The European radiology community has systematically investigated pediatric CT dose through two surveys in phantom-based studies (4,5). Their work was based on a 1984 directive of the European Union concerning protection of patients from ionizing radiation in relation to medical exposure. This resulted in regulations and guidelines (6), including the development of diagnostic reference levels (DRLs) that are based on third quartile values of mean radiation doses within hospitals (7). DRLs are advisory dose levels to be used as guidance by individual hospitals. If DRLs are “exceeded, there should be local review . . . to determine whether the protection has been adequately optimized” (7).

The diagnostic reference range (DRR) provides a minimum estimated patient dose, below which accurate interpretation of the image may be compromised, as well as an upper estimated patient dose, above which the patient dose may be in excess.

The purpose of our study was to develop DRRs and a method for an individual practice to calculate site-specific reference doses for CT scans of the abdomen or abdomen and pelvis in children that are based on body width (BW).

Materials and Methods

This multicenter retrospective study was approved by the institutional review boards of participating institutions, with waiver of the requirement for informed consent. The study was compliant with the Health Insurance Portability and Accountability Act. The Quality Improvement Registry for CT Scans in Children (QuIRCC) is a voluntary pediatric quality improvement consortium of six geographically diverse hospitals. These hospitals include Cincinnati Children’s Hospital Medical Center (Cincinnati, Ohio), Boston Children’s Hospital (Boston, Mass), Children’s Hospital of Philadelphia (Philadelphia, Pa), Duke University Medical Center (Durham, NC), Massachusetts General Hospital (Boston, Mass), and Primary Children’s Hospital (Salt Lake City, Utah). The QuIRCC data are derived from four academic children’s hospitals and two pediatric sections within adult-focused academic institutions; data reside in the National Radiology Data Registry. The consortium completed a retrospective review of consecutive patients of 1 day to 18 years of age or younger (mean age, 10 years) who underwent contrast material–enhanced CT scanning between September 1 and December 1, 2009. We reviewed CT doses from 954 scans in 939 pediatric patients (499 [53%] male and 440 [47%] female patients). All abdominal or abdominal and pelvic CT scans obtained with intravenous contrast material were included. Studies were identified through each institution’s billing database by using standardized CPT codes to find those performed for routine scan indications, such as abdominal pain, abdominal mass, or infection or inflammation. This search included diagnostic scans obtained in conjunction with positron emission tomography (PET)/CT at some sites.

Examinations that were performed without intravenous contrast material, CT angiograms, and attenuation-correction CT scans for PET/CT were excluded. Scans were obtained with 16 CT scanners at six institutions. Table E1 (online) lists specifications of CT scanners used in this study and details the use of tube current modulation. No iterative reconstruction was used during the study. The data elements displayed in Figure 1 were manually recorded for each patient from data on the scanner, in the dose report or Digital Imaging and Communications in Medicine structured report, or in the hospital information system. Patients were scanned by using either fixed tube current or AEC. Only the scans obtained immediately after contrast material administration were evaluated (single phase). No nonenhanced or delayed phase scans were included. Prior to closure of the study, the principal investigator at each site randomly reviewed at least 10% of the cases to verify accuracy of the recorded data, according to the data elements in Figure 1, by cross-checking against the source of the data.

Figure 1:

Figure 1: Chart shows data elements. AEC = automatic exposure control, CTDIvol = volumetric CT dose index, DLP = dose-length product, GE = GE Healthcare (Milwaukee, Wis).

The CTDIvol, DLP, and tube voltage in kilovolts were recorded for each patient as a function of BW. The BW measurement was obtained with electronic calipers from the axial image at the level of the splenic vein after completion of the CT scan. The association between tube voltage in kilovolts and BW group was examined. The CTDIvol, which was based on the 32-cm CT dose index (CTDI) phantom, was used for all calculations. The definition of CTDIvol, with respect to weighted CT dose index (CTDIw), was used to convert between these two metrics with the following calculation thus: CTDIvol = CTDIw/P (8), where P is pitch.

Estimated Patient Dose Calculations

The CTDIvol is used to measure the radiation output of the CT scanner to a defined phantom. While it can be used to track relative changes to patient dose as technique factors change, it does not reflect changes in patient dose because of a change in patient size. Because patient dose increases as patient size decreases for a constant radiation output of the scanner (9), sizable underestimation of pediatric patient dose occurs when CTDIvol is used as a pediatric patient dose index (10). To avoid this pitfall, we used an improved estimate of patient dose, the size-specific dose estimate (SSDE) that is based on body size (11).

