Objective: In vitro and in vivo comparisons of bone mineral density (BMD) and body composition between GE/Lunar pencil (DPXL) and fan-beam (PRODIGY) absorptiometers.
Design: Comparison of BMD, bone mineral content (BMC) and area of lumbar spine (L2–L4), femoral neck and total body. Total body composition compartments tissue (TBTissue), fat (TBF), lean tissue (TBLean) and %TBF were also compared.
Setting: Centre for Bone and Body Composition Research, University of Leeds.
Phantoms/subjects: A range of spine phantoms, a variable composition phantom (VCP) and total body phantom. A total of 72 subjects were included for the in vivo study.
Results: In vitro: A small significant underestimation of BMD by the Prodigy compared to the DPXL ranging from 0.7 to 2% (p<0.05–0.001) for the spine phantoms. The Prodigy underestimated the VCP %Fat. Although the Prodigy underestimated phantom TBBMD by 1.1±1.0%, TBBMC and area were reduced by 8.2±1.4 and 7.3±1.0%, respectively. The Prodigy overestimated TBTissue 1508 g (2.2%), TBLean 588 g (1.2%), TBF 919 g (4.8%) and %TBF (0.8%).
In vivo: BMD cross-calibration was only required in the femoral neck, DPXLBMD=0.08+0.906*PRODIGYBMD. The Prodigy had higher estimates for TBTissue 1360 g (2.3%), TBLean 840 g (2.0%), TBF 519 g (3.4%), TBBMC 32.8 g (1.3%) and %TBF (0.3%). Cross-calibration equations were required for TBTissueDPXL=−1158+0.997*TBTissuePRODIGY and TBBMCDPXL= 89.7+0.949*TBBMCPRODIGY.
Conclusions: Small differences between the two absorptiometers for both BMD and body composition can be made compatible by use of cross-calibration equations and factors. The discrepancy in body composition compartments requires further research.
Pencil-beam dual-energy X-ray absorptiometry (DXA) is an established technique used for studies of bone mineral and body composition (Mazess et al, 1990). A technical development of DXA has been the introduction of fan-beam technology with resultant reduced scanning times and improved image resolution (Eiken et al, 1994,1994).
It has been verified that differences in bone mineral and body composition measurements occur when pencil-beam DXA machines from the same (Paton et al, 1995: Tataranni et al, 1996; Economos et al, 1997; Oldroyd et al, 1998b) and also from different manufacturers (Van Loan & Mayclin, 1992; Tothill et al, 1994) are compared. Recent studies have also shown similar differences when pencil- and fan-beam DXA machines from the same manufacturer are compared (Blake et al, 1993; Faulkner et al, 1993; Abrahamsen et al, 1995; Franck et al, 1995; Bouyoucef et al, 1996; Barthe et al, 1997; Ellis & Shypailo, 1998; Ruetsche et al, 2000; Kolta et al, 2000; Tothill et al, 2001).
We have recently replaced a GE/Lunar pencil-beam densitometer (DPXL) with a GE/Lunar fan-beam densitometer—the PRODIGY. It was therefore essential to perform cross-calibration studies between the two densitometers to determine if any systematic differences in calibration occur since different scanning technologies are used. The pencil beam utilizes a rectilinear scan to provide a projection of bone on a single detector. The fan beam has an array of multiple detectors and the projected area of bone imaged on the detector is magnified depending on the bone distance from the X-ray source and detector array. The Prodigy narrow angle fan beam eliminates beam distortion at the ends of the beam path and software corrects for magnification. Both the DPXL and Prodigy densitometers have the X-ray source located below and the detector(s) above the patient bed (Mazess et al, 2000; Mazess & Barden, 2000; Nord et al, 2000).
This study was initiated to examine if any intermachine differences occurred when a DPXL pencil beam absorptiometer was upgraded to a Prodigy fan-beam. Since reference data will have to be obtained on pencil beam densitometers and longitudinal studies need to be continued on the fan-beam densitometer, it is imperative to identify any differences between them. In vitro studies involved a range of semianthropometric spine phantoms, a variable composition phantom (VCP) and a total body phantom. Determination of short- and long-term in vitro precision, estimates of BMD, BMC, area and body composition compartments were compared between the two densitometers. Where data were available from the manufacturer for bone mineral density (BMD), bone mineral content (BMC), and area of their phantoms these data were compared to the values obtained from both densitometers. Comparisons of short-term precision and estimations of bone and body composition between the two densitometers were repeated in vivo.
