Radiation Dose for Pediatric CT: Comparison of Pediatric versus Adult Imaging Facilities
The American College of Radiology Dose Index Registry for CT enables evaluation of radiation dose as a function of patient characteristics and examination type. The hypothesis of this study was that academic pediatric CT facilities have optimized CT protocols that may result in a lower and less variable radiation dose in children.
Materials and Methods
A retrospective study of doses (mean patient age, 12 years; age range, 0–21 years) was performed by using data from the National Radiology Data Registry (year range, 2016–2017) (n = 239 622). Three examination types were evaluated: brain without contrast enhancement, chest without contrast enhancement, and abdomen-pelvis with intravenous contrast enhancement. Three dose indexes—volume CT dose index (CTDIvol), size-specific dose estimate (SSDE), and dose-length product (DLP)—were analyzed by using six different size groups. The unequal variance t test and the F test were used to compare mean dose and variances, respectively, at academic pediatric facilities with those at other facility types for each size category. The Bonferroni-Holm correction factor was applied to account for the multiple comparisons.
Pediatric radiation dose in academic pediatric facilities was significantly lower, with smaller variance for all brain, 42 of 54 (78%) chest, and 48 of 54 (89%) abdomen-pelvis examinations across all six size groups, three dose descriptors, and when compared with that at the other three facilities. For example, abdomen-pelvis SSDE for the 14.5–18-cm size group was 3.6, 5.4, 5.5, and 8.3 mGy, respectively, for academic pediatric, nonacademic pediatric, academic adult, and nonacademic adult facilities (SSDE mean and variance P < .001). Mean SSDE for the smallest patients in nonacademic adult facilities was 51% (6.1 vs 11.9 mGy) of the facility’s adult dose.
Academic pediatric facilities use lower CT radiation dose with less variation than do nonacademic pediatric or adult facilities for all brain examinations and for the majority of chest and abdomen-pelvis examinations.
© RSNA, 2019
See also the editorial by Strouse in this issue.
Academic pediatric facilities use lower radiation dose than do nonacademic pediatric or adult facilities for all brain CT examinations and for the majority of chest and abdomen-pelvis CT examinations.
■ For abdomen-pelvis CT (patient size range, 14.5–18 cm), size-specific dose estimates (SSDEs) were 3.6, 5.4, 5.5, and 8.3 mGy for academic pediatric, nonacademic pediatric, academic adult, and nonacademic adult facilities, respectively (P < .001 for comparison).
■ Academic pediatric facilities had protocols that resulted in lower radiation dose at chest CT than did 78% of other facilities (nonacademic pediatric, academic adult, and nonacademic adult facilities).
■ Academic pediatric facilities had protocols that resulted in lower radiation dose for abdomen-pelvis CT than did 89% of other facilities (nonacademic pediatric, academic adult, and nonacademic adult facilities).
Prior to the early 2000s, a distinction between pediatric and adult patients was rarely made, leading to poor management of radiation dose and image quality during CT scanning (1–3). During this time, acquisition techniques were rarely adjusted based on body size (3). Initial recommendations for radiation dose reduction appeared (2,4,5) and were followed by amplification by the Image Gently Alliance in the United States (6,7) and throughout the world (8). In 2010, the Image Wisely campaign addressed similar concerns as they pertain to adult imaging (9).
Advances in radiation awareness have occurred through education, hardware and software development, and available radiation dose indexes. These advances have stemmed from a collective desire among facilities, national and international associations, and CT manufacturers to improve health care for all patients, especially pediatric patients undergoing CT. In 2011, the size-specific dose estimate (SSDE) (10) was developed to enable patient dose estimation with improved accuracy and precision in patients of every size (11). The American College of Radiology initiated its dose index registry (DIR) (12,13), which contained dose indexes for more than 45 000 000 CT examinations as of July 2017 (14). These data enable the development of size-based diagnostic reference levels that imaging sites can use to develop targeted pediatric CT doses.
