Original article
Comparison of gamma (Anger) camera systems in terms of detective quantum efficiency using Monte Carlo simulation Ida Erikssona,b, Sven-A˚ke Starckd,e and Magnus Ba˚thb,c Aim The aim of the present study was to perform an extensive evaluation of available gamma camera systems in terms of their detective quantum efficiency (DQE) and determine their dependency on relevant parameters such as collimator type, imaging depth, and energy window using the Monte Carlo technique. Materials and methods The modulation transfer function was determined from a simulated 99mTc point source and was combined with the system sensitivity and photon yield to obtain the DQE of the system. The simulations were performed for different imaging depths in a water phantom for 13 gamma camera systems from four manufacturers. Results Except at very low spatial frequencies, the highest DQE values were found with a lower energy window threshold of around 130 keV for all systems. The height and shape of the DQE curves were affected by the collimator design and the intrinsic properties of the gamma camera systems. High-sensitivity collimators gave the highest DQE at low spatial frequencies, whereas the high-resolution and ultrahigh-resolution collimators showed higher DQE values at higher frequencies. The intrinsic resolution of the
Introduction Gamma camera investigations have a relatively long history, dating back to the late 1950s when the first Anger camera was introduced [1,2]. The Anger camera, consisting of a NaI scintillation crystal, a collimator, a set of photomultiplier tubes and electronic circuits for determination of the coordinates of scintillation events, is still the standard for most gamma camera systems used today. Although the image quality has improved over the years, for example, because of further development of photomultiplier tubes and electronic circuits [3], the quality of gamma camera images is still limited by the large amount of quantum noise, which mainly originates from the statistical nature of radioactive decay and from fluctuations in the detection process in the camera. A reduction in quantum noise can be obtained either by increasing the imaging time or by increasing the amount of injected activity so that a larger number of photons are available for the system. The user also has the opportunity to influence the performance of the system by the choice of size and position of the energy window and by the choice of collimator. The properties of the collimator, together with the intrinsic resolution, determine the finest spatial resolution that can be obtained in the final image. Large hole lengths and small hole c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0143-3636
system mainly affected the DQE curve at superficial depths. Conclusion The results indicate that the manufacturers have succeeded differently in their attempts to design a system constituting an optimal compromise between sensitivity and spatial resolution. Nucl Med Commun c 2014 Wolters Kluwer Health | Lippincott 35:405–415 Williams & Wilkins. Nuclear Medicine Communications 2014, 35:405–415 Keywords: collimator, detective quantum efficiency, gamma camera, Monte Carlo technique a Department of Medical Physics, Karlstad Hospital, Karlstad, bDepartment of Radiation Physics, Institute of Clinical Sciences, the Sahlgrenska Academy at University of Gothenburg, cDepartment of Medical Physics and Biomedical Engineering, Sahlgrenska University Hospital, Gothenburg, dDepartment of Natural Science and Biomedicine, School of Health Sciences, Jo¨nko¨ping University and eHospital Physics, Department of Oncology, County Hospital Ryhov, Jo¨nko¨ping, Sweden
Correspondence to Ida Eriksson, MSc, Department of Medical Physics, Karlstad Hospital, SE-651 85 Karlstad, Sweden Tel: + 46 54 619 908; fax: + 46 54 616 370; e-mail: [email protected] Received 19 June 2013 Revised 4 October 2013 Accepted 25 October 2013
diameters increase the resolution of the system. In contrast, large hole diameters and small hole lengths increase the sensitivity of the system, allowing more photons to pass through the collimator and interact within the crystal, leading to lower noise. Historically, the parameters resolution and noise have mainly been analyzed separately in nuclear medicine [4–7]. However, in 2005 Starck et al. [8] proposed the use of detective quantum efficiency (DQE) as a relevant measure of the imaging properties of a gamma camera system. The DQE describes to what extent a system utilizes the information given as input as a function of spatial frequency and is dependent both on system sensitivity and on spatial resolution. The DQE is a concept widely used in, for example, digital projection radiography, as a measure of the performance of X-ray detectors [9–16]. Despite its less frequent use in nuclear medicine [17–19], there are some aspects that indicate that DQE actually may be of higher relevance for gamma camera systems than for X-ray detectors. First of all, DQE is of highest relevance when stochastic noise sources, such as quantum noise, have a large impact on the image quality, which is the case in nuclear medicine images. In contrast, in radiography the observer mostly performs DOI: 10.1097/MNM.0000000000000053
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
406 Nuclear Medicine Communications 2014, Vol 35 No 4
tasks that are mainly limited by the large amount of anatomical background contained in the images [16,20–30]. Although a ‘generalized DQE’ has been proposed, which, to some extent, takes this into account [31], DQE in its original form does not account for anatomical background. Second, the influence of the user on the DQE of the detector is more limited in radiography than in nuclear medicine, as the user in the former case can affect the DQE indirectly by changing the exposure and/or the beam quality, whereas in the latter case the DQE of the detector can be directly altered by, for example, adjusting the energy window. In a previous work by Eriksson et al. [32], it was shown that it is possible to use the Monte Carlo software SIMIND [33] for the determination of DQE of gamma camera systems. The Monte Carlo technique enables the simulation of complex systems and a time-efficient exploration of the relationship between the design and configuration of a gamma camera system and the resulting DQE. Therefore, the aim of the present study was to perform an extensive evaluation of available gamma camera systems in terms of their DQE and determine their dependency on relevant parameters such as collimator type, imaging depth, and energy window using the Monte Carlo technique.
