A study of mean glandular dose during diagnostic mammography in ...

The British Journal of Radiology, 76 (2003), 238–245 DOI: 10.1259/bjr/66428508

E

2003 The British Institute of Radiology

A study of mean glandular dose during diagnostic mammography in Malaysia and some of the factors affecting it 1

N JAMAL, MSc, 2K-H NG, PhD, DABMP, MIPEM and 3D MCLEAN, PhD

1

Malaysian Institute for Nuclear Technology Research (MINT), 43000 Kajang, Selangor, Malaysia, 2Department of Radiology, University of Malaya Medical Centre, 59100 Kuala Lumpur, Malaysia and 3School of Medical Radiation Sciences, University of Sydney, Lidcombe NSW, 1825 Australia

Abstract. The primary objective of this study was to determine the mean glandular dose (MGD) during diagnostic mammography in Malaysia. The secondary objective was to evaluate some of the factors affecting MGD. A survey of standard MGD was performed, based upon quality control records for the period October 1999 to August 2001. This covered 30 mammography units from 9 manufacturers. MGD was also measured for a series of patients attending mammography examinations at three other mammography units. MGD per film was estimated from recorded radiographic factors, the compressed breast thickness (CBT) and X-ray unit calibration data. MGD per woman was calculated by summing the MGDs for all films, and averaging it over both breasts. 300 women drawn equally from three major ethnic groups, namely Malay, Chinese and Indian, took part in the study. The difference of MGD per woman between ethnic groups was tested for significance using non-parametric Kruskal–Wallis and median tests. The factors affecting MGD per woman were tested for significance using a multivariate analysis of variance. The MGD for the phantom was 1.23 mGy (range 0.22– 2.39 mGy) while the mean patient based MGD per film was 1.54 mGy and 1.82 mGy for the craniocaudal and mediolateral oblique views, respectively. The mean MGD per woman was 3.37 mGy. It was also found that there is no significant difference in MGD per woman among the ethnic groups (p.0.05, Kruskal–Wallis test). However, on the multivariate test two factors, namely half value layer of the X-ray beam and (CBT), had a significant effect on MGD per woman (p,0.05). No significant relationships were seen between MGD per woman with respect to ethnicity, body mass index or age.

Malaysia is a health care level II country according to the definition by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) [1]. The Malaysian Ministry of Health has implemented a programme to encourage women to participate in secondary prevention or early detection of breast cancer. Under Malaysian public health, mammography is offered for free at government centres only for high-risk women. These are women who have had a previous breast cancer or suspicious lump in one breast, and women who are advised by the attending doctor to have mammography, such as those on hormone replacement therapy or women with a strong family history of cancer. Other women who wish to have mammography can attend clinics run by private agencies. At present it is generally assumed that the glandular tissue is the most vulnerable type of breast tissue [2]. Therefore, the suggestion that the mean glandular dose (MGD) is the most appropriate dosimetric quantity to predict the risk of radiation-induced carcinogenesis has been widely accepted [2, 3]. The two main methods for the assessment of MGD from mammography are the use of a standard breast phantom and patient-based measurement. Standard breast or phantom measurement is utilized to define MGD limits [4–6] and is well suited for quality control and inter-system comparison to ensure that all units are capable of achieving acceptable doses. Such measurements, however, do not indicate the actual dose received by the individual woman [7, 8]. Received 17 December 2001 and in final form 9 December 2002, accepted 6 January 2003.

238

The primary objective of this study was to determine the MGD during diagnostic mammography in Malaysia. The secondary objective was to evaluate some of the factors affecting MGD, namely half value layer (HVL) of the X-ray beam, compressed breast thickness (CBT), ethnic group, body mass index (BMI) and age. Such information is necessary in order to formulate recommendations to minimize radiation doses without compromizing image quality, and for the development of national reference doses as recommended by the International Atomic Energy Agency (IAEA) [9]. This information will, in turn, allow comparisons to be made with other parts of the world.

