This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rea, J A Articles by Fogelman, I Search for Related Content PubMed PubMed Citation Articles by Rea, J A Articles by Fogelman, I

Development of a phantom for morphometric X-ray absorptiometry

Osteoporosis Screening and Research Unit, 16th Floor Guy's Tower, Guy's Hospital, St Thomas' Street, London SE1 9RT, UK

	Abstract

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Morphometric X-ray absorptiometry (MXA) has recently been developed to assess vertebral deformity status using dual energy X-ray absorptiometry (DXA) machines. In contrast to bone densitometry, a vertebral morphometry phantom is not supplied by any machine manufacturer. The aim of this study was to develop a suitable phantom to quantify the accuracy and precision of the vertebral measurement software on three DXA scanners in vitro and to perform a weekly quality control (QC) scan over a 30-month period to evaluate any drift or changes in measurement accuracy over time. The phantom was constructed from Perspex and aluminium to simulate soft tissue and bone, respectively. 13 aluminium rectangles (each 30 mm wide, 25 mm high and 3 mm thick, with edges ("endplates") 6 mm thick) were set into one side of a solid Perspex block to represent the vertebral bodies from the fourth thoracic (T4) to the fourth lumbar (L4). The phantom was scanned on both the Hologic QDR2000plus and the QDR-4500A as well as the Lunar Expert-XL. Three consecutive lateral MXA scans were acquired on the Hologic machines using each of the scan modes available. On the QDR-2000plus, the lateral scan modes available are fast, array and high definition, which are all dual energy modes. These three scan modes are also available on the QDR-4500A, with the addition of a single energy scan mode. Four lateral scans were acquired on the Expert-XL machine using the single scan mode available. Each MXA scan was analysed twice by a trained operator using the standard software supplied by each manufacturer. A QC scan was performed approximately weekly over a 30-month period on only the QDR-4500A machine, and total phantom height was measured from the inferior edge of L4 to the superior edge of T4. Accuracy of "vertebral" height measurement varied between the three DXA machines and between the scan modes available. All underestimated "true" vertebral height by between 0.4% and 8.6%, with the scan modes using finer collimation producing the most accurate results. Repeat analysis precision of vertebral height measurement was best on the QDR-4500A, followed by the Expert-XL, and was poorest on the QDR-2000plus. The QC scans acquired on the QDR-4500A suggested that it was a highly stable machine, little affected by even major repairs. It must be remembered that these in vitro phantom results may not be representative of the true in vivo situation. The MXA phantom appears to be a useful tool for documenting the stability of the mechanical instruments and for checking the long-term consistency of operator precision.

	Introduction

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Prevalent vertebral deformities are a strong independent predictor of a significantly increased risk of further vertebral and non-vertebral fracture [1]. Recently, a number of studies have highlighted the potential of using morphometric X-ray absorptiometry (MXA) scans, acquired on dual energy X-ray absorptiometry (DXA) machines used conventionally to measure bone density, to identify vertebral deformities in patients affected by osteoporosis [2–7]. In contrast to bone densitometry [8], a vertebral morphometry phantom is not supplied as standard by any DXA machine manufacturer to perform long-term quality control measurements or to evaluate the accuracy of the morphometric software installed. As the identification of vertebral deformities via morphometric techniques is totally reliant on accurate, reproducible measurement of vertebral heights from the lateral aspect, it would seem judicious to develop such a phantom. Although several studies have used a selection of phantoms constructed from either excised human vertebrae [9, 10], synthetic representations of vertebrae [9, 11–13] or various combinations of Perspex and metal [13–19], to assess the accuracy and precision of morphometric measurement on various DXA machines and on conventional radiographs, none has investigated long-term quality control (QC).