Materials and methods Monte Carlo simulations
In the present work, SIMIND was used for DQE determination. SIMIND is a software system that describes a gamma (Anger) camera single-photon emission computed tomography (SPECT) system [33]. The user is able to model a gamma camera system by defining important imaging parameters of the desired system, such as collimator type, crystal thickness, intrinsic resolution, energy resolution, etc. However, the realism of the modeled system is limited and the camera’s spatial uniformity, energy response uniformity, and linearity are assumed to be ideal. Different sources and phantom setups can be modeled and the user chooses the type of acquisition (i.e. planar or SPECT imaging). After the Monte Carlo simulation, SIMIND produces result files that include the obtained image itself as well as a file containing other relevant measures of the simulation of the gamma camera system, such as the energy spectrum, the detector efficiency (the fraction between the number of photons detected in the energy window and the number of photons impinging on the crystal), and the sensitivity (the count rate in the energy window divided by the activity defined for the particular simulation). Values used in the simulations for a number of relevant parameters in SIMIND are given in Table 1. In contrast to previous work by Eriksson et al. [32], septal penetration and collimator scatter were taken into account in the
Description of parameter settings in the Monte Carlo simulations
Table 1
Parameters Collimator Crystal thickness Energy window Energy resolution Intrinsic resolution Pixel size Matrix size Number of photon histories Number of maximum scatter orders Phantom
Settings Varying (Table 2) 3/8 inch (0.95 cm) Upper energy threshold: 200 keV Lower energy threshold: varying Varying, 9.5–9.9% (Table 2) Varying, 0.33–0.38 cm (Table 2) 0.6 mm 1024 1 108 5 Rectangular water phantoms, width and height 50 cm respectively, varying thickness (20, 30, and 40 cm).
present work to increase the validity of the simulations. The photons were emitted in a wide solid angle of 901 to ensure that the whole camera surface was covered. The effect of changing the size of the energy window was studied by changing the lower energy threshold stepwise. The upper energy threshold was set to 200 keV for all simulations. With regard to the DQE determination, Eriksson et al. [32] showed that the complete frequency-dependent DQE can be determined from a single simulation of a point source, assuming that only uncorrelated quantum noise is present in the system. A 99mTc point source was therefore modeled and placed in a rectangular water phantom, which was positioned with a separation of 1 mm from the collimator. To study the effect of imaging depth on the DQE, the source was positioned in the middle of phantoms of different thicknesses (20, 30, and 40 cm). In addition, the source was placed superficially at 2 cm depth in the 20 cm phantom. At this depth, the collimator was randomly shifted in the imaging plane during the simulations to average out any effects on the DQE due to the lateral position of the source relative to the septal walls. Simulations were performed for 13 low-energy parallel hole collimators with different properties from four different manufacturers. The collimators were combined with the specific intrinsic properties of the gamma camera systems, as provided by the manufacturers. Table 2 describes the relevant parameters of the modeled systems. To investigate the effect of the intrinsic resolution on the DQE, additional simulations were performed in which the intrinsic resolution was altered for a given system. To estimate the size of the stochastic error in the DQE determination from the Monte Carlo simulation, repeated simulations of selected system setups were conducted. Detective quantum efficiency determination
As described above, the DQE describes to what extent an imaging system utilizes the incoming information and takes into consideration both system sensitivity and resolution,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 407
Table 2 Description of relevant parameters of simulated low-energy parallel hole collimators of types high sensitivity, general purpose, all purpose, high resolution, and ultrahigh resolution
Company and model
Intrinsic resolution (mm)a
Intrinsic energy resolution FWHM (%)a
GE Discovery
3.8
Mediso Anyscan
Type of collimator
Hole diameter (mm)
Hole length (mm)
Septal thickness (mm)
System sensitivity for 99mTc (cps/MBq)a
System resolution FWHM (mm)a
9.5
ELEGPb LEGP LEHR
2.50 1.90 1.50
40.0 35.0 35.0
0.40 0.20 0.20
144 121 72
10.3 9.0 7.4
3.6
9.7
LEHS LEGP LEHR LEUHR
2.50 1.90 1.50 1.40
26.0 35.0 35.0 40.0
0.30 0.20 0.16 0.16
378 117 72 45
14.1 8.9 7.4 6.5
Philips Brightview
3.3
9.6
LEGP LEHR
1.40 1.22
24.7 27.0
0.18 0.15
125 76
8.9 7.4
Siemens Symbia
3.8
9.9
LEHS LEAP LEHR LEUHR
2.54 1.45 1.11 1.16
24.1 24.1 24.1 35.8
0.36 0.20 0.16 0.13
459 149 91 45
15.6 9.4 7.5 6.0
FWHM, full-width at half-maximum. a Data obtained from the manufacturers. b Extended LEGP collimator.
