C F NJEH, MSc, PhD, S B SAMAT, BSc, PhD, A NIGHTINGALE, MSc, E A ... Dual X-ray absorptiometry (DXA) is one of the most widely used techniques for non-.
T he British Journal of Radiology, 70 (1997), 719–727
© 1997 The British Institute of Radiology
Radiation dose and in vitro precision in paediatric bone mineral density measurement using dual X-ray absorptiometry C F NJEH, MSc, PhD, S B SAMAT, BSc, PhD, A NIGHTINGALE, MSc, E A McNEIL, BSc, MPhil and C M BOIVIN, MPhil Medical Physics Department, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham B15 2TH, UK Abstract. Dual X-ray absorptiometry (DXA) is one of the most widely used techniques for noninvasive assessment of bone integrity. There is a growing demand for measurement of paediatric bone status. In DXA the principal radiation risks to patients are the carcinogenic and genetic effects. Radiation dosimetry is well established for DXA in adults, but there are limited paediatric data available. We report on a study to estimate the effective doses (EDs) received by typical 5and 10-year-old children using the paediatric scan mode on the Lunar DPX-L bone mineral density scanner. Entrance surface doses (ESDs) and percentage depth doses for the total body and PA spine scan modes were measured using lithium borate thermoluminescent dosemeters (TLDs) located at the surface and distributed at various organ locations in anthropomorphic child phantoms. The EDs were calculated from the percentage depth doses, amount of each organ irradiated and tissue weighting factors. The ESDs were measured to be 6.0 and 0.12 mGy for the posteroanterior (PA) spine and total body, respectively. PA spine EDs were calculated as 0.28 and 0.20 mSv for the 5- and 10-year-old, respectively. Total body EDs were 0.03 and 0.02 mSv for the 5- and 10-year-old children, respectively. These results compare with an adult ED of 0.21 mSv for the PA spine. They are also more than two orders of magnitude lower than reported ESDs and EDs for paediatric chest X-rays. Bone mineral density (BMD) short-term in vitro precision was 0.5% and 1% in the 5- and 10-year-old phantoms, respectively. In conclusion, the Lunar DPX-L in the paediatric mode has a high precision and very low radiation doses, similar to those reported for the adult mode.
Introduction Different diseases may directly or indirectly decrease the bone mineral content (BMC) in children. These include conditions such as renal transplantation , cystic fibrosis , juvenile rheumatoid arthritis  and eating disorders . Also, some drugs like corticoids  or anticonvulsants  will have deleterious effects on bone. Therefore, the choice of the right therapeutic strategy and best follow-up of patients with disease that could affect their bone metabolism may be assisted by assessment of bone integrity. Also, bone mass accrued during childhood has been proposed as a determinant of an individual’s susceptibility to osteoporotic fractures in adulthood . Knowledge of the timing and magnitude of peak Received 4 December 1996 and in revised form 5 March 1997, accepted 18 March 1997. Address correspondence to C F Njeh, Osteoporosis and Arthritis Research Group, 533 Parnassus Avenue, Suite U368G, Department of Radiology, University of California San Francisco, California, CA 94143-1250, USA. T he British Journal of Radiology, July 1997
bone mass is important because it may be relevant to an individual’s subsequent risk of osteoporosis, especially if a disease state or delayed sexual maturation is present. Non-invasive methods of assessing bone mineral density (BMD) have been developed. These include single photon absorptiometry (SPA), dual photon absorptiometry (DPA), quantitative computed tomography (QCT) and dual X-ray absorptiometry (DXA). DXA is presently the most widely used technique for the assessment of bone integrity. DXA can also be used to study other aspects of body composition such as percentage fat and lean body mass . The use of an X-ray source rather than a gamma emission has resulted in improved image resolution, shorter scan times, better precision and a reduction in radiation exposure per scan [9, 10]. The improved performance has led to a wider use of the X-ray absorptiometry technique. Normal ranges have been acquired for adult populations across a range of DXA machines, but normal data for children are limited, in part due to ethical considerations of radiation exposure to normal children. Therefore, a detailed assessment 719
C F Njeh, S B Samat, A Nightingale et al
of the radiation exposure to children undergoing DXA measurement is required to enable proper consideration of the ethical issues. The effect of ionizing radiation in the context of DXA is considered to be stochastic and therefore any exposure carries a risk. For low doses, as encountered in DXA, the principal risks to patients are the stochastic processes of carcinogenesis and genetic effects. Manufacturers normally quote entrance skin dose (ESD), but this is not particularly useful when assessing the total radiation hazard from a particular procedure. The appropriate quantity for the assessment of the risk of radiation injury is the effective dose . Radiation risk assessment is of particular importance in the irradiation of children because of the higher expected risk of radiation-induced effects in children than in adults . The objective of the study was to measure entrance surface dose and estimate effective dose for typical 5- and 10-year-old children undergoing posteroanterior (PA) spine and total body BMD scans on a Lunar DPX-L densitometer. We also investigated short-term in vitro precision for the two scanning modes.