Traditionally, vertebral deformities have been identified on conventional lateral radiographs of the thoracolumbar spine [20–23]. However, conventional radiographic images are affected by inherent problems, such as the magnification and distortion associated with the use of a cone beam and the exposure of the subject to a relatively large radiation dose [24]. MXA performed using DXA machines has been developed as an alternative method of acquiring lateral images of thespine required for vertebral deformity evaluation [5, 7]. MXA has a number of advantages compared with conventional radiography. These include a significantly lower radiation dose to the patient, acquisition of a single image of the whole spine, and straightforward supine patient positioning [25]. However, there is more limited control of exposure factors during acquisition of the MXA scan, and image quality is inferior to that of the high resolution and low noise conventional radiograph.

The purpose of this study was to construct a phantom that could be used: (i) to quantify the accuracy of the vertebral morphometry software on three models of DXA machine from two different manufacturers; (ii) to compare the interobserver and intraobserver precision of point placement using the different software approaches of the two DXA manufacturers; and (iii) to perform QC MXA scans, over a 30-month period, on a single DXA machine to assess any drift or change in measurement accuracy as a result of services or repairs.

	Methods and materials

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Phantom construction
The phantom was constructed from Perspex and aluminium to simulate soft tissue and bone, respectively. A solid oblong block of Perspex (455 mmx50 mmx50 mm) was fixed to two small feet 20 mm high (Figure 1). 13 aluminium rectangles were set into one side of this block to represent the vertebral bodies from the fourth lumber (L4) to the fourth thoracic (T4). These "vertebrae" were a consistent size of 30 mm wide, 25 mm high (in terms of the lateral scan image) and 3 mm thick, with edges ("endplates") 6 mm thick, which is approximately equivalent to a bone mineral density value of 0.75 g cm^-2. The endplates were designed to optimize image quality at the edge of each vertebra while minimizing the amount of aluminium required to construct the phantom. Each "vertebra" was separated from its neighbour(s) by a 5 mm gap to represent the intervertebral disc space. Total phantom height (H_total), i.e. the distance from the inferior endplate of L4 to the superior endplate of T4, was 385 mm. A strip of aluminium (3 mm thick, 20 mm wide) was set into the upper surface of the phantom along its entire length to facilitate the acquisition of a posteroanterior (PA) centreline scan on the Hologic machines.

View larger version (44K): [in this window] [in a new window]   Figure 1. Morphometric X-ray absortiometry phantom construction. (a) Overview of phantom shape. (b)Top view of phantom. (c) Spacing and size of "vertebrae". (d) Side view of a phantom "vertebra" constructed from 3 mm thick aluminium with 6 mm thick raised "endplates".

The phantom was positioned longitudinally in the centre of the scan table of each machine, according to the positioning laser on the Hologic machines and by manual measurement on the Lunar machine. A PA centreline scan was performed on each Hologic machine, which confirmed that the phantom was positioned vertically (0.0°) on the scan image (Figure 2). Information from the centreline scan was used by the machine to maintain a constant distance between the spine and the X-ray tube during acquisition of the lateral image, thus correcting for the magnification effect of the fan beam, and to set the start point for the subsequent lateral scan. Three consecutive lateral scans were then acquired using each of the available lateral MXA scan modes on each Hologic machine: fast (F), array (A) and high definition (HD) on the QDR-2000plus; and single energy (SE), F, A and HD on the QDR-4500A. In all cases the scans were acquired without repositioning the phantom. In addition, a second set of three consecutive SE scans was acquired on the QDR-4500A only, 2 weeks after the initial set.

View larger version (92K): [in this window] [in a new window]   Figure 2. Morphometric X-ray absortiometry scans of the phantom acquired on the Hologic QDR-4500A machine. (A) Posteroanterior centreline scan. (b) Lateral single energy scan.