which may be understood by the following expression [34]: MTF2 ðu; vÞ ; NNPSðu; vÞSNR2in
2.5
where MTF is the modulation transfer function, NNPS is the normalized noise power spectrum, and SNRin is the signal-to-noise ratio given as input to the system. In the present work, the DQE was determined using the relation [32]: DQEðu; vÞ ¼ MTF2 ðu; vÞDQEð0; 0Þ:
DQE × 10−7
1
2
The MTF was determined as the normalized absolute value of the Fourier transform of the point spread function obtained from the simulated point source image. The DQE(0,0) was determined as the ratio between the sensitivity (in counts per second per Bq) and the photon yield of the nuclide (89% for 99mTc). This relation is true only when the source is of small size; otherwise, the limited detector area will influence the result and lead to an underestimation of the DQE [32]. Hence, the use of the point source image is ideal also for this purpose. The sensitivities of the simulated systems were obtained from the SIMIND result files. The DQE was determined according to Eq. (2) for all simulated configurations described above. For presentation, a one-dimensional representation of the DQE was obtained by radial averaging of the two-dimensional DQE.
Results With regard to the stochastic uncertainty of the results, the relative SD of the obtained DQE was proportional to the inverse square root of the DQE value, irrespective of system, imaging parameters, and spatial frequency. For the number of emitted photons used in the present work (Table 1), the
Above 1 2 4 8 16 32 64 128 256 512 1024
2.0 Spatial frequency (cycles/cm)
DQEðu; vÞ ¼
Fig. 1
1.5
1.0
0.5
0.0 0
50
100 200 150 Lower energy threshold (keV)
250
Contour plot of DQE as a function of spatial frequency and lower energy window threshold for a GE GP system at an imaging depth of 10 cm. The lightest gray part of the plot corresponds to a DQE between 10 – 7 and 2 10 – 7, and the second lightest gray part corresponds to a DQE between 2 10 – 7 and 4 10 – 7, etc. DQE, detective quantum efficiency; GE, General Electric; GP, general purpose.
uncertainty was B0.1% for DQE values in the order of 10 – 4, B1% for DQE values around 10 – 6, etc. In Fig. 1, an example of a contour plot of DQE as a function of spatial frequency and lower energy window threshold is shown for a General Electric (GE) system
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
408 Nuclear Medicine Communications 2014, Vol 35 No 4
with a general-purpose (GP) collimator at an imaging depth of 10 cm. Except for very low spatial frequencies, the highest DQE for a given frequency was obtained with a lower energy window threshold of around 130 keV. However, for the lowest frequencies (below B0.05/cm) the lower energy threshold resulting in the highest DQE for a given frequency rapidly decreased with frequency. A similar behavior was seen for all systems irrespective of collimator type and manufacturer. This is reflected in Fig. 2, in which the DQE is shown as a function of lower energy threshold for the specific spatial frequencies of 0.03/cm (a) and 0.5/cm (b) for all 13 simulated gamma camera systems. As can be seen, the dependency of the DQE on the energy threshold is similar for all systems.
between the manufacturers was reduced for the GP and all-purpose (AP) systems (Fig. 5). Specifically, the DQE curves of the GE-extended GP system (ELEGP) decreased more rapidly with frequency, having, together with Siemens AP, the highest values at low spatial frequencies (< 0.2/cm) but the lowest values of all manufacturers at higher frequencies.
The energy window resulting in the highest DQE for a given system showed a small dependency on the imaging depth, as exemplified in Fig. 3, in which DQE contour plots for a Siemens high-resolution (HR) collimator at imaging depths of 2, 10, 15, and 20 cm are shown. For all systems, it increased from slightly below 130 keV at 2 cm depth to slightly above 130 keV at 20 cm depth for most frequencies. For the Siemens HR collimator (Fig. 3), it ranged from 125 to 133 keV for any imaging depth and any frequency above 0.2/cm. For a given frequency, the range was even smaller.
For a given manufacturer, the high-sensitivity (HS) collimator resulted in the highest DQE at low spatial frequencies, whereas the HR and ultrahigh-resolution (UHR) collimators showed higher DQE at higher frequencies. This is exemplified in Fig. 6, in which the DQE results for the four Siemens collimators at 2 and 10 cm depth are shown for a lower energy threshold of 130 keV. At 10 cm depth (Fig. 6b), the DQE was highest for the HS collimator for frequencies less than 0.4/cm, but at frequencies greater than 0.5/cm all the other three collimators had higher DQE. The AP collimator had the highest DQE in the frequency interval between 0.4 and 0.6/cm and the HR collimator had the highest DQE in the interval between 0.6 and 0.8/cm. For frequencies greater than 0.8/cm the UHR collimator gave the highest DQE values. A similar relationship between the collimators was seen at 2 cm depth (Fig. 6a), but shifted to higher spatial frequencies.