Materials and methods
The GE/Lunar PRODIGY densitometer utilizes a narrow fan beam (4.5°) orientated parallel to the longitudinal axis of the body. This is in contrast to previous densitometers where the fan beam was orientated perpendicular to the longitudinal axis of the body with fan beam angles of 30° (Hologic QDR 4500) and 12°(GE/Lunar Expert). A dual-energy X-ray source with peak X-ray energy of 80 kVp and a current of 3 mA with a K-edge filter (cerium 300 mg/cm2) gives effective energies of approximately 38 and 70 keV. The detector consists of an array of energy-sensitive cadmium zinc telluride (CZT) detectors of length 5 cm (16 elements each 3 mm wide) allowing rapid photon counting. Imaging is typically over a 24 mm length with longitudinal steps of 17 mm. Pixel size for spine, femoral neck, and total body standard scan modes are 0.6 × 1.05, 0.6 × 1.05 and 4.8 × 13.0 mm compared to 1.2 × 1.2, 1.2 × 1.2 and 4.8 × 9.6 mm for the DPXL. There is a magnification effect associated with fan-beam densitometers (Griffiths et al, 1997); to reduce this effect, a software correction is utilized with the Prodigy densitometer. Scan times for standard mode spine and femoral neck are reduced to 30 s and total body to 5 min compared to 8 and 20 min for medium scan mode on the DPXL. BMD measurements were made at the PASpine (L2–L4), dual femoral neck and total body. The femoral neck positioning device allows both legs to be abducted, inwardly rotated 25°, and scanned without repositioning. Software version 4.0 was used for analysis. Skin entrance surface dose for the standard scan modes for PASpine (L2–L4), femoral neck and total body were 37, 37 and 0.4 μGy, respectively, and effective doses <1 μSv for all modes (Mazess et al, 2000).
Pencil-beam measurements were made with a Lunar DPXL densitometer which utilizes a constant potential X-ray tube at 78 kVp and a K-edge fitter (cerium) to produce a congruent beam of stable X-rays with energies at 38 and 70 keV. BMD measurements were made at the PASpine (L2–L4), right femoral neck and total body. The femoral neck positioning device allows the leg to be abducted and then inwardly rotated by 25°. Analysis was made using Version 1.34 software. Extended research analysis was used for total body composition analysis. Skin entrance surface dose for the medium scan mode for PASpine (L2–L4), femoral neck, and total body on the DPXL were 6.1, 6.1, and <0.1 μGy, respectively, with effective doses <1 μSv in all modes.
To compare the magnification effect of the Prodigy, comparisons were made with a Lunar Expert (Ge/Lunar). This densitometer also utilizes a fan beam (12°). The X-ray source and detector array are mounted on a rotating C-arm and scans are made with the X-ray source above and the detectors below the patient (Lang et al, 1997).
The in vivo study groups consisted of both patients [anorexia (n=13), cystic fibrosis (n=23)] and normal subjects (n=36) (Table 1). Using both patients and controls enabled a wide range of bone and body composition values to be studied. The ranges determined on the DPXL were: PASpine (L2–L4)=0.709–1.556 g/cm2, femoral neck=0.652–1.154 g/cm2 and total body=0.887–1.349 g/cm2. Total body composition ranges were total body tissue (TBTissue)=29.3–95.6 kg, lean tissue (TBLean)=25.3–67 kg, body fat (TBF)=1.2–45.8 kg, %body fat (%TBF)=4–50%, bone mineral content (TBBMC)=1460–3545 g and DXA-derived body weight (DXAwt=TBTissue+TBBMC)=31.4–98.6 kg. The cross-comparison measurements were made for PASpine (n=47), right femoral neck (n=45) and total body (n=70). The study was approved by the Ethics Committee of the Leeds Teaching Hospitals Trust and informed consent was obtained from the study group.
For the cross-comparison of the two densitometers, in vitro measurements with various phantoms were also made. These phantoms consisted of the following:
Lunar aluminium spine
The Lunar spine phantom is an aluminium wedge construction, simulating four lumbar vertebrae, designed for use on Lunar densitometers. The phantom is scanned in 15 cm water simulating soft tissue (%Fat=4%). The manufacturer's quoted BMD value for the spine phantom (SN 9831) used in this study for L2–L4 region=1.243±0.037 g/cm2, the area calculated from the physical dimensions=41.02 cm2 and BMC=51.0 g (Mazess et al, 1991).