Method Standard breast method Under Malaysian legislation, each radiographic unit, including mammographic equipment has to undergo annual quality control and third party verification testing to get an annual ‘‘license to use’’. One of eight parameters required to be tested for a mammographic unit is the MGD, which must be 3 mGy or less with grid [5, 6]. The Radiation Measurements Inc. (RMI) phantom, model 156 (Radiation Measurements Inc., Middleton, WI) or equivalent, described as equivalent to 50% glandular tissue and 50% adipose tissue with CBT of 42 mm [6, 10], is required [6]. The method involves measuring the entrance surface exposure, without backscatter to the phantom and applying conversion factors determined by Wu et al [9]. The British Journal of Radiology, April 2003

Mean glandular dose during mammography

A survey of standard breast MGD was made based upon records kept at the Malaysian Institute for Nuclear Technology Research (MINT) during the period October 1999 to August 2001. This covered 30 units (2 new units and 28 already in use) from nine manufacturers: Planmed OY (n51); Toshiba (n510); Hitachi (n52); Bennett (n52); Picker (n52); GE Senographe (n54); Philips (n53); Siemens (n51); and Instrumentarium (n55). 25 units were from private clinics, whilst 5 were from government centres. The distribution of standard breast MGDs and their dependence on optical density (OD) was analysed.

Patient-based method Patient information and technique factors This study involved 316 women attending mammography examinations in three mammography units other than those surveyed, at two selected government centres in Malaysia. Characteristics and radiographic parameters for the three mammography units are given in Table 1. All units normally use 18624 cm2 format films, while only two women were exposed using 24630 cm2 format films. Each unit undergoes an extensive quality assurance procedure following the recommendation of the American College of Radiology (ACR) [6]. The mean OD was between 1.2 and 1.4 for all three units. Only craniocaudal (CC) and mediolateral oblique (MLO) views were included in this study. Magnification mammography was excluded. Patients’ identification with a particular ethnic group was based upon information given by each woman. For each woman, data on age, weight and height were also recorded. The BMI, which is a useful classification scheme for the size and shape of a woman, was derived from the ratio of weight/ height2 [11]. After excluding sets of data for 14 mastectomy and previous lumpectomy patients, because they do not represent normal body size, and two women from other ethnic groups, the present study was then limited to collecting 100 women each from the three major ethnic groups; Malay, Chinese, and Indian. CBT was measured to the nearest 0.5 mm using a ruler, at a distance of 4 cm from the chest wall, as the distance between the bottom of the compression plate and the table upon which the breast rested [12, 13]. Measurement of CBT for CC views was made at left and right sides of the compression plate, and the mean value was calculated. For MLO views, however, measurement was made at only one side of the compression plate, opposite to the woman’s arm. For each breast, the force applied was also recorded. At both centres radiographers were encouraged to achieve a firm compression in order to optimize radiographic image quality, which may account for some of the low CBT. In some cases, however, firm compression could not be applied owing to patient discomfort. Mean glandular dose and compressed breast thickness MGD per film was estimated from the recorded exposure factors, CBT and X-ray unit calibration data using the method described in the Institute of Physical Sciences in Medicine Report 59/2 [4]. The method used a table of g-factors that rely on the assumption that all breasts have a composition similar to the standard breast [7]. The g-factors were those calculated by Dance [14]. The British Journal of Radiology, April 2003

MGD per woman was calculated by summing the MGDs for all films, and then averaged over both breasts [7]. Descriptive statistics including mean, median and quartile values were calculated for CBT and MGD per film and per woman for each of the patient ethnic groups. The distributions of patient age, MGD per film, and CBT (for CC and MLO) were also analysed.

Factors affecting mean glandular dose per woman The Kruskal–Wallis and median tests were used to test the significance (p50.05) of ethnic origin in MGD per woman. Multivariate analysis of variance was also used to test the significance of the factors HVL, CBT, BMI, ethnic group and age, affecting MGD per woman. Graphical presentation as a box-whisker plot was used to display the distribution of MGD per woman against ethnic group, BMI and age.

