- Email: [email protected]
A low-cost phantom for simple routine testing of single photon emission computed tomography (SPECT) cameras
ARTICLE IN PRESS Applied Radiation and Isotopes 67 (2009) 1864–1868
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Technical note
A low-cost phantom for simple routine testing of single photon emission computed tomography (SPECT) cameras A.H. Ng a,b, K.H. Ng a,, H. Dharmendra a, A.C. Perkins c a b c
Department of Biomedical Imaging, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia Engineering Services Division, Ministry of Health, Level 2-5, Block E6, Parcel E, Federal Government Administration Centre, 62590 Putrajaya, Malaysia Academic Medical Physics, Medical School, University of Nottingham, Nottingham NG7 2UH, UK
a r t i c l e in f o
a b s t r a c t
Article history: Received 21 April 2008 Received in revised form 2 September 2008 Accepted 21 October 2008
A simple sphere test phantom has been developed for routine performance testing of SPECT systems in situations where expensive commercial phantoms may not be available. The phantom was based on a design with six universal syringe hubs set in the frame to support a circular array of six glass blown spheres of different sizes. The frame was then placed into a water-filled CT abdomen phantom and scanned with a triple head camera system (Philips IRIXTM, USA). Comparison was made with a commercially available phantom (Deluxe Jaszczak phantom). Whereas the commercial phantom demonstrates cold spot resolution, an important advantage of the sphere test phantom was that hot spot resolution could be easily measured using almost half (370 MBq) of the activity recommended for use in the commercial phantom. Results showed that the contrast increased non-linearly with sphere volume and radionuclide concentration. The phantom was found to be suitable as an inexpensive option for daily performance tests. & 2008 Elsevier Ltd. All rights reserved.
Keywords: SPECT Phantom Quality control Performance test Contrast Lesion detection
1. Introduction Recent interest in medical imaging applications such as nuclear cardiology and tumour imaging has resulted in a progressive increase in the use of clinical single photon emission computed tomography (SPECT) imaging. As a consequence, there has been an increased utilisation of clinical SPECT systems. For optimal diagnostic use it is essential that routine performance evaluation is carried out as part of an ongoing quality assurance programme. The NEMA publication NU 1-2001 (NEMA, 2001) is the basic recommended standard for performance evaluation and acceptance testing of scintillation cameras. These guidelines were originally intended for use by manufacturers as a means for specifying the standards of equipment and were later modified by workers wishing to assess ongoing equipment performance. However, the methodology described in the NEMA guidelines is more complex than necessary for many departments to use on a routine basis. It is therefore of important practical value to develop a more accessible means for routine testing. More significantly in some countries, imaging equipment may be installed with little or no funding for maintenance, quality control and adequate scientific support. In such circumstances, commer-
Corresponding author. Tel.: +60 3 7949 2091.
E-mail addresses: [email protected], [email protected] (K.H. Ng). 0969-8043/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2008.10.010
cially produced phantoms can be prohibitively expensive and as a consequence quality control procedures are not implemented. This technical note describes the commercial options available and outlines the design and use of an alternative low-cost phantom for routine performance measurement of SPECT systems. We have also compared the use of the new fabricated phantom with a commercial SPECT phantom to determine to what extent the simplified methods could be used for performance testing on a routine basis in hospital with limited resources.