Hologic spine phantom
The Hologic spine phantom, designed for use on Hologic densitometers, is composed of four calcium hydroxyapatite vertebrae of uniform density embedded in an epoxy resin block simulating soft tissue (%Fat=70%). Estimated BMD, BMC and area from the manufacturers data for the three vertebrae representing L2–L4 were 0.978 g/cm2, 45.3 g and 46.3 cm2, respectively (Wahner et al, 1988; Blake et al, 1998).
European spine phantom (ESP)
This phantom was developed for cross-calibration of DXA devices (Kalender et al, 1995). It utilizes three semianthropomorphic calcium hydroxyapatite vertebrae, representing BMD values: low (0.5 g/cm2), medium (1.0 g/cm2), and high (1.5 g/cm2), embedded in epoxy resin simulating soft tissue (%Fat=9.0%). The manufacturer's calibrated BMD for the phantom (ESP-106) used in this study simulating L2–L4=1.000 g/cm2 with assigned values for BMC and area of 30.36 g and 30.36 cm2, respectively (Lees et al, 1997).
Leeds prototype spine phantom
The spine phantom is composed of five simulated vertebrae (L1–L5) constructed from epoxy resin and bone analogue grade hydroxyapatite. The phantom is scanned in 15 cm rice to simulate soft tissue (%Fat=24%) (Milner et al, 2000). Currently, the phantom is uncalibrated for BMD, BMC and area.
Bona fide spine phantom
This phantom is intended primarily for quality control of GE/Lunar densitometers. It is a calcium hydroxyapatite step wedge with curved sides embedded in acrylic simulating soft tissue (26%Fat). It has four vertebrae with BMD ranges of 0.6–1.4 g/cm2. The curved sides provide a test of edge detection algorithms in software analysis. Nominal value for BMD (L2–L4)=1.232±0.0025 g/cm2 (Nord et al, 1997).
Lunar variable composition phantom (VCP)
The phantom consists of four acrylic blocks, two thin PVC sheets and four thin vinyl sheets. The acrylic blocks simulate fat and PVC/vinyl sheet combinations lean tissue. Total dimensions=28 cm (length) × 20 cm (width) × 15 cm (thick). The sheets are used in various combinations to simulate five different soft tissue compositions. Standard scan mode is recommended for analysis since the VCP is calibrated using this mode. The various %Fat compositions for VCP-09 used in this study are (mean±s.d.): 44.4±0.9, 36.8±0.7, 28.1±0.6, 21.2±0.4 and 16.0±0.3. The phantom is scanned in the total body mode (Diessel et al, 2000).
Leeds total body phantom
The total body phantom is constructed with aluminium representing the skeleton and a number of elliptical/circular cross-section polythene containers filled with distilled water. Polythene represents fat and distilled water lean tissue. The aluminium skeleton is placed on the scanning bed and the polythene containers placed on top. The area and volume of the aluminium skeleton=1800 cm2 and 1310 cm3, respectively. Using the BMC value of 1.53 g/cm3 for aluminium (Mazess et al, 1991), the BMC of the phantom=2004 g and total body BMD=1.113 g/cm2. Nominal %TBF= 28% (Oldroyd et al, 1998a).
Assessment of short-term in vitro precisions (same day) was made for all phantoms. In vitro long-term precisions (1 y) were measured to monitor any systematic changes with time because of scanner drift, recalibrations or software changes.
Comparison of the two densitometers was made using linear regression analysis. The ideal parameters for comparison of regression are: slope=1, intercept=0, correlation coefficient (r)=1 and standard error of estimate (s.e.e.)=0. Cross-calibration data can be analysed by two methods. In the first method, equations are derived so that measurements on the ‘old’ densitometer can be converted to agree with the ‘new’ densitometer. In this case, the ‘new’ measurements are plotted as the dependent variable and is normally used for longitudinal studies that will be continued on the ‘new’ densitometer.
In the second method, used in this study, the ‘old’ measurements are plotted as the dependent variable. This correction is required if the calibration file on the ‘new’ densitometer is to be edited to force agreement with the in vivo BMD calibration of the old scanner. This step is used by manufacturers to ensure continuity of BMD calibration on successive generations of equipment; in particular, fan and pencil beam densitometers. The statistical error of the intercept of each regression line is evaluated and if the 95% CI includes zero, the regression analysis is repeated with the intercept forced through the origin. The slope and the 95% CI of the regression line through the origin for each scan site should be examined to determine if it is significantly different from unity. If the slope is significantly different from unity, then BMD results on the ‘old’ densitometer should be multiplied by this factor to put a consistent BMD scale with the ‘new’ system (Blake et al, 1998).
