Sultan Qaboos University Med J, May 2015, Vol. 15, Iss. 2, pp. e292–296, Epub. Submitted 11 Jun 14 Revision Req. 21 Oct 14; Revision Recd. 10 Nov 14 Accepted 7 Dec 14 1Diagnostic Imaging & Radiotherapy Programme, Faculty of Health Sciences, University Kebangsaan Malaysia, Kuala Lumpur, Malaysia; 2KPJ Healthcare University College, Nilai, Malaysia *Corresponding Author e-mail: izdiharkamal@yahoo.com تقدير اجلرعة الغدية املتوسطة بواسطة طرز ذاتية للتعرض لألشعة السينية يف الرتكيب املقطعي الرقمي للثدي اإزدهار كمال، كاناجا ك �سيال، ونوال م�سطفى abstract: Objectives: The aim of this research was to examine the average glandular dose (AGD) of radiation among different breast compositions of glandular and adipose tissue with auto-modes of exposure factor selection in digital breast tomosynthesis. Methods: This experimental study was carried out in the National Cancer Society, Kuala Lumpur, Malaysia, between February 2012 and February 2013 using a tomosynthesis digital mammography X-ray machine. The entrance surface air kerma and the half-value layer were determined using a 100H thermoluminescent dosimeter on 50% glandular and 50% adipose tissue (50/50) and 20% glandular and 80% adipose tissue (20/80) commercially available breast phantoms (Computerized Imaging Reference Systems, Inc., Norfolk, Virginia, USA) with auto-time, auto-filter and auto-kilovolt modes. Results: The lowest AGD for the 20/80 phantom with auto-time was 2.28 milliGray (mGy) for two dimension (2D) and 2.48 mGy for three dimensional (3D) images. The lowest AGD for the 50/50 phantom with auto-time was 0.97 mGy for 2D and 1.0 mGy for 3D. Conclusion: The AGD values for both phantoms were lower against a high kilovolt peak and the use of auto-filter mode was more practical for quick acquisition while limiting the probability of operator error. Keywords: Thermoluminescent Dosimetry; Breast; Mammography; Radiation Dosage. امللخ�ص: الهدف: يهدف هذا البحث لدرا�سة اجلرعة الغدية املتو�سطة )ج غ م( يف خمتلف مكونات الثدي بوا�سطة اختيار عامل الطرز الذاتية للتعر�ض. واأجريت هذه الدرا�سة بجمعية ال�رشطان القومية يف كواال المبور مباليزيا بني فرباير 2012م وفرباير 2013م، وذلك با�ستخدام الن�سف قيمة وطبقة الهواء �سطح مدخل املادة( يف املنطلقة الطاقة )وحدة كريما حتديد ومت الرقمي. اال�سعاعي الثدي ت�سوير جهاز با�ستعمال املقيا�ض اللمعي احلراري للجرعات ال�سعاعية )H 100( يف ثدي �سبحي )50/50 و20/80( متوفر جتاريا ب�رشكة اأنظمة الت�سوير املحو�سب-نورفولك-فريجينا-الواليات املتحدة االأمريكية، والذي يعمل بطراز ذاتي بالن�سبة للتوقيت والرت�سيح ونظام K-V. وتو�سلت الدرا�سة اإىل اأن اأقل ج غ م لل�سبح 20/80 ذاتي التوقيت كان 2.28 ميلي قري لبعدين و2.48 ميلي قري لثالثة اأبعاد. بينما كان اأقل ج غ م لل�سبح 50/50 ذاتي التوقيت هو 0.97 ميلي قري لبعدين و1.0 ميلي قري لثالثة اأبعاد. ووجد اأن قيم ج غ م لل�سبحني كانت اأقل على خلفية ذروة كيلو فولت، واأن ا�ستخدام طراز املر�سح الذاتي كان اأكرث عملية يف عملية االكت�ساب، واأقل احتماال حلدوث خطاأ من قبل م�سغل اجلهاز. مفتاح الكلمات: مقيا�ض اجلرعات اال�سعاعية اللمعي احلراري؛ ثدي؛ ت�سوير الثدي ال�سعاعي؛ اجلرعة اال�سعاعية. Estimates of Average Glandular Dose with Auto-modes of X-ray Exposures in Digital Breast Tomosynthesis *Izdihar Kamal,1 Kanaga K. Chelliah,1 Nawal Mustafa2 It is generally assumed that the gland-ular tissue of the breast is most vulnerable to the induction of cancer by ionisation radiation.1 In order to calculate the average glandular dose (AGD) of radiation in different breast compositions, parameters such as the half-value layer (HVL) and entrance surface air kerma (ESAK) have to be measured.2 The average glandular tissue dose in mammography is generally determined from published tables with knowledge of the breast entrance skin exposure, X-ray tube target material, beam quality (HVL), breast thickness and breast composition.3 In addition, the beam spectral quality, scatter control with compression and grid, detector characteristics, image processing and radiation dosage must be taken into consideration in order to optimise the AGD.4 Clinical trials and scientific investigations have found that, in digital mammography, a tungsten X-ray tube with rhodium (Rh) and argentum (Ag) filters is optimal for measuring all breast thicknesses, and will allow for important dosage reductions