is smaller and the time required to complete the scan is reduced with consequently lower radiation exposure. The DPX-L has a highly stable potential X-ray generator with a K-edge cerium filter. The two X-ray energies required to differentiate between soft tissue and bone are produced using the filter. The X-ray tube is operated at 76 kVp and has effective beam energies of 38 and 70 keV . The DPX-L supports various current settings, i.e. 3.0, 0.75, 0.3 and 0.15 mA. Adult scans are usually carried out using 0.75 mA, and the paediatric mode uses either 0.3 or 0.15 mA. These lower current settings limit photon pile-up in the detector. Various scanning modes can be used depending on the weight/thickness of the abdomen of the patient and resolution required. The paediatric mode does not have the proximal femur scan option which is available in the adult software. This is because it is believed that the skeletal anatomy of the femoral region in a developing skeleton is ill-defined and therefore does not readily lend itself to BMD determination [Fowler C. Personal communication. Lunar Corporation, 1996]. Details of the scan modes investigated are presented in Table 1.
Materials and methods The Lunar DPX-L (Lunar Corporation, Madison, USA) DXA machine and two anthropomorphic child phantoms (ATOM Ltd, Riga, Latvia) were used in this investigation.
DPX-L densitometer The Lunar DPX-L is currently one of the most widely used DXA machines. It uses a ‘‘pencil’’ X-ray beam which passes through the patient and is detected by a single sodium iodide (NaI) detector. Both the X-ray source and detector traverse the scanning area with rectilinear motion. An ‘‘auto width’’ facility is used which automatically narrows the width of the scan path when it locates the patient’s bone mass. As a result, the area scanned
Anthropomorphic paediatric phantoms We borrowed two anthropomorphic heterogeneous phantoms (ATOM Ltd, Riga, Latvia) representing average 5- and 10-year-old children. The skeleton and lungs are made of bone-equivalent and lung-equivalent plastics. The skeleton is impregnated with P2D-MBT soft tissue equivalent material (TEM) and is encased in TEM moulded to model average 5- and 10-year-old children . The moulded phantom is sliced into 26 and 32 slices for the 5- and 10-year-old, respectively, as shown in Figure 1. For both phantoms, each slice is 2.5 cm thick and is numbered, starting from the top of the head. Each slice has several 3 mm diameter holes which are identified with numbers and can accommodate thermoluminescent dose-
Table 1. Paediatric scanning modes available on the Lunar DPX-L and manufacturer’s quoted and measured ESD Mode
Pixel size (mm)
Patient sizea or weight
Typical scan time (min)e
PA spine Total body large Total body medium Total body small
0.6×0.6 3.6×7.2 3.6×7.2 2.4×4.8
Medium Medium Fine Fine
300 150 300 300
6–16 cm 25–35 kg 15–25 kg 5–15 kg
9.6 0.3 0.1 0.3
6.0 0.12 nm nm
5.9 0.09 nm nm
5 12 9 10
a Patient size refers to the body thickness in the scan region. b Manufacturer’s quoted ESD. c Measured ESD (TLD). d Measured ESD (chamber). e Scan times are dependent on scan width and length. nm, not measured.
T he British Journal of Radiology, July 1997
Radiation dose in paediatric DXA
Figure 1. (a) Front view and (b) back view of the phantoms. Labels: (a) 5-year-old phantom and (b) 10-year-old anthropomorphic phantom.
meters (TLDs). The number of holes per slice varies depending on the number of organs of interest in it. The phantoms were 15.0 and 16.0 cm thick at the abdomen with a weight of 19.0 and 31.9 kg for the 5- and 10-year-old, respectively.