To determine the compatibility of the pencil and fan-beam absorptiometers, the differences between the two devices were also analysed using Bland and Altman analysis (1986). In this analysis, the difference between two measurements are plotted as the dependent variable against the mean of the two measurements plotted as the independent variable. The mean difference (bias), 95% CI for the difference (±2 s.e. of the mean difference), and limits of agreement (±2 s.d. of mean difference) are calculated. The mean difference is considered to be significant if the 95% CI of the mean difference does not include zero. Correlation of the differences and mean values is calculated to determine if the mean difference is dependent on the magnitude of the measurements. In this study, the DPXL was considered to be the reference densitometer and the difference to be Prodigy−DPXL.
Short-term in vitro precision measurements were made by scanning the various phantoms 10 times, with repositioning between the measurements, on the same day. In vitro precision was expressed as
In vivo precision was determined by scanning 10 subjects twice on the same day, with repositioning between scans. For in vivo precision, the error standard deviation (s.d.) was derived from the following equation:
where n is the number of patients and xi and yi are paired measurements for i=1 to n.
The %CV is derived from the equation:
Paired Student's t-tests were applied to test for significant differences between bone and body composition compartments between the two densitometers.
Statistical analysis was performed using ‘Analyse-it’, the statistical software package with Microsoft Excel 97. In this study, a P-value of <0.05 was considered significant.
The comparison between the fan and pencil-beam densitometers for the various spine phantoms used in the study is shown in Table 2. Compared to the Lunar phantom values the Prodigy underestimated BMD (1.1%) and overestimated BMC (0.7%) and area (1.6%), while the DPXL overestimated BMD (1.1%), BMC (1.6%) and area (0.4%). For the ESP phantom, a BMD overestimation of 14.1 and 13.8% occurred with the Prodigy and DPXL, with area underestimations of 5.5 and 8.1% and BMC overestimations of 8 and 4.7%, respectively. Both densitometers overestimated BMD of the Hologic phantom by approximately 25%, which may be because of differences in edge detection algorithms and calibration of Hologic and Lunar densitometers. Both the Prodigy and DPXL overestimated the nominal BMD of the Bona fide phantom by 1.3 and 0.4%, respectively. Compared to the DPXL, the Prodigy underestimated the Leeds spine phantom by 0.011 g/cm2 (1.2%). All in vitro short-term precision (%CV) were below 1%. Long-term in vitro (1.0 y) BMD precisions for the Lunar, Hologic, and ESP phantom had similar values to the short-term precision for both densitometers. No systematic trend of BMD with time was observed indicating the stability of the densitometer.
The comparison between the two densitometers for the five different soft tissue compositions of the VCP are given in Table 3a. For each composition, the fan beam is significantly lower (0.5–1.2%) than the pencil beam except for the nominal 44.4% fat composition (P=0.07). The comparison of in vitro precision (%CV) between the densitometers shows comparable values except for the nominal 16% fat composition. The fan-beam densitometer measurements for %Fat showed the closest agreement with the nominal %Fat values.
Long-term precision (1.0 y) showed an increase for all %Fat values for both fan and pencil beam. The fan beam had the highest values ranging from 2.9 to 10.3%, corresponding pencil-beam values ranged between 1.5 and 4.0%. All precision values increased with reducing %Fat values.
The comparisons of bone and body composition compartments for the total body phantom are shown in Table 3b. Compared to the estimated values, the Prodigy accurately measured TBBMD, but underestimations of 3.5% occurred for both TBBMC and area. In contrast, the DPXL overestimated TBBMD (1%), TBBMC (5.1%) and area (4.0%). The body composition compartments were all significantly higher on the Prodigy with a highly significant increase in TBTissue=1508 g (2.2%). The Prodigy yielded significantly higher values for TBLean (588±657 g, P<0.05) and TBF (919±732 g, P<0.005), but only a 0.8±1.0% increase in %TBF. In vitro short-term precision values were all below 1% for the bone compartments, and for the body composition compartments precision values were highest for the fat compartments. Long-term (1.0 y) precision values for TBBMD and %TBF for the Prodigy were 0.8 and 2.5% compared to 0.5 and 4.3% for the DPXL with no systematic trends with time.
To observe the effect of magnification of the Prodigy, a comparison was made with a Lunar Expert densitometer. The VCP was utilized with the nominal 24% fat combination, that is (4) acrylic blocks, (2) vinyl and (2) PVC sheets. The aluminium spine phantom was then placed at the bottom and top of the VCP phantom and also between the acrylic blocks so that the height of the aluminium spine phantom could be varied from 0, 5, 10, 15 and 20 cm above the bed. When the results were normalized to the 10 cm height above the bed, the per cent variations observed for the Prodigy and Expert were BMD: 0.02±0.6/0.3±0.6%; BMC: −0.1±0.9/0.7±7.9% and area: −0.1±0.2/0.7±8.1%, respectively (Table 4). The magnification effect observed with the Prodigy on BMC and area resulted in only small variations compared with the Expert.