A PA scan was not required on the Lunar machine; thus, four lateral scans were acquired using the default MXA scan mode. To produce an analysable scan image, a water-bath (22.5 cm deep) was placed between the phantom and the detectors to attenuate the X-rays, as in previous phantom studies on this machine [10]. Owing to the restricted length of the water-bath compared with the phantom, only 10 of the 13 "vertebrae" could be imaged successfully on each lateral scan and therefore four lateral scans instead of the three acquired on the Hologic machines were performed to provide a comparable total number of "vertebrae" (40 vs 39).

Scan analysis
Each MXA scan was analysed using the standard software supplied by each manufacturer for the appropriate DXA machine. Six points were placed along the superior and inferior edges of each phantom "vertebra", to characterize its shape. The lowest aluminium block on the scan was labelled as L4 and the highest was labelled as T4 on the Hologic machines and T7 on the Expert-XL owing to the restricted length of the analysable image. From these six points, three vertebral heights were measured on each "vertebral body", nominally termed anterior height; mid height and posterior height. On both the standard and the QC QDR-4500A scans, the total height of the phantom (H_total) was measured from the inferior edge of the lowest block (equivalent to L4) to the superior edge of the highest block on the scan (equivalent to T4). Three heights were measured, a total anterior height, a mid height and a posterior height and these were averaged to produce the final H_total.

Data analysis
Accuracy
Each lateral MXA scan was analysed independently. The mean vertebral height for each vertebral level using each scan mode on each DXA machine as well as the overall mean vertebral height using all vertebral levels for the scan mode/machine were calculated. These were compared with the known "true" height (25 mm) of each phantom "vertebra".

Precision
To assess intraobserver precision on the QDR-4500A, each phantom MXA scan was independently analysed twice by a single experienced MXA operator (observer 1). Interobserver precision was calculated as the difference between the first analysis of observer 1 and an analysis by a second observer (observer 2) who was a trained DXA operator; this was only calculated for the scan mode selected for the QC element of the study. Similarly, the four MXA scans acquired on the Expert-XL were analysed twice by observer 1 to evaluate intraobserver precision and the scans were also analysed independently by a trained Lunar DXA operator (observer 3) to evaluate interobserver precision. The MXA scans acquired on the QDR-2000plus were analysed twice, again by observer 1, to evaluate intraobserver precision. However, no second observer was available at the appropriate juncture to perform the analysis required for interobserver precision on this machine. Both intraobserver and interobserver precision were calculated as the root-mean-square standard deviation (rms SD in mm) and the coefficient of variation (CV in %) [28].

Quality control
A single MXA scan was acquired using the QDR-4500A on an approximately weekly basis for a period of 30 months. The total height of the phantom H_total was evaluated each week. Three heights were measured on each scan, a total anterior height, a mid height and a posterior height and these were averaged to produce the final H_total for each week. Dates of any maintenance on the machine by engineers (repairs and services) were noted, as well as any software updates. Visual evaluation and linear regression were used to look for any trends in the measurements.

	Results

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Accuracy of vertebral height measurement
As a result of the different acquisition parameters used by each DXA machine and scan mode,the images of the phantom "vertebra" varied, with the best image resolution being observed using the SE scan mode on the QDR-4500A (Figure 3). It was also obvious that the Expert-XL image demonstrated some lateral distortion of vertebral shape compared with the QDR-4500A images. The accuracy of vertebral height measurement varied between the different machines and scan modes overall (Table 1) and when split to the individual vertebral levels (Figure 4). The Expert-XL machine and the SE and HD scan modes on the QDR-4500A machine demonstrated the highest measurement accuracy. However, the accuracy of vertebral height measurement on the Expert-XL fluctuated at different vertebral levels, with an obvious decrease in mean vertebral height measured from the equivalent of T10 upwards, resulting in a 2.0% decrease in the vertebral height measured at T7 compared with L4. Although all the scan modes on both Hologic machines demonstrated a more consistent level of accuracy over the entire spine, there were obvious differences between the scan modes. The SE and HD scan modes on the QDR-4500A, which utilize a 0.5 mm collimator, demonstrated superior accuracy compared with the F and A scan modes on the same machine and the F, A and HD scan modes on the QDR-2000plus, all of which utilize a 1.0 mm collimator. All scan modes on both Hologic machines underestimated "true" vertebral height by between 2.1% and 8.6%.