by approximately 30% while still maintaining excellent image quality.5 In February 2011, the Food and Drug Adminis- tration (FDA) approved tomosynthesis machines as safe for usage in the screening and diagnosis of breast cancer, as their two dimensional (2D) and three dimensional (3D) tomosynthesis images can reveal the online brief communication Izdihar Kamal, Kanaga K. Chelliah and Nawal Mustafa Online Brief Communication | e293 internal breast architecture without distortion from tissue overlapping.6,7 In current clinical practice, auto- filter is often used rather than the auto-time and auto- kilovolt (kV) modes as the latter require manual set- up by the operator.8 However, there has been a need for further study to determine the best practice for screening mammography. The aim of this study was to assess the AGD of radiation in two different breast compositions (50% glandular and 50% adipose [50/50] and 20% glandular and 80% adipose [20/80]) with auto-modes (auto-time, auto-filter and auto-kV) of exposure factor selection. Table 1: Parameters used for the assessment of the 6 cm thick 20% glandular and 80% adipose tissue (20/80) breast phantom Mode of exposure HVL Projection kVp mAs Target/ filter ESAK in mGy AGD ± SD in mGy Auto-time† 0.550 CC/2D 28 163 W/Ag 5.62 2.62 ± 0.10 0.580 CC/2D 30 114 W/Ag 5.48 2.60 ± 0.19 0.610 CC/2D 32 78 W/Ag 4.76 2.30 ± 0.06 0.520 CC/2D 28 228 W/Rh 9.16 2.82 ± 0.01 0.530 CC/2D 30 199 W/Rh 7.94 2.67 ± 0.14 0.540 CC/2D 32 142 W/Rh 6.83 2.28 ± 0.01* 0.580 CC/3D 32 59 W/Al 6.93 2.48 ± 0.11* Auto-filter 0.535 CC/2D 31 171 W/Rh 7.01 2.44 ± 0.13 0.590 CC/3D 33 62 W/Al 7.47 2.5 ± 0.16 Auto-kV 0.535 CC/2D 31 150 W/Rh 7.12 2.31 ± 0.40 0.580 CC/3D 32 62 W/Al 7.34 2.61 ± 0.12 HVL = half-value layer; kVp = kilovolt peak; mAs = milliampere per second; ESAK = entrance surface air kerma; mGy = milliGray; AGD = average glandular dose; SD = standard deviation; CC = craniocaudal; 2D = two dimensional; W = tungsten; Ag = argentum; Rh = rhodium; 3D = three dimensional; Al = aluminium; auto-kV = auto-kilovolt. *Lowest AGD values for 2D and 3D images, respectively. †Only 32 kVp available for the auto-time 3D image. Table 2: Parameters used for the assessment of the 4 cm thick 50% glandular and 50% adipose tissue (50/50) breast phantom Mode of exposure HVL Projection kVp mAs Target/ filter ESAK in mGy AGD ± SD in mGy Auto-time 0.550 CC/2D 28 55 W/Ag 2.26 1.20 ± 0.20 0.580 CC/2D 30 65 W/Ag 1.88 1.12 ± 0.04 0.610 CC/2D 32 33 W/Ag 1.91 1.01 ± 0.15 0.520 CC/2D 28 96 W/Rh 2.98 1.16 ± 0.03 0.530 CC/2D 30 45 W/Rh 2.84 1.10 ± 0.03 0.540 CC/2D 32 33 W/Rh 2.31 0.97 ± 0.32* 0.510 CC/3D 28 60 W/Al 4.37 1.80 ± 0.136 0.550 CC/3D 30 38 W/Al 3.32 1.28 ± 0.15 0.580 CC/3D 32 27 W/Al 2.73 1.00 ± 0.01* Auto-filter 0.520 CC/2D 28 91 W/Rh 3.16 1.17 ± 0.01 0.522 CC/3D 29 58 W/Al 3.74 1.33 ± 0.32 Auto-kV 0.520 CC/2D 28 46 W/Rh 3.21 1.32 ± 0.14 0.530 CC/3D 29 48 W/Al 3.76 1.68 ± 0.17 HVL = half-value layer; kVp = kilovolt peak; mAs = milliampere per second; ESAK = entrance surface air kerma; mGy = milliGray; AGD = average glandular dose; SD = standard deviation; CC = craniocaudal; 2D = two dimensional; W = tungsten; Ag = argentum; Rh = rhodium; 3D = three dimensional; Al = aluminium; auto-kV = auto-kilovolt. *Lowest AGD values for 2D and 3D images, respectively. Estimates of Average Glandular Dose with Auto-modes of X-ray Exposures in Digital Breast Tomosynthesis e294 | SQU Medical Journal, May 2015, Volume 15, Issue 2 the original tabulation due to the use of a different X-ray spectrum. Furthermore, in order to estimate the AGD for breast tomosynthesis, Dance et al. introduced T-factors for the calculation of breast dose from a single projection and T-factors for a complete exposure series from -7.5–7.5 degrees for 15 projection in less than four seconds of scanning time.13 The T-factor value for this study was between 0.93 and 1.0. The statistical analysis of the data was performed with Statistical Package for the Social Sciences (SPSS) Version 18.0 (IBM Corp., Chicago, Illinois, USA). The statistical significance of the overall differences between auto-modes was analysed using the one-way analysis of variance. Results Table 1 and Table 2 show