Patient data were grouped into one of five BWs (lateral BWs), as follows: BW of less than 15 cm (infant), BW of 15–19 cm (toddler), BW of 20–24 cm (child), BW of 25–29 cm (older child), and BW of 30 cm or greater (adolescent) (Table 1). Conversion of CTDIvol to SSDE is dependent on patient size and on which CTDI phantom was used with the CT scanner to calculate the CTDIvol (10). Because patients of the same age may have a wide range of BWs (12), the patient’s size at a defined landmark (level of the splenic vein) was used to improve estimates of patient dose.

Table 1 Patient Characteristics for CT of the Abdomen or Abdomen and Pelvis in 939 Patients

Table 1

*Numbers in parentheses are percentages. Percentages were rounded.

The effective dose, which is based on both International Commission on Radiological Protection publication 60 (7) and International Commission on Radiological Protection publication 103 (13) weighting factors, and SSDE were calculated. The 75th percentile of the dose estimates within each BW group was assigned as the upper boundary of the DRR. The lower limit of the DRR within each BW group was chosen as the 25th percentile on the basis of the results of the image quality analysis described below.

The calculated SSDE for each patient was also evaluated independently of the BW groups. A linear regression fit of SSDE as a function of BW was calculated. If one assumes the BW for an “adult-sized” patient is 34 cm (12), one can calculate an “adult” dose from the QuIRCC data, which is the adult SSDE (SSDEadult),, by using the fitted equation. A linear regression fit of SSDE values as a function of BW normalized by the constant SSDEadult was calculated. This fit allows the calculation of the fraction of the adult dose used for a patient of any size.

Image Quality Analysis

A total of 120 scans (60 scans that represented the median [two for each BW range and 10 from each site], 30 representative scans from the first quartile from all sites, and 30 scans with the lowest SSDE from all sites) were considered for this retrospective review. Six scans overlapped, as they were both at the median SSDE for each site and in the first quartile of scans from all sites. This redundancy resulted in 114 potential scans for review. A coauthor (A.J.T., with 4 years of postfellowship experience), who was not involved with image analysis, used a random-number generator and removed identification information from the 114 remaining scans. Each “scan” consisted of three serial images from each CT scan (not the entire CT scan) and included the splenic vein, the gallbladder-liver interface, and the liver-kidney or kidney-spleen interface. The three clinical images could be electronically manipulated at the workstation by each reviewer, as would simulate clinical practice. Prior to image selection for subjective quality analysis, the images had to meet study criteria. A total of eight scans were excluded from analysis by one coauthor (A.J.T.). He removed three scans that were out of the date range and five scans that were for the nonenhanced phase of a dual-phase examination, for a total of 106 scans for subjective review. Sample scans that were not included in the study were provided in a group setting for instructional purposes to develop consensus for ranking and terminology among the reviewers.

The subjective image quality analysis of 106 scans was independently and blindly performed with identical viewing conditions by six pediatric radiologists, (M.J.C., K.D., D.J.P., D.P.F., S.J.W., J.S.P., with 8–22 years of pediatric radiology experience). Survey questions with a five-point scale are shown in Figure 2 and were used by the reviewers for the subjective image analysis. In addition, the reviewers used a five-point assessment scale to evaluate image quality by using expanded RadLex® terminology (14). Five reference images of the same anatomy with different levels of added quantum mottle simulating different radiation doses were compared with each clinical image to qualitatively evaluate image quality. The five reference images were created by using a validated computer-simulation technique that simulated tube-current reductions by adding Gaussian noise to the image (15). The first image simulated 70% of the baseline dose. Each of the subsequent four reference images simulated one-half of the radiation dose of the previous image (Fig 3). The five reference images could not be manipulated. Each of the six reviewers looked at the three clinical images for each scan and rated the clinical images compared with the five reference-level images with a single score. Finally, each reviewer rated each scan as diagnostic or nondiagnostic. Scans were assigned by the statistician (L.P.C.) to a nondiagnostic ranking if three or more of the six reviewers ranked them as such. The subjective image quality analysis comprised all studies, including those later ranked as nondiagnostic. After subjective image analysis was complete, the reason for images rated as nondiagnostic was reviewed by the principal investigator.

Figure 2:

Figure 2: Image Quality Evaluation Form. This online secure data entry form was used by the six reviewers to perform subjective image quality evaluation. Reviewers used the five-point reference scale noted in Figure 3 to complete question 2, section 3. The results from question 2 were used to determine whether a scan was rated as diagnostic or nondiagnostic. ACR = American College of Radiology, NRDR = National Radiology Data Registry, PI = principal investigator.

