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 Vassileva, J Search for Related Content PubMed PubMed Citation Articles by Vassileva, J

A phantom for dose-image quality optimization in chest radiography

Applied Physics Department, Konstantin Preslavsky University, 9712 Shumen, Bulgaria

	Abstract

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
Optimization in chest radiography requires evaluation of patient dose and image quality. This study is aimed at proposing a simple geometrical phantom that realistically simulates the important anatomical regions of the thorax. For this purpose, the standard LucAl chest phantom is modified by adding an "anthropomorphic" insert and image quality test plate. Different test objects are arranged on the plate in three important anatomical areas; lung, cardiac, and subdiaphragmal regions. The aim is to simultaneously find two types of image quality index, objective and subjective, which can be used to compare different images in order to select the better image. Two objective indices are proposed, areal contrast index C_a and scatter fraction P_s and two subjectively estimated, a low contrast visualization index P_low and a high contrast visualization index P_high. To demonstrate the potential of this phantom method it was applied to an X-ray unit to find the optical film density that ensures optimal visualization in different anatomical areas. It was found for the X-ray system under investigation that the automatic exposure control could be set to produce an optical density of about 1.8 in the lung field. The reported method is easily implemented in any clinical situation where optimization of chest radiography is needed.

	Introduction

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
The purpose of optimization in diagnostic radiology is to find the technical parameters needed to produce high image quality for minimum patient dose. Owing to the variety of X-ray units used, radiological examinations cannot be standardized. Thus, optimization is necessary for each particular X-ray unit and for each X-ray examination [1]. The optimization procedure requires evaluation of patient dose and image quality.

Image quality can be investigated in two ways, one based on physical measurements and the other on psychophysical assessments [2]. Many authors show that objective physical measurements alone are not sufficient to demonstrate the clinical advantage of one imaging technique over another [2, 3]. Subjective clinical evaluation of patients' radiographs, although being a powerful method, needs considerable resources and is often unacceptable for radiation protection reasons. The method of choice therefore involves a complex investigation of images of appropriate phantoms.

Various phantoms of different complexity are in use to simulate the human thorax. Anthropomorphic phantoms simulating shape, size and tissue composition are not available in each X-ray department because of the relative complexity of manufacture and high cost. Simpler geometrical phantoms are also designed for imaging and dosimetric purposes [4–6].

One of the most frequently used geometrical phantoms is the LucAl phantom (Standard Dosimetric/Calibration Phantom; Center for Devices and Radiological Health, Carson, CA), described by Conway et al [4]. The LucAl phantom consists of 250 mm x 250 mm polymethyl-methacrylate (PMMA) plates and 1100 aluminum (Al) alloy with specified thicknesses (Figure 1a). The overall thickness of the phantom is 267 mm, with 4.1 mm Al, 73 mm PMMA and a 190 mm air gap. The thicknesses and relative positions of the different components have been designed to accurately simulate primary and scatter transmission through the lung-field regions of a patient-equivalent anthropomorphic chest phantom (Humanoid Systems, Carson, CA) for up to 150 kVp X-rays typically used in chest radiography. Good spectral equivalence between LucAl and Alderson-Rando male and female anthropomorphic phantoms has also been found [6]. Clinical tests have shown that the LucAl phantom reliably approximates a 22.5 cm thick patient for a posteroanterior (PA) chest projection controlled by the lateral automatic exposure control (AEC) chambers [4]. The phantom was originally designed for dose measurements during chest radiography with AEC [4]. Subsequently, it was also used for image quality evaluation by inserting different test objects into the phantom [7, 8]. However, the image of this homogeneous phantom does not provide information regarding image quality under the clinically important mediastinal region of the thorax. At the same time it was shown that approximately 26% of the lung volume and 43% of the lung area are obscured by organs [9].

