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 Honey, I D Articles by Evans, D S Search for Related Content PubMed PubMed Citation Articles by Honey, I D Articles by Evans, D S

Investigation of optimum energies for chest imaging using film–screen and computed radiography

Kings Centre for the Assessment of Radiological Equipment (KCARE), Kings College Hospital, London SE5 9RS, UK

	Abstract

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
The purpose of the study was to compare the image quality of film–screen (FS) and computed radiography (CR) for adult chest examinations across a range of beam energies. A series of images of the CDRAD threshold contrast detail detection phantom were acquired for a range of tube potential and exposure levels with both CR and FS. The phantom was placed within 9 cm of Perspex to provide attenuation and realistic levels of scatter in the image. Hardcopy images of the phantom were scored from a masked light-box by two scorers. Threshold contrast indices were used to calculate a visibility index (VI). The relationships between dose and image quality for CR and for FS are fundamentally different. The improvements in VIs obtained using CR at 75 kVp and 90 kVp were found to be statistically significant compared with 125 kVp at matched effective dose levels. The relative performance of FS and CR varies as a function of energy owing to the different k-edges of each system. When changing from FS to CR, the use of lower tube potentials may allow image quality to be maintained whilst reducing effective dose. A tube voltage of 90 kVp is indicated by this work, but may require clinical verification.

	Introduction

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
Digital imaging is gradually replacing the well established technology of conventional film–screen (FS) imaging. This is primarily due to the perceived advantages of digital image processing, electronic archiving, teleradiology and the potential for optimization of image acquisition and display independently. The wide dynamic range of digital systems make them particularly attractive for imaging the widely differing attenuating structures of the chest [1]. The Ionising Radiation (Medical Exposure) Regulations (2000) [2] require that radiographic techniques be optimized for each examination type. An optimization study should investigate tube voltage and the receptor dose required to achieve the minimal acceptable level of image quality.

The Council of European Communities have published guidelines on good radiographic technique for a range of standard FS imaging examinations [3]. The suggested posteroanterior chest technique is to use 125 kVp, 180 cm focus to film distance (FDD) with an antiscatter grid in place. There is no accepted guidance for digital radiography. Fundamental differences exist between FS imaging and computed radiography (CR) and the optimum technique may not necessarily be the same. First, CR has a wide dynamic range and the ability to perform image processing operations on details in the lung and mediastinum in order to enhance their visibility. This reduces the need to minimize the contrast between signals transmitted from lung to mediastinum by selection of a high tube voltage value, as is the practice for FS which has a narrow dynamic range. With digital systems such as CR the optimum tube voltage can be obtained by maximizing the signal to noise ratio whilst maintaining an effective dose level [4–6]. Second, photo-stimulable phosphor (PSP) plates have different absorption edge characteristics to rare-earth screens owing to their different elemental compositions (see Figure 1). In the diagnostic energy range BaF(Br_0.85I_0.15), which is the bulk constituent of the Agfa photo-stimulable phosphor plate (Agfa, Peissenberg, Germany) [7, 8], has a k-edge at 37 keV, whilst Gd₂O₂S (used for rare earth screens) has a k-edge at 50 keV. It can be seen from Figure 1 that between these energies (37 keV to 50 keV) the attenuation performance of the photo-stimulable phosphor will be better relative to the FS than at other energies. For Figure 2, the fraction of the total incident transmitted beam photons, that interact with the detector, F, has been approximated using the following equation:

where E=the maximum energy of the spectrum used; =a discrete point in the energy spectrum; N()=the relative number of photons at a discrete energy level, , in the primary beam after transmission through 10 cm of Perspex and 2.5 mm Al equivalent (obtained using Institute of Physics and Engineering in Medicine (IPEM) report 78 [9]); F()=the fraction of photons of an energy, , that will interact within the detecting medium [10].

View larger version (12K): [in this window] [in a new window]   Figure 1. The mass attenuation coefficients of Gd2O2S (thin solid line) and BaFB0.85I0.15 (thick solid line) as a function of energy [10].