Figure 4 shows DQE curves for HR systems from four different manufacturers (GE, Mediso, Philips, and Siemens) at imaging depths of 2, 10, 15, and 20 cm with an energy window of 130–200 keV. At a given depth, the DQE curves from the four manufacturers had similar shape, whereas the Siemens and the Philips system resulted in slightly higher DQE compared with Mediso and GE. The similarity in the shape of the DQE curves
As described above, the DQE reached its maximum for most spatial frequencies and for all systems at a lower energy threshold of B130 keV. In Fig. 7, the DQE curves at 10 cm depth for Philips GP and Philips HR collimators with lower energy thresholds of 126 and 130 keV can be seen. Except at very low frequencies, the smaller energy window starting at 130 keV gave slightly higher DQE for both collimator types. It can also be seen that the HR
Fig. 2 (a)
(b) 2.5E−04
9.0E−06
Philips GP Siemens AP Mediso GP Philips HR GE LEGP Siemens HR Mediso HR GE HR GE ELEGP Mediso UHR Siemens UHR Mediso HS Siemens HS
8.0E−06 2.0E−04
7.0E−06 6.0E−06 DQE
DQE
1.5E−04
1.0E−04
5.0E−06 4.0E−06 3.0E−06 2.0E−06
5.0E−05
1.0E−06 0.0E+00
0.0E+00 0
50
100
150
Lower energy window threshold (keV)
200
0
50
100
150
200
Lower energy window threshold (keV)
DQE as a function of lower energy window threshold for 13 LE parallel hole collimators at 10 cm imaging depth at spatial frequencies of 0.03/cm (a) and 0.5/cm (b). AP, all purpose; DQE, detective quantum efficiency; GP, general purpose; HR, high resolution; HS, high sensitivity; LE, low-energy; UHR, ultrahigh-resolution.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 409
Fig. 3
(a)
(b)
2.5
2.5
DQE × 10−7
DQE × 10−7 2.0 Spatial frequency (cycles/cm)
Spatial frequency (cycles/cm)
2.0
1.5
1.0
0.5
1.5
1.0
0.5
0.0
0.0 0
50
150 100 Lower energy threshold (keV)
200
(c) 2.5
50
0
250
250
(d) 2.5
DQE × 10−7
DQE × 10−7
2.0 Spatial frequency (cycles/cm)
2.0 Spatial frequency (cycles/cm)
200 100 150 Lower energy threshold (keV)
1.5
1.0
1.5
1.0
0.5
0.5
0.0
0.0 0
50
100 150 Lower energy threshold (keV)
200
250
50
0
Above 1 2 4 8 16 32
100 150 Lower energy threshold (keV)
200
250
64 128 256 512 1024
Contour plots of DQE as a function of spatial frequency and lower energy window threshold for a Siemens HR system at imaging depths of 2 cm (a) and 10 cm (b) in a 20-cm-thick phantom, 15 cm in a 30-cm-thick phantom (c), and 20 cm in a 40-cm-thick phantom (d). The lightest gray part of the plot corresponds to a DQE between 10 – 7 and 2 10 – 7 and the second lightest gray part to a DQE between 2 10 – 7 and 4 10 – 7, etc. DQE, detective quantum efficiency; HR, high-resolution.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
410 Nuclear Medicine Communications 2014, Vol 35 No 4
Fig. 4
9.0E−05
3.5E−05
8.0E−05
3.0E−05
7.0E−05
2.5E−05
5.0E−05
DQE
DQE
6.0E−05
4.0E−05
2.0E−05 1.5E−05
3.0E−05
1.0E−05
2.0E−05
5.0E−05
1.0E−05 0.0E+00
0.0E+05 0
1
0.5
1.5
2
0
Spatial frequency (cm−1)
2
1.5
Spatial frequency (cm−1)
1.8E−05
1.0E−05
1.6E−05
9.0E−06 8.0E−06
1.4E−05
7.0E−06
1.2E−05
6.0E−06
1.0E−05
DQE
DQE
1
0.5
8.0E−05
5.0E−06 4.0E−06
6.0E−05
3.0E−06
4.0E−05
2.0E−06
2.0E−05
1.0E−06
0.0E+00
0.0E+00 0
0.5
1
1.5
0
Spatial frequency (cm−1)
0.5
1
1.5
Spatial frequency (cm−1) Siemens HR
Mediso HR
Philips HR
GE HR
DQE as a function of spatial frequency for HR collimators from different manufacturers (GE, Mediso, Philips, and Siemens) with energy window 130–200 keV at (a) 2 cm depth in a 20-cm-thick water phantom, (b) 10 cm depth in a 20-cm-thick phantom, (c) 15 cm depth in a 30-cm-thick phantom, and (d) 20 cm depth in a 40-cm-thick phantom. DQE, detective quantum efficiency; HR, high-resolution.
collimator resulted in higher DQE compared with the GP collimator at frequencies greater than 0.6/cm.
DQE between the two systems, whereas at 10 cm the DQE difference was small (< 5 10 – 7 at all frequencies).
