Paediatric CT optimisation utilising Catphan(R) 600 and age-specific anthropomorphic phantoms

Paediatric CToptimisation utilising Catphanw 600 and age-specificanthropomorphic phantomsJoana Santos1,*, Maria do Carmo Batista2, Shane Foley3, Graciano Paulo1, Mark F. McEntee4

and Louise Rainford31Instituto Politecnico de Coimbra, ESTESC-Coimbra Health School, Radiologia, Rua 5 de Outubro,S. Martinho do Bispo, 3046-854 Coimbra, Portugal2Departamento de Fısica Medica, Dr. Campos Costa, Consultorio de Tomografia Computorizada S.A,Porto, Portugal3School of Medicine &Medical Science, Health Science Centre, University College Dublin, Belfield Dublin 4,Ireland4Faculty of Health Sciences, The University of Sydney, Cumberland Campus, Sydney, Australia

*Corresponding author: [email protected]

Received 23 October 2013; revised 15 January 2014; accepted 29 January 2014

The purpose of the study is to perform phantom-based optimisation of paediatric computed tomography (CT) protocols andquantify the impact upon radiation dose and image noise levels. The study involved three Portuguese paediatric centres.Currently employed scanning protocols for head and chest examinations and combinations of exposure parameters were appliedto a Catphanw600 phantom to review the CT dose impact. Contrast–noise ratio (CNR) was quantified using Radia Diagnosticw

tool. Imaging parameters, returning similar CNRs (<1) and dose savings were applied to three paediatric anthropomorphicphantoms. OsiriX software based on standard deviation pixel values facilitated image noise analysis. Currently employed proto-cols and age categorisation varied between centres. Manipulation of exposure parameters facilitated mean dose reductions of 33and 28 % for paediatric head and chest CT examinations, respectively. The majority of the optimised CT examinations resultedin image noise similar to currently employed protocols. Dose reductions of up to 33 % were achieved with image qualitymaintained.

INTRODUCTION

As with all medical procedures, computed tomog-raphy (CT) examinations present both clinical bene-fits and potential radiation risks. In the past 10 y, theemployment of CT examinations for paediatricpatients increased ≏700 % worldwide and increas-ingly being the preferred method in daily practiceand emergency departments(1, 2). The clinical applic-ability of CT for paediatric diagnosis is unquestion-able; however, the potential risk of high radiationexposure associated with CT should not be ignored(3).Recent studies have suggested that CT examinations inchildren can deliver examination doses of ≏50–60 mGythat might almost triple the risk of leukaemia or braincancer, respectively(4, 5).

The clinical benefits of prescribed examinationsshould outweigh the risks(4, 6). Justification of all CTexaminations is a legislative requirement; however,recent studies have indicated that many paediatric CTscans are unnecessary(7–9). Optimisation of practice isessential to ensure the minimisation of radiation doseto the patient whilst ensuring diagnostic efficacy is main-tained(10). Several studies have reported that optimisationprocesses need to consider numerous factors includ-ing body region, clinical information, the CT scanner

technology available and image processing(8, 11–14).The Image Gently campaign clearly states that ‘Childrenare not just smaller adults, their bodies are different andrequire a different approach to imaging’. Paediatric CTprotocols should be defined by patient size taking intoaccount their high radiosensitivity and longer lifetimeexpectancy(15–17).

Multi-slice CT (MSCT) technology evolution overthe last decade increased the CT image quality byallowing the use of finer slice thicknesses and de-creased the examination time(18). However, to main-tain acceptable noise when using finer slice thicknesssettings, patient doses must be increased(19). The doselevels in CT examinations are determined by CTscanner technology and by exposure parameter selec-tion(20). The number of detectors, beam shape and fil-tration, the data acquisition system and the tubecurrent modulation are the principal differencesbetween CT manufacturers(21). A thorough under-standing of CT scanner design characteristics isessential to aid optimisation of practice(15).

Several research studies have focused upon optimisa-tion in paediatric CT. Nievelstein et al.(10) indicated anumber of strategies for paediatric CT dose reduction,for example: selective organ shielding, minimising the

# The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Radiation Protection Dosimetry (2014), pp. 1–11 doi:10.1093/rpd/ncu018

Radiation Protection Dosimetry Advance Access published February 23, 2014 at U

niversity College D

ublin on February 25, 2014http://rpd.oxfordjournals.org/

Dow

nloaded from

http://rpd.oxfordjournals.org/


number of examination phases and exposure para-meter changes, to include increased pitch values, ap-propriate tube voltage selection and automatic tubecurrent modulation. Strauss et al.(15) defined ten stepsto optimise paediatric CT dose beyond the indicationsfor tube voltage, tube current and pitch already out-lined. These included: centring the patent in the middleof the gantry, reducing the dose during scout views,selecting the acquisition mode according to the bodyregion and reducing the detector size in z direction.Lifeng Yu et al.(14) analysed tube current and tubevoltage techniques in phantoms and patients to reducepaediatric CT dose and reported that the use of lowertube potential should be carefully selected according topatient sizes and the diagnosis task being performed.Other authors(8) have presented the impact of exposureparameter manipulation and image processing in chestand abdominal paediatric CT examinations, identify-ing potential reductions in dose of up to 30 % and themanagement of noise levels through post-processingtechniques.