Results Standard breast mean glandular dose The operating conditions of 30 mammography units surveyed are summarized in Table 2. These can be considered as representative of the general situation and constitute 50% of the total number of units in the country. The phantom MGD varies between units by a factor of approximately 10. 97% of surveyed units operated using molybdenum anodes and filters, with 28 kV the most common tube potential (62.1%) followed by 27 kV (20.7%), with the remainder of units at either 26 kV or 25 kV. The mean phantom MGD of 1.23 mGy is consistent with studies from other countries (Table 3). The distribution of standard breast MGD is shown in Figure 1. The mean standard OD of 1.28 is just outside the ACR recommended range of 1.4–1.8 [6]. The expected dependence of standard breast MGD on OD is shown in Figure 2.

Patient-based mean glandular dose Patient information and technique factors Details of patient information and technique factors for the study sample are shown in Table 4. Median age of the study sample is 51 years (range 31–87 years) and coincides with the age of menopause (median 51 years in the UK), during which significant changes in composition of the breast are known to occur [3]. Mean weight of the study sample is 61¡11 kg, which is 3 kg greater than the weight of the standard woman (58 kg) [15]. Mean height is 154¡8 cm, which is 6 cm shorter than the standard height of 160 cm [15]. 4.3% of the study sample was in the age range 30–39 years, increasing to 36.7%, 47.0%, 10.7% for each subsequent decade, respectively, with 1.3% of women 70-years-old or above (see Figure 3). Mean glandular dose and compressed breast thickness A histogram showing the percentage of films as a function of the MGD per film for CC and MLO views is shown in Figure 4. The distribution of CBT was symmetrical ranging from 5 mm to 75 mm and 5 mm to 85 mm for CC and MLO views, respectively, as shown in 239

240 Table 1. Typical characteristics and radiographic parameters for the three mammography units. Periods of study: Centre 1, June–August 2000; and Centre 2, June and August 2001, and May–July 2000

Focus–film distance (cm) Anode material(s) Filtration (mm material) Exposure techniques (kV:mAs)

Tube voltage Output (mGy/mAs) b HVL (mm Al) Standard breast MGD (mGy) Grid ratio Mode of operation a

Centre 2

GE Senographe DMR (unit 1)

GE Senographe DMR (unit 2)

Siemens Mammomat 2 (unit 3)

66 Mo and Rh Mo(30mm) and Rh(25mm) Mo:Mo (26–28:110–137) Mo:Rh (26–27:108–205) Rh:Rh (26–29:105–297) 25–31 (by 1 kV step) 2.0 0.37 1.22 5:1 Variable kV protocol (Dose mode), AEC 5 KONICA SRX-201 35 90 KODAK Min-R MIN2K (small) and MIN-RM (large) 127 (Malay531, Chinese558 and Indian538)

66 Mo and Rh Mo(30mm) and Rh(25mm) Mo:Mo (25–28,30:14–355) Mo:Rh (26–27:149–384)

60 Mo Mo(30mm) Mo:Mo (28:28–698)

25–31 (by 1 kV step) 1.9 0.35 1.44 5:1 Variable kV protocol (Contrast mode), AEC 4 AGFA CP 100 34 120 KODAK Min-R FUJI UM-MA HC 139 (Malay555, Chinese532 and Indian552)

28 1.2 0.38 1.35 4:1 Fixed kV protocol, AEC

Mo, molybdenum; Rh, Rhodium; AEC, automatic exposure control; MGD, mean glandular dose; CBT, compressed breast thickness. The tube output (for Mo/Mo) was measured at a distance of 50 cm from the focus (with compression plate in place). b Half value layer (HVL) was measured at 28 kV (for Mo/Mo) following the methodology recommended by the ACR. a

4 AGFA CP 100 34 120 KODAK Min-R FUJI UM-MA HC 34 (Malay514, Chinese510 and Indian510)

N Jamal, K-H Ng and D McLean

The British Journal of Radiology, April 2003

Measured display CBT accuracy (mm) Film processor Processing temperature ( ˚C) Processing time (s) Screen Film No. of women