The intrinsic spatial resolution affected the DQE mainly at superficial depths, which is illustrated in Fig. 8. In the figure, a standard Siemens HR system with intrinsic resolution of 3.8 mm is compared with a modified system for which the intrinsic resolution was set to 3.3 mm (the highest intrinsic resolution of any of the included systems, cf. Table 2). All other parameter values for the modified systems were equal to those of the standard Siemens HR system. At 2 cm, there was an evident difference in the
Discussion In the present work, a large number of gamma camera systems have been evaluated in terms of their DQE using the Monte Carlo technique. Using only a Monte Carlo simulation of the response of a system to a point source, the entire DQE of the system could be determined. This enabled numerous systems and parameter settings to be evaluated in an efficient way, leading to high precision in the results at modest computational times.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 411
Fig. 5
(a)
(b)
1.6E−04
6.0E−05
1.4E−04 5.0E-05 1.2E−04 4.0E−05 DQE
DQE
1.0E−04 8.0E−05
3.0E−05
6.0E−04 2.0E−05 4.0E−05 1.0E−05
2.0E−05 0.0E+05
0.0E+00 0
(c)
0.5 1 1.5 Spatial frequency (cm−1)
2
0
(d)
3.0E−05
0.5 1 Spatial frequency (cm−1)
1.5
0.4
0.8
1.6E−05 1.4E−05
2.5E−05 1.2E−05 1.0E−04 DQE
DQE
2.0E−05
1.5E−05
8.0E−06 6.0E−06
1.0E−05
4.0E−05 5.0E−05
2.0E−05 0.0E+00
0.0E+00 0
0.2
0.4
0.6
0.8
1
0
Spatial frequency (cm−1)
0.2
0.6
Spatial frequency (cm−1) Siemens AP
GE LEGP
GE ELEGP
Mediso GP
Philips GP DQE as a function of spatial frequency for extended low energy GP (ELEGP), GP and AP collimators from different manufacturers (GE, Mediso, Philips, and Siemens) with energy window 130–200 keV at (a) 2 cm depth in a 20-cm-thick water phantom, (b) 10 cm depth in a 20-cm-thick phantom, (c) 15 cm depth in a 30-cm-thick phantom, and (d) 20 cm depth in a 40-cm-thick phantom. AP, all-purpose; DQE, detective quantum efficiency; GP, general purpose.
The height and shape of the DQE curve was clearly affected by the collimator design. As illustrated in Figs 4 and 5, the more sensitive collimators resulted in higher DQE at low frequencies but the DQE decreased more rapidly as the frequency increased. This effect could also be seen for the ELEGP collimator, mainly intended for
use with 123I [35], when compared with regular, less sensitive, GP and AP collimators (Fig. 5). The resolution properties of the system became more important as the imaging depth increased, and at 10 cm depth the DQE curves for the HR and UHR collimators crossed the HS and GP curves at lower frequencies than at 2 cm imaging
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
412 Nuclear Medicine Communications 2014, Vol 35 No 4
Fig. 6
1.8E−04
5.0E−05
4.0E−04
4.5E−05
1.6E−04
3.5E−04
4.0E−05
1.4E−04
3.5E−05
3.0E−04 DQE
(b)
4.5E−04
3.0E−05
2.5E−04
2.5E−05 2.0E−04
7.0E−06 6.0E−06 5.0E−06
1.2E−04 DQE
(a)
1.0E−04
4.0E−06
8.0E−05
3.0E−06
2.0E−05
1.5E−04
1.5E−05
6.0E−05 2.0E−06
1.0E−04
1.0E−05
4.0E−05
5.0E−05
5.0E−06
2.0E−05
0.0E+06
0.0E+00
0.0E+00 0
2
0.5 1 1.5 Spatial frequency (cm−1)
1.0E−06 0.0E+06 0
Siemens HS
Siemens HR
Siemens AP
Siemens UHR
0.5 1 Spatial frequency (cm−1)
DQE as a function of spatial frequency for Siemens high-sensitivity (HS), all-purpose (AP), high-resolution (HR), and ultrahigh-resolution (UHR) collimators at (a) 2 cm depth and (b) 10 cm depth in a 20-cm-thick water phantom, energy window 130–200 keV. The figure uses two y-axes; the left axis corresponds to DQE values at frequencies below (a) 0.8/cm and (b) 0.5/cm and the right axis corresponds to DQE values at higher frequencies. DQE, detective quantum efficiency.
Fig. 7
6.0E−05
GP 126-200 keV GP 130-200 keV
5.0E−05
HR 126-200 keV
8.0E−06 7.0E−06
Buvat et al. [36] that despite its inferior spatial resolution the LEHR collimator is better than the LEUHR collimator for detection of small tumors in planar scintimammography.
DQE as a function of spatial frequency for Philips general-purpose (GP) and high-resolution (HR) collimators at 10 cm depth in a 20-cm-thick water phantom with energy windows 126–200 and 130–200 keV. The figure uses two y-axes; the left axis corresponds to DQE values at frequencies below 0.5/cm and the right axis corresponds to DQE values at higher frequencies. DQE, detective quantum efficiency.
A narrow energy window (corresponding to an B15% energy window) was preferable for all gamma camera systems because of the reduction in scattered radiation independent of imaging depth (Figs 2b and 3). A narrow energy window setting requires a good intrinsic energy resolution of the system. Although the intrinsic energy resolution varied from 9.5 to 9.9% for the different manufacturers, the variation in lower energy threshold resulting in the highest DQE was small, and all systems had sufficient energy resolution for a 15% energy window. The intrinsic energy resolution and the intrinsic spatial resolution are imaging parameters that are determined by the construction of the gamma camera system and cannot be changed by the user of the system. The effect of intrinsic resolution could mainly be seen at superficial depths (Fig. 8). Deeper in the phantom the resolution properties of the collimator became more important because of the larger distance between the source and the detector and also because of the increased amount of scattered radiation in the phantom.