Reports of improved pediatric and adult CT technique changes based on patient size appeared beginning in 2008 with the Image Gently Alliance (15–16) and were followed by publications from the American Association of Physicists in Medicine in 2016 and 2017 (17). Despite these available recommendations, opportunities for improvement in the management of pediatric CT doses remain. Published reports of study periods prior to 2014 comparing pediatric CT dose indexes at adult community facilities support the need for this improvement (18–23). However, these studies focused on one type of CT study in specific populations. A national survey was needed. The DIR, a dose registry that collects deidentified CT examination data from participating facilities throughout the United States, is a good tool to use to assess the national differences between types of practices, patient characteristics, and examination types. We hypothesized that mean dose index levels and variances, respectively, were lower and less varied in academic pediatric facilities than in other facility types for each dose category. The purpose of this study was to compare DIR pediatric CT dose indexes as a function of patient size and to analyze differences between academic pediatric, nonacademic pediatric, academic adult, and nonacademic adult facilities during pediatric CT imaging.
Materials and Methods
Dose Index Registry
The DIR is compliant with the Health Insurance Portability and Accountability Act. The DIR has institutional review board assurance of status of exemption from patient written informed consent. No industry support was provided. The authors had total control of the data and information submitted for publication. This study included data from January 2016 through December 2017 for three types of single-scan pediatric CT examinations: brain CT without intravenous contrast material, chest CT without intravenous contrast material, and abdomen-pelvis CT with intravenous contrast material in patients 21 years of age or younger. Examinations without effective diameter calculations (10,24) necessary to estimate SSDE, examinations performed at facilities that contributed data from fewer than 30 total examinations, and examinations in which multiple scans were performed were excluded. The dose indexes used in this study consisted of volume CT dose index (CTDIvol), dose-length product (DLP), and SSDE. The DIR provided all three indexes for the chest and abdomen-pelvis examinations. An SSDE method for brain CT examinations does not currently exist (10).
The 519 facilities participating in this study were grouped into four categories: (a) academic pediatric, (b) nonacademic pediatric, (c) academic adult, and (d) nonacademic adult institutions. A teaching facility linked to a medical school was considered an academic institution; the remaining facilities were nonacademic. Pediatric facilities were defined according to The Children’s Hospital Association list of approximately 250 pediatric facilities in the United States, and 107 (43%) of these 251 facilities were academic. Dose indexes at academic pediatric facilities were used as benchmarks; dose indexes in the other three facility groups were compared with the benchmark by using appropriate statistical tests.
Patient Size Grouping
Dose records for the three CT examinations were analyzed based on patient effective diameter (10) of the anatomy in the direct radiation beam calculated from localizer scan images of the examination (24). The six size groups for chest and abdomen-pelvis examinations and the five size groups for the brain examinations are listed in Table 1. While target doses are best compared with the thickness (path length of primary x-rays) through the patient’s irradiated anatomy (effective diameter), some facilities use a patient’s age or weight out of convenience. However, the largest 3-year-old patient may have the same abdominal thickness as the smallest 18-year-old patient (25). Likewise, endo- and ectomorphic patients of the same weight have different thicknesses. Despite these facts, average age (25) and weight (26) associated with each size group are provided in Table 1 as a tool. The lateral thicknesses associated with the effective diameters are also provided for technologists who measure patient thickness mechanically with calipers or electronically on the localizer image to verify the acquisition technique selected on the scanner prior to the actual examination.
Size-specific Mean Dose Analysis
The mean dose level in academic pediatric facilities () was compared with that in the other three facility groups—the nonacademic pediatric (), academic adult (), and nonacademic adult () facility groups—by performing an unequal variance two-sample t test (Welch test) (27) individually for all three dose indexes (CTDIvol, DLP, and SSDE) for each examination type. The unequal variance t test was chosen since the radiation dose data did not satisfy the assumption of the standard Student t test, which required equal variance between the facility groups. The assumptions of the unequal variance t test were (a) normality and (b) independence. The radiation dose records were considered independent of each other by assuming that no patient underwent multiple examinations on the same day. Two years of data were analyzed to satisfy the normality assumption by using the central limit theorem. For each examination type and size group, a one-sided t test was performed to evaluate the hypothesis that mean dose levels in the academic pediatric facilities were lower than mean radiation doses in the other facility types (), where the null hypothesis would indicate no significant mean dose difference between the four facility types. By using the academic pediatric facility as the baseline, the relative percentage difference between and was calculated for the defined category for dose indexes and patient sizes.