Depth dose determination 80 TLDs were placed at the entrance surface and also inserted at different depths in the 5-yearold anthropomorphic phantom. Lithium borate (Li B O 5Mn) TLDs were used because of their 2 4 7 improved energy response compared with lithium fluoride (LiF). Lithium borate has better tissue equivalence than lithium fluoride, with effective atomic number of 7.4, compared with 7.42 for tissue and 8.14 for LiF. For the dosimetry of X-rays with energies below 100 keV as used in DXA, the difference in effective atomic number between TLDs and tissue leads to a difference in mass attenuation coefficient and thus errors in measured dose . Therefore, although the low doses involved in DXA scanning could benefit from the high sensitivity of TLDs such as calcium sulphate (CaSO 5Mn) the latter is not suitable because of 4 its high atomic number. TLDs were supplied, calibrated and read out by the National Radiological Protection Board (NRPB, UK). For the batch of TLDs used the minimum detectable T he British Journal of Radiology, July 1997
dose was 50 mSv at 95% confidence interval with an overall accuracy of ±7%. This is within the ±10% non-random uncertainty tolerance recommended by the dosimetry working party of the Institute of Physical Sciences in Medicine . Due to the small values of ESD quoted by the manufacturer (Table 1), multiple exposure was necessary in order to obtain a sufficiently large integrated dose for good measurement accuracy. Accordingly, for the PA spine and total body large modes the TLDs were exposed to 60 and 167 scans, respectively. The number of exposures for the total body large and PA spine was determined so that the integrated dose at the entrance and exit of the phantom was more than 50 mGy (detection limit of TLDs). Depth dose was measured only for the PA spine mode because the tube potential remains the same irrespective of the mode and therefore the rate of absorption of the X-rays is the same. Depth dose for the PA spine was then applied to calculate the EDs for all modes. ESD was measured for total body large mode, but for total body paediatric medium mode the ESD was so low that it would have required more than 500 scans to acquire enough dose to be measurable by the lithium borate TLDs, so we used the manufacturer’s quoted value. ESDs were further estimated using an MDH-2025 180 cm3 ionization chamber (Radcal 721
C F Njeh, S B Samat, A Nightingale et al
Corporation, Monrovia, California) and the water phantom which is routinely used for quality assurance measurements. Exposure readings were obtained for the PA spine and total body large modes. In order to improve the measurement precision, an average of five exposure readings was made in each case. Exposure readings (mR), including backscatter, were converted to units of air kerma (mGy) using the conversion factor 8.76 mGy mR−1 . Accuracy on the chamber calibration is ±4% at the 95% confidence level for calibration free-in-air in the reference beam. Hence, the overall uncertainty in the chamber readings is expected to be less than ±8%.
sectional area in each slice by the slice thickness (Tables 2 and 3). There was good agreement with organ volumes provided by Cristy  for both the 5- and 10-year-old children. The highest discrepancies were observed for the smaller organs due to the slice thickness of 2.5 cm. For large or widely distributed organs such as bone, red bone marrow, skin and muscle the use of mean depth would result in an underestimate of organ dose because the dose varies exponentially rather than linearly with depth. For these organs the body was subdivided into sections. Data on the distribution of red bone marrow as a function of age were taken from Cristy .
EVective dose calculation
The effective dose (ED) is defined as the sum of the absorbed doses to each irradiated organ weighted for the radiation type and radiosensitivity of that organ. It was calculated according to the recommendations in the International Commission on Radiological Protection (ICRP) Publication 60 . ED is the sum of (i) the organ doses to the 12 specified organs multiplied by their respective tissue weighting factor, and either (ii) the average dose to the remainder organs multiplied by 0.05, or (iii) if one of the remaining organs receives a dose higher than the highest dose in any of the organs for which tissue weightings are specified, then the dose for that organ multiplied by 0.025 plus the mean dose to the rest of the remainder organs multiplied by 0.025. ED can be estimated as:
Short-term BMD in vitro precision was measured on the two anthropomorphic phantoms for the PA spine only. This entailed scanning the phantoms 20 times on the same day, without repositioning. Total body precision was not investigated because the phantom did not have complete appendicular regions, which would have had an unknown effect on the analysis. Repositioning was not deemed useful because the in vivo positioning error could not be simulated using the phantoms. BMD analysis was carried out according to the manufacturer’s recommended protocol. BMC and BMD precision was then expressed as the percentage coefficient of variation (%CV), defined as:
ED=∑ W PDD ESD f (1) T T T T where W is the ICRP 60 weighting factor for T tissue or organ T , PDD is the percentage depth T dose to the centre of the organ, ESD is the entrance surface dose and f is fraction of organ (T ) T irradiated.
