Advertisement

Patient size as a parameter for determining Diagnostic Reference Levels for paediatric Computed Tomography (CT) procedures

Published:September 18, 2022DOI:https://doi.org/10.1016/j.ejmp.2022.09.004

      Highlights

      • SSDE evaluations of paediatric head, chest and abdomen CT were performed for the first time in Sri Lanka.
      • The geometric and attenuation-based methods were used to determine SSDE.
      • DRLs quantities based on SSDE were defined per age and size group.
      • The method suggested in AAPM report-204 to determine the body diameters using the age underestimates the size of the local children.
      • Suggested DRLs can be used as a reference for dose optimisation until a comprehensive dose audit is carried out for Sri Lanka.

      Abstract

      Introduction:

      The paediatric radiation dose has never been studied in Sri Lanka, nor has a national diagnostic reference level (NDRL) established. Therefore, the primary aim of this study was to propose diagnostic reference levels (DRL) and achievable dose (AD) values for paediatric CT examinations based on size.

      Methods:

      A total of 658 paediatric (0–15 years) non-contrast-enhanced (NC) studies of head, chest and abdomen regions performed during six months in two dedicated paediatric hospitals (out of the three such institutions in the country) were included. For head examinations, the dose indexes were analysed based on age, while for body examinations, both age and effective diameter (Deff) were used. The median and the third quartile of the pooled dose distribution were given as AD and NDRL, respectively.

      Results:

      The AD ranges for the head, chest and abdomen regions based on CTDIvol were 45.8–57.2 mGy, 2.9–10.0 mGy and 3.8–10.3 mGy. The corresponding NDRL ranges were 45.8–95.8 mGy, 3.5–14.1 mGy and 4.5–11.9 mGy. The AD ranges based on SSDEdeff and deff were 3.5–9.6 mGy and 4.1–10.3 mGy in chest and abdomen regions. The corresponding NDRL were 4.5–14.1 mGy and 6.1–10.6 mGy.

      Conclusion:

      Other institutions can use the present study DRLs as a reference dose for paediatric CT. The AD values can be used as a baseline for target dose optimisations, reducing doses up to 90%.

      Keywords

      1. Introduction

      Computed Tomography (CT) imaging accounts for the highest contributor of the imaging radiation dose in children, which is estimated to be 40%–70% even if it only accounts for 5%–10% of total imaging procedures [
      • Sivit C.J.
      Contemporary imaging in abdominal emergencies.
      ]. Therefore, it is essential to control the CT dose to avoid the potential risks [
      • Khong P.
      • Ringertz H.
      • Donoghue V.
      • Frush D.
      • Rehani M.
      • Appelgate K.
      • et al.
      ICRP publication 121: radiological protection in paediatric diagnostic and interventional radiology.
      ]. This radiation risk is of even greater in children where they have a significant risk from ionising radiation compared to adults [
      • Brody A.S.
      • Frush D.P.
      • Huda W.
      • Brent R.L.
      • et al.
      Radiation risk to children from computed tomography.
      ]. Therefore, reasonable radiation dose estimation is necessary to optimise the exposure used for a particular CT procedure. However, current CT dose descriptors, volume CT dose index (CTDIvol) and dose length product (DLP) are based on a standard size homogeneous phantom only. Hence, these parameters fail to reflect the dose variation due to variable body sizes of the patient and tissue inhomogeneity, and were found to underestimate the absorbed dose of average size adult and paediatric patients by 40%–70% [
      • McCollough C.H.
      • Leng S.
      • Yu L.
      • Cody D.D.
      • Boone J.M.
      • McNitt-Gray M.F.
      Ct dose index and patient dose: they are not the same thing.
      ,
      • Huda W.
      • Vance A.
      Patient radiation doses from adult and pediatric ct.
      ]. To resolve this issue, the American Association of Physicists in Medicine (AAPM) introduced a new dose indicator, size-specific dose estimate (SSDE), taking into account both X-ray output and patient size [
      • Boone J.M.
      • Strauss K.J.
      • Cody D.D.
      • McCollough C.H.
      • McNitt-Gray M.F.
      • Toth T.L.
      AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
      ]. The anteroposterior diameter (DAP), lateral diameter (DLAT), a combination of AP and LAT diameters (DAP+LAT), and effective diameter (Deff) methods were introduced for SSDE calculations. Based on the above methods, a set of conversions factors were derived to translate CTDIvol into SSDE. In addition, age-based conversion coefficients were also suggested by AAPM in cases where geometric data are unavailable [
      • Boone J.M.
      • Strauss K.J.
      • Cody D.D.
      • McCollough C.H.
      • McNitt-Gray M.F.
      • Toth T.L.
      AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
      ]. However, SSDE values vary between different approaches used in the calculation [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. According to a study reported by Brady & Kaufman [
      • Brady S.L.
      • Kaufman R.A.
      Investigation of American association of physicists in medicine report 204 size-specific dose estimates for pediatric ct implementation.
      ], SSDE derived using Deff or DAP+LAT varied 0.9%–2% where its individual measurements (DAP or (DLAT) showed comparatively larger variation of 2%–12%. Compared to size based approach, age derived SSDE varied only 2% for 0–13 years but 44% for 14–18 years. Also, weight was identified as a surrogate in determining size based conversion factors based on the strong correlation that exist between body weight and body diameter for the abdomen region [
      • Khawaja R.D.A.
      • Singh S.
      • Vettiyil B.
      • Lim R.
      • Gee M.
      • Westra S.
      • et al.
      Simplifying size-specific radiation dose estimates in pediatric ct.
      ]. Nevertheless, the size measuring location is also significantly impacted the derived SSDEs. Ozsoykal et al. [
      • Özsoykal İ.
      • Yurt A.
      • Akgüngör K.
      Size-specific dose estimates in chest, abdomen, and pelvis ct examinations of pediatric patients.
      ] showed that the mean SSDE calculated over the scan volume and central slice-SSDE values varied between 1.2%–11% for chest, abdomen and pelvis (CAP).
      Although the SSDE allows dose estimation for a given body size, it does not take into account variations in attenuation of the scanned body region. Therefore, later the SSDE based on water equivalent diameter (SSDEDw) was introduced to reflect the actual size specific dose in response to the attenuation characteristics of the scanned region. The AAPM report 220 [
      • Boone J.M.
      • Bostani M.
      • Boone J.M.
      • Christianson O.I.
      • McNitt-Gray M.F.
      • Supanich M.P.
      AAPM report 220: use of water equivalent diameter for calculating patient size and size-specific dose estimates (ssde) in ct.
      ] describes the attenuation-based method in deriving SSDE for chest and abdomen regions, while report 293 [
      • Boone J.M.
      • Strauss K.J.
      • Hernandez A.M.
      • Hardy A.
      • Appegate K.E.
      • Artz N.S.
      • et al.
      AAPM report 293: size-specific dose estimate (ssde) for head ct.
      ] continued the same for the head. However, as in any SSDE approach, the SSDEDw also varied based on the size measuring method. Abuhaimed & Martin [
      • Abuhaimed A.
      • Martin C.J.
      Estimation of size-specific dose estimates (ssde) for paediatric and adults patients based on a single slice.
      ] found that the Dwmidslice underestimate average Dw measured over the scan length by up to 13% and SSDE calculated using above two methods had root mean square difference’s (RMSD) of 1.2–4.0% for paediatrics. Also in case of truncated field of view (FOV), the calculated SSDEDw differs 0.22%, 0.0% and 2.21% form the non truncated cases for head, abdomen and chest respectively [
      • Anam C.
      • Haryanto F.
      • Widita R.
      • Arif I.
      • Dougherty G.
      The size-specific dose estimate (ssde) for truncated computed tomography images.
      ].
      Regardless of varying approaches and inherent limitations, the SSDE is now being incorporated in to the Diagnostic Reference Levels (DRL) process of dose optimisation specially in paediatrics [
      • Jaramillo-Garzón W.
      • Caballero M.
      • Alvarez-Aldana D.
      Size-specific dose estimates for pediatric non-contrast head ct scans: a retrospective patient study in Tunja, Colombia.
      ,
      • Mohammadbeigi A.
      • Khoshgard K.
      • Haghparast A.
      • Eivazi M.T.
      Local DRLs for paediatric ct examinations based on size-specific dose estimates in kermanshah, Iran.
      ,
      • Imai R.
      • Miyazaki O.
      • Horiuchi T.
      • Kurosawa H.
      • Nosaka S.
      Local diagnostic reference level based on size-specific dose estimates: assessment of pediatric abdominal/pelvic computed tomography at a Japanese national children’s hospital.
      ,
      • Bashier E.H.
      • Suliman I.
      Radiation dose determination in abdominal ct examinations of children at sudanese hospitals using size-specific dose estimates.
      ]. The DRLs were first introduced by the International Commission on Radiological Protection (ICRP) in 1990 as a tool for radiation protection and optimisation [
      • ICRP E.H.
      1990 Recommendations of the international commission on radiological protection.
      ]. Although the DRL concept was initially recommended for diagnostic radiology and nuclear medicine examinations, it is now being adopted in radiation oncology to assess the CT dose during radiotherapy treatment planning [
      • Božanić A.
      • Šegota D.
      • Debeljuh D.D.
      • Kolacio M.Š.
      • Radojčić Đ.S.
      • Ružić K.
      • et al.
      National reference levels of ct procedures dedicated for treatment planning in radiation oncology.
      ,
      • Wood T.J.
      • Davis A.T.
      • Earley J.
      • Edyvean S.
      • Findlay U.
      • Lindsay R.
      • et al.
      IPEM topical report: the first UK survey of dose indices from radiotherapy treatment planning computed tomography scans for adult patients.
      ,
      • Connor S.O.
      • Mc Ardle O.
      • Mullaney L.
      Establishment of national diagnostic reference levels for breast cancer ct protocols in radiation therapy.
      ]. Furthermore, the DRL ensures the delivery of optimum radiation dose during interventional radiology, including diagnostic and therapeutic procedures [
      • Srimahachota S.
      • Krisanachinda A.
      • Roongsangmanoon W.
      • Sansanayudh N.
      • Limpijankit T.
      • Chandavimol M.
      • et al.
      Establishment of national diagnostic reference levels for percutaneous coronary interventions (PCIs) in thailand.
      ,
      • Kim J.S.
      • Lee B.-K.
      • Ryu D.R.
      • Chun K.J.
      • Choi H.-H.
      • Roh Y.
      • et al.
      A multicentre survey of local diagnostic reference levels and achievable dose for coronary angiography and percutaneous transluminal coronary intervention procedures in Korea.
      ,
      • Ishibashi T.
      • Takei Y.
      • Sakamoto H.
      • Yamashita Y.
      • Kato M.
      • Tsukamoto A.
      • et al.
      Nationwide survey of medical radiation exposure on cardiovascular examinations in Japan.
      ]. The DRLs were typically set at the third quartile of the dose distribution so that it can be used as a reference level to identify abnormally higher exposures. Besides, median is recommended as the achievable dose (AD) for given country on the basis that 50% of the facilities already have doses below this value [
      • Le Heron J.
      Guidelines on patient dose to promote the optimisation of protection for diagnostic medical exposures: documents of the NRPB v 10 (1), 1999.
      ]. Further, the DRL values have been established locally and nationally for a procedure type or clinical indication [
      • Botwe B.O.
      • Schandorf C.
      • Inkoom S.
      • Faanu A.
      • Rolstadaas L.
      • Goa P.E.
      National indication-based diagnostic reference level values in computed tomography: Preliminary results from Ghana.
      ,
      • Damilakis J.
      • Vassileva J.
      The growing potential of diagnostic reference levels as a dynamic tool for dose optimization.
      ].
      To date, DRLs have only been defined for adult CT procedures in Sri Lanka [
      • Amalaraj T.
      • Satharasinghe D.
      • Pallewatte A.
      • Jeyasugiththan J.
      Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
      ] and the radiation dose from CT examinations and its variation in the paediatric population has never been studied. The purpose of this study was to suggest DRLs and AD values for non-contrast CT examinations of paediatric head, chest and abdomen regions based on CTDIvol, DLP and SSDE which can be used as a national reference until a comprehensive dose audit being carried out nationally. In addition, attenuation and geometric based approaches of SSDE are compared with the CTDIvol to determine its appropriateness in reflecting the actual dose received by local pediatric population.