The majority of CT scanners use automatic exposurecontrol based on tube current modulation. Iterative re-construction and the adjustment of tube voltage basedon patient size are the most recent and promising tech-nologies for optimisation in CT; however, these are notwidely available(11, 22, 23). Reductions in tube voltageand current will result in lower CT dose levels;however, this in turn impacts upon the contrast–noiseratio (CNR) and the balance of the amount of noisepresent in resultant image. Therefore, image qualityconsideration is required to ensure whether diagnosticquality is not lost and radiology confidence is main-tained(9, 24, 25).

This study is focussed upon head and chest CTexaminations being the most common paediatric CTexaminations across Europe(26–30). It has also beennoted that dose optimisation has higher relevance onpopulation effective dose levels if applied on mostcommon CT procedures(31).

The aim of this research was to investigate methodsof optimisation for paediatric head and chest exami-nations following review of the CT protocols currentlyemployed in the three national paediatric centres inPortugal. Optimisation was based on the manipula-tion of exposure parameters applied to anthropo-morphic paediatric phantoms and included imagenoise evaluation based on standard deviation mea-surements within defined homogenous regions ofinterest (ROIs) following initial experimental testingon a Catphanw 600 phantom.

MATERIALS ANDMETHODS

Optimisation tests were carried out in the threededicated regional, public paediatric centres, inPortugal (A, B and C), each performing ≏3000 CTexaminations annually. All 3 centres have MSCT

models, 2 manufactured by SiemensTM (64 multidetec-tor rows and 6 multidetector rows) and 1 by PhilipsTM

(16 multidetector rows). CT protocols were collectedfrom each centre and reviewed.

The CT Dose Index (CTDIvol—mGy) of the threeCT scanners was verified using a calibrated RaysafeTM

Xi CT ionisation chamber and a PMMA (Polymethylmethacrylate) CTDI phantom. The Statistical Packagefor Social Sciences (SPSS—version 20) software wasemployed to complete descriptive statistical analysis ofthe data collected. Three separate phases of experimen-tal work formed the research.

Phase 1: optimisation using Catphanw 600

A Catphanw 600 (The Phantom Laboratory, Salem,USA) CT quality assurance (QA) phantom (Figure 1)was employed to optimise the existing protocols. TheCatphanw 600 is internationally recognised as CT QAphantom, for use in axial, spiral and multi-slice CTscanners. This phantom is constructed by solid castmaterial modules (Figure 2) suitable for testing lowcontrast with supra-slice and sub-slice contrast targets(module CTP 515)(32). The existing protocols wereapplied in alignment with paediatric age categoriesused internationally(26–29): these being: newborns,5- and 10-y-old children.

Head and chest CT protocols were used to scan theCatphanw600 in the three paediatric centres, exposure

Figure 1. CT quality assurance phantom, Catphan w 600,positioning to test head and chest paediatric CT

examinations.

J. SANTOS ETAL.

Page 2 of 11

at University C

ollege Dublin on February 25, 2014

http://rpd.oxfordjournals.org/D

ownloaded from



parameters were systematically lowered from current-ly employed parameters: tube voltage, tube currentand slice thickness were decreased and pitch wasincreased; acquisition mode and dose modulationbased on tube current was also tested. The scannedlength remained constant (20 cm), and CT scannerdose reports were used to obtain CTDIvol per headand chest CTexamination.

Phase 2: image quality evaluation with RadiaDiagnosticw software

The Catphanw 600 images generated were evaluatedby Radia Diagnosticw Imaging QC software fromRadiological Imaging Technology (RIT), Inc., CO,USA. This software scores images per AmericanCollege of Radiology guidelines and generates ana-lysis reports per module with measures, plots(Figure 3) material values and CT number linear-ity(33). The CNR results, of module CTP 515 of theCatphanw 600, was the parameter considered forimage evaluation as noise is the principal limitingfactor in CT image quality and is directly influencedby radiation dose.

Phase 3: optimisation with age-specificanthropomorphic phantoms

Following review of the dose and image noise findingsobtained with the Catphanw 600, the experimentalprotocols, according to scanner type, were definedand applied to head and chest CTexaminations usingCIRSw anthropomorphic phantoms (ATOM dosim-etry verification phantoms—model 703, 705 and706), which simulate 0- (3.5 kg, 51 cm), 5- (19 kg, 110cm) and 10- (32 kg, 140 cm) y-old children (Figure 4),respectively. CTDIvol findings were recorded from theCT scanner dose reports. DLP values were not consid-ered for the anthropomorphic phantoms (APs) due tothe consistency of range length used.

Phase 4: anthropomorphic image quality evaluationwith OsiriXw software

Anthropomorphic images were analysed usingOsiriXw Imaging software (Antoine Rosset, Geneva)version 4.0 32 bit, using the standard deviation of 1cm2 homogenous ROIs. For head CT examinations,ROIs were established in the supra-tentorial region,

Figure 2. CT images of Catphan w 600 modules. (A) CTP 404, (B) CTP 515, (C) CTP 486, (D) CTP 591.