Organ depth calculation From Equation (1), it can be seen that the knowledge of the depth of each relevant organ is required. Where only part of an organ is included in the scanning field the fraction irradiated is also required for the calculation of effective dose. Calculations were based mainly on information provided by the phantom manufacturer, which consisted of the organ location of each TLD hole in the phantom. The centre of the named organ in ICRP 60  was calculated from the hole locations. Using this information and an adult cross-sectional atlas, we were able to estimate the position of these organs. The volume of the organs was estimated by multiplying the measured cross722
standard deviation ×100 mean
Results The depth dose curve is presented in Figure 2. The different scanning modes with the measured ESDs are presented in Table 1. Tables 2 and 3 give the estimated depth and volume of some of specified organs in different slices of the phantoms. Doses for organs specified in ICRP 60  and EDs for different scanning modes are presented in Table 4. The stomach was the greatest contributor to the PA spine dose and the red bone marrow to the total body dose. Adrenals and kidneys were the highest contributors to the remainder dose for both the PA spine and the total body. For the PA spine neither the gonads nor the breasts are included in the scanning field, so the dose will apply to both male and female. However, for the total body both the gonads (ovaries or testes), uterus and breasts are included in the field. Hence the total body EDs in Table 2 apply to females and the male values will be slightly lower. For comparison, Table 5 includes reported ESDs (per T he British Journal of Radiology, July 1997
Radiation dose in paediatric DXA Table 2. Estimation of depth, area, volume and fractions of organs on each slice of the 5-year-old phantom Organ
Area or volume fraction
15 16 17 18 Total Mean
8.8 9.3 8.1 7.8 34.0 8.5
78.21 62.50 54.24 54.46 249.41
195.53 156.25 135.60 136.15 623.53
0.31 0.25 0.22 0.22 1.00
15 16 17 Total Mean
5.0 5.0 5.2 15.2 5.1
3.08 7.30 3.74 14.12
7.70 18.25 9.35 35.30
0.22 0.52 0.26 1.00
15 16 17 Total Mean
8.1 9.8 10.0 27.9 9.3
1.07 16.39 30.50 47.96
2.68 40.97 76.25 119.90
0.02 0.34 0.64 1.00
15 16 Total Mean
3.5, 3.3 3.4, 3.4 13.6 3.4
7.69 7.58 15.27
19.23 18.95 38.18
0.50 0.50 1.00
16 Total Mean
16 17 18 Total Mean
2.37 18.85 9.46 30.68
5.93 47.13 23.70 76.70
0.08 0.61 0.31 1.00
7.2 7.2 7.2 2.5, 2.7 2.4, 2.5 2.5, 2.8 15.4 2.6
radiograph) and calculated EDs for paediatric chest and abdomen X-ray examinations. EDs were calculated from the reported ESDs using NRPBSR279 Monte Carlo data . The phantom in vitro precision data are presented in Table 6.