      2. Material and methods

      2.1 Overview

      According to ICRP, since paediatric examinations are done less frequently than adult examinations, surveys to establish paediatric DRL values may primarily focus on the main institutions providing paediatric imaging services in the country [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ]. Therefore, in this retrospective cross-sectional study, data from two hospitals (h1 and h2) out of three dedicated paediatric hospitals in the country were included. Hence, hereafter the dose levels suggested in the present study will be referred to as NDRL. Ethical approval was given by the Ethics Committee of the Institute of Biology, Sri Lanka (ERC IOBSL 214 07 2020), and the requirement to obtain informed consent was waived as no direct patient was involved in this study.

      2.2 Data collection procedure

      The data collection was undertaken for six months, from May–December 2021. Dosimetric and morphometric data from consecutive paediatric patients (1 day–16 years) undergoing head, chest and abdomen CT scans were recorded. The dosimetric data such as CTDIvol, DLP, tube potential (kVp), average and total of the product of tube current and time (mAs) of entire scan volume, scan time, phantom size and use of dose reduction algorithms were extracted from the dose report stored in the Digital Imaging and Communications in Medicine (DICOM) header, and the morphometric data such as age and scan type were taken from the patients’ scan request form. Patients with pathological conditions such as hydrocephalus, space-occupying mass lesions or cysts, ascites etc., whose actual head or body size was altered were excluded during the data collection since they influenced the SSDE calculations.

      2.3 Patient grouping

      Mainly the patients were categorised into four age groups of 0–1, 1–5, 5–10 and 10–15 years which has been reported in the literature as the most common age stratification used for pediatric dose evaluation studies [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ,
      • Vassileva J.
      • Rehani M.
      • Kostova-Lefterova D.
      • Al-Naemi H.M.
      • Al Suwaidi J.S.
      • Arandjic D.
      • et al.
      A study to establish international diagnostic reference levels for paediatric computed tomography.
      ]. This also facilitate the comparison between the results of similar studies. In addition, patients who underwent chest and abdomen CT procedures were further grouped in to four size groups based on their effective body diameter (0–14.5, 14.5–18.0, 18.0–22 and 22.0–25.0 cm). The DRLs were calculated for each size and age group per procedure type based on CTDIvol, DLP, SSDE based on effective diameter (SSDEDeff) and SSDE based on water equivalent diameter (SSDEDw).

      2.4 Size specific dose calculation

      The dose descriptor CTDIvol refers to the average radiation dose per slice, and DLP refers to the total radiation dose for the scan [
      • Amalaraj T.
      • Satharasinghe D.
      • Pallewatte A.
      • Jeyasugiththan J.
      Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
      ,
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. SSDE reflects the dose calculated at the mid-CT slice of an individual patient corrected for the patient’s size and tissue inhomogeneity [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. In this work, the AAPM report 204 [
      • Boone J.M.
      • Strauss K.J.
      • Cody D.D.
      • McCollough C.H.
      • McNitt-Gray M.F.
      • Toth T.L.
      AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
      ] and 220 [
      • Boone J.M.
      • Bostani M.
      • Boone J.M.
      • Christianson O.I.
      • McNitt-Gray M.F.
      • Supanich M.P.
      AAPM report 220: use of water equivalent diameter for calculating patient size and size-specific dose estimates (ssde) in ct.
      ] was followed when calculating SSDE. The SSDE is defined by Eq. (1) where f is the conversion coefficient expressed by Eq. (2). The water-equivalent diameter (Dw) and effective diameter (Deff) were the choice of size metrics and estimated using Eqs. ((3), (4)). A region of interest (ROI) was drawn manually at the central axial slice of the head and chest scans. For abdomen scans, two ROIs were drawn at the level of the end plate of the second lumbar vertebrae, and the iliac crest, respectively and the average size was determined. The grey-scale windowing was adjusted to facilitate the visualisation of the skin margin to minimise the inclusion of air and table within the ROI. The areas (AROI) and the average CT numbers (CT(x,y)) were recorded for each ROI. In addition, the anteroposterior (DAP) and lateral (DLAT) diameters were measured for chest and abdomen scans using the distance measuring tool available with the post-processing software.
      SSDE=CTDIvol×f
      (1)


      f=aeb×D
      (2)