Figure 3a:

Figure 3a: Axial images show how the five-point image reference scale was used for the subjective image analysis. Noise was added to the (a) first image that was equivalent to 70% of the baseline dose (baseline image at full dose not shown). Image was rated as grade 5 of five (exemplary quality). (b) Image was rated as grade 4 of five (very good quality) and was 35% of baseline dose. (c) Image was rated as grade 3 of five (good quality) and was 17% of baseline dose. (d) Image was rated as grade 2 of five (diagnostic, limited) and was 9% of baseline dose. (e) Image was rated as grade 1 of five (nondiagnostic) and was 4% of baseline dose. Reference scale was based on modifications to a clinical image of a 5-year-old CT scan of the abdomen.

Figure 3b:

Figure 3b: Axial images show how the five-point image reference scale was used for the subjective image analysis. Noise was added to the (a) first image that was equivalent to 70% of the baseline dose (baseline image at full dose not shown). Image was rated as grade 5 of five (exemplary quality). (b) Image was rated as grade 4 of five (very good quality) and was 35% of baseline dose. (c) Image was rated as grade 3 of five (good quality) and was 17% of baseline dose. (d) Image was rated as grade 2 of five (diagnostic, limited) and was 9% of baseline dose. (e) Image was rated as grade 1 of five (nondiagnostic) and was 4% of baseline dose. Reference scale was based on modifications to a clinical image of a 5-year-old CT scan of the abdomen.

Figure 3c:

Figure 3c: Axial images show how the five-point image reference scale was used for the subjective image analysis. Noise was added to the (a) first image that was equivalent to 70% of the baseline dose (baseline image at full dose not shown). Image was rated as grade 5 of five (exemplary quality). (b) Image was rated as grade 4 of five (very good quality) and was 35% of baseline dose. (c) Image was rated as grade 3 of five (good quality) and was 17% of baseline dose. (d) Image was rated as grade 2 of five (diagnostic, limited) and was 9% of baseline dose. (e) Image was rated as grade 1 of five (nondiagnostic) and was 4% of baseline dose. Reference scale was based on modifications to a clinical image of a 5-year-old CT scan of the abdomen.

Figure 3d:

Figure 3d: Axial images show how the five-point image reference scale was used for the subjective image analysis. Noise was added to the (a) first image that was equivalent to 70% of the baseline dose (baseline image at full dose not shown). Image was rated as grade 5 of five (exemplary quality). (b) Image was rated as grade 4 of five (very good quality) and was 35% of baseline dose. (c) Image was rated as grade 3 of five (good quality) and was 17% of baseline dose. (d) Image was rated as grade 2 of five (diagnostic, limited) and was 9% of baseline dose. (e) Image was rated as grade 1 of five (nondiagnostic) and was 4% of baseline dose. Reference scale was based on modifications to a clinical image of a 5-year-old CT scan of the abdomen.

Figure 3e:

Figure 3e: Axial images show how the five-point image reference scale was used for the subjective image analysis. Noise was added to the (a) first image that was equivalent to 70% of the baseline dose (baseline image at full dose not shown). Image was rated as grade 5 of five (exemplary quality). (b) Image was rated as grade 4 of five (very good quality) and was 35% of baseline dose. (c) Image was rated as grade 3 of five (good quality) and was 17% of baseline dose. (d) Image was rated as grade 2 of five (diagnostic, limited) and was 9% of baseline dose. (e) Image was rated as grade 1 of five (nondiagnostic) and was 4% of baseline dose. Reference scale was based on modifications to a clinical image of a 5-year-old CT scan of the abdomen.

Statistical Analysis

For the objective component of the study, descriptive statistics included the following: means; standard errors; first, second, and third quartiles; and ranges. These descriptive statistics were calculated for each of the technical variables (CTDIvol, SSDE, DLP, and effective dose) according to patient size. The association between tube voltage in kilovolts and BW group was tested by using Spearman ρ. Categorical variables were compared by using a χ2 test. All tests were two sided. P values of .05 or less were considered to indicate a significant difference. A mixed model was used to determine whether there were significant differences in mean SSDE and CTDIvol across institutions after adjusting for patient size. With use of software (Proc Mixed, SAS, version 9.2; SAS, Cary, NC), we fit a one-way analysis of covariance model with random effects.

To analyze the subjective component for each of the 106 scans, we calculated the number of reviewers that rated the scan as nondiagnostic. We then compared the CTDIvol and SSDE for the diagnostic and nondiagnostic scans.

Results

Table 1 lists characteristics of the patients included in the study. Nine hundred fifty-four abdominal or abdominal and pelvic CT scans were obtained in 939 patients. Thirty-eight of 954 (4%) scans were of the abdomen only. Of 954 scans, 939 (98%) scans were single phase, 14 (1%) were dual phase, and one (0%) was triple phase. Twenty-eight percent (263 of 939) of studies were performed with a fixed tube current, and 72% (676 of 939) of studies were performed with tube current modulation. The mean age was 10 years (range, 1 day to18 years). The patients’ weight ranged from 2 to 167 kg. The BW ranged from 10 to 45 cm. One site provided 39% of the patients, while the two pediatric sections within primarily adult-focused academic medical centers each provided approximately 6% of the patients. The most common indications were abdominal pain, right lower quadrant pain, and appendicitis.