Centre 1

Mean glandular dose during mammography Table 2. Standard optical density (OD) and standard breast mean glandular dose (MGD) for 30 mammography units surveyed

Mean Standard deviation Median 1st quartile 3rd quartile Minimum Maximum

Standard ODa

Entrance surface dose (mGy)

Standard breast MGD (mGy)b

1.28 0.16 1.30 1.17 1.43 0.94 1.47

7.00 3.17 6.50 8.57 4.61 1.34 15.05

1.23 0.44 1.19 0.99 1.44 0.50 2.39

a

OD was measured at centre of RMI 156 phantom image. Standard breast MGD was estimated following the methodology recommended by the American College of Radiology.

b

Table 3. Comparison of estimated mean glandular dose (MGD) with other studies Data source

Technique used

Protocol

Heggie, Australia [8]

estimation

Eklund et al, Sweden [17]

estimation

Victorian Wu et al [9] 490 State Rosenstein et al 1350

Moran et al, Spain [22] Klein et al, Germany [2]

Conversion factor

No. of Mean women CBT (mm)

Mean MGD standard per film breast (mGy) MGD (mGy)

(CC +MLO): 52

median: 2.17 median: 4.42 mean: 2.26 mean: 4.6 mean: 1.25

50

estimation, TLD IPSM

Dance [14]

1596 1496 350

estimation

Klein et al [2]

1678

55.9

mean: 1.59

Wu et al [9]

945 4400

50.8 CC: 45

mean: 2.07 mean: 1.49

IPSM

Dance [14]

4633

CC: 52

8745

MLO,: 54 CC: 51.5

Gentry and TLD DeWerd, USA [21] Burch and estimation Goodman, UK [20] Young and Burch, UK [7]

estimation

IPSM

Dance [14]

Present study

estimation

ACR IPSM

Wu et al [9] Dance [14]

54 62 52

MGD per woman (mGy)

mean: 0.68 mean: 0.76 a

1.3

a

1.52

a

1.34

MLO: 54.3

mean: 1.6

median: 1.4

median: 3.3

median: 1.7 median: 1.65 mean: 3.72 median: 2.03

b

1.23

300

CC: 37.5 MLO: 44.5

median: 1.44 mean: 3.37 median: 1.65 median: 3.21

CBT, compressed breast thickness; CC, craniocaudal; MCO, mediolateral oblique; TLD, thermoluminescent dosemeter; IPSM, Institute of Physical Sciences in Medicine; ACR, American College of Radiology. a Using 4 cm Perspex. b Using RMI 156 phantom (Radiation Measurements Inc., Middleton, WI).

Figure 1. Distribution of standard breast mean glandular dose (MGD) for 30 mammography units surveyed. The British Journal of Radiology, April 2003

Figure 2. Standard breast mean glandular dose (MGD) against standard optical density (OD) for 30 mammography units surveyed. 241

N Jamal, K-H Ng and D McLean Table 4. Details of patient information and technique factors for the patient study sample. Mean value and standard deviation (range in parentheses) are given Ethnic Group

Age (Years)

Weight (kg)

Height (cm)

BMI

Technique factors

Malay

50¡8 (35–87) 62¡12 (33–94)

154¡7 (140–182) 26¡5 (11–39) LCC RCC LMLO RMLO

26¡1 26¡1 26¡1 26¡1

(25–30) 118¡59 (25–30) 122¡75 (25–30) 162¡86 (25–30) 170¡105

(14–334) (16–562) (15–676) (15–819)

17¡3 17¡3 17¡3 17¡3

(10–20) (7–20) (9–20) (9–20)

Chinese 52¡8 (31–81) 56¡9 (39–88)

154¡7 (132–178) 24¡4 (13–37) LCC RCC LMLO RMLO

27¡1 27¡1 27¡1 27¡1

(25–28) (25–31) (25–29) (25–31)

(23–257) (19–315) (21–306) (15–362)