Size-specific Dose Variance Analysis
A one-sided F test was used to test the hypothesis that variance in the academic pediatric facilities () was smaller than that in the other facility types () for all the dose indicators applicable to the examination where the null hypothesis would indicate no significant difference in dose variance between the four facility types. The same assumptions as the unequal variance t test also hold for the F test.
Correction for Multiple Comparisons
The significance level of the tests was corrected to suppress the familywise error rate (mean and variance) that could lead to false-positive findings. The Holm-Bonferroni correction was used to modify the level of the tests by providing protection against type I error rate while resulting in a slightly higher type II error rate (28). The corrected level of the test, was based on the following equation:where N is the number of comparisons (or individual hypothesis) made, (equal to 0.05) is the overall level of the test, and the formula is applied after sorting the P values of all the tests in ascending order and testing the null hypothesis for every kth test, sequentially. The total number of tests (N) was the product of the three facility types compared to academic pediatric facilities and the number of different dose indexes. A failure of the t tests indicated no evidence for a difference in the mean dose level between academic pediatric facilities and the compared facility type. Similarly, a failure of the F tests indicated no evidence of a smaller variance in mean dose level between academic pediatric facilities and the compared facility type. All percentage relative differences in Table 2 were calculated with the following formula:where represents the mean for one of the three facility types compared with academic pediatric facility mean ().
Table 3 lists the number of each facility type, the number of examinations from each facility type, and the demographic distribution of the study population. While the number of total pediatric facilities participating in the study was less than 10% of the study’s total, pediatric facilities contributed more than 24% of the total number of examinations. A total of 239 622 submitted examinations were analyzed, including 151 386 brain, 4258 chest, and 84 657 abdomen-pelvis examinations. There were 120 819 (50.4%) and 118 803 (49.6%) studies in female and male subjects, respectively. While the age range for male and female patients at each facility type was 0–21 years, the mean age ranged from 9 to 10 years at pediatric facilities and was 15 years at adult facilities.
Figures 1–3 show the CT dose indexes as a function of patient size, facility type, and examination type. The overall values of SSDE for abdomen-pelvis examinations are approximately 50% greater than the respective values for chest examinations. The overall DLP values for abdomen-pelvis CT are approximately double those for chest CT. Tables 4–6 lists size-specific mean values and standard deviations for the SSDE (where applicable), CTDIvol, and DLP. The P values of the mean (t test) and variance (F test) of the comparisons of academic pediatric facilities to the other three types of facilities are listed in Tables 4–6. The sample size for chest examinations was less than 50 studies for seven of 24 size groups primarily in adult facilities for smaller patients. The sample sizes of all abdomen-pelvis and brain size groups were larger than 50 examinations. Figure 4 shows examples of acquired images of the three types of CT studies analyzed.
For chest and abdomen-pelvis examinations, respectively, across all six size groups, three dose descriptors, and three facilities compared with the academic pediatric facility, 42 (78%) and 48 (89%) of 54 of the unequal variance t tests with Holm-Bonferroni correction applied indicated the academic pediatric facilities had lower mean doses (P < .01). For the same examinations, 96% (52 of 54) and 94% (51 of 54), respectively, of F tests indicated that academic pediatric facilities had smaller variance in dose levels (P < .01). For all t and F tests for brain examinations, the null hypothesis was rejected. Overall, these results suggest that mean dose is lower and less variable in academic pediatric facilities than in the other three facility types. The majority of cases in which we failed to reject the null hypothesis occurred in comparisons with a sample size of fewer than 50 examinations, which occurred in patients who generally had effective diameters of 0–18 cm (Tables 4–6). Because we used the Holm-Bonferroni correction, there were instances in which P was less than 0.05 but greater than and hence was not statistically significant.