The variation in the use of tube voltage in kilovolts in the patients studied is shown in Table 2. As expected, there was a significant decrease in tube voltage in kilovolts in smaller-sized patient groups (P < .0001).

Table 2 Patient Size Compared with Tube Voltage

Table 2

Note.—Data are numbers of patients. Numbers in parentheses are percentages, and percentages were rounded.

The distribution of SSDE in milligrays within each BW group is summarized in Table 3. By definition, the analysis included the six scans ranked as nondiagnostic. The 25th and 75th percentiles represent the lower and upper limits for the DRR. The DRRs are as follows: 5.8–12.0 mGy for an infant (BW of <15 cm), 7.3–12.2 mGy for a toddler (BW of 15–19 cm), 7.6–13.4 mGy for a child (BW of 20–24 cm), 9.8–16.4 mGy for an older child (BW of 25–29 cm), and 13.1–19.0 mGy for an adolescent (BW of ≥30 cm).

Table 3 Distribution of SSDE

Table 3

The distribution of CTDIvol, DLP, and effective dose are listed in Tables 46, respectively. These distributions establish DRRs on the basis of CTDIvol, DLP, and effective dose.

Table 4 Distribution of CTDIvol Measured with 32-cm CTDI Phantom

Table 4

Table 5 Distribution of DLP Measured with 32-cm Phantom

Table 5

Note.—DLP was measured in milligray-centimeters. Ten scans could not be included because equipment could not provide DLP.

Table 6 Distribution of Effective Dose

Table 6

Note.—ICRP pub = International Commission on Radiological Protection publication. Ten scans could not be included because equipment could not provide DLP.

Figure 4a is a scatterplot of SSDE (in milligrays) as a function of BW for the QuIRCC data. The coefficients of the linear fit to the data are provided in Figure 4a; the P value for the slope of the regression was less than .0001. The SSDE from this fit for an adult-sized patient, SSDEadult with BW of 34 cm, is 16.6 mGy. Figure 4b normalizes the fitted curve of Figure 4a with the SSDEadult; the P value for the slope of the regression was less than .0001. This normalized fit indicates that the fractions of the SSDEadult values used within QuIRCC for patients with BWs of 10, 14, 18, 22, 26, and 30 cm were 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9, respectively. For example, if a site wishes to use the consortium’s data as a guide for a conversion factor for their site-specific protocols for a patient with a BW of 18 cm (along the horizontal axis), and the site uses 20 mGy for an adult abdomen, the site would select 0.6 · 20 = 12 mGy as a target SSDE for a patient with a BW of 18 cm.

Figure 4a:

Figure 4a: (a) Scatterplot of SSDE (in milligrays) as a function of BW shows the coefficients of the linear fit to the data. (b) Plot shows normalization of the fitted curve of image a to the SSDE of 34-cm BW, an average size adult (12). The P value for the slope of the regression of both fits was less than .0001. The normalized curve allows a facility to establish pediatric reference radiation doses if the SSDE of their standard adult patient for abdominal or abdominal and pelvic CT studies with contrast is known. For example, if a site wishes to use the consortium’s data as a guide for a conversion factor for their site-specific protocols for a patient of a BW of 18 cm (along the horizontal axis), and the site uses 20 mGy for an adult abdomen, the site would select 0.6 ⋅ 20 = 12 mGy that could be used by the site as a target SSDE for a patient with a BW of 18 cm. * = multiplied by.

Figure 4b:

Figure 4b: (a) Scatterplot of SSDE (in milligrays) as a function of BW shows the coefficients of the linear fit to the data. (b) Plot shows normalization of the fitted curve of image a to the SSDE of 34-cm BW, an average size adult (12). The P value for the slope of the regression of both fits was less than .0001. The normalized curve allows a facility to establish pediatric reference radiation doses if the SSDE of their standard adult patient for abdominal or abdominal and pelvic CT studies with contrast is known. For example, if a site wishes to use the consortium’s data as a guide for a conversion factor for their site-specific protocols for a patient of a BW of 18 cm (along the horizontal axis), and the site uses 20 mGy for an adult abdomen, the site would select 0.6 ⋅ 20 = 12 mGy that could be used by the site as a target SSDE for a patient with a BW of 18 cm. * = multiplied by.