2. Materials and methods The design of the low cost phantom was based on the common parameters measured using commercial phantoms. The most important of these parameters are spatial resolution, contrast and sensitivity. After reviewing the currently available phantoms and their performances in measuring these parameters, the low cost phantom was designed and then fabricated. 2.1. Descriptions of the phantom The in house phantom was based on the design of Perkins et al. (2007). The phantom consisted of an acrylic support and base constructed in an ‘‘L’’ shape originally designed with a circular array of ‘‘blind hubs’’ to hold Luer lock syringes in a perpendicular
ARTICLE IN PRESS A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
array. The angle between each holder was 601 and the diameter across two opposite holders was 100 mm. When filled with radioactivity the syringes would form standard sources of radioactivity in a reproducible configuration, so that they could be seen as parallel rods during image reconstruction. For this study the syringes were replaced with spherical glass spheres. Glass was being used due to its superior durability, chemical and heat resistance. It is considered relative economic as the glass can be shaped according to the laboratory and clinical requirements. The issues of the fragility of the glass and difficulty of making reproducible spheres are relatively insignificant in this project as the cost and feasibility of the phantom are the main concerns. Spheres are generally used as a three-dimensional structure suitable for measurement in the three common imaging planes. Whilst glass does have problems with fragility, this was considered to be the most widely available and inexpensive method of production. Cylinders could be used for a more crude assessment in a single plane (Perkins et al., 2007); however, spheres were considered to be more representative of the clinical situation, when imaging small tumour, as for example. The spheres were constructed by using glass blowing technique. The manufacturing technique will depend upon the skill and dexterity of the glass blower. The volumes can be accurately measured by filling and measuring the amount of water contained in each sphere. Manufacturing reproducibility may be an issue, but in practice, this is not a problem since these phantoms are not intended for mass production. Once a range of object sizes have been produced and measured with calipers, these will form a basis for repeated measurements within a given centre or intercomparison between a small number of different SPECT cameras. The tolerance for the size, wall thickness is estimated to be 10–20% from its original measurement. Fig. 1 shows the phantom constructed for this study. Fig. 2 shows the fabricated glass spheres for the sphere test phantom. This was considered to be an economical method of production using techniques that would be widely available. The resulting product was of acceptable durability and resistant to chemical and heat effects. The stem of each single glass sphere was attached to a solid glass rod by a conically tapered ground glass joint with the inner (female) and outer (male) diameter of 5 mm. This type of joint is commonly used as a stopper for laboratory glassware. The outer diameters of six spheres varied from 10.20 to 24.10 mm. The length from the centre of each sphere to the end of the stem was equal, ensuring all the spheres were in the same plane for image acquisition and transverse reconstruction. In order to keep the spheres hold in place, an acrylic Perspex plate was added to support the spheres perpendicular to the acrylic holder wall. To simulate the clinical situation the assembled phantom was inserted into a CT abdomen phantom (Fig. 1). This was made from acrylic with a dimension of 333 mm w 183 mm h 262 mm depth, which was the approximate size of an adult abdomen. The volume of this empty D-shaped cylinder was 10 L. The phantom was filled with water to simulate attenuation and scattering effects on image quality. Once the sphere phantom was inserted the abdomen phantom was closed with acrylic end plate to cover the top of the phantom. A plastic O-ring was inserted as a seal between the plate and abdomen phantom to prevent any leakage of fluid and the plate was secured with 18 screws distributed equally around the edge.
2.2. SPECT performance tests All imaging was carried out at the Nuclear Medicine Unit, Department of Biomedical Imaging, University of Malaya Medical
1865
Centre, Kuala Lumpur, using a triple detector SPECT camera (Philips IRIXTM, USA). The scanning time and injected activity of the performance test evaluation recommended by NEMA NU-1 (2001) and Graham et al. (1995) were modified for optimum scanning. Due to the lower activity concentration in the Jaszczak phantom, the radioactivity injected was higher than low-cost phantom. Thus, the scanning time was increased in order to complete a performance test. The spheres were filled with a range of concentrations of Tc-99 m between 58 and 101 MBq ml1. The phantom was placed on the couch and SPECT acquisition was performed using 1201 continuous scanning mode with incremental view of 31 each. This acquisition took a data collection time of 90 min. Images were reconstructed on an Odyssey LX computer using filtered back projection, 0.11 cm1 attenuation correction and a Butterworth filter 1.20 cut-off for a 128 128 matrix (Picker International, 1999). For comparison, the Deluxe Jaszczak phantom (Data Spectrum Corporate, USA) was filled with 750 MBq Tc-99 m (giving a concentration of 0.18 MBq ml1). The image contrast was measured by defining regions of interest on the computer images in order to obtain the average pixel count rates. An ROI was drawn approximately 5% of the vicinity from the target volume. The final contrast was calculated by using the formula given by Groch et al. (2000) and Sprawls (1995): ‘‘Cold lesion’’ phantom contrast;
‘‘Hot lesion’’ phantom contrast;
C cold ¼
C hot ¼
CB CS CB
CS CB CS
(1)
(2)
where CS is the average pixel count in the sphere region and CB is the average pixel count in the background area. One initial concern regarding the construction of the sphere phantom was the possibility of radioactivity leaking from the ground glass joints. Several leak tests were carried out by monitoring the activity in the water inside the abdominal chamber and coloured ultramarine blue was injected into the cavities so that any leak could be observed. There was no evidence of any leakage.