depth (Fig. 6). However, the UHR collimator required relatively high frequencies before it resulted in the highest DQE. This corresponds well with the finding by
While designing a collimator, a compromise between sensitivity and spatial resolution has to be made. It can be argued that the optimal compromise is reached for a collimator for which the DQE is maximized. As such, the
HR 130-200 keV
4.0E−05
6.0E−06
DQE
5.0E−06
3.0E−05
4.0E−06 3.0E−06
2.0E−05
2.0E−06 1.0E−05
1.0E−06
0.0E+00 0
0.2
0.4
0.6
0.8
1
0.0E+06 1.2
Spatial frequency (cm−1)
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 413
Fig. 8
(a)
(b)
9.0E−05 8.0E−05
3.5E−05
3.0E−05
7.0E−05 2.5E−05
2.0E−05
5.0E−05
DQE
DQE
6.0E−05
4.0E−05
1.5E−05
3.0E−05 1.0E−05 2.0E−05 5.0E−06
1.0E−05
0.0E+00
0.0E+00 0
0.5 1 1.5 Spatial frequency (cm−1)
2
0
Intrinsic resolution 3.3 mm
0.2
0.4 0.6 Spatial frequency (cm−1)
0.8
1
Intrinsic resolution 3.8 mm
DQE as a function of spatial frequency for Siemens HR systems with intrinsic spatial resolution of 3.8 and 3.3 mm with energy window 130–200 keV at (a) 2 cm depth and (b) 10 cm depth in a 20-cm-thick water phantom. DQE, detective quantum efficiency; HR, high-resolution.
success of the compromise is dependent on, for example, imaging depth and imaging task (as different spatial frequencies are of different relevance for different imaging tasks). Therefore, it may be difficult to compare systems resulting in DQE curves with different shapes that cross each other. However, if the DQE curves differ only in magnitude, comparison is simplified and can be used to assess the design of the system. For example, Fig. 4 indicates that especially Philips, and to some degree also Siemens, has succeeded slightly better than GE and Mediso in designing an efficient HR collimator, whereas the differently shaped DQE curves in Fig. 5 complicate the assessment of the GP collimators. Although the higher DQE at superficial depths is partly due to the higher intrinsic resolution, the difference at larger depths is mainly due to differences in collimator design (cf. Fig. 8). In the simulations, the point source was located at different depths in uniform rectangular water phantoms of different thicknesses. In the clinical situation, the reviewer is disturbed by background noise in the region of interest, originating from uptake of activity in other organs, blood, and soft tissue. Any effects from the background noise are not taken into account by the DQE concept. As previously described, the importance of the anatomical background compared with the importance of the quantum noise is lower in nuclear medicine than in
radiography. Nevertheless, the extent to which the background noise affects the clinical interpretation of the results obtained in the present work should be investigated. A further limitation in the present work is the fact that the Monte Carlo simulations account only for the quantum noise. Other noise sources or sources of nonhomogeneity in the images, such as nonuniformity of the gamma camera, are not included. For example, at spatial frequencies below 0.5/cm, the noise power spectrum may be dominated by nonstochastic noise [37]. From this it can be understood that the DQE of the systems in the present study is overestimated at low spatial frequencies. Alternatively, the results can be seen to represent the ideal DQE for perfectly calibrated gamma camera systems for which sources to nonstochastic noise have been removed. The extent to which the excluded noise sources affect the final results is not investigated here, but the influence can be expected to differ depending on the energy window and collimator type being used. For example, the nonuniformity of the gamma camera has a larger effect on an image acquired with a narrow energy window [38]. However, in the previous work by Eriksson et al. [32] a comparison between Monte Carlo-generated DQE curves and the experimental determinations of DQE performed by Starck et al. [8] was made, and a good agreement was found irrespective of energy window, collimator, and spatial frequency. This indicates that, for a
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
414 Nuclear Medicine Communications 2014, Vol 35 No 4
well-calibrated gamma camera system, the quantum noise is the dominating noise source affecting the DQE, in turn indicating the potential of the Monte Carlo technique for determination of the DQE. In addition to the potential systematic errors discussed here, a Monte Carlo simulation is always associated with stochastic errors. The large amount of photons included in the present simulations resulted in small stochastic errors for relevant spatial frequencies. Although the DQE takes both sensitivity and resolution into account, it may be difficult to compare different systems if the DQE curves have different shapes and cross each other, as noted above. Despite these difficulties, the DQE concept has recently, to some extent, been used for optimization in nuclear medicine imaging. For example, Robert et al. [18,19] used the DQE concept to find the most suitable collimator to improve the high frequency content in CdZnTe gamma camera scintimammography. Furthermore, Jeon and Gyuseong [17] compared the performance of a coded aperture collimator and pinhole collimators with the DQE for a small gamma camera system. Taking into account the arguably higher relevance of DQE in nuclear medicine than in radiography and the already wide use of the concept in radiography, there is potential for an increase in the use of DQE for evaluating the performance of gamma camera systems. In summary, the present paper gives an extensive evaluation of the DQE of gamma camera systems, which was performed using the Monte Carlo technique. The dependency of the DQE on imaging properties such as imaging depth, energy window, and collimator type has been investigated. The results indicate that the manufacturers have succeeded differently in their attempt at designing a system constituting an optimal compromise between sensitivity and spatial resolution.