17¡3 18¡3 18¡3 18¡3

(8–20) (8–20) (9–20) (4–20)

Indian

50¡8 (35–71) 64¡11 (41–102) 153¡10 (120–195) 27¡4 (16–38) LCC RCC LMLO RMLO

26¡1 26¡1 26¡1 26¡1

(25–29) 126¡64 (19–372) (25–29) 123¡58 (25–295) (25–31) 180¡110 (30–800) (25–31) 185¡119 (21–800)

18¡3 18¡3 18¡2 18¡2

(10–21) (8–21) (10–20) (11–21)

All

51¡8 (31–87) 61¡11 (33–102)

View

154¡8 (120–195) 26¡5 (11–39) CC MLO

kV

mAs

106¡53 105¡54 131¡69 133¡74

Force (N)

26¡1 (25–31) 117¡61 (14–372) 17¡3 (7–21) 26.5¡1 (25–31) 160¡97.5 (15–819) 18¡3 (4–21)

BMI, body mass index; CC, craniocaudal; MLO, mediolateral oblique; RCC, right craniocaudal; LCC, left craniocaudal; RMLO, right mediolateral oblique; LMLO, left mediolateral oblique.

Figure 5. Descriptive statistics for MGD and CBT for the study sample are summarized in Table 5.

Factors affecting mean glandular dose per woman On multivariate testing, two factors, namely HVL (p,0.001) and CBT (p50.045) were found to be significantly associated with MGD. There was no significant link between MGD and ethnic group, BMI or age with p-values of 0.588, 0.537 and 0.426, respectively. Figures 6–8 describe the distribution of MGD with ethnic origin, BMI and age, using box-whisker plots.

Discussion Standard breast mean glandular dose

Figure 3. Age distribution of the study sample according to ethnic groups.

The mean standard breast MGD of 1.23 mGy is below the reference dose of 3 mGy when using a grid [5, 6], as expected. Figure 1 shows 93.3% of the surveyed units operated in the range 0.6–2.0 mGy and only 3.3% above 2.0 mGy. From this survey it appears that a new reference dose for standard breast MGD using the RMI 156 with grid of 2 mGy can be adopted in Malaysia without any compromise in existing image quality.

Figure 4. Histogram showing the percentage of films as a function of the mean glandular dose (MGD) per film for craniocaudal (CC) and mediolateral oblique (MLO). Axis labels represent the upper bound of alternate 0.2 mGy bands. 242



Figure 5. Histogram showing the percentage of films as a function of compressed breast thickness (CBT) for craniocaudal (CC) and mediolateral oblique (MLO) views. Axis labels represent the midpoint of 5 mm bands. Table 5. Distribution of mean glandular dose (MGD) per woman, MGD per film and compressed breast thickness (CBT) for the study sample. Median (lower quartile and upper quartile in parentheses) and mean values are given Ethnic Group

MGD per woman (mGy)

View

MGD per film (mGy)

Median

Mean

Median

Malay

3.32(2.35, 4.28)

3.36

RCC LCC RMLO LMLO

1.44(0.93, 1.45(1.05, 1.58(1.12, 1.63(1.12,

Chinese

3.19(2.53, 3.68)

3.31

RCC LCC RMLO LMLO

Indian

3.2(2.59, 4.17)

3.44

All

3.21(2.50, 4.01)

3.37

CBT (mm) Mean

Median

Mean

2.04) 2.01) 2.08) 2.52)

1.54 1.58 1.73 1.88

39(30, 39(29, 46(37, 48(38,

46) 46) 56) 55)

38 38 46 46

1.45(1.06, 1.50(1.14, 1.61(1.25, 1.64(1.28,

1.53) 1.51) 1.75) 1.82)

1.84 1.72 2.07 2.09

35(25, 35(24, 39(26, 42(30,

42) 43) 49) 49)

33 33 38 39

RCC LCC RMLO LMLO

1.45(0.45, 1.34(0.47, 1.65(0.31, 1.69(0.31,

1.79) 1.83) 2.33) 2.36)