3. Results Costs: A comparison of the capabilities and costs between the sphere test phantom and other commercially available phantoms is given in Table 1. For the ‘‘in house’’ sphere phantom, the acrylic holder was fabricated in the Department of Medical Physics and Clinical Engineering at University of Nottingham Hospitals NHS Trust, Nottingham, UK, at a cost in the order of US$ 100.00. The CT abdomen phantom was built at a cost of US$ 500.00. The spheres and supports were fabricated by The Glass Blowing Group of the Malaysian Nuclear Agency at a cost of US$ 120. Fabrication of the Perspex cover plate was done in the workshop of the Engineering Faculty, University of Malaya, at a cost of US$ 10.00. This resulted in a total cost of the phantom of US$ 730. 3.1. SPECT performance tests The first and obvious difference between the phantoms was that the sphere test was designed to measure hot spot detection, whereas the Jaszczak phantom demonstrated cold spot detection. The Jaszczak phantom required a nominal 750 MBq or higher activity for use, while the sphere phantom could be used with less
ARTICLE IN PRESS 1866
A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
CT abdomen Phantom Cap
CT abdomen Phantom
O-ring
Nylon Handler
Slip Type Luer Lock
Knob
Holder hole
Sphere
Perspex plate
Plastic screw
Acrylic Holder
Fig. 1. Photograph of the low-cost sphere test phantom.
4. Discussion
Fig. 2. Photograph showing glass spheres with outer diameters of 10.2, 11.1, 16.2, 18.6, 20.0 and 24.1 mm.
than 50% of that activity (of the order of 370 MBq). When using the same activity concentration of 0.18 MBq ml1 in each phantom, all six of the hot spheres could be visualised in the sphere phantom, whereas only four cold spheres could be visualised for the Jaszczak phantom. Fig. 3 shows typical images obtained from each phantom. The concentrations in both phantoms were different, where 34.48 MBq ml1 was in spheres of the phantom and 0.18 MBq ml1 in Jaszczak phantom. Contrast measurements for each phantom in these conditions are shown in Fig. 4. This shows that lower activity was required to perform the test. To further assess the observed contrast from each sphere the concentration was increased from 58 to 101 MBq ml1. The results are shown in Fig. 5. Increasing sphere volumes from 0.55 to 2.60 ml. resulted in increase in the contrast; however, a saturation effect was observed for higher volumes.