View larger version (12K): [in this window] [in a new window]   Figure 2. The fraction of photons interacting with a BaF(Br0.85I0.15) computed radiography (CR) plate (solid line) and a Gd2O2S phosphor screen (dashed line). The values were calculated using Equation 1, and published tables of data [9, 10]. Both curves have been normalized to their peak point.

It has been shown that for adult chest radiography, the performance of CR and FS is comparable [11–14]. It should be noted that these studies did not use FS systems of the same speed. Conversely, for paediatric chest imaging where lower tube potential is used and the scatter contribution is less, it has been shown that a dose reduction of approximately 30% is possible with CR whilst maintaining the same image quality as with FS [15].

Studies of optimum tube voltage to be used for adult chest radiography using CR have resulted in somewhat conflicting results. When assessing the visibility of low contrast opacities in human volunteers [4] and low contrast disks within a phantom [16], in the range of 80–140 kVp, it was concluded that lower tube potential would result in significantly better image quality. Conversely, using anthropomorphic chest phantoms, signal to noise ratio and image quality assessed using the Commission of the European Communities (CEC) guidelines [3] were not significantly affected by tube voltage [5, 17]. It should be noted that as well as having different methods of assessing image quality these studies have different methods of matching dose at different tube potentials.

The variation of image quality with tube potential will be a complex function of the primary beam spectrum, the scatter spectrum and the variation of receptor absorption efficiency with energy. The aims of this work were to determine the optimum tube potential for chest imaging using CR, and to assess the comparative image quality of FS and CR as a function of tube voltage.

	Theory

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
National guidelines for the assessment of radiological equipment [18] recommend the use of threshold contrast detail detection (TCDD) for the annual routine testing of image intensifiers. TCDD has been suggested for routine testing of CR [19, 20]. A TCDD test object consists of either a series of circular holes drilled into a material, or a series of circular objects positioned on a base material. Typically the number of details of each diameter visible in the image is recorded and used to calculate values of threshold contrast index, H_T(A),

where A is the detail area and C_T is the minimum detectable contrast. C_T must be calculated for the specific beam conditions used (energy and filtration). The threshold contrast index data can be viewed graphically, or summarized as a single quality index (QI) [21]. However, comparison of H_T(A) values between images obtained at different tube potential values is not useful, since equal detail visibility would give rise to a superior H_T(A) values at higher tube voltages. We propose to use a visibility index (VI) which is determined solely from the physical size of the detected objects rather than their radiographic contrasts,

where N is the number of different groups of details with fixed diameters in the phantom used, d(A)_ref is the smallest visible depth of hole for a given detail diameter in the reference image and d(A)_image is the corresponding depth in the image of interest. Two similar depth-based image quality indices have been used in a study of chest radiographs [22]. The form of the visibility index is based on the quality index suggested by Gallacher et al [21]. The use of a reference curve taken from some "standard" image has the effect of normalizing both the quality index suggested by Gallacher and the visibility index suggested here, such that a value above 1.0 will be indicative of better-than-standard image quality.

For a quantum noise limited imaging system the detectable threshold contrast, C_T, of holes of depth, d, in an otherwise uniform background is related to receptor dose, D, as follows,

where

where µ is the attenuation coefficient of air in the holes in the phantom. It follows that the threshold hole depth d is proportional to –ln D, and as a consequence VI must also be proportional to –ln D for a quantum noise limited detector. The mean VI value should be less affected by variations in the noise and observer performance than the individual d(A)_image values. Also, it is simpler to compare two images using a single measure of image quality, rather than compare 10–15 values of d(A). However, it is still important to assess whether the system's relative performance is affected by detail size by examining the full TCDD curve.