PAEDIATRIC CTOPTIMISATION UTILISING PHANTOMS

Page 3 of 11

at University C



ownloaded from



Figure 3. Example of RIT’s Radia Diagnosticw software reports of the four different Catphanw 600 modules. (A) CTP 404,(B) CTP 515, (C) CTP 486, (D) CTP 591.

J. SANTOS ETAL.

Page 4 of 11

at University C



ownloaded from



orbits and infra-tentorial region, and for chest CT, theROIs were positioned at points in the region of theshoulder and heart (Figure 5). A total of 9 ROIs weredefined for head CT examinations and 6 ROIs forchest, totalling 465 ROIs.

RESULTS


A total of 99 CTexaminations were performed on theCatphanw 600 phantom involving the application ofexperimental head and chest imaging protocols; thesewere tailored to the scanner model: Centre A (n¼38),Centre B (n¼30) and Centre C (n¼31).

For head CT examinations, Centre A reported twocurrently employed protocols (0- to 3-y-olds and .3-y-olds), Centre B identified three protocols (newborns,1-y-olds and 2- to 10-y-olds) and Centre C using threedifferent categories (0- to 18-month-olds, 18-month- to6-y-olds and .7-y-olds). For chest CT examinations,the currently employed protocol categorisation is thesame as outlined for head examinations except forCentre C, which had one currently employed protocol,and reported that this was adjusted in practice to indi-vidual patients.


Following a review of the Catphanw 600 findings, theselection of protocols to be applied to the APs waslimited to a variation of 1.3 in CNR, when comparedwith the currently employed. The protocols weredivided per age categorisation, in order to be per-formed on age-specific APs.

Phase 3: optimisation with age-specific APs

A total of 61 CTexaminations were performed to testthe impact of exposure parameter manipulation fromthe currently employed values, for the three scannermodels. The CNR findings obtained by the RadiaDiagnosticw software and the resulting dose valuesfrom the 0-, 5- and 10-y-old APs (AP) for the threepaediatric centres are presented in Tables 1–3.

In comparison with the currently employed proto-col, the overall mean percentage dose reductionachieved was 42, 31 and 25 % for head CT examina-tions and 38, 39 and 6 % for chest CT examinations,respectively, for 0-, 5- and 10-y-old APs.

The mean dose reduction per paediatric centre was36, 25 and 32 % for head CT examinations and 9, 29and 40 % for chest CT examinations, respectively, forCentres A, B and C in comparison with the currentlyemployed protocol.


The results of the ROI standard deviation and CTDIvolreduction analysis are summarised in Tables 4 and 5.

The majority of the optimised CT protocols, acrossthe three paediatric centres resulted in reduced imagenoise when compared with the currently employedprotocols, returning a mean pixel value standard devi-ation of 10+6.4. In comparison with the currentlyemployed parameters for head CT examinations, thehighest variation in mean pixel value standard devi-ation of the experimental protocols were 35, 13 and58 % for Centres A, B and C, respectively. The highestvariation in mean pixel value standard deviationrecorded for chest examinations were 8, 54 and 36 %for Centres A, B and C, respectively.

DISCUSSION


The variation in CTequipment across the three centresdetermined the need to tailor optimisation to equip-ment models(25).

A number of exposure parameters impeded theability to standardise across the centres, these included:variation in tube voltage selection and tube currentdose modulation options, slice thickness combinations

Figure 4. CIRSw Anthropomorphic phantoms (ATOMModel 706, 705 and 703) used to perform head and chest

paediatric CTexaminations.


Page 5 of 11

at University C



ownloaded from



Figure 5. Example of ROI locations, for image analyses with OsiriXw software, for head (A–C) and chest (D and E)CTexaminations on the AP (Model 705).

Table 1. A summary of CNR findings using the RIT’s Radia Diagnosticw software and the CT dose values (CTDIvol) of thecurrently employed and experimental protocols for head and chest paediatric CTexaminations performed on APs in Centre A.

BodyRegion

n8 Tubevoltage(kV)

Rotationtime (s)

Tubecurrent–timeproduct (mAs)

Slicethickness(mm)

Mode Pitch Tube currentmodulation

CNR AP CTDI(mGy)

Head 1a 120 1 230 4.8 A — No 2.51 0 38.252 120 1 200 4.8 A — No 2.39 0 33.413 100 1 230 4.8 A — No 1.63 0 23.464 80 1 260 4.8 A — No 1.50 0 13.425a 120 1 310/290 2.4/4.8 A — No 1.38 5 49.956 100 1 310/290 2.4/4.8 A — No 1.81 5 30.647 100 1 250 2.4/4.8 A — No 2.01 5 25.518 120 1 310/290 2.4/4.8 A — No 1.38 10 41.579 100 1 310/290 2.4/4.8 A — No 1.81 10 25.0310 100 1 250 2.4/4.8 A — No 2.01 10 20.82

Chest 11a 80 0.5 50 5 H 0.8 No 0.77 0 0.8112 80 0.5 50 5 H 0.8 Yes 0.72 0 0.6513 80 0.5 40 5 H 0.9 No 2.50 0 0.6814a 100 0.5 50 5 H 0.8 No 1.46 5 1.7615 100 0.5 40 5 H 0.8 No 1.09 5 1.4116 100 0.5 50 5 H 0.9 No 1.17 5 1.7617 100 0.5 50 5 H 0.8 Yes 1.46 10 1.9718 100 0.5 50 5 H 0.8 No 1.09 10 2.2219 100 0.5 50 5 H 0.9 No 1.17 10 2.22

aCT currently employed protocol.