A range of commercial phantoms is now available for the quality control of SPECT cameras. These may offer a wide range of performance measures; however, from our survey no phantom appears to be comprehensive. It can therefore be an expensive undertaking to purchase phantoms capable of measuring all the quality control parameters. In many departments, financial restrictions can dictate that this may not be feasible and, therefore, quality control testing is not carried out. This has obvious implications for the standard of the clinical service. There is clearly a need for an inexpensive and practical option for routine performance measurements. This technical note provides a platform for enabling this approach. Based on a modified design previously published (Perkins et al., 2007) a low-cost phantom with glass blown spheres was successfully fabricated and tested to evaluate SPECT camera performance. A simplified method was suggested for the performance test. The triple-head SPECT system was able to detect six spheres in the sphere test phantom. The ability to detect ‘‘cold’’ regions in the commercial phantom tested was lower than the detection of ‘‘hot’’ regions in the sphere test phantom. This was as expected, since gamma cameras are known to be superior at detecting small discrete hot areas rather than similarly sized cold areas. For the hot regions the results showed that the contrast increased non-linearly with the increase of sphere volume. It is also important to appreciate that the sphere test phantom requires considerably less radioactivity than the cylindrical phantom and consequently reduces the exposure to the users, this being an important consideration during regular use. The most important aspect of this note, however, is the significant cost difference between the sphere test phantom and the commercial phantom which was approximately US$ 1300. This technical note provides description of a simplified method for routine performance testing, that could be undertaken by most departments with limited scientific and financial resources. Adopting this approach should enable implementation of basic quality control measures in a wider range of centres.
ARTICLE IN PRESS A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
1867
Table 1 Comparison of commercially available SPECT/PET performance test phantoms and low cost sphere test phantom. Parameter/name
PET/SPECT performance phantom
Jaszczak SPECT phantom
NEMA IEC body phantom set
Sphere performance test phantom
Main features
Source tank, seven ‘‘cold’’ lesions insert, linearity/uniformity insert, ‘‘hot’’ lesion insert
Cylinder, cold rod, six solid spheres or hollow spheres
Body phantom, a lung insert, six spheres
Adult abdomen phantom, six borosilicate spheres, acrylic holder and Perspex plate
Yes Yes Yes Yes No Yes Yes Yes 1870.00
Yes Yes Yes Yes No Yes Yes Yes 2050.00
Yes No No No Yes Yes Yes Yes NA
Yes Yes Yes No Yes Yes Yes Yes 730.00
Performance test availability Spatial resolution Linearity Uniformity Centre of rotation Hot contrast Cold contrast Lesion detectability Object size Approximate cost (USD$)
Fig. 4. Contrast of different radionuclide concentration in various sphere volumes.
1.00 0.99 0.98
Contrast
0.97 Radionuclide Concentration (MBq ml-1) 101 92 81 71 58 57 46 35
0.96 0.95 0.94 0.93 0.92 0.91 0.00
Fig. 3. Reconstructed axial images of the phantoms. (a) Hot spheres in low cost sphere phantom. (b) Cold spheres in the commercial phantom.
2.00
4.00 Sphere Volume, ml
6.00
8.00
Fig. 5. Contrast comparison of images by sphere test phantom (34.48 MBq ml1) and commercial Deluxe Jaszczak phantom (0.18 MBq ml1).
ARTICLE IN PRESS 1868
A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
Acknowledgements The authors wish to thank the lecturers, physicists, technologists and engineers at the University of Malaya Medical Centre for support with this work. We would also wish to thank Mr. D. Clay of the Medical Physics and Clinical Engineering Department at Nottingham University Hospital NHS Trust for fabrication acrylic support and base in the sphere test phantom. This project was supported in part by University of Malaya Postgraduate Research Fund (PPP) (P0118/2006C). References Graham, L.S., Fahey, F.H., Madsen, M.T., Aswegen, A.V., Yester, M.V., 1995. Quantitation of SPECT performance, AAPM Report no. 52, report of AAPM Nuclear Medicine Committee Task Group 4. Med. Phys. 22 (4), 401–409.
Groch, M.W., Erwin, W.D., 2000. SPECT in the year 2000: basic principles. J. Nucl. Med. Technol. 28, 233–234. National Electrical Manufacturers Association (NEMA), 2001. Performance Measurements of Scintillation Cameras. NEMA Standards Publication NU 1-2002, NEMA, USA. Perkins, A.C., Clay, D., Lawes, S.C., 2007. A simple low cost phantom for the quality control of SPECT cameras. World J. Nucl. Med. 6, 35–39. Picker International, 1999. Manual of AXIS/IRIX Quality Assurance and Performance Testing (unpublished). Sprawls, P., 1995. Physical Principles of Medical Imaging, second ed. Medical Physics Publishing, Wisconsin.