1.55 1.54 1.85 1.92

39(32, 38(32, 47(39, 46(39,

48) 46) 56) 58)

40 39 48 48

CC MLO

1.44(1.04, 1.86) 1.65(1.26, 2.21)

1.54 1.82

37(29, 45) 45(36, 54)

37 44

Patient-based mean glandular dose

Figure 6. Box-whisker plot of mean glandular dose (MGD) per woman for different ethnic groups; Malay (n5100), Chinese (n5100), and Indian (n5100). The 25th and 75th percentile marks the box and whiskers extend to the range outliers excluded. The median is marked in the box. The median, which is not central, represents the study population, which is not normally distributed. The open circles (o) are outliers that represent cases that have values of more than 1.5 box-lengths from the 75th percentile. The asterisks (*) are extreme values that represent cases that have values of more that three boxlengths from the 75th percentile.


Mean glandular dose and compressed breast thickness Table 5 shows that the mean MGD per film for the MLO (1.82 mGy) was 14.6% higher than for the CC view (1.54 mGy). This could be explained by the presence of pectoral muscle in the oblique view [7]. The mean CBT for CC and MLO films was 37 mm and 44 mm, respectively. These two thicknesses correspond closely to the standard breast equivalent thickness as used in the mammography dosimetry protocol [6] in Malaysia. CBT values for CC and MLO films in the present study are somewhat less than the values recorded by others [7, 12]. While no image quality assessment was made, the associated films were judged to be clinically acceptable. The difference in CBT of 18.7% between MLO and CC views (Table 5) is larger than that found in UK (5.4%) and US (9.1%) based studies [7, 12] for asymptomatic women. This may represent a real difference between the populations, a difference in the compression force used, or a difference in the thickness measurement technique. However, Highnam et al [13] found that mean CBT for MLO was 9% less than that for the CC view (50.9 mm vs 55.5 mm) when MLO was for symptomatic women and CC was for asymptomatic women. In the current study the greatest difference was found in Indian, followed by Malay and then Chinese, women with values of 20.8% (38.5 mm vs 243

N Jamal, K-H Ng and D McLean

and consequently, again, the energy absorption. This highlights the principle that minimizing HVL will maximize MGD. It also indicates that vigorous compression reduces MGD. No significant relationships were seen between MGD per woman with respect to ethnicity, BMI or age. We found that the MGD per woman is almost constant with age from 45 years to 69 years (Figure 8). This trend is similar to that reported by Young [16]. However, this is contrary to the results reported by Beckett and Kotre [3], where MGD per woman declines with increasing age. This could be explained by the fact that our study is limited to diagnostic (referral) mammography, and by the fact that our MGD calculation did not include glandular content, which is commonly accepted to decrease with age [8, 17].

Figure 7. Box-whisker plot showing variations of mean glandular dose (MGD) per woman with BMI between 21 and 35.

Figure 8. Box-whisker plot showing variations of mean glandular dose (MGD) per woman age 45 years to 69 years.

46.5 mm), 20.5% (39 mm vs 47 mm) and 15.7% (35 mm vs 40.5 mm), respectively. These figures indicate that the breast is thinner on the CC than the MLO view, with slightly less compression force generated in the CC than in the MLO view (Table 4), and is similar to what has been reported earlier [12].

Factors affecting mean glandular dose per woman Table 4 and Figure 6 show a close agreement of MGD per woman for Malay, Chinese and Indian groups. The MGD per woman varies greatly within the Chinese group, which could be due to the large number of younger women in the sample (Figure 3). The p-values show no significant difference (at significance level of 0.05) of MGD per woman between ethnic groups. A possible explanation for this could be that the study samples came from the same geographical area and all underwent diagnostic mammography. As expected, two factors have a significant effect (p,0.05) on MGD per woman, namely HVL and CBT. HVL is an indirect measure of the X-ray energy and influences the amount of energy deposited, while CBT influences the length of X-ray passage through the breast 244