View larger version (64K): [in this window] [in a new window]   Figure 3. Images of phantom "vertebral bodies" acquired using (A) the single energy scan mode on the QDR-4500A, (B) the fast scan mode on the QDR-4500A, (C) the array scan mode on the QDR-4500A, (D) the high definition scan mode on the QDR-4500A and (E) the Expert-XL.

View larger version (30K): [in this window] [in a new window]   Figure 4. Accuracy of phantom "vertebral" height measurement compared with the true height of 25 mm using three different dual energy X-ray absorptiometry machines and various morphometric X-ray absortiometry scan modes. (a) QDR-2000plus. (b) QDR-4500A. (c) Expert-XL. Scan modes: •, single energy; , fast; , array; , high definition.

MXA quality control
A combination of factors resulted in the SE scan mode on the QDR-4500A being selected for weekly QC. It demonstrated both accuracy and precision values superior to the F and A scan modes and similar to the HD mode. However, the SE scan takes 12 s to acquire compared with 12 min for the HD scan. The machine was given a standard preventative maintenance service every 6 months (during weeks 6, 32, 58, 85, 112). Minor repairs were carried out on the machine to correct small faults, such as a faulty optical drive, during weeks 54, 58, 91, 124. More substantial repairs, involving replacement of the X-ray tank owing to oil leakage (week 61) and replacement of the motor encoders (week 115), were required only twice during the study period. No software updates that affected MXA were released by Hologic Inc. during the study.

QC scans were not performed on 39 (30.0%) of the 130 weeks owing to limited availability of machine time or the absence of the single operator who performed and analysed all the scans. The mean H_total measured each week was plotted against time to evaluate any fluctuation or drift (Figure 5). Over the 130 weeks, the mean H_total was 384.86 mm (SD 0.59 mm, CV 0.15%). 82/91 (90.1%) measurements were within 1 mm (0.26%) of the true H_total (385 mm). Visual evaluation of the QC plot seemed to show a slow decline in H_total measured over the first year of the study. Linear regression confirmed a significant decline in H_total during the first 52 weeks of the study (y=385.07–0.013x, p<0.05). Although the mean H_total was lower for weeks 1–52 of the study, 384.73 mm (SD 0.60 mm, CV 0.16%) compared with 384.96 mm (SD 0.57 mm, CV 0.15%) for weeks 53–130, this difference failed to reach statistical significance (p>0.05).

View larger version (14K): [in this window] [in a new window]   Figure 5. Mean total phantom height measured using the single energy scan mode on the QDR-4500A machine over a period of 130 weeks.

	Discussion

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

A number of studies have investigated the accuracy and precision of vertebral height measurement of in vitro specimens or phantoms using DXA scanners with the MXA facility or on conventional radiographs [9–19]. They have used: synthetic phantoms or test tools constructed from soft tissue and bone substitutes such as Perspex and aluminium [13–19]; excised vertebrae from cadavers [9, 10]; simulated versions of the spine [13]; and DXA phantoms supplied for densitometry calibration and QC [9, 11–13]. Comparison between these studies are difficult not only owing to the wide variation in phantom construction, but also owing to the use of different versions of MXA software, different machines and different scan modes.