2 3 4
5
6
7
8
9
10
11
12
13
14 15
16
17
18
Acknowledgements
19
The authors thank Michael Ljungberg for answering questions and concerns regarding the simulations in SIMIND.
20
¨ rebro Regional This work was supported by the Uppsala-O Research Council (RFR-314501), (RFR-229021), (RFR153131), the Centre for Clinical Research Va¨rmland (LIVFOU-299291), (LIVFOU-217791), (LIVFOU-110761, LIVFOU-149561), and the Swedish Federal Government under the LUA/ALF agreement (ALFGBG-147491).
21
22
23
24
Conflicts of interest
There are no conflicts of interest.
25
References
26
1
Anger HO. A new instrument for mapping gamma-ray emitters. Biol Med Quart Rep 1957; UCRL-3653:8.
Anger HO. Scintillation camera. Rec Sci Instrum 1958; 29:27. Petersen TE, Furenlid LR. SPECT detectors: the Anger camera and beyond. Phys Med Biol 2011; 56:R145–R182. Imbert L, Poussier S, Franken PR, Songy B, Verger A, Morel O, et al. Compared performance of high-sensitivity cameras dedicated to myocardial perfusion SPECT: a comprehensive analysis of phantom and human images. J Nucl Med 2012; 53:1897–1903. Rasouli M, Takavar A, Ay MR, Saber S, Loudos G. Effects of crystal pixel size and collimator geometry on the performance of a pixelated crystal g-camera using Monte Carlo simulation. J Nucl Med Technol 2010; 38: 199–204. A¨rlig A˚, Jacobsson L, Larsson A, Ljungberg M, Wikkelso¨ C. Selection of collimator for rCBF studies and evaluation of triple-headed SPET using noise-resolution plots. Nucl Med Commun 1997; 18:655–661. Weinmann AL, Hruska CB, O’Connor MK. Design of optimal collimation for dedicated molecular breast imaging systems. Med Phys 2009; 36: 846–856. Starck S-A˚, Ba˚th M, Carlsson S. The use of detective quantum efficiency (DQE) in evaluating the performance of gamma camera systems. Phys Med Biol 2005; 50:1601–1609. Dobbins JT III, Ergun DL, Rutz L, Hinshaw DA. DQE(f) of four generations of computed radiography acquisition devices. Med Phys 1995; 22: 1581–1593. Kengyelics SM, Davies AG, Cowen AR. A comparison of the physical imaging properties of Fuji ST-V, ST-VA and ST-VN computed radiography imaging plates. Med Phys 1998; 25:2163–2169. Flynn MJ, Samei E. Experimental comparison of noise and resolution for 2k and 4k storage phosphor radiography systems. Med Phys 1999; 26: 1612–1623. Moy JP. Signal-to-noise ratio and spatial resolution in X-ray electronic imagers: is the MTF a relevant parameter. Med Phys 2000; 27: 86–93. Granfors PR, Aufrichtig R. Performance of a 41 41-cm2 amorphous silicon flat panel X-ray detector for radiographic imaging applications. Med Phys 2000; 27:1324–1331. Samei E, Flynn MJ. An experimental comparison of detector performance for computed radiography systems. Med Phys 2002; 29:447–459. Ba˚th M, Sund P, Ma˚nsson LG. Evaluation of the imaging properties of two generations of a CCD-based system for digital chest radiography. Med Phys 2002; 29:2286–2297. Sund P, Ba˚th M, Kheddache S, Ma˚nsson LG. Comparison of visual grading analysis and determination of detective quantum efficiency for evaluating system performance in digital chest radiography. Eur Radiol 2004; 14: 48–58. Jeon H, Gyuseong C. The detective quantum efficiency (DQE) for evaluating the performance of a small gamma camera system with uniformly redundant array (URA) collimator. Nucl Instrum Methods Phys Res A 2008; 591: 279–281. Robert C, Monte´mont G, Rebuffel V, Buvat I, Gue´rin L, Verger L. Simulation-based evaluation and optimization of a new CdZnTe gammacamera architecture (HiSens). Phys Med Biol 2010; 55:2709–2726. Robert C, Monte´mont G, Rebuffel V, Verger L, Buvat I. Optimization of parallel hole collimator/CdZnTe gamma-camera architecture for scintimammography. Med Phys 2011; 38:1806–1819. Samei E, Flynn MJ, Eyler WR. Detection of subtle lung nodules: relative influence of quantum and anatomic noise on chest radiographs. Radiology 1999; 213:727–734. Bochud FO, Valley J-F, Verdun FR, Hessler C, Schnyder P. Estimation of the noisy component of anatomical backgrounds. Med Phys 1999; 26:1365–1370. Burgess AE, Jacobson FL, Judy PF. Human observer detection experiments with mammograms and power-law noise. Med Phys 2001; 28:419–437. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Kheddache S, Grahn A, Ruschin M, et al. Nodule detection in digital chest radiography: introduction to the radius chest trial. Radiat Prot Dosimetry 2005; 114:85–91. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Kheddache S, Grahn A, Bochud FO, et al. Nodule detection in digital chest radiography: part of image background acting as pure noise. Radiat Prot Dosimetry 2005; 114: 102–108. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Hoeschen C, Tischenko O, Kheddache S, et al. Nodule detection in digital chest radiography: effect of anatomical noise. Radiat Prot Dosimetry 2005; 114:109–113. Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Flinck A, Ullman G, Ma˚nsson LG. Nodule detection in digital chest radiography: effect of nodule location. Radiat Prot Dosimetry 2005; 114:92–96.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 415
27 Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Johnsson A˚A, Ma˚nsson LG. Nodule detection in digital chest radiography: effect of system noise. Radiat Prot Dosimetry 2005; 114:97–101. 28 Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Grahn A, Ruschin M, et al. Nodule detection in digital chest radiography: summary of the radius chest trial. Radiat Prot Dosimetry 2005; 114:114–120. 29 Tingberg A, Ba˚th M, Ha˚kansson M, Medin J, Besjakov J, Sandborg M, et al. Evaluation of image quality of lumbar spine images: a comparison between FFE and VGA. Radiat Prot Dosimetry 2005; 114:53–61. 30 Ruschin M, Timberg P, Ba˚th M, Hemdal T, Svahn T, Saunders R, et al. Dose dependence of mass and microcalcification detection in digital mammography: free response human observer studies. Med Phys 2007; 34:400–407. 31 Richard S, Siewerdsen JH, Jaffray DA, Moseley DJ, Bakhtiar B. Generalized DQE analysis of radiographic and dual-energy imaging using flat-panel detectors. Med Phys 2005; 32:1397–1414. 32 Eriksson I, Starck S-A˚, Ba˚th M. Determination of the detective quantum efficiency of gamma camera systems: a Monte Carlo study. Radiat Prot Dosimetry 2010; 139:219–227.