      Dw=211000CT(x,y)ROI+1×AROIπ
      (3)


      Deff=(DAP)×(DLAT)
      (4)


      The measured body diameter, D (Dw or Deff) and values a and b extracted from the AAPM report 204 [
      • Boone J.M.
      • Strauss K.J.
      • Cody D.D.
      • McCollough C.H.
      • McNitt-Gray M.F.
      • Toth T.L.
      AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
      ] were used to calculate the corresponding conversion coefficients. The values a and b for 16 cm diameter phantom (head) were 1.9852 and −0.0486 while 4.3781 and −0.0433 for 32 cm diameter phantom (chest and abdomen).
      Table 1Descriptive statistics of the scan parameters and body region dimensions for different CT scan regions and various age groups. Dw: water equivalent diameter, Deff: effective diameter, mAs: product of tube current and time, kVp:tube voltage. The range within and mean outside the brackets.
      RegionAgeCountDwDeffkVp
      Mode is given instead of mean.
      mAsmAsScan
      (years)(n)(cm)(cm)(total)(average)time (s)
      Head0–1649.8 (6.9–13.4)120 (100–140)1086 (104–1713)256 (240–335)5.7 (3.4–6.8)
      1–55410.8 (9.0–14.1)120 (100–140)1265 (247–1665)2505.0 (3.9–8.3)
      5–107311.3 (8.4–14.5)120 (120–140)1497 (1281–1750)249 (240–250)6.3 (5.2–7.0)
      10–156511.4 (9.1–14.4)120 (120–140)1501 (656–1823)2506.5 (6.1–7.3)
      Chest0–15210.7 (6.9–14.1)11.4 (8.8–14.2)100/120232 (117–392)139 (99–179)2.8 (1.5–3.9)
      1–54513.8 (7.5–16.0)13.3 (10.6–15.3)120 (100–120)436 (245–748)192 (179–205)2.4 (2.2–4.8)
      5–105514.6 (10.6–21.0)12.5 (7.4–20.1)120 (100–120)837 (394–1776)179 (179–179)3.9 (2.3–6.8)
      10–154019.5 (10.6–26.1)19.7 (8.4–23.6)120 (100–120)1127 (195–1901)296 (197–395)4.7 (2.6–6.8)
      Abdomen0–1479.1 (8.8–11.3)11.7 (10.9–12.8)120909 (789–1000)5.8 (5.8–5.8)
      1–54310.6 (8.9–13.2)13.6 (10.7–15.4)100 (100–120)627 (500–760)138 (119–187)5.4 (4.4–6.5)
      5–105111.9 (10.2–14.5)15.7 (13.1–18.2)120 (100–120)926 (623–1245)142 (127–158)5.9 (5.5–6.2)
      10–156914.6 (11.9–20.4)18.9 (17.2–22.0)120 (100–120)1418 (823–1776)155 (149–160)6.3 (5.5–6.6)
      a Mode is given instead of mean.
      Figure thumbnail gr1
      Fig. 1Distribution of effective diameter (Deff) and water equivalent diameter (Dw) of paediatric patients (0–15 years) against different age groups for (a) abdomen, (b) chest and (c) head regions.

      2.5 Statistical analysis

      Descriptive and inferential statistics were performed using Minitab® 17.1.0 statistical software. Graphical analysis was done using Originpro®2021 graphical software. The median and third quartile of the pooled dose distribution for each age category were proposed as the AD and NDRL. In addition, the relationship between body dimensions and age was determined using the fitted function. Moreover, the results of this study were compared with similar studies in other countries.
      Figure thumbnail gr2
      Fig. 2Relationship of various body dimensions with age for abdomen region. The Pearson’s correlation coefficients (r) for age against Deff, Dw, DAP, DLAT, and D(AP+LAT) were 0.89, 0.84,0.81,0.90 and 0.89.
      Figure thumbnail gr3
      Fig. 3Relationship of various body dimensions with age for chest region. The Pearson’s correlation coefficients (r) for age against eff, Dw, DAP, DLAT, and D(AP+LAT) were 0.90, 0.80, 0.83, 0.85, and 0.90.
      Table 2Descriptive statistics of the resultant dosimetric parameters for different CT scan regions. third quartile is given as the national diagnostic reference level (NDRL) while median is given as the achievable dose (AD), Dw: water equivalent diameter, Deff: effective diameter. The range within and mean outside the brackets.
      ProcedureAge or deffCTDIvol (mGy)DLP (mGy cm)SSDE (deff)SSDE (dw)
      MedianNDRLMedianNDRLMedianNDRLMedianNDRL
      (range)(range)(range)(range)
      Head NC0–1 years45.8 (2.2–95.8)45.8850.0 (36.7–1652)85056.8(2.8–117.9)60.5
      1–5 years50.4 (4.1–95.8)95.81057 (80–2035)105760.7 (5.0–107)96.1
      5–10 years57.2 (21.6–95.8)95.81163 (259–1843)116367.3 (24.6–108.5)93.8
      10–15 years57.2 (22.9–95.8)57.21206 (106–1891)120666.1 (26.4–96.0)68.9
      Chest NC0–1 years2.9 (2.1–7.3)3.550.6 (25.3–166)69.17.4 (5.5–16.7)8.77.3 (5.5–16.7)8.4
      1–5 years4.5 (3.0–7.7)7.4137 (52.0–271.0)154.010.3 (7.4–16.4)16.310.0 (7.2–16.2)15.4
      5–10 years5.4 (3.7–14.0)7.6157 (74.0–632)240.511.2(8.1–25.4)15.611.4 (9.2–20.8)15.4
      10–15 years10.0 (6.4–15.2)14.1316 (23–620)600.815.5 (11.3–25.8)23.216.6 (12.3–24.7)21.0
      0–14.5 cm3.4 (2.1–7.3)4.568 (25–166)1247.8 (5.4–16.7)10.27.5 (5.5-16.7)10.0
      14.5–18.0 cm5.6 (4.8–13.2)7.6165 (117–632)23911.2 (10.3–16.1)16.111.4 (10.0–19.615.4
      18.0–22.0 cm8.8 (6.4–15.5)11.2292 (206–612)44115.4 (11.3–18.9)18.915.9 (12.3–24.7)20.1
      22.0–25.0 cm10.0 (8.0–14.0)14.1316 (203–620)61215.5 (12.7–23.2)23.015.9 (13.1–21.0)20.9
      Abdomen NC0–1 years3.8 (3.1–6.3)4.5103 (88.0–200.0)143.59.1 (7.4–14.9)11.08.4 (8.2–16.8)12.0
      1–5 years4.1 (3.1–9.0)6.1126.2 (95.5–386.0)142.09.1 (6.9–22.2)13.58.4 (8.0–22.7)13.9
      5–10 years6.5(3.4–12.3)10.3221 (65.8–300.0)247.013.6 (7.6–23.3)19.612.8 (7.5–29.9)24.5
      10–15 years10.3 (3.4–14.1)11.9436 (76.4–603.0)603.019.5 (5.7–26.8)22.117.6 (5.9–30.0)27.0
      0–14.5 cm4.1 (3.1–9.0)5.2115 (65–386)1429.1 (6.9–22.2)12.68.4 (6.5–21.8)11.5
      14.5–18.0 cm8.9 (3.9–14.0)11.3247 (102–1036)43617.2 (8.2–26.8)21.915.7 (8.0–21.7)20.4
      18.0–22.0 cm10.3 (3.4–12.3)10.6436 (76.0–1036)43618.8 (5.7–23.4)19.617.0 (5.5–23.5)19.0
      22.0–25.0 cm16.1 (14.1–18.1)17.6605 (603–607)60625.2 (23.2–27.2)26.722.9 (20.9–24.9)24.4
      Figure thumbnail gr4
      Fig. 4Box and Whisker plot displaying the distribution of CTDIvol, SSDEDeff and SSDEDw for head, chest and abdomen regions. A,B C and D represent age ranges 0–1, 1–5, 5–10 and 10–15 years respectively.
      Figure thumbnail gr5
      Fig. 5Box and Whisker plot displaying the distribution of CTDIvol and SSDEDw for head region. A,B C and D represent age ranges 0–1, 1–5, 5–10 and 10–15 years respectively.