	Materials and methods

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
Tests were performed in a general X-ray room equipped with a Opti 150/30/50C tube and Polydoros LX50 generator (Siemens Medical Solutions, Erlangen, Germany), vertical bucky with moving antiscatter grid (focused at 1.50 m, 70 lines cm^–1, grid ratio 7:1), and ADC Solo CR reader with MD10 phosphor plates (Agfa Healthcare, Peissenberg, Germany). A second set of tests were performed in a general X-ray room equipped with a ROT1750 tube and Optimus 50 generator (Philips Medical Systems, Eindhoven, The Netherlands), vertical bucky with a moving antiscatter grid (focused at 1.80 m, 36 lines cm^–1, grid ratio 12:1), and a 400 speed HT-G film–screen cassette, processed using an Agfa CIC system.

Image quality was assessed using the CDRAD (University Hospital Nijmegen, St Radboud, The Netherlands) threshold contrast detail detectability test object. The CDRAD is a 1 cm thick Perspex plate containing circular holes of 15 different diameters (ranging from 0.3 mm to 8 mm) and 15 different depths (ranging from 0.3 mm to 8 mm) [23]. The detail depths are such that within a reasonable range of exposures some, but not all, of the details should be visible at all detail diameters.

The phantom was placed between 4 cm of Perspex (front) and 5 cm of Perspex (behind) to provide scatter and give attenuation similar to an adult chest. All exposures were made in the vertical bucky with a moving antiscatter grid, and a focal spot to detector distance of 180 cm (see Figure 3), as suggested in the CEC guidelines [3].

View larger version (17K): [in this window] [in a new window]   Figure 3. A schematic diagram of the experimental setup used.

The images were scored using a masked light box and consistent low level ambient light conditions. The counting of "half visible" details was permitted, and this was accounted for in the analysis by taking a mean of the adjacent hole depths. VI values for each image were then calculated (Equation 2). The values from the 90 kVp CR automatic exposure control (AEC) images, with a receptor dose of 2.9 µGy, are used as the reference for calculating the VI values quoted throughout this report (N.B. this choice of normalization image is arbitrary). The images were scored by two experienced scorers, and the average of the two VI for each image was taken. Scoring was performed allowing variable viewing distance protocol, and standardized ambient light conditions.

Entrance surface dose was measured using a calibrated Radcal 2025 electrometer with 6 cm³ ionization chamber (Radcal Corporation, Monrovia, CA). The entrance surface doses and beam conditions were used to calculate effective doses for each set of exposure conditions (using the "Effdose" calculation package [24, 25]).

Receptor dose with the phantom in place was measured using a Dali 120 cm³ ion (PTW, Freiberg, Germany) chamber placed directly inside the bucky. Optical densities of the film–screen images were measured in a region away from any visible details using a Parry DT1405 densitometer (Parry, Newbury, UK).

	Results

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
Figures 4 and 5 show the variation of VI with effective dose for the three tube potential values investigated with CR and FS, respectively. Table 1 lists the receptor doses and optical densities of the points for the FS data. Figure 6 plots the ratio of FS VI to CR VI for each tube voltage tested. With FS there is a point beyond which increasing exposure results in a decrease in detail visibility owing to optical density saturation. However, with CR at each tube potential the VI rises with exposure as lower contrast details to become more visible over the decreasing quantum noise. Consequently the image quality performance of FS relative to CR decreases beyond the peak associated with the optimum film blackening (Figure 6).

View larger version (11K): [in this window] [in a new window]   Figure 4. Plot of visibility index (VI) versus effective dose for sets of computed radiography (CR) images obtained at 75 kVp (thick solid line with data points marked by horizontal lines), 90 kVp (dashed line with data points marked by crosses) and 125 kVp (thin solid line with data points marked by diamond shapes).

View larger version (13K): [in this window] [in a new window]   Figure 5. Plot of visibility index (VI) versus effective dose for sets of film–screen (FS) images obtained at 75 kVp (thick solid line with data points marked by horizontal lines), 90 kVp (dashed line with data points marked by crosses) and 125 kVp (thin solid line with data points marked by diamond shapes).

View larger version (11K): [in this window] [in a new window]   Figure 6. A plot of the ratio of film–screen (FS) visibility index (VI) trend-line to the computed radiography (CR) VI trend-line at 75 kVp (thick solid line), 90 kVp (dashed line) and 125 kVp (thin solid line).