J. SANTOS ETAL.

Page 6 of 11

at University C



ownloaded from



Table 2. A summary of CNR findings using the RIT’s Radia Diagnosticw software and the CT dose values (CTDIvol) of thecurrently employed and experimental protocols for head and chest paediatric CTexaminations performed on APs in centre B.

BodyRegion

n8 Tubevoltage(kV)

Rotationtime (s)


Slicethickness(mm)


CNR AP CTDI(mGy)

Head 1a 80 2.5 200/130 5 A — Yes 1.74 0 6.042 80 2.5 180 5 A — No 1.49 0 10.983 80 2.5 130 5 A — No 1.51 0 7.934 80 2.5 180 5 H 0.9 Yes 1.71 0 10.985a 130 2.5 140 4 A — No 2.42 5 33.046 110 2.5 140 4 A — No 2.14 5 22.827 110 2.5 130/183 4 H 0.9 Yes 2.49 5 32.208 110 2.5 130 4 H 0.9 No 2.33 5 24.709 130 2.5 140 4 A — No 2.42 10 30.50

10 110 2.5 140 4 A — No 2.14 10 21.06Chest 11a 110 0.8 40 5 H 1.5 Yes 1.33 0 1.90

12 110 0.8 40 5 H 1.5 No 1.18 0 3.4813 80 0.8 40 5 H 1.5 Yes 0.89 0 0.9314 80 0.8 40 5 H 1.5 No 1.58 0 1.4015a 110 0.8 80 5 H 1.5 Yes 1.00 5 7.6216 110 0.8 80 5 H 1.5 No 1.44 5 6.9817 110 0.8 50 5 H 1.5 Yes 1.53 5 3.9418 110 0.8 50 5 H 1.5 No 1.23 5 4.3619a 110 0.8 80 5 H 1.5 Yes 1.00 10 2.6420 110 0.8 80 5 H 1.5 No 1.44 10 6.0021 110 0.8 50 5 H 1.5 Yes 1.53 10 2.6422 110 0.8 50 5 H 1.5 No 1.23 10 3.76


Table 3. A summary of CNR findings using the RIT’s Radia Diagnosticw software and the CT dose values (CTDIvol) of thecurrently employed and experimental protocols for head and chest paediatric CTexaminations performed on APs in Centre C.

BodyRegion

n8 Tubevoltage(kV)

Rotationtime (s)


Slicethickness(mm)


CNR AP CTDI(mGy)

Head 1a 120 0.75 300 3 A — No 2.42 0 45.602 90 0.75 300 3 A — No 2.15 0 21.103 120 0.75 250 3 A — No 1.51 0 38.004 90 0.75 230 3 A — No 1.96 0 16.205 120 1 300 3 A — No 1.23 5 44.906a 120 0.75 350 3 A — No 2.45 5 52.407 120 0.75 250 3 A — No 3.49 5 37.408 90 0.75 250 3 A — No 2.20 5 17.509 120 0.75 375/300 3/6 A — No 2.45 10 42.7510 120 0.75 350/250 3/6 A — No 2.20 10 36.55

Chest 11a 120 0.5 50 3 H 0.688 No 0.56 0 3.5012 90 0.5 50 3 H 0.688 No 0.59 0 1.5013 90 0.5 50 3 H 1 No 0.51 0 1.5014 90 0.5 50 5 H 1 No 0.54 0 1.5015a 120 0.5 50 3 H 0.688 No 0.56 5 3.5016 90 0.5 50 3 H 0.688 No 0.59 5 1.5017 90 0.5 50 3 H 1 No 0.51 5 1.5018a 120 0.5 50 3 H 0.688 No 0.56 10 3.7019 120 0.5 50 5 H 0.688 No 0.37 10 3.3020 120 0.5 50 3 H 1 No 0.55 10 3.70



Page 7 of 11

at University C



ownloaded from



and pitch(15, 34–36). The Catphanw 600 CT QAphantom has been employed by previous researchersto analyse the influence of exposure parameters onimage quality(22, 24, 25, 37); however, none of thesestudies included paediatric protocols.


CNR is considered a good method of evaluating CTimages(14, 20, 37) as noise is the primary limitingparameter of CT image quality. According to imageanalysis manufacturer, clinical studies comparingimage quality levels show that trained observers

(subjective evaluation) and RIT’s Radia Diagnosticw

software (objective evaluation) render similarresults(33), adding confidence to the methodologyused here. Radia Diagnosticw software proved to be agood method for protocols selection based on CNR.

Phase 3: optimisation with age-specific APs

The CT protocols currently employed were found tobe categorised by children’s age; however, age subsetsused varied across the three centres and these were notaligned to international paediatric radiography guide-lines recommendations(38). Authors recommend theformulation of age-categorised protocols(39–41), as

Table 4. Percentage of dose reduction (CTDIvol) and ROI’s mean and percentage standard deviation for head CT protocols,per paediatric APs, analysed by individual centres, in comparison with currently employed protocols.