      3. Results

      3.1 Study sample

      A total of 658 non-contrast CT examinations (head (256), chest (192), and abdomen (210)) of patients (age 0–15 years) belonging to two paediatric hospitals were included for size-specific dose determination. The majority (68%) of the studies were extracted in the h1. Automatic exposure control (AEC) was used for 91.4% of abdomen CT scans and 93.5% of chest CT scans. None of the head CT scans used the AEC of any type.

      3.2 CT exposure parameters and patient characteristics

      Table 1 summarises the descriptive analysis of the scan parameters and the patient body dimensions. Fig. 1, illustrates the distribution of body diameters (Deff and Dw) against different age groups for head, chest and abdomen regions. Due to the large attenuation variation of the brain tissue and skull vault of the head, the most appropriate parameter to determine size is Dw [
      • Fahmi A.
      • Anam C.
      • Ali M.H.
      The size-specific dose estimate of paediatric head CT examinations for various protocols.
      ].
      The average Deff increases from 0–15 years by 72.8% and 61.5% for the chest and abdomen, respectively. The Dw increases by 16.3%, 82.2% and 60.4% for head, chest and abdomen regions, respectively. According to Table 1, the average Dw of the head does not change much after the first year of life. A nearly similar trend was observed in the geometric measurements (The AP and lateral diameters) of Kleinman et al. [
      • Kleinman P.L.
      • Strauss K.J.
      • Zurakowski D.
      • Buckley K.S.
      • Taylor G.A.
      Patient size measured on ct images as a function of age at a tertiary care children’s hospital.
      ] study (22.8% and 20.1%). Therefore, the average Dw of the head can be similar within different age categories even if narrow age groups were used. Also, it is clearly noted that the increase in Dw is higher than the increase in Deff for the chest region. Moreover, the diameter of the abdomen region (Dw) of children aged 10–15 years showed a wide variation. This finding was consistent with the recent study done by Densie et al. [
      • Bos D.
      • Zensen S.
      • Opitz M.K.
      • Haubold J.
      • Nassenstein K.
      • Kinner S.
      • et al.
      Diagnostic reference levels for chest computed tomography in children as a function of patient size.
      ].
      The correlations of body diameters (DAP, DLAT, DAP+LAT, Deff and Dw) with patient’s age for abdomen and chest procedures are illustrated in Fig. 2, Fig. 3. The lines indicate the linear fit of the function, y=a+bx. The corresponding Pearson’s correlation coefficients (r) for age against DAP, DLAT, DAP+LAT, Deff and Dw were 0.81/0.83, 0.90/0.85, 0.89/0.90, 0.89/0.90 and 0.84/0.80 in abdomen and chest regions respectively.

      3.3 Variation of dosimetric parameters

      Table 2 displays the descriptive statistics of the CT dose indexes CTDIvol, DLP, SSDEDeff and SSDEDw extracted for the most common age stratification hierarchy for children [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. In addition to the age groups provided in Table 1, four size groups based on the patient Deff has also been given for chest and abdominal CT examinations. As expected, the dose indexes increased with age and size. Fig. 4, Fig. 5 illustrate the distribution of CTDIvol, DLP, SSDEDeff and SSDEDw using Box and Whisker plots. The average CTDIvol values for the chest and abdomen regions were lower than that of SSDEDeff and SSDEDw, irrespective of the age ranges. Also, the average CTDIvol of the head is lower than SSDEDw for all age groups. Fig. 6, illustrates the relationship of water equivalent diameter (Dw) against volume computed tomography dose index (CTDIvol) and dose length product (DLP) for chest and abdomen regions. The CTDIvol and DLP increases with increasing body diameter (Dw).

      3.4 Dose comparison between institutions.

      Fig. 7 compares the resultant institutional DRLs (medians) based on CTDIvol, SSDEDeff and SSDEDw for head, chest and abdomen regions in order to identify the degree of optimisation required for institutions, h1 and h2 The head CT typical values based on SSDEDw for h2 (66.1 mGy–99.9 mGy) were significantly higher compared to h1 (56.8 mGy–65.7 mGy) for all age ranges. This variation can be attributed to the
      high tube voltages (140 kVp) used during head CT scans of the children in that specific institution. The h1 reported the highest typical values of the abdomen region based on SSDEDw for 5–10 years (12.8 mGy) and 10–15 years (18 mGy) compared to h2 reported doses, 10.4 mGy and 6.6 mGy. The age group 0–1 year was not presented for comparison due to the unavailability of sufficient data from h2. Also, the chest CT typical values (SSDEDw) were highest in h2 for all age ranges (12.1 mGy–19.2 mGy), while the resultant range of typical values for h1 was 7.3 mGy–16.6 mGy. The higher doses in h2 for the chest can be due to the use of 120 kVp for all age ranges, while the choice of kVp in h1 was shifted between 100 kVp and 120 kVp due to the routine use of age-specific scanning protocols.
      Figure thumbnail gr6
      Fig. 6Scatter plots of water equivalent diameter (Dw) against volume computed tomography dose index (CTDIvol) and dose length product (DLP) for chest and abdomen regions.

      3.5 International DRL comparison

      Fig. 8, Fig. 9 illustrate the comparison of present study DRLs based on CTDIvol and SSDEDeff with internationally published DRLs. The CTDIvol for head and torso studies were defined for 16 cm and 32 cm diameter phantoms which were similar in both institutions. A wide variation in DRLs based on both CTDIvol and SSDEDeff was observed between the similar age groups in different countries. The head DRLs based on CTDIvol were higher than most of the compared countries. The abdomen and chest DRL based on SSDE were comparable with the Iran 2019 [
      • Mohammadbeigi A.
      • Khoshgard K.
      • Haghparast A.
      • Eivazi M.T.
      Local DRLs for paediatric ct examinations based on size-specific dose estimates in Kermanshah, Iran.
      ] and the Japan 2019 [
      • Yamazaki D.
      • Miyazaki O.
      • Takei Y.
      • Matsubara K.
      • Shinozaki M.
      • Shimada Y.
      • et al.
      Usefulness of size-specific dose estimates in pediatric computed tomography: revalidation of large-scale pediatric ct dose survey data in Japan.
      ] studies. However, the DRL based on CTDIvol for the chest and abdomen region was comparable with most of the countries. In addition, Table 3 presents a DLP/CTDIvol comparison, which facilitates the evaluation of scan length variation for different scan regions and age groups.
      Table 3Comparison of present study DLP/CTDIvol ratios for head, chest and abdomen CT with the data extracted from listed countries and international reports.
      Country/OrganisationDLP/CTDIvol ratios
      HeadChestAbdomen
      0–1 years1–5 years5–10 years0–1 years1–5 years5–10 years0–1 years1–5 years5–10 years
      Present study10.511.012.119.720.831.531.723.324.0
      Japan
      • Takei Y.
      • Miyazaki O.
      • Matsubara K.
      • Shimada Y.
      • Muramatsu Y.
      • Akahane K.
      • et al.
      Nationwide survey of radiation exposure during pediatric computed tomography examinations and proposal of age-based diagnostic reference levels for Japan.
      5.36.318.820.721.825.7
      Turkey
      • Ataç G.K.
      • Parmaksız A.
      • İnal T.
      • Bulur E.
      • Bulgurlu F.
      • Öncü T.
      • Gündoğdu S.
      Patient doses from ct examinations in Turkey.
      9.3111213.315.820.589.512.5
      Italy
      • Granata C.
      • Origgi D.
      • Palorini F.
      • Matranga D.
      • Salerno S.
      Radiation dose from multidetector ct studies in children: results from the first Italian nationwide survey.
      16.519.626.5
      France
      • Célier D.
      • Roch P.
      • Etard C.
      • Le Pointe H.D.
      • Brisse H.J.
      Multicentre survey on patient dose in paediatric imaging and proposal for updated diagnostic reference levels for France. Part 1: computed tomography.
      15.520.521.6192027.83640
      Jordan
      • Rawashdeh M.
      • Abdelrahman M.
      • Zaitoun M.
      • Saade C.
      • Alewaidat H.
      • McEntee M.F.
      Diagnostic reference levels for paediatric ct in Jordan.
      15.517.422.232.325.836
      European Commission
      • EC M.
      European guidelines on diagnostic reference levels for paediatric imaging.
      13.212.612.627.827.826.034.334.327.8
      Sudan
      • Bashier E.H.
      • Suliman I.
      Radiation dose determination in abdominal ct examinations of children at sudanese hospitals using size-specific dose estimates.
      14.815141526.219.5
      Thailand
      • Kritsaneepaiboon S.
      • Trinavarat P.
      • Visrutaratna P.
      Survey of pediatric mdct radiation dose from university hospitals in thailand: a preliminary for national dose survey.
      251915.217.724.63522.330.540
      Figure thumbnail gr7
      Fig. 7Variation of typical dose values (median) between two centres (h1 and h2) based on CTDIvol, SSDEDeff and SSDEDw for head, chest and abdomen regions.
      Figure thumbnail gr8
      Fig. 8Comparison of present study DRL based on CTDIvol with internationally published DRLs. Thailand 2012 , Turkey 2015 , France 2009 , Kenya 2016 , South Korea 2017 , Japan 2019 , Sudan 2019 , Jordan 2019 , Italy 2015 , Switzerland 2008 .