View larger version (53K): [in this window] [in a new window]   Figure 1. Schematic view of the modified LucAl chest phantom. (a) Standard LucAl chest phantom. (b) Anthropomorphic insert. (c) Image quality test (IQT) plate containing low and high contrast objects as described in the text. PMMA, polymethyl-methacrylate; Al, aluminium.

	Anthropomorphic LucAl phantom

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
Any phantom used for evaluation of chest image quality should realistically simulate the mediastinum, heart and soft tissue organs in the subdiaphragmal area, i.e. the OD under these areas of the phantom image should be similar to the relevant anatomical areas on an average patient's thorax image. Therefore our purpose was to add to the original LucAl phantom parts simulating the X-ray transmission through these regions of the chest. Using data from the International Commission on Radiological Units and Measurements Report 44 [10] we selected Al as a phantom material for bones (spine and ribs) and PMMA for muscles (heart and subdiaphragm). These materials are easily available but their densities and attenuation coefficients differ to some extent from those of the tissues to be simulated. That is why the relevant thickness for each anatomical detail was chosen using the following methodology. First, on some radiographs of standard size patients the ODs of four clinically important regions were measured; the middle lung between ribs, the central mediastinum, the heart, and the subdiaphragm. Images were obtained on a conventional X-ray unit with chest stand equipped with a 7:1 moving antiscatter grid, at a 150 cm focus to film distance and 70 kVp. Using the characteristic curve of the particular screen–film combination (SFC), OD values were converted to air kerma values. For this purpose, a H&D curve was prepared using time scale sensitometry at 70 kVp, constant tube current and a focus-to-film distance of 150 cm. The spectra of the X-rays transmitted through the LucAl phantom plus added different layers of PMMA or aluminum were then derived for the same tube potential using a computer programme based on the Birch and Marshall catalogue of X-ray spectra [11]. From those data, necessary thicknesses of the respective materials to obtain the required air kerma values were determined. As a result the spine was simulated by an Al strip 35 mm wide and 5 mm thick. A slab of 10 mm PMMA was added to the central region of the phantom to give the broad beam attenuation of the upper mediastinum. Cylindrical blocks of PMMA were used for the heart and the subdiaphragmal organs, with thicknesses of 64 mm and 75 mm, respectively. Two 3 mm thick Al strips were added to the right lung field to provide information regarding OD changes from the ribs. The position of these inserts relative to the homogeneous layers of the LucAl phantom is shown in Figure 1b.

For image quality evaluation, different test objects were arranged on a 10 mm thick PMMA plate. This image quality test (IQT) plate was inserted into the phantom, substituting for one PMMA layer with the same thickness (Figure 1). The position of the IQT plate can be varied according to the imaging task. The different test objects are arranged on the IQT plate to determine image quality parameters in three important anatomical areas; lung, cardiac, and subdiaphragmal regions. The objects' positioning is shown in Figure 1c. The objects are set in two groups, according to the method of estimation of quality parameters. The test objects for subjective analysis are situated on the left side of the IQT plate and those for OD measurements are situated on the right side.