AP Centre A Centre B Centre C

Protocolnumber

SD mean % CTDIvol % SD N8 SDmean % CTDIvol % SD N8 SDmean % CTDIvol % SD

0 1a 3.85 NA NA 1a 24.31 NA NA 1a 12.89 NA NA2 3.47 213 210 2 24.06 31 21 2 3.54 254 2733 3.39 239 212 3 24.04 82 21 3 2.74 217 2794 2.96 265 3 4 26.56 82 9 4 3.82 264 270

5 5a 3.51 NA NA 5a 5.67 NA NA 5 3.42 214 286 4.51 239 29 6 6.41 23 13 6a 2.68 NA NA7 4.72 249 35 7 5.90 225 4 7 3.44 217 29

8 6.20 231 9 8 4.85 261 4410 8a 5.01 NA NA 9 4.64 NA NA 9 3.55 NA NA

9 6.00 240 20 10 5.20 231 12 10 3.55 219 5810 6.77 250 35


Table 5. Percentage of dose reduction (CTDIvol) and ROI’s mean and percentage standard deviation for chest CT protocols,per paediatric APs, analysed by individual centres, in comparison with currently employed protocols.

AP Centre A Centre B Centre C

Protocolnumber

SD mean % CTDIvol % SD N8 SDmean % CTDIvol % SD N8 SDmean % CTDIvol % SD

0 11a 12.77 NA NA 11a 10.68 NA NA 11a 10.61 NA NA12 12.39 220 23 12 10.08 83 26 12 14.30 257 3513 12.53 216 22 13 12.94 251 21 13 14.41 257 36

14 10.91 226 2 14 12.22 257 155 14a 7.95 NA NA 15a 5.05 NA NA 15a 20.58 NA NA

15 8.62 220 8 16 5.65 28 12 16 18.07 257 21216 7.54 0 25 17 7.78 248 54 17 18.82 257 29

18 5.79 243 1510 17 14.18 NA NA 19a 13.87 NA NA 18a 21.62 NA NA

18 13.22 13 27 20 11.26 127 219 19 14.49 211 23319 13.45 13 25 21 13.56 0 22 20 19.03 0 212

22 14.17 42 2


J. SANTOS ETAL.

Page 8 of 11

at University C



ownloaded from



re-iterated in the paediatric European guidelines forradiography(38). The need for age-related protocols arejustified, otherwise paediatric CT examinations willresult in some patients receiving higher than optimalradiation dose values(42, 43). A non-standardised andlimited approach to age categorisation was identified inthe three paediatric centres. These protocols weretested on the Catphanw 600 and APs and repeatedusing the age categorisations as defined by Europeanguidelines on quality criteria for diagnostic radiograph-ic images in paediatrics(38) and as found in Europeanpaediatric CT DRL’s studies(26, 27). The majority of thecurrently employed dose values were higher than theEuropean CT DRL’s studies(27–29, 44) and the opti-misation tests allow similar or lower CT dose values.


Anthropomorphic phantoms have been employed inseveral previous studies that were aimed at optimisingCT practice, and in addition, OsiriX software hasbeen incorporated to evaluate image analyses usingmean pixel values and the standard deviation ofpixels in an ROI(36, 45–48). This method was appliedin this work in addition to the inclusion of age-specif-ic categorisation and more realistic tissue-equivalentfor paediatric CT dose studies(26, 27, 38). In image pro-cessing low standard deviation values in a homogenousarea represents a low level of noise, normally associatedwith higher dose values applied during imaging(49).

For chest CT examinations, the majority of thestandard deviation differences were found to be lowerwith the new parameters than with the currentlyemployed parameters (Table 5); these results corres-pond to reduced image noise levels following optimisa-tion. It was noted that as dose reductions rose to 50 %,increases in standard deviation values were recorded.

Similar findings occurred across the three paediatriccentres for the newborn head CTexaminations (Table 4).For 5- and 10-y-old phantom head CTexaminations, thestandard deviation percentage varied between 4 and44% and 12 to 58%, respectively. These findings demon-strated that the currently employed head CT protocolwas more suitable for imaging 10-y-old patients than for5-y-olds, suggesting substantial potential to optimiseimaging parameters for 5-y-old patients (mean dose de-crease of 20 %), with a reduced impact on image noisewhen comparedwith findings of 10-y-olds.

Global discussion

This study identified that minor exposure parametermanipulations, involving tube current–time productand tube voltage reduction (,60 mAs and 30 kV forhead CTexaminations; ,30 mAs and 30 kV for chestCT examinations), resulted in paediatric CT meandose reductions of up to 26 %. The use of reduced

tube voltages for head CTexaminations resulted in in-ferior but not significantly different CNRvalues (meanvariation + 0.2), and the CTDIvol levels decreased by31 % across the centres, the least variation in CNRvalues being recorded for smaller patients. The ma-nipulation of tube current proved of increased benefitfor the head CTexamination as the currently employedhigh values allowed manipulation to aid optimisationfor all three centres. For chest CT, the manipulation oftube voltage resulted in a mean variation of 0.15 inCNR, across the centres. The CTDIvol decreased by23 %, and the pre-optimisation tube current was lessvaried across the three centres, than demonstrated forhead examinations. The exposure parameter manipula-tions were undertaken with consideration of both theequipment design and patient size, allowing a reduc-tion of CT dose levels with no significant impact inimage quality.