      4. Discussion

      Sri Lanka has a national policy that provides free healthcare for all citizens, and most hospitals are therefore state-owned. It is reported that there are about 63 CT machines in the country, where only three are dedicated paediatric units [
      • Amalaraj T.
      • Satharasinghe D.
      • Pallewatte A.
      • Jeyasugiththan J.
      Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
      ]. More than 70% of the CT scanners are 16-slice or lesser, having primitive dose reduction capabilities [
      • Amalaraj T.
      • Satharasinghe D.
      • Pallewatte A.
      • Jeyasugiththan J.
      Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
      ]. Therefore, routine monitoring of radiation doses is essential to ensure patients receive the lowest possible doses always. The first known study conducted by a group of independent researchers in Sri Lanka had identified dose optimisation requirements for some of the CT units in the country [
      • Amalaraj T.
      • Satharasinghe D.
      • Pallewatte A.
      • Jeyasugiththan J.
      Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
      ]. However, a formal dose evaluation program has not been implemented in Sri Lanka which can be due to the limited physical and human resources in the country. Also, paediatric CT practices have not been evaluated in Sri Lanka. Unlike adult DRLs, which are set for an average subject, pediatric DRL requires collecting sufficient data for several weight groups per procedure [
      • Almén A.
      • Guðjónsdóttir J.
      • Heimland N.
      • Højgaard B.
      • Waltenburg H.
      • Widmark A.
      Establishing paediatric diagnostic reference levels using reference curves–A feasibility study including conventional and ct examinations.
      ]. This makes it inherently a mammoth task to establish a pediatric DRL. Therefore, the present study evaluated dose data from two paediatric units to represent the current national practice based on age and body size.
      The patient dose indicator CTDIvol is usually defined with respect to a cylindrical PMMA (polymethylmethacrylate) phantom of 16 cm or 32 cm in diameter. In the present study, 16 cm diameter phantom was used to define CTDIvol of the head, while 32 cm diameter phantom was used to define CTDIvol of the chest and abdomen. However, the resultant Deff of paediatric chest and abdomen regions (0–15 years) ranged between 7.4 cm to 23.6 cm, which is lower than the diameter of the phantom (32 cm) used to define the CTDIvol. Therefore, CTDIvol will underestimate the paediatric chest and abdomen CT, which emphasises the importance of a dose index where the size of the individual patient is taken into account.
      Nevertheless, when patient body size is not taken into consideration, the CTDIvol underestimates the dose by a factor between 1.8 to 2.8 for chest and abdomen regions. Moreover, CTDIvol underestimate the dose by a factor of 1.6–2.8 for chest and abdomen regions when changes in attenuation differences are not considered. However, the CTDIvol underestimates the dose in the head region only by a factor of 1.1–1.3 when changes in attenuation differences are not considered. This suggests that the CTDIvol considerably underestimates the paediatric radiation dose in CT.
      The SSDE based on body diameters facilitates the demonstration of size-dependent dose variation for CT scanning. However, it is not always possible to measure the body diameters in the clinical routine to determine SSDE. The age, weight and body mass index (BMI) were proposed as an effective alternative quantity to successfully determine the body diameters given their ready availability [
      • Brady S.L.
      • Kaufman R.A.
      Investigation of American association of physicists in medicine report 204 size-specific dose estimates for pediatric ct implementation.
      ,
      • Khawaja R.D.A.
      • Singh S.
      • Vettiyil B.
      • Lim R.
      • Gee M.
      • Westra S.
      • et al.
      Simplifying size-specific radiation dose estimates in pediatric ct.
      ,
      • Mohammadbeigi A.
      • Khoshgard K.
      • Haghparast A.
      • Eivazi M.T.
      Local DRLs for paediatric ct examinations based on size-specific dose estimates in kermanshah, Iran.
      ,
      • Iriuchijima A.
      • Fukushima Y.
      • Nakajima T.
      • Tsushima Y.
      • Ogura A.
      Simple method of size-specific dose estimates calculation from patient weight on computed tomography.
      ,
      • Parikh R.A.
      • Wien M.A.
      • Novak R.D.
      • Jordan D.W.
      • Klahr P.
      • Soriano S.
      • et al.
      A comparison study of size-specific dose estimate calculation methods.
      ]. Therefore, age was suggested as a surrogate in estimating SSDE [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ,
      • Hwang J.-Y.
      • Do K.-H.
      • Yang D.H.
      • Cho Y.A.
      • Yoon H.-K.
      • Lee J.S.
      • et al.
      A survey of pediatric ct protocols and radiation doses in South Korean hospitals to optimize the radiation dose for pediatric ct scanning.
      ]. Khawaja et al. [
      • Khawaja R.D.A.
      • Singh S.
      • Vettiyil B.
      • Lim R.
      • Gee M.
      • Westra S.
      • et al.
      Simplifying size-specific radiation dose estimates in pediatric ct.
      ] reported age and DAP,DLAT and Deff correlations of 0.69, 0.68 and 0.67 respectively for abdomen region whereas the corresponding coefficients resulted in the present study were 0.66, 0.82 and 0.79. A similar study done by Mohammadbeigi et al. [
      • Mohammadbeigi A.
      • Khoshgard K.
      • Haghparast A.
      • Eivazi M.T.
      Local DRLs for paediatric ct examinations based on size-specific dose estimates in kermanshah, Iran.
      ] had resulted moderate correlations of age with Dw (0.62) and Deff (0.54) for chest and abdomen regions respectively.
      Figure thumbnail gr9
      Fig. 9Comparison of present study DRL based on SSDE with internationally published DRLs. Sudan 2019 , Japan 2018 , Iran 2019 , Japan 2015 , USA 2017 , USA 2013 .
      The AAPM report 204 [
      • Boone J.M.
      • Strauss K.J.
      • Cody D.D.
      • McCollough C.H.
      • McNitt-Gray M.F.
      • Toth T.L.
      AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
      ] recommended a set of age-based Deff values for the abdomen region that can be used whenever the geometric data are unavailable. However, the corresponding age-based Deff values derived from the fitted function in Fig. 2, Fig. 3 were found to be lower (2%–32%) than the AAPM suggested values. This is attributed to the rather smaller physique of the Sri Lankan children compared to European children. Therefore, the function y=a+bx can be used to estimate the Deff for a particular age whenever an Asian child is concerned for SSDE calculation. The corresponding values of a and b for abdomen and chest regions were 11.62, 0.59 and 10.20, 0.72 respectively.
      Dose monitoring at a regular interval is essential for a proper quality control program. DRL is a tool to identify radiation doses, whether abnormally high or low. According to the DRL guideline, the quantities used to define DRL should adequately reflect the amount of ionising radiation used for a particular medical imaging task [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ]. Besides, dose variation due to the size and composition of different patient groups and body regions must be considered effectively during the dose optimisation. Therefore, the present study attempted to establish DRLs for children aged 0–15 years based on SSDE. According to the literature and ICRP-135, the weight or size-specific paediatric DRLs are the most appropriate [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ,
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. In the past, age has been used to group children, and DRLs have been defined accordingly. The frequently used age groups in the literature were (0–1, 1–5, 5–10, and 10–15 years) [
      • Vassileva J.
      • Rehani M.
      • Kostova-Lefterova D.
      • Al-Naemi H.M.
      • Al Suwaidi J.S.
      • Arandjic D.
      • et al.