• H₁: The distributions of the CR VI at 125 kVp and 90 kVp are identical
  
• H₂: The distributions of the CR VI at 125 kVp and 75 kVp are identical
  
• H₃: The distributions of the CR VI at 90 kVp and 75 kVp are identical
  

The data were grouped into pairs of visibility indices obtained with approximately matched effective doses at different tube voltages. The data were matched such that for each pair the higher dose point was taken from the apparently lower image quality distribution (i.e. 125 kVp for H₁ and H₂, and 75 kVp for H₃). Differences of up to 30% in dose were accepted for matched pairs. The results are summarized in Table 2. The Wilcoxen statistical test is not appropriate for testing the film–screen data, although clearly there is a strong tube voltage dependence shown in Figure 5.

	Discussion

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

With FS it should be noted that the clinical details of interest in the lung and mediastinum will often not be displayed at the optimum optical density. Consequently the true performance of FS relative to CR will generally be poorer than indicated here, since CR VI improve relative to FS away from the peak of the FS curves (Figure 6).

Optimum tube voltage for CR chest examinations
Figure 4 demonstrates that the image quality for the CR system was poorest at 125 kVp across the range of effective doses investigated. This is due to a combination of the variation of detector response with photon energy (see Figures 1 and 2) and the decreased signal contrast at high tube voltage.

It can be seen from Figure 2 that over the range of beam energies investigated in this study the photostimulable phosphor suffers a much greater reduction in primary beam absorption efficiency than the gadolinium oxysulphide screen. This is in broad agreement with the results of Fetterly and Handiandreou [26] who showed that the DQE of a Fuji CR system decreased with increasing tube voltage from 70 kVp to 95 kVp and 120 kVp. The different behaviour is due to the different k-edge positions of the two materials. The higher k-edge of Gd₂O₂S is beneficial for detecting high tube potential beams which have a greater percentage of photon energies above this threshold. It should be noted that the depth at which interactions occur in both the CR plate and intensifying screen will have some energy dependence. The efficiency of the transfer of light from the phosphors to either the film or the photomultiplier tube (in the case of CR) will be dependent on the depth of its origin. For CR, the average depth of interaction will be greater for the higher tube voltages, resulting in a greater spread of laser light and attenuation of photostimulated luminescence. For the FS system the relationship between depth of absorption and image quality is more complex owing to the dual screen geometry.

The sensitivity to the primary beam photons is a partial explanation for the comparatively poor performance of the CR system at high tube voltages. The scattered photon absorption efficiency of the two receptors should also be considered. Unfortunately this effect cannot be quantified as easily as the effect of the primary beam spectrum. Both detectors will in general be more sensitive to the scattered radiation incident on them than to the primary beam due to its lower energy. It can be seen from Figure 2 that the sensitivity of the photostimulable phosphor increases more sharply with decreasing tube potential than the Gd₂O₂S screen. The scatter spectrum can be crudely considered to the equivalent to a beam of lower tube potential than the primary beam. At the higher energies the sensitivity to scatter and primary will be broadly similar for the Gd₂O₂S screen. For the photostimulable phosphor however, the sensitivity to scatter may be significantly higher than the primary beam.

The results of this study of the optimum tube voltage for CR support the results of other investigations [4, 16]. Both these studies used Fuji CR plates which use BaF(Br_0.85I_0.15) phosphors similar to that used by the Agfa system investigated in this work. These studies are therefore all comparable, and all suggest an improvement in detail visibility at tube voltages lower than the 125 kVp suggested for FS imaging [3]. It should be noted that the rationale behind the choice of high tube potential for FS is to maintain an acceptable range of radiographic densities across both the lung and mediastinum, rather than to maximize detail visibility in one or other of these regions. With CR the wide dynamic range and the ability to process data post acquisition suggest that minimization of contrast between mediastinum and lung should not be necessary [4]. The use of a tube voltage lower than 125 kVp is strongly suggested. The CEC guidelines require that chest exposures should be no longer than 20 ms to minimize cardiac motion artefact. It is recommended from the results of this work that a tube voltage in the range of 75 kVp to 90 kVp is used provided that the exposure time can be limited below 20 ms. Clinical trials a required to verify this result.