Discussion with relevant clinical parties followingreview of the optimisation process and experimentaltesting has supported the introduction of age-cate-gorised protocols in alignment with European guide-lines(38) across the three centres, and local review of theprotocols was enacted. One centre opted for a softwareupgrade to define new protocols with the most recenttechnological potentialities for CT optimisation, astube current and tube voltage modulation, this wasinitiated following consultation involving the researchgroup and the clinical management team.

LIMITATIONS

Due to the variation in CTequipment across the threecentres, it was not appropriate to test matching proto-cols across centres. Further investigation involving anincreased number of centres with comparable equip-ment may facilitate the potential to investigate agreater range of protocol options currently being clin-ically applied on scanner models.

The experimental work involved phantom imagesand the use of an objective image noise measures. ForCatphanw 600 images, the CNR was obtained withthe RIT’s Radia Diagnosticw software, and for APimages, the image noise of homogenous ROIs wasmeasured with OsiriXw Imaging software; however, itwas not possible to convert these noise measurementsto Hounsfield Units. The incorporation of optimisedclinical images for image quality review would alsoaid in demonstrating the impact of noise levels uponradiology image interpretation.

CONCLUSIONS

Catphanw 600 phantom, CNR analysis, age-categorised APs and image noise analysis wereemployed with the aim of identifying optimisationstrategies for head and chest paediatric CT examina-tions. The manipulation of tube current–time product,


Page 9 of 11

at University C



ownloaded from



tube voltage, pitch, slice thickness and acquisitionmode facilitated mean percentage dose reductions of36, 25 and 32 % for head CT examinations and 9, 29and 40 % for chest CT examinations, respectively, forCentres A, B and C. Paediatric CT dose reductionpotential was identified with minimal impact on imagenoise.

In line with the European recommendations, thisstudy included four age categories and recommendsthis for use in clinical practice and for considerationwhen developing imaging scanning protocols. Thefindings of this work have demonstrated the potentialto lower tube voltage and current for newborns. Themanipulation of scanning parameters should be care-fully considered on individual bases with regard topatient age/size as indicated by this research.

The optimisation experiments performed in theclinical centres raised awareness locally with regard toCT protocol selection. The experimental optimisationfindings are currently being used to aid protocol reviewlocally in all three centres. This research group is con-tinuing work in the centre that upgraded its CT soft-ware as a direct impact of the findings of this research.Comparisons of radiation dose and image qualitylevels within patient images, pre- and post-software up-grading, using anatomical criteria is proceeding.

It is recommended that further communicationbetween specialised Paediatric centres regarding imagingprotocols and methods of age categorisation couldfurther promote a decrease in paediatric CTexposure.

ACKNOWLEDGEMENTS

The authors thank the Head of Radiology Departmentsof the paediatric centres and all radiographers whocontributed to this study.

REFERENCES

1. Brenner, D. J. and Hall, E. J. Computed tomography—anincreasing source of radiation exposure. N. Engl. J. Med.357(22), 2277–2284 (2007).

2. Sun, Z., Ng, K. H. and Vijayananthan, A. Is utilisationof computed tomography justified in clinical practice?Part I: application in the emergency department.Singapore Med. J. 51(3), 200–206 (2010).

3. Shah, N. B. and Platt, S. L. ALARA: is there a cause foralarm? Reducing radiation risks from computed tomog-raphy scanning in children. Curr. Opin. Pediatr. 20(3),243–247 (2008).

4. Pearce, M. S., Salotti, J.a., Little, M. P., McHugh, K.,Lee, C., Kim, K. P. et al. Radiation exposure from CTscans in childhood and subsequent risk of leukaemia andbrain tumours: a retrospective cohort study. Lancet380(9840), 499–505 (2012).

5. Brenner, D. J. and Hall, E. J. Cancer risks from CTScans?: now we have data, what next? Radiology 265(2),330–331 (2012).

6. American College of Radiology. ACR–SPR practiceguideline for performance of pediatric computed

tomography. (2008) p. 1–13. Available on: http://www.acr.org/Quality-Safety/Standards-Guidelines/Practice-Guidelines-by-Modality/CT

7. Nigrovic, L. E., Schunk, J. E., Foerster, A., Cooper, A.,Miskin, M., Atabaki, S. M. et al. The effect of observa-tion on cranial computed tomography utilization forchildren after blunt head trauma. Pediatrics 127(6),1067–1073 (2011).

8. Singh, S., Kalra, M. K., Thrall, J. H. and Mahesh, M.Pointers for optimizing radiation dose in pediatric CTprotocols. J. Am. Coll. Radiol. 9(1), 77–79 (2012).

9. IAEA. (International Atomic Energy Agency) Safetyreports series no71 radiation protection in paediatric radi-ology—IAEA safety standards and related publications(2012). Available on: http://www-ns.iaea.org/standards/

10. Nievelstein, R. A. J., Dam, I. M. and Molen, A. J.Multidetector CT in children: current concepts and dosereduction strategies. Pediatr. Radiol. 40(8), 1324–1344(2010).

11. Mccollough, C. H., Primak, A. N., Braun, N., Kofler,J. M., Yu, L. and Christner, J. Strategies for reducing radi-ation dose in CT. Natl. Inst. Health 47(1), 27–40 (2009).