The in vivo short-term precision values determined for the Prodigy fan beam for both bone and body composition show an improved precision compared to the pencil-beam precision determined in a previous study (Woodrow et al, 1996) (Table 5).
Comparison of the fan/pencil-beam densitometers results for the lumbar spine, femoral neck, and total body BMD shows no significant differences. However, comparisons of BMC and area for the femoral neck and total body show significantly higher values for the fan beam densitometer of 0.09 g (2%), 0.08 cm2 (2%) and 32.8 g (1.4%), 30 cm2 (1.4%), respectively (Table 6).
The results of the linear regression analysis are summarized in Table 7. The pencil- and fan-beam measurements were highly correlated (r2=0.94–0.98) for the three sites. Only for the femoral neck was the intercept significantly different from zero (P<0.05) and the slope significantly different from one (P<0.05). For all three sites, the %s.e.e. results (1.1–3.2%) were comparable to the short-term precision (1.1–1.7%) results.
When the regression analysis is made with the intercept forced through zero, the slope of the regression lines changed from 0.968 to 0.998 for the PASpine and 0.994 to 1.001 for total-body BMD.
The Bland–Altman analysis of the three sites indicated no significant offsets. The largest individual differences were observed at the femoral neck (maximum 6.7%). Only for the femoral neck BMD did the correlation between the differences and the mean value approach significance (P=0.07) (Table 9).
Comparison of body composition compartments indicated that the fan beam was significantly higher for all compartments, ranging from 1360 g (2.3%) for TBTissue to 0.3% (1.2%) for %TBF (Table 8). When compared with scale measured body weight, the pencil-beam-derived DXAwt was significantly lower, 0.45 kg (P<0.0001) and fan-beam-derived DXAwt significantly higher, 0.94 kg (P<0.0001).
Linear regression analysis of the compartments indicated that only TBTissue and TBBMC had an intercept significantly different from zero and TBBMC also had a slope significantly different from one. Similar to the BMD regression analysis, the %SEE (0.7–5.7%) were comparable to the short-term precision (0.3–2.9%) results (Table 7).
Forcing the regression lines for TBLean, TBF, %TBF through zero resulted in the following correction factors:
The Bland–Altman analysis for the body composition compartments indicated significant offsets for all compartments (Table 9).
The feasibility of using the ESP for PASpine (L2–L4) cross-calibration between the two densitometers was determined by estimating the mean BMD values of the individual simulated vertebrae. They were estimated from the 10 measurements used in the short-term precision study. The Prodigy BMD values were 0.614, 1.087 and 1.626 compared to 0.611, 1.098 and 1.598 g/cm2 for the DPX-L. Differences were 0.003 g/cm2 (0.5%), 0.011 g/cm2 (1.0%) and 0.028 g/cm2 (1.7%), respectively. The in vitro and in vivo regressions were compared and an in vitro cross-calibration equation was derived from the ESP (Figure 1).
Since the intercept was not significantly different from zero, the regression was forced through zero to obtain a correction factor.
This factor is in close agreement to the in vivo factor of 0.998.
Similarly, an in vitro cross-calibration equation was derived for %Fat using the five %Fat configurations of the VCP (Figure 2).
Since the intercept was not significantly different from zero, the intercept was forced through zero to obtain a correction factor.
The VCP-derived %Fat correction factor (1.024) shows a different relationship between the two densitometers than that observed in vivo (0.986) because of the pencil beam measuring higher %Fat values for all VCP configurations compared to the fan beam.
In the in vitro spine phantoms study, the BMD comparisons showed that the fan-beam densitometer tended to measure slightly lower values of BMD than the pencil beam; for the semi-anthropometric phantoms these differences were all within 1%. The largest difference occurs with the aluminium spine phantom (2%). Neither densitometer showed an improvement over the other in measuring known area values of the phantoms.