Discussion Dose Measured ESDs were 6.00 mGy for the PA spine and 0.12 mGy for the total body paediatric large modes using TLDs. These values are smaller than the manufacturer’s quoted values of 9.6 and 0.3 mGy, respectively. Because the doses involved in total body scanning are so low, they approach the detection limit of the lithium borate TLDs, even with multiple scans. Hence the uncertainty associated with the ESD measurements is large. For this batch of TLDs it was ±22% for doses around 100 mGy. The mean accumulated dose for multiple exposures was 400 mGy for AP spine and 20 mGy for total body. The low total body dose could explain the differences from the manufacturer’s quoted ESD. However, ESDs measured using the ionization chamber were similar to those measured using TLDs. T he British Journal of Radiology, July 1997
The EDs were estimated to be 0.3 mSv for the 5-year-old and 0.2 mSv for the 10-year-old PA spine. Total body EDs were 0.03 and 0.02 mSv, respectively. These values are much lower than the average daily natural background in the United Kingdom of 7 mSv . For the PA spine, the ED for the 5-year-old was slightly higher than that of the 10-year-old as would be expected with increasing thickness of the abdomen. Red bone marrow is one of the most radiosensitive organs and the radiation dose to it in children depends on the body part scanned and the age-dependent bone marrow distribution. For the lumbar spine there is a gradual increase in active marrow with age from 6.8% at age 5 to 8.4% at age 10 . This did not have a large effect on the difference in estimated ED between the 5- and 10-year-old child. Total body EDs are a factor of 10 lower than the PA spine EDs. Doses reported here are estimates which will vary depending on the child’s weight, height and thickness at the abdomen. We found no previously reported EDs for paediatric DXA using the Lunar machine. However, the results reported in this work are of the same order of magnitude as those reported for adults using the Lunar machines [26, 27]. The other major 723
C F Njeh, S B Samat, A Nightingale et al Table 3. Estimation of depth, area, volume and fractions of organs on each slice of the 10-year-old phantom Organ
Area or volume fraction
17 18 19 20 21 Total Mean
12.0 11.0 10.5 9.5 8.8 51.8 10.4
65.76 85.71 102.90 73.14 35.99 363.50
164.40 214.28 257.25 182.85 89.98 908.75
0.18 0.24 0.28 0.20 0.10 1.00
18 19 20 Total Mean
6.6 6.4 6.0 19.0 6.3
10.64 14.66 14.65 39.95
26.60 36.65 36.63 99.88
0.27 0.37 0.36 1.00
19 20 21 Total Mean
10.7 11.5 11.0 33.2 11.1
15.91 31.56 40.64 88.11
39.78 78.90 101.60 220.28
0.18 0.36 0.46 1.00
19 20 Total Mean
5.23 4.07 9.30
13.07 10.18 23.25
0.56 0.44 1.00
19 20 Total Mean
9.0 8.8 17.8 8.9
10.01 12.28 22.29
25.03 30.70 55.73
0.45 0.55 1.00
20 21 22 23 Total Mean
3.5, 4.9 4.3, 4.2 4.0, 3.6 4.2 28.7 4.1
7.93 27.28 26.36 7.26 68.83
19.83 68.20 65.90 18.15 172.08
0.12 0.40 0.38 0.10 1.00
3.5, 3.0 2.9 9.4 3.1
Figure 2. The depth dose curve for the lunar DPX-L.
manufacturer of DXA machines is Hologic (Waltham, USA). Hologic uses voltage switching to produce the two X-ray energies as opposed to K-edge filtration employed by Lunar. This affects the dose received by the patient. For the pencil beam first generation scanners, reported adult values are higher for Hologic than for Lunar 724
[26, 28]. Koo et al  reported an ESD of 3 mGy for paediatric total body using the Hologic 1000/W which is larger than the 0.12 mGy measured here. Faulkner et al  reported 10 mGy ESD and effective dose equivalent of 2.7 mSv for total body mode using Hologic QDR-2000 (fan beam mode). However, for the adult mode fan T he British Journal of Radiology, July 1997
Radiation dose in paediatric DXA Table 4. ICRP 60  weighted tissue/organ doses for different paediatric scanning modes and their respective effective doses Tissue/organ
Ovaries Red bone marrow Lower large intestine Lung Stomach Bladder Breast Liver Oesophagus Thyroid Skin Bone surfaces Remainder
0.20 0.12 0.12 0.12 0.12 0.05 0.05 0.05 0.05 0.05 0.01 0.01 0.05
Weighted tissue dose (mSv) PA spine 5-year-old
PA spine 10-year-old
Total body mediuma
Total body largeb
0.0000 0.0194 0.0378 0.0000 0.1164 0.0000 0.0000 0.0451 0.0000 0.0000 0.0023 0.0033 0.0580
0.0000 0.0214 0.0280 0.0000 0.0775 0.0000 0.0000 0.0203 0.0000 0.0000 0.0023 0.0032 0.0484
0.0059 0.0057 0.0019 0.0039 0.0023 0.0008 0.0008 0.0011 0.0017 0.0015 0.0006 0.0006 0.0016
0.0019 0.0051 0.0029 0.0033 0.0016 0.0005 0.0004 0.0008 0.0016 0.0011 0.0005 0.0006 0.0014
a Appropriate for average 5-year-old. b Appropriate for average 10-year-old.