Hospital Variation

Figure 5 demonstrates variation in the mean SSDE among hospitals. The unadjusted mean SSDE for individual hospitals ranged from 8.6 to 16.5 mGy. From the mixed model, after adjusting for patient BW, using shrinkage estimates of the random coefficient, the estimates of the adjusted means for SSDE ranged from 9.6 to 16.8 mGy. The adjusted mean for facility A was significantly lower than the overall mean (P = .007). Facility E had a significantly higher adjusted mean compared with the overall mean for SSDE (P = .0011). For all other facilities, the means were not significantly different from the overall adjusted mean.

Figure 5:

Figure 5: Box and whisker plot with summary statistic data show variation in the mean SSDE among hospitals. The mean SSDE for facility A was significantly lower than the overall mean (P = .0068) of the other sites. The adjusted mean SSDE for facility E was significantly higher (P = .0011). The mean SSDEs for the other facilities were not significantly different from the overall adjusted mean. Two boxes were clipped, as they were outside of the displayed range. Max = maximum, Med = median, Min = minimum, N = number of CT scans obtained at each facility.

Subjective Image Analysis

Six of 106 (6%) scans were ranked nondiagnostic. Table E2 (online) provides the results of the subjective ranking. Figure 6 illustrates a diagnostic and nondiagnostic scan, as judged by the reviewers. Four of six (67%) nondiagnostic scans were below the 10th percentile on the basis of the SSDE. Five of six (83%) nondiagnostic scans had an SSDE less than the 25th percentile. After data analysis was closed, it was found that, in these scans, imaging was performed by using a specific PET CT with diagnostic CT research protocol in addition to nonenhanced attenuation-correction CT with PET. However, the complete scans had been fully diagnostic clinically, as judged by local radiologists. The unacceptable scan with SSDE above the 25th percentile was caused by a subcutaneous metal implant with artifact.

Figure 6a:

Figure 6a: (a) Representative axial image from a CT scan of the abdomen at the level of the gallbladder used during image quality analysis subjectively ranked as diagnostic by the investigators. The patient BW was 19 cm (infant), and the SSDE was 4.7 mGy (120 kV). (b) Representative axial image from a CT scan of the abdomen and pelvis during the same image quality analysis was ranked as nondiagnostic by site investigators. The patient’s BW was 19 cm (infant), and the SSDE was 3.1 mGy (90 kV).

Figure 6b:

Figure 6b: (a) Representative axial image from a CT scan of the abdomen at the level of the gallbladder used during image quality analysis subjectively ranked as diagnostic by the investigators. The patient BW was 19 cm (infant), and the SSDE was 4.7 mGy (120 kV). (b) Representative axial image from a CT scan of the abdomen and pelvis during the same image quality analysis was ranked as nondiagnostic by site investigators. The patient’s BW was 19 cm (infant), and the SSDE was 3.1 mGy (90 kV).

Discussion

This consortium developed DRRs that were based on the body size of pediatric patients undergoing abdominal or abdominal and pelvic CT for routine clinical indications. The upper limit of the DRR, standard DRL values (75th percentile), provides a target dose that a facility should strive not to exceed, thus satisfying the “need to improve optimization of this high-dose imaging modality for this especially vulnerable section of the population” (16). The lower limit of the DRR, which is based on subjective image quality analysis, provides a minimum target dose that should be exceeded and satisfies the “need for robust studies that determine minimum exposure” (17). The DRR addresses the balance between the patient’s radiation dose (risk) and image quality necessary to provide an accurate diagnosis (benefit) (18).

An American College of Radiology Blue Ribbon Panel on Radiation Dose in Medicine recommended “the development of a national database for radiation dose indices to address the actual range of exposures” for imaging examinations (19). The DRR addresses that recommendation. In this study, while we also developed DRRs that were based on CTDIvol, DLP, and effective dose, the SSDE DRR provides the best estimate of the patient’s radiation dose (10). CTDIvol and DLP are doses delivered to standard phantoms, not individual patients (9). Because of uncertainties and oversimplifications, the effective dose should not be used for epidemiologic studies or for estimating population risks (20).

The only U.S. data (75th percentile) for pediatric CT uses CTDIvol data from the CT Accreditation Program of the American College of Radiology for a “5 year old abdomen,” 20 mGy based on a 16-cm CTDI phantom (21). The SSDE was 21 mGy for a “patient” with the equivalent BW, 18.5 cm, corresponding to the ACR pediatric CT accreditation data (11). The 75th percentile SSDE of the 15–19-cm patient group in our study was 12.2 mGy (12.2 of 21), which was 58% of the pediatric CT accreditation data of the American College of Radiology by using the 16-cm CTDI phantom.