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
Technical note
A low-cost phantom for simple routine testing of single photon emission computed tomography (SPECT) cameras A.H. Ng a,b, K.H. Ng a,, H. Dharmendra a, A.C. Perkins c a b c
Department of Biomedical Imaging, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia Engineering Services Division, Ministry of Health, Level 2-5, Block E6, Parcel E, Federal Government Administration Centre, 62590 Putrajaya, Malaysia Academic Medical Physics, Medical School, University of Nottingham, Nottingham NG7 2UH, UK
a r t i c l e in f o
a b s t r a c t
Article history: Received 21 April 2008 Received in revised form 2 September 2008 Accepted 21 October 2008
A simple sphere test phantom has been developed for routine performance testing of SPECT systems in situations where expensive commercial phantoms may not be available. The phantom was based on a design with six universal syringe hubs set in the frame to support a circular array of six glass blown spheres of different sizes. The frame was then placed into a water-filled CT abdomen phantom and scanned with a triple head camera system (Philips IRIXTM, USA). Comparison was made with a commercially available phantom (Deluxe Jaszczak phantom). Whereas the commercial phantom demonstrates cold spot resolution, an important advantage of the sphere test phantom was that hot spot resolution could be easily measured using almost half (370 MBq) of the activity recommended for use in the commercial phantom. Results showed that the contrast increased non-linearly with sphere volume and radionuclide concentration. The phantom was found to be suitable as an inexpensive option for daily performance tests. & 2008 Elsevier Ltd. All rights reserved.
Keywords: SPECT Phantom Quality control Performance test Contrast Lesion detection
1. Introduction Recent interest in medical imaging applications such as nuclear cardiology and tumour imaging has resulted in a progressive increase in the use of clinical single photon emission computed tomography (SPECT) imaging. As a consequence, there has been an increased utilisation of clinical SPECT systems. For optimal diagnostic use it is essential that routine performance evaluation is carried out as part of an ongoing quality assurance programme. The NEMA publication NU 1-2001 (NEMA, 2001) is the basic recommended standard for performance evaluation and acceptance testing of scintillation cameras. These guidelines were originally intended for use by manufacturers as a means for specifying the standards of equipment and were later modified by workers wishing to assess ongoing equipment performance. However, the methodology described in the NEMA guidelines is more complex than necessary for many departments to use on a routine basis. It is therefore of important practical value to develop a more accessible means for routine testing. More significantly in some countries, imaging equipment may be installed with little or no funding for maintenance, quality control and adequate scientific support. In such circumstances, commer-
Corresponding author. Tel.: +60 3 7949 2091.
E-mail addresses: [email protected], [email protected] (K.H. Ng). 0969-8043/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2008.10.010
cially produced phantoms can be prohibitively expensive and as a consequence quality control procedures are not implemented. This technical note describes the commercial options available and outlines the design and use of an alternative low-cost phantom for routine performance measurement of SPECT systems. We have also compared the use of the new fabricated phantom with a commercial SPECT phantom to determine to what extent the simplified methods could be used for performance testing on a routine basis in hospital with limited resources.