Studies evaluating the QDR-2000plus have identified a 4.9% overestimation of vertebral height [9, 11]. This is far higher than the 5–7% underestimate found in our study. It is difficult to explain such a large difference, however, it seems likely to result from differences in phantom construction, point placement protocols, phantom positioning and interobserver variation. No data are available from other studies to compare with our QDR-4500A results. Accuracy of vertebral height measurement on the Expert or Expert-XL have varied from an underestimate of 4.9% [15] to an overestimate of 2.3% [11], compared with the very small underestimate of 0.4% found in this study. This large difference may partly be the result of the fluctuation found on this machine in the accuracy of vertebral height measurement across the vertebral levels, which may be accentuated by measurement of different vertebral levels in each study. It has been suggested that this fluctuation in accuracy at different vertebral levels on the Expert machines is probably the result of accumulating saturation of the detectors as the scan acquisition progresses from L4 up the spine [18].

Variation in point placement protocols may also play an important role in the differences between the various studies. Points are placed using a diamond shape cursor in the Expert software. These points may be placed according to three distinct protocols: above, on, or below the appropriate endplate (Figure 6). In this study, points were placed on the endplate. However, the method of point placement is not specified in the other studies and it would seem likely that this may explain the varying overestimation and underestimation of vertebral height in other phantom studies.

View larger version (10K): [in this window] [in a new window]   Figure 6. Schematic representation of the possible protocols for point placement on a phantom "vertebral body" using the Expert-XL software.

No studies have assessed the precision of vertebral height measurement in vitro using any type of phantom on the QDR-2000plus or the QDR-4500A. Expert-XL intraobserver repeat analysis precision of vertebral height measurement of 0.6–1.0% has been found for a Perspex and aluminium phantom [15], compared with 2.0% (with repositioning) on a phantom constructed from excised human vertebrae [10]. Similarly, Felsenberg et al [11] identified a precision of approximately 1.0% for the wedge and mid-wedge ratios using the European Spine Phantom phantomon an Expert machine, and this value doubled when excised specimens of human vertebrae were used [9]. These values correspond well with the intraobserver repeat analyses precision value of 1.2% found in this study.

MXA quality control
The QC aspect of this study suggests that the QDR-4500A is a highly stable machine in terms of MXA scanning, little affected by even relatively major repairs. Although differences were observed on a week to week basis, in real terms they were small, equivalent to a maximum of 0.40% (1.55 mm) above and a minimum of 0.44% (1.71 mm) below the true phantom height (385 mm). As a result of the accuracy measurements, it was expected that the SE scan mode would continuously slightly underestimate the true phantom height by 0.12 mm. These weekly fluctuations are too small to adversely affect quantitative vertebral deformity identification in vivo.

Measuring the phantom on a weekly basis allows identification of even a small drift in theaccuracy of vertebral height measurement, which will only become apparent when QC measurements are collected over many weeks or months. Although these differences are currently small, a continuing drift over the long time period often employed in longitudinal studies may potentially create errors in the vertebral heights measured and perhaps require retrospective correction of the data. The slight fluctuations in MXA scan magnification seen on a week to week basis currently cannot be explained. This may be the result of a hardware problem or some kind of environmental variation. Further work is required to try and explain this phenomenon. No other studies have published data on the utility of a phantom for long-term MXA quality control on DXA machines to compare with the results of this study.

Study limitations
The phantom was constructed from Perspex and aluminium to simulate bone and soft tissue, respectively, and therefore was not an accurate representation of the in vivo situation as it does not adequately simulate the edge characteristics of in vivo human vertebrae and their overlying soft tissue. A phantom such as ours will produce images with sharper discrimination of "bone" and "soft tissue" and therefore more clearly defined endplates than might be seen for in vivo vertebrae, especially vertebrae with decreased density as a result of osteopenia or osteoporosis or those covered by a very thick layer of adipose tissue. It would therefore be expected that the phantom would produce more accurate measurement of "vertebral" heights and lower precision errors than the real in vivo situation. A previous study has shown that there is little difference in the vertebral heights measured in vivo by the different scan modes [5]. The majority of interobeserver and intraobserver precision figures calculated in this study are from repeat analyses of a single set of MXA scans acquired without repositioning. Repositioning was only investigated for the SE scan mode on the QDR-4500A and led to a slight increase in precision errors when measuring "vertebral" height.