Because U.S. data are limited, our results are compared with results from other countries to identify whether CT pediatric radiation dose estimates are increasing or decreasing. The earliest DRL for pediatric abdominal CT in the European literature is based on a survey of over 40 CT scanners in seven European countries by using a plastic, cylindrical phantom with an effective diameter of 17.1 cm to model a 5-year-old abdomen (4). The researchers in this 1999 study (4) reported a CTDIw32 of 12.5 mGy. In comparison, the 75th percentile for CTDIw32 for our 15–19-cm patient group was lower, with 61% (7.6 of 12.5) of the European reference level.

In another more recent international study published in 2008 by researchers in a single large pediatric hospital in Canada, they calculated DLP and effective dose for 5-, 10-, and 15-year-old patients. The median effective doses in our study for 5-, 10-, and 15-year-old patients were 5.8, 6.3, and 7.1 mSv, which correspond to 69% (5.8 of 8.4), 71% (6.3 of 8.9), and 120% (7.1 of 5.9) of the doses in the Canadian study, respectively (22).

There are several reasons why the dose estimates reported in this study were approximately 61% of the values from the pediatric CT accreditation data of the American College of Radiology and the data in previously published European studies. The use of AEC for CT scans may reduce patient dose by as much as 40% (23). Because 72% of our CT scans were completed with AEC and this feature was not available during the 1999 European study (24), our dose levels should be lower. Educational programs on radiation protection of the Society for Pediatric Radiology (25) and the Image Gently campaign of the Alliance for Radiation Safety in Pediatric Imaging (26) have urged health care providers, including radiologists and CT technologists, to carefully manage radiation doses during the CT examination in children. Because the previously discussed 2008 Canadian study was conducted approximately 2 years prior to our study, dose reductions are more comparable because both studies benefited from the same advances in the field.

An additional purpose of our study was to develop a method for an individual practice to calculate site-specific reference doses for abdominal or abdominal and pelvic CT scans as a function of BW influenced by consortium data. The calculated fractions of patient dose (SSDE) used within the consortium as a function of BW provide guidance to establish reference doses for the individual scanners of an individual facility. The product of the QuIRCC fraction for a given BW and the adult dose used by the facility (SSDE) gives a reference dose for a given BW. The adult dose used by the facility expressed as SSDE can be calculated from the annual CTDIvol measurements provided by the site’s medical physicist. Use of the facility’s adult dose in the calculation, as opposed to the adult dose of the consortium, adjusts the facility reference dose to the specific radiation output of the individual CT scanner and the image quality requirements of the radiologists at the facilitiy. While our study findings suggest that experienced pediatric radiologists can successfully image the smallest pediatric patients with approximately one-half of the patient radiation dose used for an adult, individual facilities may elect to use less aggressive dose reductions for the smallest patients. A facility’s radiologists, technologists, and medical physicist should work together to establish reference doses and generate technique charts with reduced tube current (24) and/or high voltage values that result in a patient dose (SSDE) less than the reference dose for a given BW.

There were several limitations to this study. Because 39% of the data came from one of the six facilities in the consortium, this facility is overrepresented in the data set while the other five facilities are underrepresented in the study.

While in this study we created a DRR for standard abdominal CT scans for routine clinical indications that required intravenous contrast material, other types of specific clinical indications (eg, renal stone disease) will probably require reduced DRRs owing to increased conspicuity of stones (27), while tasks of increased complexity may require an increased radiation dose (28). Other body regions, such as the thorax and head, with specific attenuation characteristics will require examination- and body part–specific DRRs.

Of the patients in our study, 72% were scanned using the AEC feature of the CT scanner. The SSDE estimate of patient dose was derived from fixed–tube current scans. However, the SSDE should be correctly calculated for CT scans using AEC, because the CTDIvol value reported by the scanner is assumed to be based on the average tube current along the z-axis of the patient, which should result in an average CTDIvol.

This study was completed with CT scanners that did not have the capability to perform iterative reconstruction, a technique that has been shown to reduce patient doses substantially (29). Furthermore, the small radiation doses associated with the initial projection scan of each study (30) were not included in our reported patient doses. Finally, this study was limited to six institutions, four of which specialize in the imaging of children and have radiologists that primarily read pediatric images. Our reported values may be lower than values in practices that are primarily adult focused and perform imaging in pediatric patients less of the time. CT DRRs provide an upper estimated patient dose (DRL), above which the dose may be in excess, and a minimum estimated patient dose, below which accurate interpretation of the associated image may be compromised. The concept of DRR begins to address the balance between the patient’s risk (radiation dose) and benefit (image quality necessary for an accurate diagnosis). The conservative use of CT scans in children and the DRRs foster improved dose management for pediatric CT. Until estimates of organ dose for individual pediatric patients become available (31), the SSDE should be adopted as an interim estimate of patient dose.