Table 2 shows the means of the selected dose indexes for three group sizes (average ages were <2 years, approximately 10 years, and >21 years) for the other three facility classifications as compared with those in the academic pediatric facility. SSDE and DLP are included for the abdomen-pelvis and chest examinations. CTDIvol is listed for the brain examinations. In general, the mean values of the dose indexes increase in the following order: academic pediatric, nonacademic pediatric, academic adult, and nonacademic adult, which results in the relative differences in Table 2. For example, for the smallest size group, mean CTDIvol of the brain retrospectively has values of 20, 21, 23, and 29 mGy, respectively. For the medium size group of the abdomen-pelvis, the mean SSDE has values of 4.3, 5.5, 5.6, and 6.7 mGy, respectively.
The analysis of dose descriptors from a national survey of three types of pediatric CT examinations showed that academic pediatric facilities deliver a significantly lower radiation dose with a significantly smaller variance to pediatric patients when compared with nonacademic pediatric, academic adult, and nonacademic adult facilities. This result held for all brain examinations and for 42 of 54 (78%) chest and 48 of 54 (89%) abdomen-pelvis examinations across all six size groups, three dose descriptors, and three facilities when compared with radiation dose at the academic pediatric facility. Also, variance for the academic pediatric facilities was significantly less for all brain examinations and the vast majority of chest and abdomen-pelvis examination size groupings among the different facilities (52 of 54 [96%] and 51 of 54 [94%], respectively). Reduction of both mean dose index and variance through careful practice improvement better serves the patient population than does a program that only reduces the mean dose index.
The previously reported calculated SSDE values in academic pediatric facilities (15,16) and those in this study were dependent on the year of the study for all sizes of patients. The mathematical fits in Figures 4 (15) and 1c (16) enable calculation and comparison of results from 2009 (abdomen CT) and 2013 (chest CT), respectively, with the results of this study (Tables 4, 5) in 2016 and 2017. For the abdomen, the SSDE in this study was 4.2 of 8.4 (50%), 4.3 of 11.5 (37%), and 5.8 of 15.6 (37%) of the study results in 2009 for the smaller than 14.5 cm, 18–22 cm, and larger than 28.5 cm size groups, respectively. For the same size groups, the SSDE of the chest in this study was 2.3 of 2.6 (88%), 2.6 of 3.8 (68%), and 4.5 of 6.3 (71%) of the study results in 2013. In our study, the majority of examinations used iterative reconstruction, while this dose reduction technique was not available in the five academic pediatric facilities in 2009, but was available in some academic pediatric facilities in 2013. In addition, when patient doses were compared between two different time periods since 2000, the more recent dose reports were lower than the prior dose reports (1–3,29,30). These reductions are most likely due to technologic improvements in the CT scanners as opposed to any practice differences in different types of facilities.
In a previous study, five academic pediatric hospitals reported that pediatric radiologists with 9—23 years of experience were comfortable interpreting images of 1-year-old patients with half the radiation dose used for a standard-size adult in the same facility (15,16). However, radiologists with less experience interpreting pediatric images may require increased doses to decrease image noise to provide equivalent care for pediatric patients. The relative differences in SSDE (trunk) or CTDIvol (head) in Table 2 indicate that adult or nonacademic pediatric facilities use one to two times the radiation dose used in academic pediatric departments at the time of this study.
A limited number of studies have compared dose indexes between academic pediatric and other facilities. A study from 2008 of 40 adult community hospitals (18) found high CTDIvol and DLP values with a wide variance of 2–42 mGy and 58–2030 mGy·cm, respectively, compared with mean values of 3 mGy ± 2 and 130 mGy·cm ± 96 for an academic pediatric hospital. A 2014 study compared 233 CT examinations performed at community hospitals with 287 examinations performed at a pediatric facility and found that CTDIvol in the community hospitals was 75% greater (8.6 vs 4.9 mGy) than in the pediatric hospital (20). A 2012–2013 brain CT survey including 250 hospitals suggested that dose indexes did not vary significantly by region of the country, trauma level, teaching status, CT accreditation, number of CT scanners, or use of dedicated pediatric CT protocols. However, the reported CTDIvol at dedicated children’s hospitals was 19% (22.3 vs 27.6 mGy) lower than that at the general hospitals (22).