Table 5. Reported mean ESDs per radiograph and calculated EDs using Monte Carlo simulation for two common radiological examinations Ref.
   (a) (b)   (a) (b)  (a) (a)
Age (years) 1–5 1–4 1–5 1–5 1–5 6–10 5–9 6–10 6–10 10–15 10–14 10–14
50 70 200 260 110 80 60 310 150 80 400 220
5.2 7.3 21.6 37.8 12.7 7.9 6.1 43.2 16.7 5.7 38.8 17.2
470 420 750 1200 1620 770 670 1280 1990 1060 — 2290
89.8 73.2 137 242 308 132 106 245 378 125 — 302
(a) and ( b) refer to data from two different centres.
Table 6. In vitro precision from 20 repeated measurements on paediatric anthropomorphic phantoms 5-year-old
Mean Stdv CV(%)
27.50 0.144 0.52
1.243 0.006 0.48
27.47 0.281 1.01
1.273 0.013 1.02
beam has been reported to give higher doses than pencil beam [26, 31]. Few studies have reported the effective doses for common paediatric radiological examinations. However, ESDs for common paediatric radiological examinations have been reported [21–24] (Table 5). ESDs are in the range 50–400 mGy for PA chest X-ray and 420–1990 mGy for abdomen T he British Journal of Radiology, July 1997
X-ray examinations. We used these values and assumed a total X-ray filtration of 2.5 mm Al equivalent and 70 kVp in cases where the actual values were not provided to estimate the EDs using NRPB-SR279 Monte Carlo data . From Table 5 it can be seen that the doses encountered in DXA are more than two orders of magnitude less than common radiological examinations. Exposure of children to ionizing radiation gives more grounds for concern than adults, because for children there is a longer period available for the delayed effects of radiation to be expressed. This means that the lifetime risk of radiation-induced cancer per unit dose to the child is likely to be higher than for the adult. The probability of aggregated radiation detriment can be related to ED and for children this is age and sex related. For a 10-year-old female, the total lifetime risk of 725
C F Njeh, S B Samat, A Nightingale et al
radiation-induced cancer is estimated as 16×10−2 Sv−1 compared with 5.9×10−2 Sv−1 for a 30-yearold adult . A paediatric subject referred for bone densitometry in our centre will normally receive a PA spine and total body scan. The total effective dose will then be 0.31 mSv giving a lifetime risk of 5×10−8 for fatal cancer. This is negligible compared with the possible benefits from the scan. In addition to carcinogenesis, genetic effects are also of concern in the total body examination because of the irradiation of the gonads. The estimated dose to the ovaries was 0.002 mSv which is 9% of the total dose. Even though these doses are quite low, good radiological practice should be implemented. Any exposure to radiation should produce sufficient benefit to the exposed individual or to the society to offset the radiation detriment it causes. The dose should be kept as low as reasonably achievable. Using the ‘‘auto scan width’’ option and the shortest possible scan length will help to reduce the dose. In carrying out paediatric radiography, staff and accompanying adult doses should also be considered since there is a tendency to stay close to the patient. The scattered radiation dose rates for Lunar DPX and DPX-L machines have been reported for adult scanning modes [26, 32]. The dose rate at 1 m from the centre of the scanning table was reported to be less than 1 mSv h−1. It could be expected that the scattered radiation from similar machines in paediatric mode will be of the same order of magnitude and hence no additional protection is required.