Advances in Knowledge
  • • The diagnostic reference range (DRR) is a newer quality improvement tool that provides a minimum estimated patient radiation dose, below which reduced image quality may not be diagnostic, and an upper estimated patient dose, above which the dose may be in excess.

  • • The CT patient dose estimates for a 5-year-old patient at the consortium hospitals in this study are 61% (7.6 of 12.5 mGy, by using volumetric CT dose index as the dose index) of the dose estimates to a 5-year-old phantom derived from a survey of 40 CT scanners in Europe a decade earlier.

  • • With respect to abdominal or abdominal and pelvic CT images obtained with contrast medium, experienced pediatric radiologists were confident in interpreting images of small patients with 14-cm body width (BW) obtained with one-half the estimated dose (5.8 of 12 mGy, based on size-specific dose estimate [SSDE]) used for an adult-sized 34-cm–BW patient; subjectively evaluated cases of 14-cm–BW patients with a dose of 6 mGy or more were judged to be of diagnostic quality.

Implication for Patient Care
  • • The DRRs that were based on SSDEs and the method to calculate reference doses for a particular facility provided in this article allow substantial dose reduction to be achieved while maintaining acceptable image quality.

Disclosures of Conflicts of Interest: M.J.G. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: received payment for three to six visiting professorships a year at academic institutions. Other relationships: none to disclose. K.J.S. No relevant conflicts of interest to disclose. L.P.C. No relevant conflicts of interest to disclose. K.E.M. No relevant conflicts of interest to disclose. A.J.T. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: received royalties for a book from Amirsys; holds stock in and was paid for travel, accommodations, and meeting expenses unrelated to activities listed from Merge Healthcare. Other relationships: none to disclose. D.B.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: author and institution were paid for patents through intellectual property license agreement by Radimetrics. Other relationships: none to disclose. M.J.C. No relevant conflicts of interest to disclose. K.D. No relevant conflicts of interest to disclose. D.J.P. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: received travel reimbursement for user group meeting by GE Healthcare and Philips, received payment for service on professional speakers bureau from Toshiba, received royalties for textbook chapter from Amirsys. Other relationships: none to disclose. D.P.F. No relevant conflicts of interest to disclose. S.J.W. No relevant conflicts of interest to disclose. J.S.P. No relevant conflicts of interest to disclose.

Author Contributions

Author contributions: Guarantor of integrity of entire study, M.J.G.; 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, M.J.G., K.J.S., K.D., D.J.P.; clinical studies, M.J.G., A.J.T., K.D., D.J.P., S.J.W.; experimental studies, K.J.S., A.J.T., D.P.F., J.S.P.; statistical analysis, L.P.C., D.B.L.; and manuscript editing, all authors

From the 2011 RSNA Annual Meeting.

Supported by the Derek Harwood-Nash/Harvey and Jean Picker Education Scholar Grant from the Radiological Society of North America Research and Education Foundation and by the Society for Pediatric Radiology Research and Education Foundation.