No significant differences in BMD were observed for the in vivo lumbar spine measurement site and no magnitude dependence was observed. The ESP-derived in vitro cross-calibration factor=0.991 was in close agreement with the in vivo cross calibration factor=0.998, confirming the feasibility of using this phantom for cross-calibration of absorptiometers as recommended by Genant et al (1994). In the femoral neck in vivo comparison, although the Prodigy measured significantly higher values of BMC and area, there was no significant difference in BMD. The femoral neck was the only site with an intercept significantly different from zero and slope significantly different from one, and was also the only site to indicate a small magnitude dependence on BMD. The femoral neck has inferior precision values (1.5 and 1.7%) compared to the other sites, reflecting the difficulty in obtaining identical rotation of the femoral neck for measurement comparison. Additionally, in this in vivo study, different positioning devices were used for the pencil- and fan-beam measurements which may have introduced a measurement error.
The total body phantom, with aluminium representing bone, did not reflect the differences observed in vivo. The in vitro results indicated that the Prodigy measured significantly lower values of TBBMD (1%), TBBMC (8.2%) and area (7%) compared to the DPXL. However, in the in vivo comparison, although there was no significant difference in TBBMD, the Prodigy measured significantly higher values of TBBMC (1.3%) and area (1.4%). However, in a previous study using the total body phantom to compare the DPXL with a DPX, only small differences were observed in BMD (0.8%), BMC (0.7%) and area (1.5%) (Oldroyd et al, 1998b). It was also observed with the aluminium spine phantom that the Prodigy measured significantly lower values of BMC and area compared to the DPXL. This may reflect a difference in the way aluminium and bone are analysed by the densitometers. The TBBMD in vivo study indicated no significant differences and no magnitude dependence.
The results of the BMD comparisons at the three sites are similar to those observed by Mazess et al (2000), who reported no significant offsets or magnitude dependence when the Prodigy was compared with a DPX-IQ. However, differences in the slopes of the regression equations were observed between the two studies: DPXL:DPX-IQ vs Prodigy, lumbar spine 0.968:1.003, femoral neck 0.906:1.023 and total body 0.993:1.000.
The comparison of body composition parameters from the total body phantom indicated higher values for the Prodigy-derived TBTissue, 1.5 kg (2.1%), %TBF 0.8%, TBLean 690 g (1.4%) and TBF 930 g (4.9%) compared to 1.36 kg (2.3%), 0.3%, 840 g (2%) and 520 g (3.4%) observed in vivo. The largest difference observed in the in vivo body composition comparison was in the higher estimation of TBTissue with the Prodigy. These differences were not related to the magnitude of the parameter.
Although the mean difference for %TBF was only 0.3% because of the higher value of TBTissue, significantly higher values of TBLean and TBF were observed. Therefore, the derived cross-calibration equations and factors should be applied for longitudinal body composition studies continued on the Prodigy. The difference observed in TBTissue in this study was not observed by Mazess et al (2000), whose values for Prodigy TBTissue and %TBF were 67.3 kg and 35.3% compared with DPX-IQ values of 67.0 kg and 35.5%, indicating that the Mazess patient population was heavier with a larger fat mass than the patients in this study. Further research is necessary to determine if the disparity of TBTissue observed in this study is due to a machine miscalibration or a population effect.
The use of the VCP to derive an in vitro cross-calibration factor for %TBF resulted in a different factor being derived: 1.024 than that observed in vivo: 0.986. This was because of the DPXL measuring higher values for %Fat for all five combinations of the VCP. One factor that may contribute to this is that standard analysis mode should be used for the DPXL analysis of the VCP, as recommended by the manufacturers. In light of these findings, we would recommend the use of the in vivo correction factor derived in equation (6) above. Further work needs to be carried out in order to make the VCP more reflective of these in vivo findings. In this study, the extended research analysis mode was used as this mode is incorporated in the Prodigy analysis mode.
The total body phantom demonstrated a different relation for bone than observed in vivo; however, soft tissue analysis was comparable to that of the subject group. The inability of the VCP to reflect the in vivo analysis of %TBF indicates that both of these phantoms may only be useful in the cross-calibration of densitometers of the same design and manufacturer.
In conclusion, no correction factors need to be applied for lumbar spine and total body BMD; however, a small correction is required for femoral neck BMD. Hence, reference data accumulated on pencil-beam densitometers can be used as reference data for the Prodigy. Body composition correction factors need to be applied until the difference observed in TBTissue can be clarified. Futher cross-calibration studies are required to evaluate a comparison of the paediatric measurement modes between the two densitometers.
Abrahamsen B, Gram J, Hansen TB & Beck-Nielsen H (1995): Cross calibration of QDR-2000 and QDR-1000 dual-energy X-ray densitometers for bone mineral and soft-tissue measurements. Bone 16, 385–390.