the parameters used in the experiment. The results were divided according to composition and thickness (6 cm 20/80 and 4 cm 50/50). AGD values obtained from the experiment increased with the thickness of the phantom. The lowest AGD for the fatty breast tissue (20/80) on auto-time was 2.28 mGy with tungsten (W)/Rh anode/filter for 2D images and 2.48 mGy for 3D images [Table 1]. Meanwhile, for normal breast tissue (50/50), the lowest AGD for the 2D images on auto-time was 0.97 mGy with W/Rh anode/ filter and 1.0 mGy for 3D images [Table 2]. With auto-time mode, the dosage was minimal compared with the two other modes (auto-filter and auto-kV). However for the 3D image, a tube voltage of 28 kVp and 30 kVp was insufficient to penetrate the 20/80 phantom due to the existence of a backup timer, which automatically stops the exposure when the time ends. Therefore only the 32 kVp was able to penetrate the phantom.14 Analysis of the 50/50 phantom was statistically significant, indicating that the AGD value was influenced by different types of auto-modes (auto- filter, auto-time and auto-kV) (F [2, 15] = 6.918; P <0.05). However, for the phantom 20/80, the analysis was not statistically significant, indicating that the AGD values were not influenced by the different types of auto-modes (F [2, 15] = 0.495; P >0.05). Discussion For the 20/80 breast phantom, the AGD with auto- time produced a lower dose with the Rh filter, correlating with the manufacturer’s suggestion to use the W/Rh combination for a breast phantom thickness between 4 and 7 cm.15 The Ag filter should be considered for use on patients with fatty tissue Methods Between February 2012 and February 2013 a phantom study was conducted using a tomosynthesis digital mammography X-ray machine, the Hologic® Selenia® AWS 5000 (Hologic, Bedford, Massachusetts, USA). Two types of commercially available breast phantoms (Computerized Imaging Reference Systems, Inc., Norfolk, Virginia, USA) were used. The first tissue equivalent breast phantom was composed of 50/50 tissue with a thickness of 4 cm, while the second phantom was composed of 20/80 tissue with a thickness of 6 cm. A 100H thermoluminescent dosimeter (TLD) was used to obtain the ESAK value. The experiment was repeated three times for each parameter in order to reduce any errors in measurement. The first experiment was conducted using the auto-filter mode to enable the automatic choice of the target, filter, kV peak (kVp) and milliampere per second (mAs). For auto-time, the system automatically selected the mAs but the operator manually selected the filter and kVp. For 2D images, two types of filter were used, Ag and Rh, while for 3D images only one filter was used, aluminium (Al). Each filter was used with 28, 30 and 32 kVp, respectively. Finally, for auto-kV, the system automatically selected the kVp and mAs while the operator selected the filter. A sachet containing three TLDs was placed on the surface of the phantom, with the centre of the sachet 40 mm from the chest wall edge and centred with regards to the lateral direction. A further sachet of unexposed TLDs was retained for background reading.9 TLDs were calibrated in terms of air kerma at mammographic energies, using all anode/filter combinations as used in normal practice. In this study, an ionisation chamber was used to calibrate the TLDs. The ionisation chamber was placed at the same effective point of the measurements and exposed to the same dose as the TLDs. Thus, the energy responses were consistent because they were calibrated at the same energies where the measurements were taken.10 In order to calculate the AGD, the HVL for each kVp was also measured. The experimentally estimated AGDs were compared with theoretically calculated values based on methods indicated in earlier studies.11–13 According to methods proposed by Dance et al. the AGDs in the current study were calculated using this equation: AGD = K x G x C x S x T.11–13 In this equation, K is the ESAK value from the TLD reading; factor G is the conversion factor of ESAK to AGD; factor C corrects any differences in breast composition from 50% glandularity, and factor S corrects any differences from Izdihar Kamal, Kanaga