Our study had some limitations that may have led to biased results: data in Table 3 indicate only 10 and 19 of 519 total facilities were academic or nonacademic pediatric facilities, respectively; thus, fewer than 59 000 of the total 240 000 examinations in our study came from pediatric facilities; only the three most common types of pediatric CT examinations—brain, chest, and abdomen-pelvis—are included in our study. Because of the voluntary nature of data submission to DIR and the inability to randomly sample data due to a limited number of patients and facilities in each size category, this study may not have completely mitigated participation bias; finally, for chest and abdomen-pelvis examinations, only sites that chose to submit localizer scan images, which were needed to calculate effective diameter used to calculate SSDE, were included in our study. In addition, despite the inclusion of 2 full years of national data, the sample size of six of 68 (9%) of our defined size ranges of effective diameter failed to exceed a sample size of 50, which weakened the power of the statistical analysis. The calculated means in our study can only be used as an estimate of the median value (achievable dose) for each size group, patient examination, and facility type; the medians for this type of dose study typically are less than their corresponding mean. Finally, analysis of different patient populations, indications, image quality, and practices of different types of facilities—all of which affect differences in patient doses—was beyond the scope of our study.
Substantial progress has been made in pediatric CT dose reduction. Abdominal CT dose in the smallest pediatric patients in nonacademic adult facilities in our study was 51% (6.1 vs 11.9 mGy) of the adult dose (Table 4). In 2001, staff were not trained to size adjust their CT radiographic techniques for small children (3), which would have delivered at least double the dose to the smallest patients when compared with the largest. Despite this progress, careful management of pediatric patient dose is not complete. Our study found patient doses in adult facilities to be up to double the dose (Table 2) of the academic pediatric facility for the same size patient. Our study found reduced variance of pediatric patient CT dose in academic pediatric facilities when compared with nonacademic, academic adult, and nonacademic adult facilities for all patient sizes. It is recommended that each facility charge a team of their radiologists, technologists, and medical physicists to compare patient CT dose indexes with the values in this national survey. A serious dose analysis and protocol review should be conducted to identify possible ways to better manage CT doses.Disclosures of Conflicts of Interest: K.J.S. disclosed no relevant relationships. E.S. disclosed no relevant relationships. D.S. disclosed no relevant relationships. J.R.M. disclosed no relevant relationships. S.L.B. disclosed no relevant relationships.
The authors thank Joanne Lovelace for her expertise in managing the references for this article and J.R. Wells, Y. Zhang, and E. Samei of the Duke University Clinical Imaging Physics Group for allowing use of their code to automatically estimate the effective diameter (patient size) from localizer images within the ACR’s DIR.
Author contributions: Guarantors of integrity of entire study, K.J.S., E.S., D.S., S.L.B.; 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; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, K.J.S., E.S., J.R.M., S.L.B.; clinical studies, E.S., S.L.B.; statistical analysis, E.S., D.S., S.L.B.; and manuscript editing, all authors
- 1. . Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol 2001;176(2):289–296. Crossref, Medline, Google Scholar
- 2. . Minimizing radiation dose for pediatric body applications of single-detector helical CT: strategies at a large children’s hospital. AJR Am J Roentgenol 2001;176(2):303–306. Crossref, Medline, Google Scholar
- 3. . Helical CT of the body: are settings adjusted for pediatric patients? AJR Am J Roentgenol 2001;176(2):297–301. Crossref, Medline, Google Scholar
- 4. . Lessons we have learned from our children: cancer risks from diagnostic radiology. Pediatr Radiol 2002;32(10):700–706. Crossref, Medline, Google Scholar