Objects for subjective estimation of image quality
The objects to be simulated, their size and structure were selected following clinical requirements for chest radiography [1, 3]. Phantom materials were chosen with respect to their tissue equivalence and accessibility [10]. Low contrast objects, such as subtle lung nodules or small tumours with diffuse edges, are simulated by three groups of eight paraffin wax hemispheres (N₁, N₂, N₃) with diameters from 3 mm up to 10 mm. Four polyethylene cylinders (M) with diameters from 7 mm to 20 mm with different heights are used to simulate lesions with more defined edges. Pulmonary infiltrates are represented by large flakes of paraffin wax 1–2 mm thick (I₁, I₂). Small vessels are simulated by nylon cords with a diameter of 1.5 mm and 2.0 mm (V₁, V₂). Groups of these objects were positioned in the lung, cardiac and subdiaphragmal areas of the phantom. Two groups of grain, one of 0.5 mm diameter copper (G₁, G₂) and one of 1 mm diameter polytetrafluoroethylene (F₁, F₂), simulating fibrosis, are situated in the lung and the retrocardiac areas. Two polyvinylchloride cylinders with drilled holes of 2.5 mm and 5 mm in diameter and 2 mm high (L₁, L₂), positioned in the lung region, are used to determine low contrast visualization. To assess high contrast resolution, two groups of wire meshes are positioned over the "lung" and "heart", respectively (W₁, W₂). Each group consists of four meshes with 9 wires, 13 wires, 15 wires and 30 wires per mm, respectively, the first two with 0.2 mm diameter wire, the third with 0.15 mm diameter wire and the fourth with 0.1 mm diameter wire. Two groups of copper wires of 0.3 mm, 0.2 mm and 0.14 mm in diameter (P₁, P₂) are used to simulate pneumothorax. Additionally, a resolution test object consisting of 11 line groups with spatial frequency from 0.6 mm^-1 up to 3.3 mm^-1 (H) is positioned in the lung field to check the system's high contrast resolution capability. The positions of all these objects can be seen in Figure 1c.

Objects for physical measurements
20 areas for OD measurement are arranged on the IQT plate to enable evaluation of objective image quality indices. To this end, polyethylene cylinders 20 mm in diameter are arranged in five groups of two cylinders, one 4 mm and one 8 mm high in each group. These groups are placed in the lung, cardiac, subdiaphragmal and upper mediastinal areas and over the "rib" (Figure 2). An additional cylinder is placed on the crossing of two ribs. To assess radiographic contrast, round reference areas are marked near the test objects. Lead stoppers 7 mm thick and 10 mm x 10 mm in size, to measure the scatter fraction, are arranged in three areas; lung, cardiac and subdiaphragmal regions. Using the electronic catalogue of spectra [11] it was found that this lead thickness corresponds to approximately 25 half value layers of X-rays generated at 120 kV, and practically fully stops the primary beam.

View larger version (132K): [in this window] [in a new window]   Figure 2. Radiograph of the anthropomorphic LucAl chest phantom.

	Method for phantom image evaluation

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
The anthropomorphic phantom with inserted IQT plate can be exposed with an imaging system that needs optimization, changing one of the exposure parameters, e.g. tube voltage, filtration, SFC, mAs, antiscatter device, and keeping all other parameters unchanged. During exposure, the dosimetric detector is placed on the entrance surface of the phantom to measure the incident air kerma. Images of the phantom (Figure 2) are analyzed in two ways; by measuring OD under the test objects arranged in the right half of the imaging plate and by subjective evaluation of the visualization of the objects in the left half. The aim is to simultaneously find two types of image quality index, objective and subjective, which can be used to compare different images in order to select the better image.

Objective quality indices
Two objective indices are proposed; the areal contrast index C_a and a scatter fraction P_s. Optical densities OD_i (i=1–20) are measured in the 20 fixed spots described above (Figures 1c and 2). Radiographic contrast for each object is calculated as the difference between OD_i under the test object i and OD_ref in the relevant reference area. The areal contrast index C_a (a=lung, heart, subdiaphragm, upper mediastinum, ribs) is calculated by summing the density differences for the objects located in each of the five anatomical areas.

The scatter fraction P_s, representing the contribution of scattered radiation to the image formation, is calculated using the formula:

where K_s and (K_s+K_p) are values of air kerma behind the lead stopper and close to it, respectively. These values are derived from measuring OD at these points and then transforming them into air kerma values using the SFC characteristic curve.

Subjective quality indices
Images are subjectively evaluated at optimal ambient light using a clinical viewing box with homogeneous brightness. Visualization of each object is scored using a four point scale; 3 if the object is clearly defined, 2 if it is visible, 1 if it is hardly discernible and 0 if it is fully missing. Two quality indices are then formed for each anatomical area by summing the scores of the low contrast objects N, M, I, V, F, G and L for the low contrast visualization index P_low and those of the high contrast objects P, W and H for the high contrast visualization index P_high.