      A study to establish international diagnostic reference levels for paediatric computed tomography.
      ,
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. However, large variations of body sizes exist even within these age groups. Further, Kleinman et al. [
      • Kleinman P.L.
      • Strauss K.J.
      • Zurakowski D.
      • Buckley K.S.
      • Taylor G.A.
      Patient size measured on ct images as a function of age at a tertiary care children’s hospital.
      ] have demonstrated that the individual patient size does not correlate well with patient age, even though the average patient sizes depend on age. Since body size mainly influences the CT dose under ATCM, it is recommended to use stratification based on body sizes rather than the age when evaluating CT radiation dose [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ]. Also, weight is identified as a more reliable factor to be used along with the DRL quantity than age [
      • Watson D.J.
      • Coakley K.S.
      Paediatric ct reference doses based on weight and ct dosimetry phantom size: local experience using a 64-slice ct scanner.
      ]. The European Commission (EC) [
      • EC M.
      European guidelines on diagnostic reference levels for paediatric imaging.
      ] has proposed a set of weight bands for the paediatric DRL process with their corresponding age bands. However, these age-specific body weights substantially vary with ethnicity. Given its limitations, age should only be used as a grouping parameter if it is the only available measure. Moreover, age-based DRLs will primarily facilitate the DRL comparison since most of the available paediatric DRLs are set for age groups [
      • Satharasinghe D.
      • Jeyasugiththan J.
      • Wanninayake W.
      • Pallewatte A.
      Paediatric diagnostic reference levels in computed tomography: a systematic review.
      ]. Therefore, in the present study, the DRLs were grouped based on patient age and effective body diameter regardless of the DRL quantity. Finally, comparisons were made of the DRLs defined per age group due to the substantial availability of similar studies in the literature.
      Approximately similar DRLs resulted in SSDEDeff and SSDEDw methods along the compared age and size groups. In contrast, most of the DRL based on CTDIvol were lower for all three body regions. The chest and abdomen CT DRL based on CTDIvol for all age ranges were comparable with the internationally published DRLs. However, the head CT DRLs based on CTDIvol is approximately doubled in the 1–5 and 5–10 years group compared to other countries. The use of high kVp (140 kVp) in h2 can be identified as the major contributor to the rise in CTDIvol. Although DRL serves as a margin to identify higher doses, it does not guide the lowest possible dose that can be used to obtain an adequate quality image. Therefore, DRL alone cannot be used to make decisions for optimisation. Nevertheless, AD provides a margin for target dose optimisation. In the present study, dose optimisations could lead to maximum dose reductions of 90.0%, 64.4% and 58.5% for head, chest and abdomen regions if AD values were followed. However, DRL and AD should not be applied for the individual patient dose index comparisons because DRL is defined for standard patients only [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ].
      Moreover, CTDIvol or SSDE comparisons alone cannot conclude that the delivered radiation doses are satisfactory since those reflect the dose to the central slice only. Sometimes a patient with a similar cross-sectional area might receive varying effective doses (ED) due to the differences in scan lengths [
      • Avramova-Cholakova S.
      • Dyakov I.
      • Yordanov H.
      • O’Sullivan J.
      Comparison of patient effective doses from multiple ct examinations based on different calculation methods.
      ]. Therefore, the scan length plays a critical role in limiting radiation dose. The DLP/CTDIvol ratio is the suitable reflector of the scan length used for a specific CT procedure. The DLP/CTDIvol ratios obtained for head, chest and abdomen regions were comparable with the internationally published and the European Commission (EC) recommended values. However, further reduction in head CT ratios (DLP/CTDIvol) can be achieved since some countries still reported lower ratios for various body regions and age groups.
      The CTDIvol and conversion factors will only estimate the dose at the central axial slice. Therefore, the determined SSDE can be slightly higher than the SSDE averaged over the entire scan volume. In addition to the sources of variability with the dose-measuring location, the size measuring technique can also significantly contribute to the uncertainties. In the present study, the ROI and diameter measurements were done only once by a single investigator, which can affect the accuracy of the measurements. Even though measures were taken to minimise the inclusion of air within the ROI, total exclusion of air is impossible in manual ROI selection. Also, Li et al. [
      • Li B.
      • Behrman R.H.
      • Norbash A.M.
      Comparison of topogram-based body size indices for ct dose consideration and scan protocol optimization.
      ] had reported that the CT table within the ROI increased the total attenuation by 12%. Wang et al. [
      • Wang J.
      • Duan X.
      • Christner J.A.
      • Leng S.
      • Yu L.
      • McCollough C.H.
      Attenuation-based estimation of patient size for the purpose of size specific dose estimation in ct. Part I. Development and validation of methods using the ct image.
      ] described that the careful placing of manual ROI without attempting to exclude all portions of the patient table still provides reasonably accurate results. In addition, the sub-optimal attenuation calibration and edge enhancing filters also add uncertainties to the measured Dw. However, the accuracy of pixels was ensured in the present study by the routine calibrations done at the two CT centres.
      Given the retrospective nature of this study, there were difficulties in obtaining some information such as weight and height. This limited the use of more appropriate stratification methods such as weight and BMI in DRL process. This was identified as the major limitation of this study. Also, the present study only evaluated the CT doses from two major paediatric hospitals in the country. The designated children’s hospitals are supposedly more experienced in paediatric radiology. Furthermore, the referred medical indications might be significantly different from other CT centres, and perhaps the radiation doses are not representative of the country. However, the ICRP-135 states that it is sufficient to focus primarily on the main hospitals that provide paediatric imaging during the initial attempt to provide pediatric DRL for a country with no reference to use as guidance [
      • Vañó E.
      • Miller D.
      • Martin C.
      • Rehani M.
      • Kang K.
      • Rosenstein M.
      • et al.
      ICRP publication 135: diagnostic reference levels in medical imaging.
      ]. Regardless of these limitations, the present study provides a better initiative for dose optimisation in Sri Lanka, specially the pediatric CT doses.

      5. Conclusion

      The current CT dose descriptor, CTDIvol, underestimates paediatric CT dose, and hence it will provide false DRLs leading to inadequate optimisation. The degree of underestimation depends on the size of the phantom used to define the CTDIvol. The 32 cm diameter phantom recommended for paediatric torso studies underestimates the dose by a factor of 1.8–2.8. Therefore, when defining DRLs for the children more reliable dose descriptor is required, such as SSDE. The Dw or Deff can be used to determine the corresponding conversion coefficient to translate CTDIvol to SSDE in children. However, if age is used as a surrogate in determining the size of a local child, the correction suggested in the present work should be used. Moreover, the DRLs and ADs suggested in this study can be used as a baseline for a dose optimisation program for paediatric CT procedures in Sri Lanka.