- 5. . Strategies for CT radiation dose optimization. Radiology 2004;230(3):619–628. Link, Google Scholar
- 6. . The Image Gently campaign: working together to change practice. AJR Am J Roentgenol 2008;190(2):273–274. Crossref, Medline, Google Scholar
- 7. . The ‘Image Gently’ campaign: increasing CT radiation dose awareness through a national education and awareness program. Pediatr Radiol 2008;38(3):265–269. Crossref, Medline, Google Scholar
- 8. . Image Gently campaign: making a world of difference. Med Phys Int 2015;3(2):94–108. Google Scholar
- 9. . Image Wisely: a campaign to increase awareness about adult radiation protection. Radiology 2010;257(3):601–602. Link, Google Scholar
- 10. . Size-specific dose estimates (SSDE) in pediatric and adult body CT examinations. College Park, Md: American Association of Physicists in Medicine, 2011; 30. Crossref, Google Scholar
- 11. . Dose is not always what it seems: where very misleading values can result from volume CT dose index and dose length product. J Am Coll Radiol 2014;11(3):233–237. Crossref, Medline, Google Scholar
- 12. . Using the American College of Radiology dose index registry to evaluate practice patterns and radiation dose estimates of pediatric body CT. AJR Am J Roentgenol 2018;210(3):641–647. Crossref, Medline, Google Scholar
- 13. . The ACR computed tomography dose index registry: the 5 million examination update. J Am Coll Radiol 2013;10(12):980–983. Crossref, Medline, Google Scholar
- 14. . Image Wisely: reflections of a former co-chair. JACR 2018;15(7):1055–1057. Google Scholar
- 15. . Diagnostic reference ranges for pediatric abdominal CT. Radiology 2013;268(1):208–218. Link, Google Scholar
- 16. . Pediatric chest CT diagnostic reference ranges: development and application. Radiology 2017;284(1):219–227. Link, Google Scholar
- 17. . Pediatric protocols: routine pediatric head. AAPM Web site. https://www.aapm.org/pubs/CTProtocols/documents/PediatricRoutineHeadCT.pdf. Published 2015. Accessed June 1, 2018. Google Scholar
- 18. . Variation in computed tomography radiation dose in community hospitals. J Pediatr Surg 2012;47(6):1167–1169. Crossref, Medline, Google Scholar
- 19. . Are facilities following best practices of pediatric abdominal CT scans? J Surg Res 2013;181(1):11–15. Crossref, Medline, Google Scholar
- 20. . Pediatric emergency CT scans at a children’s hospital and at community hospitals: radiation technical factors are an important source of radiation exposure. AJR Am J Roentgenol 2015;205(2):409–413. Crossref, Medline, Google Scholar
- 21. . Variation in pediatric cervical spine computed tomography radiation dose index. Acad Emerg Med 2015;22(12):1499–1505. Crossref, Medline, Google Scholar
- 22. . Variation in CT pediatric head examination radiation dose: results from a national survey. AJR Am J Roentgenol 2015;204(3):W293–W301. Crossref, Medline, Google Scholar
- 23. . Size-appropriate radiation doses in pediatric body CT: a study of regional community adoption in the United States. Pediatr Radiol 2013;43(9):1128–1135. Crossref, Medline, Google Scholar
- 24. . Automated size-specific CT dose monitoring program: assessing variability in CT dose. Med Phys 2012;39(11):7131–7139. Crossref, Medline, Google Scholar
- 25. . 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. Crossref, Medline, Google Scholar
- 26. . 2000 CDC growth charts for the United States: methods and development. Vital Health Stat 11 2002;(246):1–190. Google Scholar
- 27. . The unequal variance t-test is an underused alternative to Student’s t-test and the Mann-Whitney U test. Behav Ecol 2006;17(4):688–690. Crossref, Google Scholar
- 28. A simple sequentially rejective multiple test procedure. Scand J Stat 1979;6(2):65–70. Google Scholar
- 29. . 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.]. Crossref, Medline, Google Scholar
- 30. . Patient radiation doses from adult and pediatric CT. AJR Am J Roentgenol 2007;188(2):540–546. Crossref, Medline, Google Scholar
Article HistoryReceived: Aug 13 2018
Revision requested: Oct 8 2018
Revision received: Dec 11 2018
Accepted: Dec 18 2018
Published online: Feb 05 2019
Published in print: Apr 2019