	Conclusions

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 
CR and FS have inherently different imaging characteristics; for example different energy response and responses to variations in exposure. It is therefore unlikely to be appropriate to acquire images using the same exposure factors for both systems. For chest imaging the results reported here indicate that CR produces optimum image quality in the range of 75 kVp to 90 kVp, rather than the higher energies which are often used in practice.

	Acknowledgments

 
The authors would like to acknowledge the Medicine and Healthcare products Regulatory Agency (MHRA) for funding this projects and the staff of the imaging departments for their assistance and co-operation in providing access to the imaging systems.

	Footnotes

 
Current address for D S Evans, Imaging and Medical Physics, University Hospital Birmingham NHS Trust, Birmingham, UK.

Received for publication January 30, 2004. Revision received September 23, 2004. Accepted for publication December 1, 2004.

	References

Top Abstract Introduction Theory Materials and methods Results Discussion Conclusions References

 

1. Goodman LR, Wilson CR, Foley WD. Digital radiography of the chest: problems and promises. AJR Am J Roentgenol 1988;150:1241–52.[Free Full Text]
2. Ionising radiation (medical exposure) regulations 2000. UK: The Stationary Office Limited, 2000.
3. Council of European Communities. European guidelines on quality criteria for diagnostic radiographic images, Rep EUR 16260 EN, 1996.
4. Nobuhiro O, Hajime N, Seiichi M, Kunihiro T, Katsumi N, Akira Y. Optimal beam quality for chest radiography. Invest Radiol 1996;31:126–31.[CrossRef][Medline]
5. Tingberg A, Sjostrom D. Search for optimal tube voltage for image plate radiography. Prog Biomed Optics Imaging 2003;4:187–96.
6. Dobbins JT, Samei E, Chotas HG, Warp RJ, Baydush AH, Floyd CE, et al. Chest radiography: optimisation of x-ray spectrum for caesium iodide – amorphous silicon flat panel detector. Radiology 2003;226:221–30.[Abstract/Free Full Text]
7. Nakano Y, Gido T, Honda S, Maezawa A, Wakamatsu H, Yanagita T. Improved computed radiography image quality from BaFI:Eu photostimulable phosphor. Med Phys 2002;29:592–7.[CrossRef][Medline]
8. Rowlands JA. The physics of computed radiography. Phys Med Biol 2002;47:123–66.[CrossRef]
9. Craney K, Gilmore BJ, Fogarty GWA, Desponds A, Sutton DG. Catalogue of diagnostic x-ray spectra and other data. Institute of Physics and Engineering in Medicine report 78 (CDROM), York: IPEM, 1997.
10. Boone JM, Chavez AE. Comparison of x-ray cross sections for diagnostic and therapeutic medical physics. Med Phys 1996;23:1997–2005.[CrossRef][Medline]
11. Cook LT, Insana MF, Mcfadden MA, Hall TJ, Cox GG. Comparison of low contrast detectability of a screen film system and third generation computed radiography. Med Phys 1994;21:691–5.[CrossRef][Medline]
12. Ganten M, Radeleff B, Kampschulte A, Daniels MD, Kauffmann G, et al. Comparing image quality of flat panel chest radiography with storage phosphor radiography and film screen radiography. AJR Am J Roentgenol 2003;181:171–6.[Abstract/Free Full Text]
13. Schaefer-Prokop CM, Prokpop M, Schmidt A, Neitzel U, Galanski M. Selenium radiography versus storage phosphor and conventional radiography in the detection of simulated chest lesions. Radiology 1996;201:45–50.[Abstract/Free Full Text]
14. Bacher K, Smeets P, Bonnarens K, De Hauwere A, Verstraete K, et al. Dose reduction in patients undergoing chest imaging. AJR Am J Roentgenol 2003;181:923–9.[Abstract/Free Full Text]
15. Hufton AP, Doyle SM, Carty HML. Digital radiography in paediatrics: radiation dose considerations and magnitude of possible dose reduction. Br J Radiol 1998;71:186–99.[Abstract]
16. Dobbins JT, Rice JJ, Beam CA, Ravin CE. Threshold perception performance with computed and screen film radiography: implications for chest radiography. Radiology 1992;183:179–87.[Abstract/Free Full Text]
17. Chotas HG, Floyd CE, Dobbins JT, Ravin CE. Digital chest radiography with photostimulable phosphors: signal-to-noise ratio as a function of kilovoltage with matched exposure risk. Radiology 1993;186:395–8.[Abstract/Free Full Text]
18. IPEM CoR/NRPB. Recommended standards for the routine testing of diagnostic x-ray imaging systems, IPEM report 77. York: Institute of Physics and Engineering in Medicine, 1997;41–7.
19. Samei E, Seibert JA, Willis CE, Flynn MJ, Mah E, Junck KL. Performance evaluation of CR systems. Med Phys 2001;28:361–71.[CrossRef][Medline]
20. Finch A, Bury R, Davies G, Fife I, Hufton A, Workman A. Assurance of quality in the diagnostic imaging department. London: British Institute of Radiology, 2001;67–71.
21. Gallacher DJ, Mackenzie A, Batchelor S, Lynch J, Saunders JA. Use of a quality index in threshold contrast detail detection measurements in television fluoroscopy. Br J Radiol 2003;76:464–72.[Abstract/Free Full Text]
22. Soldt RTM, Zweers D, Van Den Berg L, Geleijns J, Jansen JThM, Zoetelief J. Survey of posteroanterior chest radiography in The Netherlands: patient dose and image quality. Br J Radiol 2003;76:398–405.[Abstract/Free Full Text]
23. Thijssen MAO, Bijkerk KR, Van der Burght RJM. Manual CDRAD-phantom type 2.0. University Hospital Nijmegen, 1998.
24. Hart D, Jones DG, Wall BF. Estimation of effective dose in diagnostic radiology from entrance surface does and dose area product measurements. NRPB Report NRPB-R262 Chilton, UK: National Radiological Protection Board, 1994.
25. Effdose software V1.04. National Board of Health. Denmark: National Institute of Radiation Hygiene, 1995.
26. Fetterley KA, Hangiandreou NJ. Effects of the x-ray spectra on the DQE of a computed radiography system. Med Phys 2001;28:241–8.[CrossRef][Medline]