	Conclusions

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
This phantom is unlikely to be helpful in assessing intersystem variation of vertebral height measurement or in providing cross-calibration values between machines for large scale studies, as has been suggested by others [15], because it does not truly reflect the in vivo situation. However, the phantom constructed for this study appears to be a useful tool for documenting the stability of the mechanical instruments and checking the long-term consistency of operator precision.

Received for publication July 6, 2000. Revision received October 19, 2000. Accepted for publication October 24, 2000.

	References

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 

1. Black DM, Arden NK, Palermo L, Pearson J, Cummings SR, for the Study of Osteoporotic Fractures Research Group. Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. J Bone Miner Res 1999;14:821–8.[Medline]
2. Steiger P, Cummings SR, Genant HK, Weiss H. Morphometric X-ray absorptiometry of the spine: correlation in vivo with morphometric radiography. Study of Osteoporotic Fractures Research Group. Osteoporos Int 1994;4:238–44.[Medline]
3. Chappard C, Kolta S, Fechtenbaum J, Dougados M, Roux C. Clinical evaluation of spine morphometric X-ray absorptiometry. Br J Rheum 1998;37:496–501.[Abstract/Free Full Text]
4. Harvey SB, Hutchison KM, Rennie EC, Hukins DWL, Reid DM. Comparison of the precision of two vertebral morphometry programs for the Lunar Expert-XL imaging densitometer. Br J Radiol 1998;71:388–98.[Abstract]
5. Rea JA, Steiger P, Blake GM, Fogelman I. Optimizing data acquisition and analysis of morphometric X-ray absorptiometry. Osteoporos Int 1998;8:177–83.[Medline]
6. Rea JA, Steiger P, Blake GM, Potts E, Smith IG, Fogelman I. Morphometric X-ray absorptiometry: reference data for vertebral dimensions. J Bone Miner Res 1998;13:464–74.[Medline]
7. Rea JA, Chen MB, Li J, Blake GM, Steiger P, Genant HK, et al. Morphometric X-ray absorptiometry and morphometric radiography of the spine: a comparison of prevalent vertebral deformity identification. J Bone Miner Res 2000;15:564–74.[Medline]
8. Miller CG. Instrument quality control and phantoms for DXA scanners. In: Blake GM, Wahner HW, Fogelman I, editors. The evaluation of osteoporosis: dual energy X-ray absorptiometry and ultrasound in clinical practice. London: Martin Dunitz, 1999:173–92.
9. Gowin W, Diessel E, Mews J, Hoja T, Touby F, Felsenberg D. Vertebral morphometry: a comparison of accuracy and reproducibility of radiographs with new DXA devices. J Bone Miner Res 1995; 10(Suppl. 1):S266.
10. del Rio L. Morphometric absorptiometry: an in vivo and in vitro study on accuracy and reproducibility of a new DXA device. J Bone Miner Res 1997; 12(Suppl. 1):S260.
11. Felsenberg D, Gowin W, Diesel E, Armbrust S, Mews J. Recent developments in DXA. Quality of new DXA/MXA devices for densitometry and morphometry. Eur J Radiol 1995;20:179–84.[Medline]
12. von Stetten E, Steiger P. Accuracy in vitro of vertebral dimensions obtained by scanning fan-beam compared to radiographs. Osteoporos Int 1995;6:79.
13. Rawlings DJ, Warren H, Faulkner K. Performance and acceptance testing of morphometric software. In: Ring EFJ, Elvins DM, Bhalla AK, editors. Current research in osteoporosis and bone mineral measurement IV: 1996. Proceedings at the Fifth Bath conference on Osteoporosis and Bone Mineral Research Measurement; 1996, June 24–26; Bath, UK. London: British Institute of Radiology, 1996:97.
