A Critical Comparison of Three Full Field Digital ...

1 Med J Malaysia Vol 65 No 2 June 2010119A Critical Comparison of Three Full Field Digital mammography Systems using Figure of MeritA Critical Comparison of Three Full Field DigitalMammography Systems using Figure of MeritK C Kanaga BSC(APPSc), PG DIP(Edu), MSC(Med Physics); H H Yap; S E Laila; T Sulaiman; M Zaharah; A A ShantiniDiagnostic Imaging and Radiotherapy Programme, Faculty of Allied Health Sciences, National University of Malaysia, Jalan Raja MudaAbd Aziz, 50300 Kuala Lumpur, Women s Cancer Detection and Breast Clinic, Ground Floor 66, Jalan Raja Muda Abdul Aziz, 50300 KualaLumpur, Selayang Hospital, Lebuh Raya Selayang-Kepong, 68100 Batu Caves, Selangor, Diagnostic Imaging Department, Kuala LumpurGeneral Hospital, Jalan Pahang, 50586 Kuala LumpurORIGINAL ARTICLESUMMARYFull Field Digital mammography (FFDM)

2 Has beenprogressively introduced in medical centers in recent , it is questionable which exposure parametersare suitable in order to reduce the glandular breast dosesas they are related to induced carcinogenesis. The goal ofthis study was to compare the average glandular doses(AGD) and image quality of Three FFDM systems namelySiemens Mammomat NovationDR, Hologic Lorad Selenia andGeneral Electric Senographe Essential using a Figure ofMerit. A Computerized Imaging Reference Systems (CIRS)tissue equivalent breast phantom which consists ofphototimer compensation plate with different thicknessand glandularity was exposed in fully automatic exposurecontrol mode in the cranio-caudal projection similar toclinical settings.

3 Thermoluminescent dosimeter 100H (TLD-100H) was used to measure the entrance surface air kerma(ESAK), the AGD was calculated using European protocolwhilst the image quality was assessed quantitatively bymeasuring the contrast to noise ratio (CNR) value. Theobtained values were used to calculate the Figure of Merit(FOM) to analyze the effectiveness of the system . RepeatedMeasures ANOVA analysis showed that there is a significantdifference (p< ) in the mean value of AGD and CNRbetween the Three FFDM systems. Hologic Lorad Seleniasystem contrbuted the highest AGD value while GeneralElectric Senographe Essential had the highest CNR and FOMvalue.

4 In conclusion, this study may provide an objectivecriterion during the selection of a mammography unit byusing the figure of merit for screening or :full Field Digital mammography , average glandular dose,contrast-to noise ratio, figure of is the gold standard used for the earlydetection of breast cancer although there are complementarymodalities such as breast-self and clinical examination sinceit is able to detect abnormalities that are not palpable. Theintroduction of Digital mammography has been very slowcompared to other imaging modalities due to the highdemands on image quality and low dose in is generally assumed that the glandular tissue of the breastis most vulnerable to the induction of cancer by ionizationradiation2.

5 As stated in the European Protocol on Dosimetryin mammography , the average glandular dose is the quantityof radiation risk related to induced carcinogenesis3. Periodicalquality control of the system is essential for obtaining highquality mammograms to detect any lesions ormicrocalcifications that suggest potentially optimization of the exposure parameters withrespect to dose and image quality using the automaticexposure control (AEC) has become one of the current issuesin Digital mammography as the AEC is one the keycomponents of full Field Digital mammography (FFDM).

6 For qualitative assessment of image quality, several types ofparameters and scoring systems using a breast phantomhave been employed4,5,6. However, variance in the evaluationof the images may exist depending on the evaluationprocedures or viewing conditions, even when the phantomimages are viewed by the same person7. Besides, it is generallyacknowledged that image quality can be assessedquantitatively by contrast-to-noise ratio (CNR), howeverCNR as commonly defined for mammography equipments8is not an absolute quantity, but its value is a range and ismanufacturer and system dependent.

7 Several papers werepublished showing that square root of CNR divided byaverage glandular dose (AGD), termed the Figure of Merit(FOM), is a good parameter for AEC optimization of digitalradiography (DR) systems9,10,11,12, goal of this study was to compare the AGD and imagequality of Three FFDM systems, namely Siemens MammomatNovationDR, Hologic Lorad Selenia and General ElectricSenographe Essential by using AND METHODSS everal Digital images were acquired using the Three FFDM systems: Siemens Mammomat NovationDR; the HologicLorad Selenia and the General Electric Senographe Essentiawith a set of phototimer compensation plate made of breastequivalent material (Computerized Imaging ReferenceSystems (CIRS), Inc.)

8 , Norfolk, Virginia). Fifteen differentphantoms were assembled and imaged, simulating breastsof five different thickness (4cm, , 5cm, and6cm) for Three different fibroglandular/adipose tissue ratios(70/30, 50/50 and 30/70). CIRS resin material mimics thephoton attenuation coefficients of a range of breast average elemental composition of the human breastThis article was accepted: 29 June 2010 Corresponding Author: Kanaga Chelliah, Diagnostic Imaging and Radiotherapy Porgramme, Allied Health Sciences Faculty, University KebangsaanMalaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur Email.

9 J Malaysia Vol 65 No 2 June 2010 Original Article120being mimicked is based on the individual elementalcomposition of adipose and glandular tissue as reported byHammerstein et al14. In each phantom stack assembly, thesignal block contains spec groups ( to indiameter) that simulate microcalcification, nylon fibers ( ) that simulate fiber and hemispheric masses( to in thickness) that simulate masses. Testobjects within the phantom range in sizes from those thatshould be visible on any system to objects that would bedifficult to resolve on even the best mammographic systemavailable of mammograms was done in the cranio-caudal projection in fully automatic exposure control(AEC) mode with the phantoms positioned at the chestwall edge of the receptor, centered left to right.

10 The sourceto image distance was 65cm, the focal spot size was broadand the compression force was 10 N. Thermoluminescentdosimeter 100-H (TLD-100H) was placed on top of thephantom to measure the entrance surface air kerma(ESAK). Half value layer (HVL) for each exposure wasmeasured using an ionization chamber (Model 9095 Radcal Corporation) with the compression paddleremoved. The AGD was calculated for each combinationparameter of those settings (anode/filter combination,kVp and mAs values) chosen by, the AEC by applyingpublished conversion factors to the ESAK previouslymeasured using the European = ESAK.