K. Chelliah and Nawal Mustafa Online Brief Communication | e295 and a thickness greater than 6 cm for dose reduction purposes. It is recommended that, for fatty tissue, all factors such as kVp, filter and compression should be optimised in order to obtain the lowest dose possible without compromising the image quality.16 With the breast phantom 50/50, the pattern was similar as the lowest AGD was obtained using auto-time, followed by auto-filter and auto-kV. When the efficacy of different exposure modes are compared, the following factors should be considered: the radiation dose received by the patient, the image quality and the ease of use of the modes for the operator.17 In this study there was a significant dose change between the different modes of exposure, particularly for the 50/50 breast phantom. For 2D and 3D images, the 32 kVp with auto-time mode produced a low radiation dose, but when the kVp was high, the scatter radiation increased and therefore reduced the image contrast.18 Therefore, unnecessary changes to the kVp should be avoided, especially for breasts with reduced thicknesses. Furthermore, several research studies have shown that 28 kVp is the optimum kVp for the 50/50 breast phantom since it produces a high-quality image.4,19 In the current study, the auto- filter mode automatically calculated 28 kVp as the tube output. The results of this study indicate that the auto- filter mode is preferable for normal breast tissue (50/50) since this mode is automatic, fast and easy to operate. The 20/80 AGD value showed no significant difference between the exposure modes. The difference between the kVp values was minimal for the 2D images. The auto-time used 32 kVp, the auto-filter used 31 kVp and the auto-kV used 31 kVp, meaning that the difference was only 1 kVp. Therefore, there was no significant difference in the AGD values when the exposure modes used were different. This finding was also found to be consistent with the 3D imaging. Therefore the auto-mode did not have an obvious role in reducing the dose for fattier breast tissue. This is contrary to the findings of Myung et al., which found differences in the AGD values.17 However, their findings may have been due to the mammography system used, which could operate in three different modes of automatic optimisation of parameter: contrast (CNT), standard (STD) and dose modes, which vary between low dose and high image quality.17 The CNT mode emphasised higher contrast and, therefore, higher image quality. The dose mode focused on dose reduction with acceptable image quality. The STD mode was balanced between good contrast and dose reduction.17 The AGD, ESAK, and mAs were ordered as CNT mode had a higher AGD than STD mode and dose mode produced the lowest AGD. For the phantom 50/50, the total AGD value for the 2D and 3D combination mode was below the recommended dose limit of 3 mGy per projection. In order to ease the operator’s task, the auto-filter is the recommended mode for 50/50 breast types for both 2D and 3D images.20 On the other hand, the results obtained in the current study showed that the auto-kV mode produced the highest AGD and, therefore, was not advisable for use with this type of breast. Similarly, for the fatty type of breast tissue (20/80), auto-filter could be considered the mode of choice for both a 2D and 3D image as the kVp is high enough to penetrate thicker breast tissue and the mode is automatic. A limitation to this study was that the findings focused only on the dosage of radiation. Further research on auto-modes of exposure and image quality is therefore required. Conclusion This study demonstrated that the AGD values for 50/50 and 20/80 breast phantoms for both 2D and 3D imaging were lower with a high kVp while using auto- time mode. However, the use of auto-filter mode was found to be more practical as it was fast to use and reduced the risk of operator error. Alternatively, for 20/80 breast types, it would be more suitable to choose either auto-time or auto-filter modes interchangeably since the dose does not change. However, this can be risky if the operator has insufficient training or knowledge of utilising different modes. References 1. Dronkers DJ, Hendriks J, Holland R, Rosenbusch G, Eds. The Practice of Mammography: Pathology, technique, interpretation, adjunct modalities. 