12. Tement, S. T. A. White paper on radiation protection bythe European Society of Radiology. Insights Imaging2(4), 357–362 (2011).

13. Reid, J., Gamberoni, J., Dong, F. and Davros, W.Optimization of kVp and mAs for pediatric low-dose simu-lated abdominal CT: is it best to base parameter selectionon object circumference? AJR. Am. J. Roentgenol. 195(4),1015–1020 (2010).

14. Yu, L., Bruesewitz, M. R., Thomas, K. B., Fletcher, J. G.,Kofler, J. M. and Mccollough, C. H. Optimal tube poten-tial for radiation dose reduction in pediatric CT: principles,clinical implementations, and pitfalls. Radiographics (2011).Available on: http://radiographics.rsna.org/content/31/3/835.full.pdf

15. Strauss, K. J., Goske, M. J., Kaste, S. C., Bulas, D., Frush,D. P., Butler, P. et al. Image Gently: ten steps you can take tooptimize image quality and lower CT dose for pediatricpatients. Am. J. Roentgenol. 194(4), 868–873 (2010).

16. Schneebaum, N., Blau, H., Soferman, R., Mussaffi, H.,Ben-Sira, L., Schwarz, M. et al. Use and yield of chestcomputed tomography in the diagnostic evaluation ofpediatric lung disease. Pediatrics 124(2), 472–479 (2009).

17. Smith, A. B., Dillon, W. P., Lau, B. C., Gould, R.,Verdun, F. R., Lopez, E. B. et al. Radiation dose reduc-tion strategy for CT protocols: successful implementationin neuroradiology section. Radiology 247(2) (2008).Available on: http://radiology.rsna.org/content/247/2/499.short

18. Edyvean, S. CTequipment and performance issues: radi-ation protection 162. Radiat. Prot. Dosim. 153(2),190–196 (2012).

19. Arthurs, O. J., Yates, S. J., Set, P. A.K., Gibbons, D. A.and Dixon, A. K. Evaluation of image quality andradiation dose in adolescent thoracic imaging: 64-slice ispreferable to 16-slice multislice CT. Br. J. Radiol.82(974), 157–161 (2009).

20. Nagel, H. D. CT Parameters that influence the radiationdose. Radiat. Dose Adult Pediatr. Multidetector Comput.Tomogr. Chapter 3 (2007). ISBN 978-3-540-28888-6.

21. Mahesh, M. MDCT Physics—The Basics, TechnologyImage Quality and Radiation Dose. Wolter KuweLW&W (2009).

J. SANTOS ETAL.

Page 10 of 11

at University C



ownloaded from

http://www.acr.org/Quality-Safety/Standards-Guidelines/Practice-Guidelines-by-Modality/CT






http://www-ns.iaea.org/standards/





http://radiographics.rsna.org/content/31/3/835.full.pdf








http://radiology.rsna.org/content/247/2/499.short









22. Van de Casteele, E., Parizel, P. and Sijbers, J.Quantitative evaluation of ASiR image quality: an adap-tive statistical iterative reconstruction technique. In: Pelc,N. J., Nishikawa, R. M. and Whiting, B. R., Eds. (2012)Feb 23; 83133F-83133F-5. Available on: http://proceedings.spiedigitallibrary.org/proceeding.aspx?doi=10.1117/12.911283

23. Lee, K. H., Lee, J. M., Moon, S. K., Baek, J. H., Park,J. H., Flohr, T. G. et al. Attenuation-based automatictube voltage selection and tube current modulation fordose reduction at contrast-enhanced liver CT. Radiology265(2), 437–447 (2012).

24. Tang, K., Wang, L., Li, R., Lin, J., Zheng, X. and Cao,G. Effect of low tube voltage on image quality, radiationdose, and low-contrast detectability at abdominal multidetec-tor CT: phantom study. J. Biomed. Biotechnol. 130–169(2012). Available on: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3347747&tool=pmcentrez&rendertype=abstract

25. Zarb, F., Rainford, L. and McEntee, M. F. Developingoptimized CT scan protocols: phantom measurements ofimage quality. Radiography 17(2), 109–114 (2011).

26. Shrimpton, P. C. and Hillier, M. C. NRPB-W67 dosesfrom computed tomography (CT) examinations in theUK—2003 review (2005). Available on: http://www.biophysicssite.com/Documents/NRPB_W67/NRPB_W67.pdf

27. Medical Council. Diagnostic reference levels—the medicalcouncil regulates the medical profession in Ireland (2004).Available on: http://www.medicalcouncil.ie/About-Us/Legislation/Medical-Ionising-Radiation/Diagnostic-Referance-Levels-03-12-2004.pdf

28. Galanski, M., Nagel, H. D. and Stamm, G. PaediatricCTexposure in the Federal Republic of Germany—resultsof a Nation-wide Survey in 2005/06 (2006). Available on:http://www.mh-hannover.de/fileadmin/kliniken/diagnostische_radiologie/download/Report_German_Paed-CT-Survey_2005_06.pdf.