	Application of the anthropomorphic LucAl phantom

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
As an example, the method was applied in the adjustment of AEC settings of an X-ray diagnostic unit (TUR D800-3; VEB Hermann Mattern, Dresden, Germany). With this unit chest radiography is performed on a chest stand equipped with a moving grid with a grid ratio of 7:1 and a grid strip frequency of 28 mm^-1, at 150 cm focus-to-film distance. The grid was checked for focusing, centreing and artefacts. A SFC with speed class 400 was used with quanta fast detail screens (Sterling Diagnostic Imaging, Bad Homburg, Germany) combined with blue Agfa CP-BU New films (Agfa-Gevaert AG, Morstel, Belgium). The films were processed with an automatic processor (Optimax; Protec Medizintechnik, Oberstenfeld, Germany) with Agfa chemistry according to the manufacturer's specifications. Consistency of film processing was checked by light sensitometry. Densitometry was performed using a DensiX densitometer (PTW, Freiburg, Germany) with a measuring aperture of 7 mm² and accuracy of ±0.02 at ODle;1 and ±1.5% at OD>1.

Standard PA chest radiography was simulated on the chest stand with the anthropomorphic LucAl phantom. Eight films were exposed with a standard soft-beam technique of 70 kVp and 1 mm Al added filter. The only varied parameter was mAs (between 3.2 mAs and 16 mAs). The same radiographic cassette was used for each exposure. The quality indices C_a, P_low and P_high were evaluated. The incident air kerma on the phantom entrance surface was measured directly during exposure with the dosemeter WD 10 (Wellhofer, Schwarzenbruck, Germany) and a diode detector, with an error of less than 5%.

Results are presented in Table 1 and Figure 3. Table 1 shows the change of optical density OD_ref in lung, cardiac and subdiaphragmal areas and in the ribs as a function of incident air kerma on the phantom surface. Areal contrast index C_a for four anatomical areas (lung, cardiac and subdiaphragmal regions, and in the ribs) as a function of OD_lung is presented in Figure 3a. Subjectively, derived low contrast visualization index P_low for lung, cardiac and subdiaphragmal areas are given in Figure 3b as a function of OD_lung. On the same graph this dependence is shown for high contrast visualization index P_high in the lung field.

View larger version (15K): [in this window] [in a new window]   Figure 3. Variation of the image quality indices with the optical density (OD) in the lung area ODlung for posteroanterior chest radiography at constant beam quality (70 kVp and 1 mm aluminium added filter). (a) Areal contrast index Ca in the lung, cardiac and diaphragmal areas and in the ribs. •, lung; , cardiac; *, subdiaphragmal; +, ribs. (b) Low contrast visualization index Plow for the lung, cardiac and diaphragmal areas and high contrast visualization index Phigh in lung field. •, Plow lung; , Phigh lung; , Plow cardiac; *, Plow subdiaphragmal.

The contrast in cardiac and subdiaphragmal areas increases quickly with increasing OD_lung. This is the reason that subjective image quality indices in these areas increase. Above OD_lung=2.0, however, there is a rapid deterioration in visualization of important subtle round details in the lung with a diameter of 3 mm and low contrast linear details with a diameter of 1–2 mm. This is owing to contrast reduction in the lung as well as decreased visual perception at higher ODs, which also depends on viewing box brightness.

Taking into account changes of all quality indices, we could decide that in our particular case, the AEC could be set to produce OD_lung in the lung field of approximately 1.8. The optimal OD has to be found for each particular combination of SFC, viewing box and viewing conditions.

	Conclusion

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 
This study aimed to propose a simple method based on phantom measurements to find the optimal technical parameters for a given X-ray unit. The reported method was applied for one X-ray system to find the optimal AEC settings. A forthcoming task is to use this method to find the optimal tube potential and filtration for the same X-ray system.