References

  • 1 Gray JE, Archer BR, Butler PF, et al.. Reference values for diagnostic radiology: application and impact. Radiology 2005;235(2):354–358.
  • 2 Linet MS, Kim KP, Rajaraman P. Children’s exposure to diagnostic medical radiation and cancer risk: epidemiologic and dosimetric considerations. Pediatr Radiol 2009;39(suppl 1):S4–S26.
  • 3 Hricak H, Brenner DJ, Adelstein SJ, et al.. Managing radiation use in medical imaging: a multifaceted challenge. Radiology 2011;258(3):889–905.
  • 4 Shrimpton PC, Wall BF. Reference doses for paediatric computed tomography. Radiat Prot Dosimetry 2000;90(1–2):249–252.
  • 5 Shrimpton PC, Hillier MC, Lewis MA, Dunn M. National survey of doses from CT in the UK: 2003. Br J Radiol 2006;79(948):968–980. [Published correction appears in Br J Radiol 2007;80(956):685. Dosage error in article text.]
  • 6 Goske MJ. Getting it right: are regulation and registries for CT radiation dose in children the answer? Pediatr Radiol 2011; 41(suppl 2):567–570.
  • 7 1990 Recommendations of the International Commission on Radiological Protection. Ann ICRP 1991;21(1–3):1–201.
  • 8 International Electrotechnical Commission (IEC). Medical Electrical Equipment. Part 2–44: Particular requirements for the safety of x-ray equipment for computed tomography. IEC publication no. 60601-2-44. Ed 2.1. Geneva, Switzerland: International Electrotechnical Commission Central Office, 2002.
  • 9 Strauss KJ, Goske MJ. Estimated pediatric radiation dose during CT. Pediatr Radiol 2011;41(suppl 2):472–482.
  • 10 Strauss KJ, Butler PF, Goske MJ, Ritenour ER. Image gently: reducing radiation dose in pediatric computed tomography through collaboration. Med Phys 2009;36(12):5719–5720.
  • 11 Boone JM, Strauss KJ, Cody DD, et al.. Size specific dose estimates (SSDE) in pediatric and adult body CT examinations. American Association of Physicists in Medicine Web site. http://www.aapm.org/pubs/reports/RPT_204.pdf. Published February 2011. Accessed August 30, 2012.
  • 12 Kleinman PL, Strauss KJ, Zurakowski D, Buckley KS, Taylor GA. Patient size measured on CT images as a function of age at a tertiary care children’s hospital. AJR Am J Roentgenol 2010;194(6):1611–1619.
  • 13 Huda W, Magill D, He W. CT effective dose per dose length product using ICRP 103 weighting factors. Med Phys 2011; 38(3):1261–1265.
  • 14 RadLex: a lexicon for uniform indexing and retrieval of radiology information resources. Radiological Society of North America Web site. http://www.rsna.org/RadLex.aspx. Published 2010. Accessed August 30, 2012.
  • 15 Frush DP, Slack CC, Hollingsworth CL, et al.. Computer-simulated radiation dose reduction for abdominal multidetector CT of pediatric patients. AJR Am J Roentgenol 2002;179(5):1107–1113.
  • 16 Pages J, Buls N, Osteaux M. CT doses in children: a multicentre study. Br J Radiol 2003;76(911):803–811.
  • 17 Golding SJ. Radiation exposure in CT: what is the professionally responsible approach? Radiology 2010;255(3):683–686.
  • 18 Shet N, Chen J, Siegel EL. Continuing challenges in defining image quality. Pediatr Radiol 2011;41(5):582–587.
  • 19 Amis ES, Butler PF, Applegate KE, et al.. American College of Radiology white paper on radiation dose in medicine. J Am Coll Radiol 2007;4(5):272–284.
  • 20 McCollough CH, Christner JA, Kofler JM. How effective is effective dose as a predictor of radiation risk? AJR Am J Roentgenol 2010;194(4):890–896.
  • 21 ACR Practice Guidelines for Diagnostic Reference Levels in Medical X-Ray Imaging. American College of Radiology Web site. http://www.acr.org/SecondaryMainMenuCategories/quality_safety/guidelines/med_phys/reference_levels.aspx. Published 2006. Accessed August 30, 2012.
  • 22 Thomas KE, Wang B. Age-specific effective doses for pediatric MSCT examinations at a large children’s hospital using DLP conversion coefficients: a simple estimation method. Pediatr Radiol 2008;38(6):645–656.
  • 23 McCollough CH, Bruesewitz MR, Kofler JM. CT dose reduction and dose management tools: overview of available options. RadioGraphics 2006;26(2):503–512.
  • 24 Regulatory Agency Evaluation Report 05016. United Kingdom Department of Health Medicines and Healthcare products. http://www.impactscan.org/reports/Report05016.htm. Published February 2005. Accessed August 2, 2012.
  • 25 The ALARA (as low as reasonably achievable) concept in pediatric CT intelligent dose reduction. Multidisciplinary conference organized by the Society of Pediatric Radiology. August 18-19, 2001. Pediatr Radiol 2002;32(4):217–313.
  • 26 Goske MJ, Applegate KE, Boylan J, et al.. The Image Gently campaign: working together to change practice. AJR Am J Roentgenol 2008;190(2):273–274.
  • 27 Karmazyn B, Frush DP, Applegate KE, Maxfield C, Cohen MD, Jones RP. CT with a computer-simulated dose reduction technique for detection of pediatric nephroureterolithiasis: comparison of standard and reduced radiation doses. AJR Am J Roentgenol 2009;192(1):143–149.
  • 28 ICRP publication 105. Radiation protection in medicine. Ann ICRP 2007;37(6):1–63.
  • 29 Singh S, Kalra MK, Shenoy-Bhangle AS, et al.. Radiation dose reduction with hybrid iterative reconstruction for pediatric CT. Radiology 2012;263(2):537–546.
  • 30 O’Daniel JC, Stevens DM, Cody DD. Reducing radiation exposure from survey CT scans. AJR Am J Roentgenol 2005;185(2):509–515.
  • 31 Li X, Samei E, Segars WP, Sturgeon GM, Colsher JG, Frush DP. Patient-specific radiation dose and cancer risk for pediatric chest CT. Radiology 2011;259(3):862–874.

Article History

Received April 9, 2012; revision requested June 6; final revision received September 17; accepted November 1; final version accepted January 10, 2013.
Published online: July 2013
Published in print: July 2013