This article has been cited by other articles:

				C S MOORE, J R SAUNDERSON, and A W BEAVIS Investigating the exposure class of a computed radiography system for optimisation of physical image quality for chest radiography Br. J. Radiol., September 1, 2009; 82(981): 705 - 710. [Abstract] [Full Text] [PDF]

				H GEIJER, E NORRMAN, and J PERSLIDEN Optimizing the tube potential for lumbar spine radiography with a flat-panel digital detector Br. J. Radiol., January 1, 2009; 82(973): 62 - 68. [Abstract] [Full Text] [PDF]

				C S MOORE, A W BEAVIS, and J R SAUNDERSON Investigation of optimum X-ray beam tube voltage and filtration for chest radiography with a computed radiography system Br. J. Radiol., October 1, 2008; 81(970): 771 - 777. [Abstract] [Full Text] [PDF]

				B M Moores Relative biological effectiveness and exposure of the female breast Br. J. Radiol., February 1, 2007; 80(950): 141 - 142. [Full Text] [PDF]

				H Cole and C P Lawinski Developments in radiography and fluoroscopy Imaging, June 1, 2006; 18(2): 75 - 81. [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 Honey, I D Articles by Evans, D S Search for Related Content PubMed PubMed Citation Articles by Honey, I D Articles by Evans, D S