1st ed. Germany: Georg Thieme Verlag, 2001. Pp. 59–96. 2. Ng KH, Jamal N, DeWerd L. Global quality control perspective for the physical and technical aspects of screen-film mammography: Image quality and radiation dose. Radiat Prot Dosimetry 2006; 121:445–51. doi: 10.1093/rpd/ncl051. 3. Wu X, Barnes GT, Tucker DM. Spectral dependence of glandular tissue dose in screen-film mammography. Radiology 1991; 179:143–8. doi: 10.1148/radiology.179.1.2006265. 4. Baldelli P, Phelan N, Egan G. Investigation of the effect of anode/filter materials on the dose and image quality of a digital mammography system based on an amorphous selenium flat panel detector. Br J Radiol 2010; 83:290–295. doi: 10.1259/ bjr/60404532. 5. Andrew S, Biao C, Alan S. Minimizing Dose in Digital Mammography: Tungsten x-ray tubes with rhodium and silver filters optimize image quality. From: cdn.medicexchange.com/ images/whitepaper/Minimizing%20Dose%20in%20Digital%20 Mammography.pdf?1272615617 Accessed: Oct 2014. 6. United States Food and Drug Administration. Selenia Dimensions 3D System - P080003. From: www.fda.gov/MedicalDevices/Prod uctsandMedicalProcedures/DeviceApprovalsandClearances/ Recently-ApprovedDevices/ucm246400.htm Accessed: Oct 2014. Estimates of Average Glandular Dose with Auto-modes of X-ray Exposures in Digital Breast Tomosynthesis e296 | SQU Medical Journal, May 2015, Volume 15, Issue 2 15. Young KC, Oduko JM, Woolley L. Technical Evaluation Of The Hologic Selenia Full Field Digital Mammography System NHSBSP Equipment Report 0701 January 2007.. From: www. cancerscreening.nhs.uk/breastscreen/publications/nhsbsp- equipment-report-0701.pdf Accessed: Oct 2014. 16. UnaVita - Healthcare Sustainability. Digital Mammography and the Fatty Breast. From: www.una-vita.com/publications# Accessed: Oct 2014. 17. Ko MS, Kim HH, Cha JH, Shin HJ, Kim JH, Kim MJ. Dose reduction in automatic optimization parameter of full field digital mammography: Breast phantom study. J Breast Cancer 2013; 16:90–6. doi: 10.4048/jbc.2013.16.1.90. 18. Kanaga KC, Yap HH, Laila SE, Sulaiman T, Zaharah M, Shantini AA. A critical comparison of three full field digital mammography systems using figure of merit. Med J Malaysia 2010; 65:119–22. 19. Williams MB, Raghunathan P, More MJ, Seibert JA, Kwan A, Lo JY, et al. Optimization of exposure parameters in full field digital mammography. Med Phys 2008; 35:2414–23. 20. Izdihar K, Kanaga KC, Krishnapillai V, Tamanang S. Determination of tube output (kVp) and exposure mode for breast phantom of various thicknesses/glandularity for digital mammography. Malays J Med Sci 2015; 22:40–9. 7. Darambara DG. Digital Breast Tomosynthesis. Presented at London Quality Assurance Breast Cancer Screening Study Day, June 2011, London. 8. McCullagh JB, Baldelli P, Phelan N. Clinical dose performance of full field digital mammography in a breast screening programme. Br J Radiol 2011; 84:1027–33. doi: 10.1259/ bjr/83821596. 9. International Atomic Energy Agency. Dosimetry in Diagnostic Radiology: An international code of practice. From: www-pub. iaea.org/MTCD/publications/PDF/TRS457_web.pdf Accessed Oct 2014. 10. Zeidan M. Assessment of Mean Glandular Dose in Mammography. Master thesis. Christchurch, New Zealand: University of Canterbury, 2009. 11. Dance DR. Monte Carlo calculation of conversion factors for the estimation of mean glandular breast dose. Phys Med Biol 1990; 35:1211–9. doi: 10.1088/0031-9155/35/9/002. 12. Dance DR, Skinner CL, Young KC, Beckett JR, Kotre CJ. Additional factors for the estimation of mean glandular breast dose using the UK mammography dosimetry protocol. Phys Med Biol 2000; 45:3225–40. doi: 10.1088/0031-9155/45/11/308. 13. Dance DR, Young KC, van Engen RE. Estimation of mean glandular dose for breast tomosynthesis: Factors for use with the UK, European and IAEA breast dosimetry protocols. Phys Med Biol 2011; 56:453–71. doi: 10.1088/0031-9155/56/2/011. 14. Carlton RR, Adler AM. Principles of Radiographic Imaging: An art and a science. 4th ed. Independence, Kentucky, USA: Cengage Learning, 2006. Pp. 606–60714.