29. Keen, C. E. Swiss pediatric CT survey leads to nationaldose standards. AuntMinnie.com (2008) 41–43.Available on: http://www.auntminnie.com/index.aspx?sec=prtf&sub=def&pag=dis&itemId=82961&printpage=true&fsec=sup&fsub=cto

30. Kostova-Lefterova, D. and Vassileva, J. Survey of prac-tice in paediatric computed tomography. Radiat. Prot.Dosim. 147(1–2), 156–159 (2011).

31. Mccollough, C. H. Diagnostic Reference Levels. ImageWisely (2010) (November):1–6.

32. The Phantom Laboratory. Catphanw 500/600. ThePhantom Laboratory (2000).

33. RIT’s. Radiological Imaging Technology, Inc. Radiaautomated QC phantom analysis (2012). Available on:http://www.radimage.com

34. Goldman, L. W. Principles of CT and CT technology.J. Nucl. Med. Technol. 35(3), 115–128 (2007). Quiz129–30.

35. Goldman, L. W. Principles of CT: multislice CT. J. Nucl.Med. Technol. 36(2), 57–68 (2008). Quiz 75–6.

36. Fujii, K., Aoyama, T., Yamauchi-Kawaura, C.,Koyama, S., Yamauchi, M., Ko, S. et al. Radiation doseevaluation in 64-slice CT examinations with adult andpaediatric anthropomorphic phantoms. Br. J. Radiol.82(984), 1010–1018 (2009).

37. Chan, M., Yang, J., Song, Y., Burman, C., Chan, P andLi, S. Evaluation of imaging performance of major imageguidance systems. Biomed. Imaging Interv. J. 7(2), e11(2011).

38. EUR16261. European Commission, European guidelineson quality criteria for diagnostic radiographic images inpaediatrics (1996). Available on: ftp://ftp.cordis.europa.eu/pub/fp5-euratom/docs/eur16261.pdf.

39. Hernanz-Schulman, M. Pediatric CT and Image Gently.Image Wisely (2010). p. 2006–2008.

40. Sun, Z., Ng, K. H. and Sarji, S. A. Is utilisation of com-puted tomography justified in clinical practice? Part IV:applications of paediatric computed tomography.Singapore Med. J. 51(6), 457–463 (2010).

41. Singh, S., Kaira, M. K., Moore, M. A., Shailam, R.,Liu, B., Toth, T. L. et al. Dose reduction and compliancewith pediatric CT protocols adapted to patient size, clinic-al indication, and number of prior studies. Radiology252(1), 200–206 (2009).

42. Cohen, M. D. Pediatric CT radiation dose: how low canyou go?. Am. J. Roentgenol. 192(5), 1292–1303 (2009).

43. Hart, D., Wall, B. F., Shrimpton, P. C. and Dance, D. R.The establishment of reference doses in paediatric radi-ology as a function of patient size. Radiat. Prot. Dosim.90, 235–238 (2000).

44. Shrimpton, P. C., Hillier, M. C., Lewis, M. A. andDunn, M. National survey of doses from CT in the UK:2003. Br. J. Radiol. 79(948), 968–980 (2006).

45. Feng, S-T., Law, M. W.-M., Huang, B., Ng, S., Li, Z-P.,Meng, Q-F. et al. Radiation dose and cancer risk frompediatric CT examinations on 64-slice CT: a phantomstudy. Eur. J. Radiol. 76(2), e19–e23 (2010).

46. Soderberg, M., Gunnarsson, M. and Nilsson, M.Simulated dose reduction by adding artificial noise tomeasured raw data: a validation study. Radiat. Prot.Dosim. 139(1–3), 71–77 (2010).

47. Kim, G., Jung, H-J., Lee, H-J., Lee, J-S., Koo, S. andChang, S-H. Accuracy and reliability of length measure-ments on three-dimensional computed tomography usingopen-source OsiriX software. J. Digit. Imaging 25(4),486–491 (2012).

48. Dello, S. A.W. G., Stoot, J. H.M. B., van Stiphout,R. S.a., Bloemen, J. G., Wigmore, S. J., Dejong, C. H.C.et al. Prospective volumetric assessment of the liver on apersonal computer by nonradiologists prior to partialhepatectomy. World J. Surg. 35(2), 386–392 (2011).

49. Mccollough, C. H., Bruesewitz, M. R. and Kofler, J. M.CT dose reduction and dose management tools?: overviewof available options. Radiographics 26(2), 503–513 (2006).


Page 11 of 11

at University C



ownloaded from

http://proceedings.spiedigitallibrary.org/proceeding.aspx?doi=10.1117/12.911283










http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3347747&tool=pmcentrez&rendertype=abstract









http://www.biophysicssite.com/Documents/NRPB_W67/NRPB_W67.pdf








http://www.medicalcouncil.ie/About-Us/Legislation/Medical-Ionising-Radiation/Diagnostic-Referance-Levels-03-12-2004.pdf








http://www.mh-hannover.de/fileadmin/kliniken/diagnostische_radiologie/download/Report_German_Paed-CT-Survey_2005_06.pdf








http://www.auntminnie.com/index.aspx?sec=prtf&sub=def&pag=dis&itemId=82961&printpage=true&fsec=sup&fsub=cto








http://www.radimage.com





ftp://ftp.cordis.europa.eu/pub/fp5-euratom/docs/eur16261.pdf