By modifying the standard LucAl chest phantom by adding an anthropomorphic insert and image quality test plate, we can evaluate the performance of different components of the imaging chain and their influence on the final result; image quality and patient dose. This method is easily applicable to any clinical situation in which optimization of the procedure is needed.

	Acknowledgments

 
The author is grateful to Mr M Ganchev for his careful reading of the manuscript. The helpful comments and continuous support of Miss I Castellano from the Physics Department of the Royal Marsden Hospital, London are gratefully acknowledged. Mr. G Gerov is thanked for engineering assistance.

	Footnotes

 
This work has been supported by a grant from the National Science Fund of the Republic of Bulgaria (Contract F-720).

Received for publication January 29, 2002. Revision received June 25, 2002. Accepted for publication July 5, 2002.

	References

Top Abstract Introduction Anthropomorphic LucAl phantom Method for phantom image... Application of the... Conclusion References

 

1. European Commission. European guidelines on quality criteria for diagnostic radiographic images, EUR 16260. Luxembourg: European Commission, 1996.
2. International Commission on Radiological Units and Measurements. Medical imaging—the assessment of image quality. ICRU Report 54. Bethesda, MD: ICRU, 1996.
3. Stieve F-E. Radiological requirements for the specification of image quality criteria. In: Moores BM, Wall BF, Erisckat H, Schibilla H, editors, Optimization of Image Quality and Patient Exposure in Diagnostic Radiology, BIR Report 20. London: The British Institute of Radiology, 1989:221–38.
4. Conway BJ, Butler PF, Duff JE, Fewell TR, Gross RE, Jennings RJ, et al. Beam quality independent attenuation phantom for estimating patient exposure from x-ray automatic exposure controlled chest examinations. Med Phys 1984;11:827–32.[Medline]
5. Chotas H, Floyd C, Johnson G, Ravin C. Quality Control Phantom for Digital Chest Radiography. Radiology 1997;202:111–6.[Abstract/Free Full Text]
6. Servomaa A, Tapiovaara M. Patient equivalent phantoms in chest radiography. In: Moores B, Stieve F, Eriskat H, Schibilla H, editors. Technical and Physical Parameters for Quality Assurance in Medical Diagnostic Radiology: Tolerances, Limiting Values and Appropriate Measuring Methods. BIR Report 18. London: British Institute of Radiology, 1989.
7. Kaczmarek R, Conway B, Slayton R, Suleiman O. Results of a nationwide survey of chest radiography: comparison with results of a previous study. Radiology 2000;215:891–6.[Abstract/Free Full Text]
8. Leitz W, Hedberg-Vikstrom B, Conway BJ, Showalter CK, Rueter FG. Assessment and comparison of chest radiography techniques in the United States and Sweden. Br J Radiol 1990;63:33–40.[Abstract/Free Full Text]
9. Chotas H, Ravin C. Chest radiography: estimated lung volume and projected area obscured by the heart, mediastinum, and diaphragm. Radiology 1994;193:403–4.[Abstract/Free Full Text]
10. International Commission on Radiological Units and Measurements. Tissue Substitutes in Radiation Dosimetry. ICRU Report 44. Bethesda, MD: ICRU, 1989.
11. Cranley K, Gilmore BJ, Fogarty GWA, Desponds L. Electronic version prepared by D Sutton. Catalogue of diagnostic X-ray spectra and other data, The Institute of Physics and Engineering in Medicine Report No.78. York, UK: IPEM, 1997.

This article has been cited by other articles:

				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]

				C S Moore, G P Liney, A W Beavis, and J R Saunderson A method to optimize the processing algorithm of a computed radiography system for chest radiography Br. J. Radiol., September 1, 2007; 80(957): 724 - 730. [Abstract] [Full Text] [PDF]

				J Vassileva A phantom approach to find the optimal technical parameters for plain chest radiography Br. J. Radiol., August 1, 2004; 77(920): 648 - 653. [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 Vassileva, J Search for Related Content PubMed PubMed Citation Articles by Vassileva, J