2. Materials and methods The design of the low cost phantom was based on the common parameters measured using commercial phantoms. The most important of these parameters are spatial resolution, contrast and sensitivity. After reviewing the currently available phantoms and their performances in measuring these parameters, the low cost phantom was designed and then fabricated. 2.1. Descriptions of the phantom The in house phantom was based on the design of Perkins et al. (2007). The phantom consisted of an acrylic support and base constructed in an ‘‘L’’ shape originally designed with a circular array of ‘‘blind hubs’’ to hold Luer lock syringes in a perpendicular
ARTICLE IN PRESS A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
array. The angle between each holder was 601 and the diameter across two opposite holders was 100 mm. When filled with radioactivity the syringes would form standard sources of radioactivity in a reproducible configuration, so that they could be seen as parallel rods during image reconstruction. For this study the syringes were replaced with spherical glass spheres. Glass was being used due to its superior durability, chemical and heat resistance. It is considered relative economic as the glass can be shaped according to the laboratory and clinical requirements. The issues of the fragility of the glass and difficulty of making reproducible spheres are relatively insignificant in this project as the cost and feasibility of the phantom are the main concerns. Spheres are generally used as a three-dimensional structure suitable for measurement in the three common imaging planes. Whilst glass does have problems with fragility, this was considered to be the most widely available and inexpensive method of production. Cylinders could be used for a more crude assessment in a single plane (Perkins et al., 2007); however, spheres were considered to be more representative of the clinical situation, when imaging small tumour, as for example. The spheres were constructed by using glass blowing technique. The manufacturing technique will depend upon the skill and dexterity of the glass blower. The volumes can be accurately measured by filling and measuring the amount of water contained in each sphere. Manufacturing reproducibility may be an issue, but in practice, this is not a problem since these phantoms are not intended for mass production. Once a range of object sizes have been produced and measured with calipers, these will form a basis for repeated measurements within a given centre or intercomparison between a small number of different SPECT cameras. The tolerance for the size, wall thickness is estimated to be 10–20% from its original measurement. Fig. 1 shows the phantom constructed for this study. Fig. 2 shows the fabricated glass spheres for the sphere test phantom. This was considered to be an economical method of production using techniques that would be widely available. The resulting product was of acceptable durability and resistant to chemical and heat effects. The stem of each single glass sphere was attached to a solid glass rod by a conically tapered ground glass joint with the inner (female) and outer (male) diameter of 5 mm. This type of joint is commonly used as a stopper for laboratory glassware. The outer diameters of six spheres varied from 10.20 to 24.10 mm. The length from the centre of each sphere to the end of the stem was equal, ensuring all the spheres were in the same plane for image acquisition and transverse reconstruction. In order to keep the spheres hold in place, an acrylic Perspex plate was added to support the spheres perpendicular to the acrylic holder wall. To simulate the clinical situation the assembled phantom was inserted into a CT abdomen phantom (Fig. 1). This was made from acrylic with a dimension of 333 mm w 183 mm h 262 mm depth, which was the approximate size of an adult abdomen. The volume of this empty D-shaped cylinder was 10 L. The phantom was filled with water to simulate attenuation and scattering effects on image quality. Once the sphere phantom was inserted the abdomen phantom was closed with acrylic end plate to cover the top of the phantom. A plastic O-ring was inserted as a seal between the plate and abdomen phantom to prevent any leakage of fluid and the plate was secured with 18 screws distributed equally around the edge.
2.2. SPECT performance tests All imaging was carried out at the Nuclear Medicine Unit, Department of Biomedical Imaging, University of Malaya Medical
1865
Centre, Kuala Lumpur, using a triple detector SPECT camera (Philips IRIXTM, USA). The scanning time and injected activity of the performance test evaluation recommended by NEMA NU-1 (2001) and Graham et al. (1995) were modified for optimum scanning. Due to the lower activity concentration in the Jaszczak phantom, the radioactivity injected was higher than low-cost phantom. Thus, the scanning time was increased in order to complete a performance test. The spheres were filled with a range of concentrations of Tc-99 m between 58 and 101 MBq ml1. The phantom was placed on the couch and SPECT acquisition was performed using 1201 continuous scanning mode with incremental view of 31 each. This acquisition took a data collection time of 90 min. Images were reconstructed on an Odyssey LX computer using filtered back projection, 0.11 cm1 attenuation correction and a Butterworth filter 1.20 cut-off for a 128 128 matrix (Picker International, 1999). For comparison, the Deluxe Jaszczak phantom (Data Spectrum Corporate, USA) was filled with 750 MBq Tc-99 m (giving a concentration of 0.18 MBq ml1). The image contrast was measured by defining regions of interest on the computer images in order to obtain the average pixel count rates. An ROI was drawn approximately 5% of the vicinity from the target volume. The final contrast was calculated by using the formula given by Groch et al. (2000) and Sprawls (1995): ‘‘Cold lesion’’ phantom contrast;
‘‘Hot lesion’’ phantom contrast;
C cold ¼
C hot ¼
CB CS CB
CS CB CS
(1)
(2)
where CS is the average pixel count in the sphere region and CB is the average pixel count in the background area. One initial concern regarding the construction of the sphere phantom was the possibility of radioactivity leaking from the ground glass joints. Several leak tests were carried out by monitoring the activity in the water inside the abdominal chamber and coloured ultramarine blue was injected into the cavities so that any leak could be observed. There was no evidence of any leakage.