Comparison with other studies Ng [18] evaluated MGD for a small sample of 87 country women, using a semi-empirical method based upon knowledge of X-ray tube output, exposure factors, CBT and breast type. This study gave the MGD as 1.13 mGy, comparable with 1.54 mGy per CC film in the present study. The difference can partly be explained by the fact that there has been a recent trend towards using higher film OD with an expected increase in MGD [19]. UNSCEAR [1] has recently reported data for two Health Care Level II countries, namely Iran (Islamic Republic of) and Turkey. The average entrance surface dose in Iran, based upon a patient survey, was 5.5¡1.9 mGy before the introduction of a program of quality control, and 4.23 mGy after. Similarly, a localized survey in Turkey found the entrance surface dose to be 3.3¡0.2 mGy. These values are comparable with 7.0¡3.2 mGy for the present study. The mean CBT reported in the Iranbased study was 30 mm, which is lower than the value of 37.5 mm for the CC view in the present study in Malaysia and may, in part, account for the lower entrance doses recorded. Results from the present study are compared with those of similar recent studies in Table 3, relating either to patient or phantom measurements. MGD estimates in the present study were based on the assumption that all breasts have a standard 50:50 adipose:glandular composition. Eklund et al [17], Young and Burch [7] and Burch and Goodman [20] have used the same assumption. However Heggie [8], Klein et al [2] and Gentry and DeWerd [21] estimated breast glandular content for MGD measurement. Interestingly Heggie [8] has reported that if the survey data was reanalysed assuming 50:50 adipose: glandular composition, average MGD per film was reduced from 2.3 mGy to 2.1 mGy (5.2%). Similarly, Klein et al [2] reported the influence of the actual breast composition causes variations in the order of 15%. The median MGD per film obtained from this study for both CC and MLO views, with CBT of 37.5 and 44.5 mm, respectively, are comparable with UK studies (CBTs of 52 and 57 mm) [20]. However, they are lower than values reported by Young and Burch [7] who reviewed doses for the same study sample chosen by Burch and Goodman [20], which also indicated the trend of using higher film OD and thus higher dose. By contrast, the median MGD reported by Heggie [8] seems high as the study includes both CC and MLO films and used the Victorian protocol, which would have included using 26 kVp for all procedures. The MGD per film of 1.25 mGy reported by The British Journal of Radiology, April 2003


Eklund et al [17] was much lower than values reported by other recent studies. However, no grid was used in this study.

Errors in mean glandular dose estimation Errors in MGD estimation could result from errors in the measurements of CBT and interpolation of g-values. Estimations of a conversion factor depend upon HVL measurements. Some authors [22] have demonstrated that the error in HVL measurement could be greater than ¡10%, which may result in an uncertainty in g of approximately ¡20%. It must be noted that conversion factors published by Wu et al [9] are approximately 10% higher than the corresponding factors published by Dance [14]. The assumption of a standard composition will result in overestimation of MGD to breast with higher glandular content and underestimation of MGD to breast with a lower glandular content.

Conclusion Mean standard breast MGD (with grid) for mammographic practice in Malaysia is 1.23 mGy. A new reference dose for standard breast MGD with grid of 2 mGy could be adopted in Malaysia without any compromise in existing image quality. Standard breast MGD is fairly well correlated with film OD. Mean MGD per film for CC and MLO views are 1.54 mGy and 1.82 mGy, respectively, and men CBTs are 37.5 mm and 44.5 mm, respectively. The mean MGD per woman was 3.37 mGy. The non-parametric Kruskal–Wallis test showed no significant difference in MGD per woman amongst ethnic groups (p.0.05). However, on the multivariate test two factors, namely HVL of the X-ray beam and CBT, had a significant effect on MGD per woman (p,0.05). No significant relationships were seen between MGD per woman with respect to ethnicity, BMI or age.