      References

        • Sivit C.J.
        Contemporary imaging in abdominal emergencies.
        Pediatr Radiol. 2008; 38: 675-678
        • Khong P.
        • Ringertz H.
        • Donoghue V.
        • Frush D.
        • Rehani M.
        • Appelgate K.
        • et al.
        ICRP publication 121: radiological protection in paediatric diagnostic and interventional radiology.
        Ann ICRP. 2013; 42: 1-63
        • Brody A.S.
        • Frush D.P.
        • Huda W.
        • Brent R.L.
        • et al.
        Radiation risk to children from computed tomography.
        Pediatrics. 2007; 120: 677-682
        • McCollough C.H.
        • Leng S.
        • Yu L.
        • Cody D.D.
        • Boone J.M.
        • McNitt-Gray M.F.
        Ct dose index and patient dose: they are not the same thing.
        Radiology. 2011; 259: 311-316
        • Huda W.
        • Vance A.
        Patient radiation doses from adult and pediatric ct.
        AJR Am J Roentgenol. 2007; 188: 540-546
        • Boone J.M.
        • Strauss K.J.
        • Cody D.D.
        • McCollough C.H.
        • McNitt-Gray M.F.
        • Toth T.L.
        AAPM report 204: size-specific dose estimates (ssde) in pediatric and adult body ct examinations.
        College Park (MD): American Association of Physicists in Medicine (AAPM), 2011
        • Satharasinghe D.
        • Jeyasugiththan J.
        • Wanninayake W.
        • Pallewatte A.
        Paediatric diagnostic reference levels in computed tomography: a systematic review.
        J Radiol Prot. 2021; 41: R1
        • Brady S.L.
        • Kaufman R.A.
        Investigation of American association of physicists in medicine report 204 size-specific dose estimates for pediatric ct implementation.
        Radiology. 2012; 265: 832-840
        • Khawaja R.D.A.
        • Singh S.
        • Vettiyil B.
        • Lim R.
        • Gee M.
        • Westra S.
        • et al.
        Simplifying size-specific radiation dose estimates in pediatric ct.
        AJR Am J Roentgenol. 2015; 204: 167-176
        • Özsoykal İ.
        • Yurt A.
        • Akgüngör K.
        Size-specific dose estimates in chest, abdomen, and pelvis ct examinations of pediatric patients.
        Diagn Interv Radiol. 2018; 24: 243
        • Boone J.M.
        • Bostani M.
        • Boone J.M.
        • Christianson O.I.
        • McNitt-Gray M.F.
        • Supanich M.P.
        AAPM report 220: use of water equivalent diameter for calculating patient size and size-specific dose estimates (ssde) in ct.
        College Park (MD): American Association of Physicists in Medicine (AAPM), 2014
        • Boone J.M.
        • Strauss K.J.
        • Hernandez A.M.
        • Hardy A.
        • Appegate K.E.
        • Artz N.S.
        • et al.
        AAPM report 293: size-specific dose estimate (ssde) for head ct.
        College Park (MD): American Association of Physicists in Medicine (AAPM), 2019
        • Abuhaimed A.
        • Martin C.J.
        Estimation of size-specific dose estimates (ssde) for paediatric and adults patients based on a single slice.
        Phys Med. 2020; 74: 30-39
        • Anam C.
        • Haryanto F.
        • Widita R.
        • Arif I.
        • Dougherty G.
        The size-specific dose estimate (ssde) for truncated computed tomography images.
        Radiat Prot Dosim. 2017; 175: 313-320
        • Jaramillo-Garzón W.
        • Caballero M.
        • Alvarez-Aldana D.
        Size-specific dose estimates for pediatric non-contrast head ct scans: a retrospective patient study in Tunja, Colombia.
        Radiat Prot Dosim. 2021; 193: 221-227
        • Mohammadbeigi A.
        • Khoshgard K.
        • Haghparast A.
        • Eivazi M.T.
        Local DRLs for paediatric ct examinations based on size-specific dose estimates in kermanshah, Iran.
        Radiat Prot Dosim. 2019; 186: 496-506
        • Imai R.
        • Miyazaki O.
        • Horiuchi T.
        • Kurosawa H.
        • Nosaka S.
        Local diagnostic reference level based on size-specific dose estimates: assessment of pediatric abdominal/pelvic computed tomography at a Japanese national children’s hospital.
        Pediatr Radiol. 2015; 45: 345-353
        • Bashier E.H.
        • Suliman I.
        Radiation dose determination in abdominal ct examinations of children at sudanese hospitals using size-specific dose estimates.
        Radiat Prot Dosim. 2019; 183: 444-449
        • ICRP E.H.
        1990 Recommendations of the international commission on radiological protection.
        Ann ICRP. 1991; 21: 1-201
        • Božanić A.
        • Šegota D.
        • Debeljuh D.D.
        • Kolacio M.Š.
        • Radojčić Đ.S.
        • Ružić K.
        • et al.
        National reference levels of ct procedures dedicated for treatment planning in radiation oncology.
        Phys Med. 2022; 96: 123-129
        • Wood T.J.
        • Davis A.T.
        • Earley J.
        • Edyvean S.
        • Findlay U.
        • Lindsay R.
        • et al.
        IPEM topical report: the first UK survey of dose indices from radiotherapy treatment planning computed tomography scans for adult patients.
        Phys Med Biol. 2018; 63185008
        • Connor S.O.
        • Mc Ardle O.
        • Mullaney L.
        Establishment of national diagnostic reference levels for breast cancer ct protocols in radiation therapy.
        Br J Radiol. 2016; 8920160428
        • Srimahachota S.
        • Krisanachinda A.
        • Roongsangmanoon W.
        • Sansanayudh N.
        • Limpijankit T.
        • Chandavimol M.
        • et al.
        Establishment of national diagnostic reference levels for percutaneous coronary interventions (PCIs) in thailand.
        Phys Med. 2022; 96: 46-53
        • Kim J.S.
        • Lee B.-K.
        • Ryu D.R.
        • Chun K.J.
        • Choi H.-H.
        • Roh Y.
        • et al.
        A multicentre survey of local diagnostic reference levels and achievable dose for coronary angiography and percutaneous transluminal coronary intervention procedures in Korea.
        Radiat Prot Dosim. 2019; 187: 378-382
        • Ishibashi T.
        • Takei Y.
        • Sakamoto H.
        • Yamashita Y.
        • Kato M.
        • Tsukamoto A.
        • et al.
        Nationwide survey of medical radiation exposure on cardiovascular examinations in Japan.
        Nihon Hoshasen Gijutsu Gakkai Zasshi. 2020; 76: 64-71
        • Le Heron J.
        Guidelines on patient dose to promote the optimisation of protection for diagnostic medical exposures: documents of the NRPB v 10 (1), 1999.
        Radiat Prot News Notes. 1999; : 17-18
        • Botwe B.O.
        • Schandorf C.
        • Inkoom S.
        • Faanu A.
        • Rolstadaas L.
        • Goa P.E.
        National indication-based diagnostic reference level values in computed tomography: Preliminary results from Ghana.
        Phys Med. 2021; 84: 274-284
        • Damilakis J.
        • Vassileva J.
        The growing potential of diagnostic reference levels as a dynamic tool for dose optimization.
        Phys Med. 2021; 84: 285-287
        • Amalaraj T.
        • Satharasinghe D.
        • Pallewatte A.
        • Jeyasugiththan J.
        Establishment of national diagnostic reference levels (NDRL) for computed tomographic (ct) procedures in Sri Lanka: first nation-wide dose survey.
        J Radiol Prot. 2021;
        • Vañó E.
        • Miller D.
        • Martin C.
        • Rehani M.
        • Kang K.
        • Rosenstein M.
        • et al.
        ICRP publication 135: diagnostic reference levels in medical imaging.
        Ann ICRP. 2017; 46: 1-144
        • Vassileva J.
        • Rehani M.
        • Kostova-Lefterova D.
        • Al-Naemi H.M.
        • Al Suwaidi J.S.