3. Results Costs: A comparison of the capabilities and costs between the sphere test phantom and other commercially available phantoms is given in Table 1. For the ‘‘in house’’ sphere phantom, the acrylic holder was fabricated in the Department of Medical Physics and Clinical Engineering at University of Nottingham Hospitals NHS Trust, Nottingham, UK, at a cost in the order of US$ 100.00. The CT abdomen phantom was built at a cost of US$ 500.00. The spheres and supports were fabricated by The Glass Blowing Group of the Malaysian Nuclear Agency at a cost of US$ 120. Fabrication of the Perspex cover plate was done in the workshop of the Engineering Faculty, University of Malaya, at a cost of US$ 10.00. This resulted in a total cost of the phantom of US$ 730. 3.1. SPECT performance tests The first and obvious difference between the phantoms was that the sphere test was designed to measure hot spot detection, whereas the Jaszczak phantom demonstrated cold spot detection. The Jaszczak phantom required a nominal 750 MBq or higher activity for use, while the sphere phantom could be used with less
ARTICLE IN PRESS 1866
A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
CT abdomen Phantom Cap
CT abdomen Phantom
O-ring
Nylon Handler
Slip Type Luer Lock
Knob
Holder hole
Sphere
Perspex plate
Plastic screw
Acrylic Holder
Fig. 1. Photograph of the low-cost sphere test phantom.
4. Discussion
Fig. 2. Photograph showing glass spheres with outer diameters of 10.2, 11.1, 16.2, 18.6, 20.0 and 24.1 mm.
than 50% of that activity (of the order of 370 MBq). When using the same activity concentration of 0.18 MBq ml1 in each phantom, all six of the hot spheres could be visualised in the sphere phantom, whereas only four cold spheres could be visualised for the Jaszczak phantom. Fig. 3 shows typical images obtained from each phantom. The concentrations in both phantoms were different, where 34.48 MBq ml1 was in spheres of the phantom and 0.18 MBq ml1 in Jaszczak phantom. Contrast measurements for each phantom in these conditions are shown in Fig. 4. This shows that lower activity was required to perform the test. To further assess the observed contrast from each sphere the concentration was increased from 58 to 101 MBq ml1. The results are shown in Fig. 5. Increasing sphere volumes from 0.55 to 2.60 ml. resulted in increase in the contrast; however, a saturation effect was observed for higher volumes.
A range of commercial phantoms is now available for the quality control of SPECT cameras. These may offer a wide range of performance measures; however, from our survey no phantom appears to be comprehensive. It can therefore be an expensive undertaking to purchase phantoms capable of measuring all the quality control parameters. In many departments, financial restrictions can dictate that this may not be feasible and, therefore, quality control testing is not carried out. This has obvious implications for the standard of the clinical service. There is clearly a need for an inexpensive and practical option for routine performance measurements. This technical note provides a platform for enabling this approach. Based on a modified design previously published (Perkins et al., 2007) a low-cost phantom with glass blown spheres was successfully fabricated and tested to evaluate SPECT camera performance. A simplified method was suggested for the performance test. The triple-head SPECT system was able to detect six spheres in the sphere test phantom. The ability to detect ‘‘cold’’ regions in the commercial phantom tested was lower than the detection of ‘‘hot’’ regions in the sphere test phantom. This was as expected, since gamma cameras are known to be superior at detecting small discrete hot areas rather than similarly sized cold areas. For the hot regions the results showed that the contrast increased non-linearly with the increase of sphere volume. It is also important to appreciate that the sphere test phantom requires considerably less radioactivity than the cylindrical phantom and consequently reduces the exposure to the users, this being an important consideration during regular use. The most important aspect of this note, however, is the significant cost difference between the sphere test phantom and the commercial phantom which was approximately US$ 1300. This technical note provides description of a simplified method for routine performance testing, that could be undertaken by most departments with limited scientific and financial resources. Adopting this approach should enable implementation of basic quality control measures in a wider range of centres.