Precision Precision errors in this study are due to variation in detected photon density and also to edge detection effects resulting in BMC and bone crosssectional area imprecision. For example, for the 5-year-old the measured BMC and bone crosssectional area precisions were 0.33% and 0.52%, respectively, resulting in calculated root mean square precision of 0.44% for BMD which is similar to the measured value of 0.48%. This is similar to the adult values reported previously by Lilley et al . However, in vivo precision would be expected to be poorer than in vitro because of repositioning and movement errors. Difference in precision for the 5- and 10-year-old phantoms could be due to the high BMC values in both (68% higher than age-matched values). A high BMC results in fewer photons reaching the detector and the intrinsic precision is dependent on the number of photons detected. There were also occasions when the edge detection algorithms failed to detect the edges properly. This was rectified manually. This was never observed in the 5-year-old phantom. This may indicate that the 726
precision is dependent on the ratio of BMC to soft tissue. Precision error (CV) affects the minimum detectable change from repeated measurements on a single subject. At 95% confidence, it is given by 2√2× CV . Hence, less than 3% change in BMD could be measured from repeated measurements using the paediatric mode.
Conclusion Although phantoms of two specific ages were used, it is considered that the dosimetry results obtained from this work are applicable for the age range 5–18 years. This is because there was no significant difference in EDs between the adult and 10-year-old. This extends the adult data already published by Njeh et al , Bezakova et al  and Lewis et al . The PA spine and total body doses are about two and three orders of magnitude lower, respectively, than those incurred during common radiological examinations. PA spine dose is equivalent to about 1 h of average UK background radiation and total body dose to about 6 min. In vitro precision is similar to that reported for adult modes.
Acknowledgments We are grateful to Ms Dawn Broadhead of Newcastle Regional Medical Physics Department for loan of the phantoms, to AuRa Scientific for help with transporting them and to Dr Alun Beddoe for his comments on the manuscript. SBS is grateful for the award of a joint fellowship by the Commonwealth Scholarship Commission Contract No: MYF-015, by the British Council Contract No: MAL 520-00-03, and by the Universiti Kebangsaan Malaysia Contract No: UKM(PER) 3994.
References 1. Chesney RW, Rose PG, Mazess RB. Persistence of diminished bone mineral content following renal transplantation in childhood. Pediatrics 1984;73:459–66. 2. Gibbens DT, Gilsanz V, Boechat MI, et al. Osteoporosis in cystic fibrosis. J Pediatr 1988;113:295–300. 3. Falcini F, Trapani S, Civinini R, et al. The primary role of steroids in osteoporosis in juvenile rheumatoid arthritis patients evaluated by dual X-ray absorptiometry. Endocrinol Invest 1996;19:165–9. 4. Mazess RB, Barden HS, Ohlrich ES. Skeletal and body-composition effects of anorexia nervosa. Am J Clin Nutr 1990;52:438–41. 5. Chesney RW, Mazess RB, Rose P, Jax DK. Effect of prednisone on growth and bone mineral content in childhood glomerular disease. Am J Dis Child 1978;132:768–72. T he British Journal of Radiology, July 1997
Radiation dose in paediatric DXA 6. Timperlake RW, Cook SD, Thomas KA, et al. Effects of anticonvulsant drug therapy on bone mineral density in a pediatric population. J Pediatr Orthop 1988;8:467–70. 7. Seeman E, Young N, Szmukler G, et al. Risk factors for osteoporosis. Osteoporosis Int 1993;3(Suppl. 1):40–3. 8. Laskey MA. Dual-energy X-ray absorptiometry and body composition. Nutrition 1996;12:45–51. 9. Lang P, Steiger P, Faulkner KG, et al. Current techniques and recent developments in quantitative bone densitometry. Radiol Clin North Am 1991;29:49–76. 10. Wahner HW, Fogelman I. The Evaluation of Osteoporosis: Dual Energy X-ray Absorptiometry in Clinical Practice. London: Martin Dunitz Ltd, 1994:14–87. 