        • Arandjic D.
        • et al.
        A study to establish international diagnostic reference levels for paediatric computed tomography.
        Radiat Prot Dosim. 2015; 165: 70-80
        • Satharasinghe D.
        • Jeyasugiththan J.
        • Wanninayake W.
        • Pallewatte A.
        Paediatric diagnostic reference levels in computed tomography: a systematic review.
        J Radiol Prot. 2021; 41: R1
        • Fahmi A.
        • Anam C.
        • Ali M.H.
        The size-specific dose estimate of paediatric head CT examinations for various protocols.
        Radiat Prot Dosim. 2020; 188: 522-528
        • Kleinman P.L.
        • Strauss K.J.
        • Zurakowski D.
        • Buckley K.S.
        • Taylor G.A.
        Patient size measured on ct images as a function of age at a tertiary care children’s hospital.
        AJR Am J Roentgenol. 2010; 194: 1611-1619
        • Bos D.
        • Zensen S.
        • Opitz M.K.
        • Haubold J.
        • Nassenstein K.
        • Kinner S.
        • et al.
        Diagnostic reference levels for chest computed tomography in children as a function of patient size.
        Pediatr Radiol. 2022; : 1-10
        • Mohammadbeigi A.
        • Khoshgard K.
        • Haghparast A.
        • Eivazi M.T.
        Local DRLs for paediatric ct examinations based on size-specific dose estimates in Kermanshah, Iran.
        Radiat Prot Dosim. 2019; 186: 496-506
        • Yamazaki D.
        • Miyazaki O.
        • Takei Y.
        • Matsubara K.
        • Shinozaki M.
        • Shimada Y.
        • et al.
        Usefulness of size-specific dose estimates in pediatric computed tomography: revalidation of large-scale pediatric ct dose survey data in Japan.
        Radiat Prot Dosim. 2018; 179: 254-262
        • Takei Y.
        • Miyazaki O.
        • Matsubara K.
        • Shimada Y.
        • Muramatsu Y.
        • Akahane K.
        • et al.
        Nationwide survey of radiation exposure during pediatric computed tomography examinations and proposal of age-based diagnostic reference levels for Japan.
        Pediatr Radiol. 2016; 46: 280-285
        • Ataç G.K.
        • Parmaksız A.
        • İnal T.
        • Bulur E.
        • Bulgurlu F.
        • Öncü T.
        • Gündoğdu S.
        Patient doses from ct examinations in Turkey.
        Diagn Interv Radiol. 2015; 21: 428
        • Granata C.
        • Origgi D.
        • Palorini F.
        • Matranga D.
        • Salerno S.
        Radiation dose from multidetector ct studies in children: results from the first Italian nationwide survey.
        Pediatr Radiol. 2015; 45: 695-705
        • Célier D.
        • Roch P.
        • Etard C.
        • Le Pointe H.D.
        • Brisse H.J.
        Multicentre survey on patient dose in paediatric imaging and proposal for updated diagnostic reference levels for France. Part 1: computed tomography.
        Eur Radiol. 2020; 30: 1156-1165
        • Rawashdeh M.
        • Abdelrahman M.
        • Zaitoun M.
        • Saade C.
        • Alewaidat H.
        • McEntee M.F.
        Diagnostic reference levels for paediatric ct in Jordan.
        J Radiol Prot. 2019; 39: 1060
        • EC M.
        European guidelines on diagnostic reference levels for paediatric imaging.
        in: Radiation protection. Vol. 185. European Union, Luxembourg2016
        • Kritsaneepaiboon S.
        • Trinavarat P.
        • Visrutaratna P.
        Survey of pediatric mdct radiation dose from university hospitals in thailand: a preliminary for national dose survey.
        Acta Radiol. 2012; 53: 820-826
        • Ataç G.K.
        • Parmaksız A.
        • İnal T.
        • Bulur E.
        • Bulgurlu F.
        • Öncü T.
        • et al.
        Patient doses from ct examinations in Turkey.
        Diagn Interv Radiol. 2015; 21: 428
        • Brisse H.
        • Aubert B.
        Ct exposure from pediatric mdct: results from the 2007–2008 SFIPP/ISRN survey.
        J Radiol. 2009; 90: 207-215
        • Korir G.K.
        • Wambani J.S.
        • Korir I.K.
        • Tries M.A.
        • Boen P.K.
        National diagnostic reference level initiative for computed tomography examinations in Kenya.
        Radiat Prot Dosim. 2016; 168: 242-252
        • Kim M.
        • Chang K.
        • Hwang J.
        • Nam Y.
        • Han D.
        • Yoon J.
        Radiation dose for pediatric and young adult ct: A survey to establish age-based reference levels of 2015–2016 in Korea.
        Radiat Prot Dosim. 2017; 175: 228-237
        • Bashier E.H.
        • Suliman I.
        Radiation dose determination in abdominal ct examinations of children at sudanese hospitals using size-specific dose estimates.
        Radiat Prot Dosim. 2019; 183: 444-449
        • Verdun F.R.
        • Gutierrez D.
        • Vader J.P.
        • Aroua A.
        • Alamo-Maestre L.T.
        • Bochud F.
        • et al.
        Ct radiation dose in children: a survey to establish age-based diagnostic reference levels in Switzerland.
        Phys Med Eur Radiol. 2008; 18: 1980-1986
        • Almén A.
        • Guðjónsdóttir J.
        • Heimland N.
        • Højgaard B.
        • Waltenburg H.
        • Widmark A.
        Establishing paediatric diagnostic reference levels using reference curves–A feasibility study including conventional and ct examinations.
        Phys Med. 2021; 87: 65-72
        • Strauss K.J.
        • Goske M.J.
        • Towbin A.J.
        • Sengupta D.
        • Callahan M.J.
        • Darge K.
        • et al.
        Pediatric chest ct diagnostic reference ranges: development and application.
        Radiology. 2017; 284: 219-227
        • Goske M.J.
        • Strauss K.J.
        • Coombs L.P.
        • Mandel K.E.
        • Towbin A.J.
        • Larson D.B.
        • et al.
        Diagnostic reference ranges for pediatric abdominal CT.
        Radiology. 2013; 268: 208-218
        • Iriuchijima A.
        • Fukushima Y.
        • Nakajima T.
        • Tsushima Y.
        • Ogura A.
        Simple method of size-specific dose estimates calculation from patient weight on computed tomography.
        Radiat Prot Dosim. 2018; 178: 208-212
        • Parikh R.A.
        • Wien M.A.
        • Novak R.D.
        • Jordan D.W.
        • Klahr P.
        • Soriano S.
        • et al.
        A comparison study of size-specific dose estimate calculation methods.
        Pediatr Radiol. 2018; 48: 56-65
        • Hwang J.-Y.
        • Do K.-H.
        • Yang D.H.
        • Cho Y.A.
        • Yoon H.-K.
        • Lee J.S.
        • et al.
        A survey of pediatric ct protocols and radiation doses in South Korean hospitals to optimize the radiation dose for pediatric ct scanning.
        Medicine. 2015; 94
        • Watson D.J.
        • Coakley K.S.
        Paediatric ct reference doses based on weight and ct dosimetry phantom size: local experience using a 64-slice ct scanner.
        Pediatr Radiol. 2010; 40: 693-703
        • Avramova-Cholakova S.
        • Dyakov I.
        • Yordanov H.
        • O’Sullivan J.
        Comparison of patient effective doses from multiple ct examinations based on different calculation methods.
        Phys Med. 2022; 99: 73-84
        • Li B.
        • Behrman R.H.
        • Norbash A.M.
        Comparison of topogram-based body size indices for ct dose consideration and scan protocol optimization.
        Med Phys. 2012; 39: 3456-3465
        • Wang J.
        • Duan X.
        • Christner J.A.
        • Leng S.
        • Yu L.
        • McCollough C.H.
        Attenuation-based estimation of patient size for the purpose of size specific dose estimation in ct. Part I. Development and validation of methods using the ct image.
        Med Phys. 2012; 39: 6764-6771