Acknowledgments We acknowledge the co-operation of both patients and staff at the mammography clinic of the Department of Radiology, University of Malaya Medical Centre, Kuala Lumpur, and the mammography clinic of the Department of Radiology, Kuala Lumpur Hospital for participating in this study. We thank Professor Dr S-H Ong, Statistics Consultant, Institute of Mathematical Sciences, University of Malaya for his advice regarding statistical methods. We also thank Tan Sri Datu Dr Hj Mohamad Taha bin Arif, the Director-General of the Ministry of Health, Malaysia for his permission to publish this paper and his support of this work.

References 1. United Nations Scientific Committee on the Effects of Atomic Radiation. Report to the General Assembly, with scientific annexes, Volume I: sources. United Nations Sales Publication E.00.IX.3. New York, NY: UN, 2000.


2. Klein R, Aichinger H, Dierker J, Jansen JTM, Joit-Barfub S, Sabel M, et al. Determination of average glandular dose with modern mammography units for two large groups of patients. Phys Med Biol 1997;42:651–71. 3. Beckett JR, Kotre CJ. Dosimetric implications of age related glandular changes in screening mammography. Phys Med Biol 2000;45:801–13. 4. Institute of Physical Sciences in Medicine. Commissioning and routine testing of mammographic x-ray systems Report 59(2). York, UK: IPSM, 1994. 5. International Atomic Energy Agency. International basic safety standards for protection against ionising radiation and for the safety of radiation sources. Safety Series no. 115–1. Vienna, Austria: IAEA, 1996. 6. American College of Radiology Committee on Quality Assurance in Mammography. Mammography quality control manual, medical physicist’s section. Reston, VA: ACR, 1999. 7. Young KC, Burch A. Radiation dose received in the UK Breast Screening Programme in 1997 and 1998. Br J Radiol 2000;73:278–87. 8. Heggie JCP. Survey of dose in screening mammography. Australas Phys Eng Sci Med 1996;19:207–16. 9. Wu X, Barnes GT, Tucker DM. Spectral dependence of glandular tissue dose in screen-film mammography. Radiology 1991;179:143–8. 10. McLean D. Breast composition and radiographic breast equivalence. Australas Phys Eng Sci Med 1997;20:11–9. 11. Gibson RS. Principles of nutritional assessment. New York, NY: Oxford University Press, 1990;178–82. 12. Helvie MA, Chan H-P, Adler DD, Boyd PG. Breast thickness in routine mammograms: effect on image quality and radiation dose. AJR 1994;163:1371–4. 13. Highnam RP, Brady JM, Shepstone BJ. Estimation of compressed breast thickness during mammography. Br J Radiol 1998;71:646–53. 14. Dance DR. Monte Carlo calculation of conversion factors for the estimation of mean glandular breast dose. Phys Med Biol 1990;35:1211–9. 15. International Commission on Radiological Protection. Report of the task group on reference man. ICRP 23. Oxford, UK: Pergamon Press, 1975. 16. Young KC. Radiation doses in the UK trial of breast screening in women aged 40–48 years. Br J Radiol 2002; 75:362–70. 17. Eklund S, Thilander A, Leitz W, Mattsson S. The impact of anatomic variations of absorbed radiation doses in mammography. Radiat Prot Dosim 1993;49:167–70. 18. Ng KH. Recent advances in patient dosimetry in diagnostic radiology. In: Tajuddin AA, Bradley DA, editors. Centennial of the x-ray–an account of developments in radiological physics and radiology in Malaya and Malaysia. Kuala Lumpur, Malaysia: Malaysian Institute of Physics, 1995: 49–71. 19. Faulkner K, Law J, Robson KJ. Assessment of mean glandular dose in mammography. Br J Radiol 1995;68: 877–81. 20. Burch A, Goodman DA. A pilot survey of radiation doses received in the United Kingdom Breast Screening Programme. Br J Radiol 1998;71:517–27. 21. Gentry JR, DeWerd LA. TLD measurements of in vivo mammographic exposures and the calculated mean glandular dose across the United States. Med Phys 1996;23:899–903. 22. Moran P, Chevalier M, Vano E. Comparative study of dose values and image quality in mammography in the area of Madrid. Br J Radiol 1994;67:556–63.

245