Barthe N, Braillon P, Ducassou D & Basse-Cathalinat B (1997): Comparison of two Hologic DXA systems (QDR 1000 and QDR 4500/A). Br. J. Radiol. 70, 728–739.
Blake GM, Parker JC, Buxton FMA & Fogelman I (1993). Dual X-ray absorptiometry: a comparison between fan beam and pencil beam scans. Br. J. Radiol. 66, 902–906.
Blake GM, Whaner HW & Fogelman I (1998): The Evaluation of Osteoporosis: Dual Energy X-ray Absorptiometry and Ultrasound in Clinical Practice, 2nd Edition. Martin Dunitz, London.
Bland JM & Altman DG (1986): Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1, 307–310.
Bouyoucef SE, Cullum ID & Ell PJ (1996): Cross-calibration of a fan-beam X-ray densitometer with a pencil-beam system. Br. J. Radiol. 69, 522–531.
Diessel E, Fuerst T, Njeh CF, Tylavsky F, Cauley J, Dockrell M & Genant HK (2000): Evaluation of a new body composition phantom for quality control and cross-calibration of DXA devices. J. Appl. Physiol. 89, 599–605.
Economos CD, Nelson ME, Fiatarone MA, Dallal GE, Hemsfield SB, Wang J, Yasumara S, Ma R, Vaswani AN, Russell-Aulet M & Pierson RM (1997): A multi-center comparison of dual energy X-ray absorptiometers: In vivo and in vitro soft tissue measurement. Eur. J. Clin. Nutr. 51, 312–317.
Eiken P, Kolthoff N, Barenholdt O, Hermansen F & Pors NS (1994): Switching from DXA pencil-beam to fan-beam. II: studies in vivo. Bone 15, 667–670.
Eiken P, Barenholdt O, Bjorn JL, Gram J & Pors NS (1994): Switching from DXA pencil-beam to fan-beam. I: studies in vitro at four centres. Bone 15, 671–676.
Ellis KJ & Shypailo RJ (1998): Bone mineral and body composition measurements: cross-calibration of pencil-beam and fan-beam dual-energy X-ray absorptiometers. J. Bone Miner. Res. 13, 1613–1619.
Faulkner KG, Gluer C-C, Estilo M & Genant HK (1993): Cross-calibration of DXA equipment: upgrading from a Hologic QDR 1000/W to a QDR 2000. Calcif. Tissue Int. 52, 79–84.
Franck H, Munz M & Scherrer M (1995): Evaluation of dual-energy X-ray absorptiometry bone mineral measurement — comparison of a single-beam and fan-beam design: the effect of osteophytic calcification on spine bone mineral density. Calcif. Tissue Int. 56: 192–195.
Genant HK, Grampp S, Gluer CC, Faulkner KG, Jergas M, Engelke K, Hagiwara S & Van Kuijk (1994): Universal standardization for dual X-ray absorptiometry: patient and phantom cross-calibration results. J. Bone Miner. Res. 9, 1503–1514.
Griffiths MR, Noakes KA & Pocock NA (1997): Correcting the magnification error of fan beam densitometers. J. Bone Miner. Res. 12, 119–123.
Kalender WA, Felsenberg D, Genant HK, Fischer M, Dequeker J & Reeve J (1995): The European spine phantom—a tool for standardization and quality control in spinal bonemineral measurements by DXA and QCT. Eur. J. Radiol. 20, 83–92.
Kolta S, Ravaud P, Fechtenbaum J, Dougados M & Roux C (2000): Follow-up of individual patients on two DXA scanners of the same manufacturer. Osteoporos Int 11, 709–713.
Lang T, Takada M, Gee R, Wu C, Li J, Hayashi-Clark, Schoen S, March V & Genant HK (1997): A preliminary evaluation of the Lunar Expert-XL for bone densitometry and vertebral morphometry. J. Bone Miner. Res. 12, 136–143.
Lees B, Garland SW, Walton C & Stevenson JC (1997): Evalution of the European spine phantom in a multi-centre clinical trial. Osteoporos. Int. 7, 570–574.
Mazess RB & Barden HS (2000): Evaluation of differences between fan-beam and pencil-beam densitometers. Calcif. Tissue Int. 67, 291–296.
Mazess RB, Barden HS, Bisek JP & Hanson J . (1990): Dual-energy X-ray absorptiometry for total-body and regional bone-mineral and soft-tissue composition. Am. J. Clin. Nutr. 51, 1106–1112.
Mazess RB, Trempe JA, Bisek JP, Hanson JA & Hans D (1991): Calibration of dual-enery X-ray absorptiometry for bone density. J. Bone Miner. Res. 6, 799–806.