14. Steel SA, Walker R, Howey S, Langton CM. Phantom for evaluating vertebral morphometry using the Lunar Expert. Osteoporos Int 1997;7:286.
15. Howey S, Steel SA, Thorpe JA, Walker R, Langton CM. Development and evaluation of a phantom for morphometric X-ray absorptiometry. In: Ring EFJ, Elvins DM, Bhalla AK, editors. Current research in osteoporosis and bone mineral measurement V: 1998. Proceedings of the Sixth Bath Conference on Osteoporosis and Bone Mineral Measurement; 1998 June 22–26; Bath, UK. London: British Institute of Radiology, 1998:47–8.
16. Thorpe JA, Steel SA, Langton CM. An investigation of the effect of vertebral malalignment on the accuracy of the Lunar Expert-XL MXA procedure. In: Ring EFJ, Elvins DM, Bhalla AK, editors. Current research in osteoporosis and bone mineral measurement V: 1998. Proceedings of the Sixth Bath Conference on Osteoporosis and Bone Mineral Measurement; 1998 June 22–26; Bath, UK. London: British Institute of Radiology, 1998:48–9.
17. van Bodegom JW, Kuiper JW, van Rijn RR, van Kuijk C, Zwamborn AW, Grashuis JL. Vertebral dimensions: influence of X-ray technique and patient size on measurements. Calcif Tissue Int 1998;62:214–8.[Medline]
18. Thorpe JA, Steel SA, Langton CM. A phantom based study on the effect of subjects positioning on morphometric X-ray absorptiometry using the Lunar Expert-XL. Br J Radiol 1998:1153–61.
19. Thorpe JA, Steel SA. Image resolution of the Lunar Expert-XL. Osteoporos Int 1999;10:95–101.[Medline]
20. Jensen FG, McNair P, Boesen J, Hegedus V. Validity in diagnosing osteoporosis. Eur J Radiol 1984;4:1–3.[Medline]
21. Genant HK, Wu CY, van Kuijk C, Nevitt MC. Vertebral fracture assessment using a semiquantitative technique. J Bone Miner Res 1993;8,9:1137–48.[Medline]
22. McCloskey EV, Spector TD, Eyres KS, Fern ED, O'Rourke N, Vasikaran S, et al. The assessment of vertebral deformity: a method for use in population studies and clinical trials. Osteoporos Int 1993;3:138–47.[Medline]
23. Eastell R, Cedel SL, Wahner HW, Riggs BL, Melton LJ. Classification of vertebral fractures. J Bone Miner Res 1991;6:207–15.[Medline]
24. Lewis MK, Blake GM. Patient dose in morphometric X-ray absorptiometry. Osteoporos Int 1995;5:281–2.[Medline]
25. Blake GM, Rea JA, Fogelman I. Vertebral morphometry studies using dual energy X-ray absorptiometry. Semin Nucl Med 1997;27:276–90.[Medline]
26. Bouyoucef SE, Cullum ID, Ell PJ. Cross-calibration of a fan-beam X-ray densitometer with a pencil-beam system. Br J Radiol 1996;69:522–31.[Abstract/Free Full Text]
27. Lang T, Takada M, Gee R, Wu C, Li J, Hayashi-Clark C, et al. A preliminary evaluation of the Lunar Expert-XL for bone densitometry and vertebral morphometry. J Bone Miner Res 1997;12:136–43.[Medline]
28. Gluer C-C, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK. Accurate assessment of precision errors: how to measure the reproducibility of bone densitometry techniques. Osteoporos Int 1995;5:262–70.[Medline]

This article has been cited by other articles:

				E. M. Lewiecki and A. J. Laster Clinical Applications of Vertebral Fracture Assessment by Dual-Energy X-Ray Absorptiometry J. Clin. Endocrinol. Metab., November 1, 2006; 91(11): 4215 - 4222. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rea, J A Articles by Fogelman, I Search for Related Content PubMed PubMed Citation Articles by Rea, J A Articles by Fogelman, I