11. International Commission On Radiological Protection. 1990 Recommendations of the International Commission on Radiological Protection (ICRP Publication 60, Annals of the ICRP, 21). Oxford: Pergamon Press, 1991:11–21. 12. National Radiological Protection Board. Occupational, Public and Medical Exposure (Documents of the NRPB, Vol. 4). Chilton: NRPB, 71–4. 13. Mazess R, Collick B, Trempe J, et al. Performance evaluation of a dual-energy X-ray bone densitometer. Calcif Tissue Int 1989;44:228–32. 14. Varchenya V, Gubatova D, Sidorin V, Kalnitsky S. Children’s heterogeneous phantoms and their application in rontgenology. Radiation Protection Dosimetry 1993;49:77–8. 15. Kron T. Thermoluminescence dosimetry and its applications in medicine, Part 1: Physics, materials and equipment. Australasian Phys Eng Sci Med 1994;17:175–99. 16. National Radiological Protection Board. National Protocol for Patient Dose Measurements in Diagnostic Radiology. Chilton: NRPB, 1992:26–9. 17. International Commission on Radiation Units and Measurements. Measurement of dose equivalents from external photon and electron radiation (ICRU Report 47). Bethesda, Maryland: ICRU, 1992:36. 18. Cristy M. Mathematical phantoms representing children of various ages for use in estimates of internal dose (NUREG/CR-1159, ORNL/NUREG/TM367) Oak Ridge, TN: Oak Ridge National Laboratory, 1980:103. 19. Cristy M. Active bone marrow distribution as a function of age in humans. Phys Med Biol 1981;26:389–400. 20. Hart D, Jones DG, Wall BF. Coefficient for estimating effective doses from paediatric X-ray examinations (NRPB-R279). London: HMSO, 1996. 21. Martin CJ, Farquhar B, Stockdale E, Macdonald S. A study of the relationship between patient dose and size in paediatric radiology. Br J Radiol 1994;67:864–71.
T he British Journal of Radiology, July 1997
22. Kyriou JC, Fitzgerald M, Pettett A, Cook JV, Pablot SM. A comparison of doses and techniques between specialist and non-specialist centres in the diagnostic X-ray imaging of children. Br J Radiol 1996;69:437–50. 23. Chapple CL, Faulkner K, Lee REJ, Hunter EW. Results of a survey of doses to paediatric patients undergoing common radiological examinations. Br J Radiol 1992;65:225–31. 24. Ruiz MJ, Gonza´lez L, Van˜o´ E, Martinez A. Measurement of radiation doses in the most frequent simple examinations in paediatric radiology and its dependence on patient age. Br J Radiol 1991;64:929–33. 25. Hughes JS, Shaw KB, O’Riordan MC. Radiation exposure of the UK population: 1988 review. National Radiation Protection Board (Report NRPB-R227). London: HMSO, 1989. 26. Njeh CF, Apple K, Temperton DH, Boivin CM. Radiological assessment of a new bone densitometer—the Lunar EXPERT. Br J Radiol 1996;69:335–40. 27. Bezakova E, Collins PJ, Beddoe AH. Absorbed dose measurements in dual energy X-ray absorptiometry (DXA). Br J Radiol 1997;70:172–9. 28. Lewis MK, Blake GM, Fogelman I. Patient dose in dual X-ray absorptiometry. Osteoporosis Int 1994;4:11–15. 29. Koo WWK, Walters J, Bush AJ. Technical considerations of dual-energy X-ray absorptiometry-based bone mineral measurements for paediatric studies. J Bone Miner Res 1995;10:1998–2004. 30. Faulkner RA, Bailey DA, Drinkwater DT, Wilkinson AA, et al. Regional and total body bone mineral content, bone mineral density, and total body tissue composition in children 8–16 years of age. Calcified Tissue Int 1993;53:7–12. 31. Blake GM, Patel R, Lewis MK, Batchelor S, et al. New generation dual X-ray absorptiometry scanners increase radiation dose to patients and staff. J Bone Miner Res 1996;11(Suppl 1):S157. 32. Patel R, Blake GM, Batchelor S, Fogelman I. Occupational dose to the radiographer in dual-X-ray absorptiometry: a comparison of pencil-beam and fan-beam systems. Br J Radiol 1996;69:539–43. 33. Lilley J, Walters BG, Health DA, Drolc Z. In vivo and in vitro precision for bone density measured by dual-energy X-ray absorption. Osteoporosis Int 1991;1:141–6. 34. Hassager C, Jensen B, Gotfredsen A, Christiansen C. The impact of measurement errors on the diagnostic value of bone mass measurements: Theoretical considerations. Osteoporosis Int 1991;1:250–6.