ARTICLE IN PRESS A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
1867
Table 1 Comparison of commercially available SPECT/PET performance test phantoms and low cost sphere test phantom. Parameter/name
PET/SPECT performance phantom
Jaszczak SPECT phantom
NEMA IEC body phantom set
Sphere performance test phantom
Main features
Source tank, seven ‘‘cold’’ lesions insert, linearity/uniformity insert, ‘‘hot’’ lesion insert
Cylinder, cold rod, six solid spheres or hollow spheres
Body phantom, a lung insert, six spheres
Adult abdomen phantom, six borosilicate spheres, acrylic holder and Perspex plate
Yes Yes Yes Yes No Yes Yes Yes 1870.00
Yes Yes Yes Yes No Yes Yes Yes 2050.00
Yes No No No Yes Yes Yes Yes NA
Yes Yes Yes No Yes Yes Yes Yes 730.00
Performance test availability Spatial resolution Linearity Uniformity Centre of rotation Hot contrast Cold contrast Lesion detectability Object size Approximate cost (USD$)
Fig. 4. Contrast of different radionuclide concentration in various sphere volumes.
1.00 0.99 0.98
Contrast
0.97 Radionuclide Concentration (MBq ml-1) 101 92 81 71 58 57 46 35
0.96 0.95 0.94 0.93 0.92 0.91 0.00
Fig. 3. Reconstructed axial images of the phantoms. (a) Hot spheres in low cost sphere phantom. (b) Cold spheres in the commercial phantom.
2.00
4.00 Sphere Volume, ml
6.00
8.00
Fig. 5. Contrast comparison of images by sphere test phantom (34.48 MBq ml1) and commercial Deluxe Jaszczak phantom (0.18 MBq ml1).
ARTICLE IN PRESS 1868
A.H. Ng et al. / Applied Radiation and Isotopes 67 (2009) 1864–1868
Acknowledgements The authors wish to thank the lecturers, physicists, technologists and engineers at the University of Malaya Medical Centre for support with this work. We would also wish to thank Mr. D. Clay of the Medical Physics and Clinical Engineering Department at Nottingham University Hospital NHS Trust for fabrication acrylic support and base in the sphere test phantom. This project was supported in part by University of Malaya Postgraduate Research Fund (PPP) (P0118/2006C). References Graham, L.S., Fahey, F.H., Madsen, M.T., Aswegen, A.V., Yester, M.V., 1995. Quantitation of SPECT performance, AAPM Report no. 52, report of AAPM Nuclear Medicine Committee Task Group 4. Med. Phys. 22 (4), 401–409.
Groch, M.W., Erwin, W.D., 2000. SPECT in the year 2000: basic principles. J. Nucl. Med. Technol. 28, 233–234. National Electrical Manufacturers Association (NEMA), 2001. Performance Measurements of Scintillation Cameras. NEMA Standards Publication NU 1-2002, NEMA, USA. Perkins, A.C., Clay, D., Lawes, S.C., 2007. A simple low cost phantom for the quality control of SPECT cameras. World J. Nucl. Med. 6, 35–39. Picker International, 1999. Manual of AXIS/IRIX Quality Assurance and Performance Testing (unpublished). Sprawls, P., 1995. Physical Principles of Medical Imaging, second ed. Medical Physics Publishing, Wisconsin.