Mazess RB, Hanson JA, Payne R, Nord R & Wilson M (2000): Axial and total-body bone densitometry using a narrow-angle fan-beam. Osteoporos. Int. 11, 158–166.
Milner R, Berry E, Marsden AJ, Smith AH & Smith MA (2000): Anthropomorphic quality assurance phantoms for bone densitometry. 14th International Bone Densitometry Workshop, Warnemünde, Germany. Osteoporos. Int. 11 (Suppl 3), S27.
Nord RH, Bisek JP & Miller CG (1997): A new hydroxyapatite spine phantom. 12th International Bone Densitometry Workshop, Crieff, Scotland. Osteoporos. Int. 7: 287.
Nord RH, Homuth JR, Hanson JA & Mazess RB (2000): Evaluation of a new DXA fan-beam instrument for measuring body composition. Ann. N. Y. Acad. Sci. 904, 118–125.
Oldroyd B, Milner R, Smith AH & Smith MA (1998a): A total body phantom for use with Lunar dual-energy X-ray absorptiometers. Appl. Radiat. Isotopes. 49, 525–526.
Oldroyd B, Truscott JG, Woodrow G, Milner R, Stewart SP, Smith AH, Westmacott CF & Smith MA (1998b). Comparison of in vivo body composition using two Lunar dual-energy X-ray absorptiometers. Eur. J. Clin. Nutr 52, 180–185.
Paton NIJ, Macallan DC, Jebb SA, Panzianas M & Griffin GE (1995): Dual-energy X-ray absorptiometry results differ between machines. Lancet 346, 899–900.
Ruetsche AG, Lippuner K, Jaeger P & Casez J-P (2000): Differences between dual X-ray absorptiometry using pencil beam and fan beam modes and their determinants in vivo and in vitro. J. Clin. Densit. 3, 157–166.
Tataranni PA, Pettit DJ & Ravussin E (1996): Dual energy X-ray absorptiometry: inter-machine variability. Int. J Obes. Relat. Metab. Disord. 20, 1048–1050.
Tothill P, Avenell A, Love J & Reid DM (1994): Comparisons between Hologic, Lunar and Norland dual-energy X-ray absorptiometers and other techniques used for whole-body soft tissue measurements. Eur. J. Clin. Nutr. 48, 781–794.
Tothill P, Hannan WJ & Wilkinson S (2001): Comparisons between a pencil beam and two fan beam dual energy X-ray absorptiometers used for measuring total body bone and soft tissue. Br. J. Radiol. 74: 166–176.
Van Loan MD & Mayclin PL (1992): Body composition assessment: dual energy X-ray absorptiometry (DEXA) compared to reference methods. Eur. J. Clin. Nutr. 46, 125–130.
Wahner HW, Dunn WL, Brown ML, Morin RL & Riggs L (1988): Comparison of dual-energy X-ray absorptiometry and dual photon absorptiometry for bone mineral measurements of the lumbar spine. Mayo Clin. Proc. 63, 1075–1084.
Woodrow G, Oldroyd B, Turney JH & Smith MA (1996): Influence of changes in peritoneal fluid on body composition measurements by dual-energy X-ray absorptiometry in patients receiving continuous ambulatory peritoneal dialysis. Am. J. Clin. Nutr. 64, 237–241.
We thank Dr Belinda Lees (Royal Brompton Hospital, London) for the loan of the ESP and Hologic phantoms, Dr Derek Pearson (Nottingham City Hospital) for the loan of the Bona Fide phantom, and Dr Russ Nord (GE/Lunar) for the loan of the VCP.
About this article
Cite this article
Oldroyd, B., Smith, A. & Truscott, J. Cross-calibration of GE/Lunar pencil and fan-beam dual energy densitometers—bone mineral density and body composition studies. Eur J Clin Nutr 57, 977–987 (2003). https://doi.org/10.1038/sj.ejcn.1601633
- dual-energy X-ray absorptiometers
- pencil beam
- fan beam
- body composition
- bone mineral
Comparison of obesity and metabolic syndrome prevalence using fat mass index, body mass index and percentage body fat
PLOS ONE (2021)
Does Bone Mineral Density Differ between Fan-Beam and Pencil-Beam?: A Meta-Analysis and Systemic Review
Journal of Bone Metabolism (2021)
Precision of the Hologic Horizon A dual energy X-ray absorptiometry in the assessment of body composition
Obesity Research & Clinical Practice (2020)
Veterinary Clinics of North America: Exotic Animal Practice (2019)