33
34
35
36
37
38
Ljungberg M, Strand SEA. Monte Carlo program for the simulation of scintillation camera characteristics. Comput Methods Programs Biomed 1989; 29:257–272. Dobbins JT III. Image quality metrics for digital systems. In: Beutel J, Kundel HL, Van Metter RL, editors. Handbook of medical imaging vol 1 physics and psychophysics. Bellingham, WA: SPIE Press; 2000. pp. 161–222. Larsson A, Jakobson MoS, Ljungberg M, Riklund K. Dopamine D2 receptor SPECT with 123I-IBZM: evaluation of collimator and post-filtering when using model-based compensation – a Monte Carlo study. Phys Med Biol 2010; 55:1971–1988. Buvat I, Laffont S, Le Cloirec J, Bourguet P, Di Paola R. Importance of the choice of the collimator for the detection of small lesions in scintmammography: a phantom study. Phys Med Biol 2001; 46:1343–1355. Grossman LW, Anderson MP, Jennings RJ, Kruger JB, Lukes SJ, Wagner RF, Warr CP. Noise analysis of scintillation camera images: stochastic and non-stochastic effects. Phys Med Biol 1986; 31:941–953. International Atomic Energy Agency (IAEA). IAEA quality control atlas for scintillation camera systems. Vienna: IAEA; 2003.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Comparison of gamma (Anger) camera systems in terms of detective quantum efficiency using Monte Carlo simulation Ida Erikssona,b, Sven-A˚ke Starckd,e and Magnus Ba˚thb,c Aim The aim of the present study was to perform an extensive evaluation of available gamma camera systems in terms of their detective quantum efficiency (DQE) and determine their dependency on relevant parameters such as collimator type, imaging depth, and energy window using the Monte Carlo technique. Materials and methods The modulation transfer function was determined from a simulated 99mTc point source and was combined with the system sensitivity and photon yield to obtain the DQE of the system. The simulations were performed for different imaging depths in a water phantom for 13 gamma camera systems from four manufacturers. Results Except at very low spatial frequencies, the highest DQE values were found with a lower energy window threshold of around 130 keV for all systems. The height and shape of the DQE curves were affected by the collimator design and the intrinsic properties of the gamma camera systems. High-sensitivity collimators gave the highest DQE at low spatial frequencies, whereas the high-resolution and ultrahigh-resolution collimators showed higher DQE values at higher frequencies. The intrinsic resolution of the
Introduction Gamma camera investigations have a relatively long history, dating back to the late 1950s when the first Anger camera was introduced [1,2]. The Anger camera, consisting of a NaI scintillation crystal, a collimator, a set of photomultiplier tubes and electronic circuits for determination of the coordinates of scintillation events, is still the standard for most gamma camera systems used today. Although the image quality has improved over the years, for example, because of further development of photomultiplier tubes and electronic circuits [3], the quality of gamma camera images is still limited by the large amount of quantum noise, which mainly originates from the statistical nature of radioactive decay and from fluctuations in the detection process in the camera. A reduction in quantum noise can be obtained either by increasing the imaging time or by increasing the amount of injected activity so that a larger number of photons are available for the system. The user also has the opportunity to influence the performance of the system by the choice of size and position of the energy window and by the choice of collimator. The properties of the collimator, together with the intrinsic resolution, determine the finest spatial resolution that can be obtained in the final image. Large hole lengths and small hole c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0143-3636
system mainly affected the DQE curve at superficial depths. Conclusion The results indicate that the manufacturers have succeeded differently in their attempts to design a system constituting an optimal compromise between sensitivity and spatial resolution. Nucl Med Commun c 2014 Wolters Kluwer Health | Lippincott 35:405–415 Williams & Wilkins. Nuclear Medicine Communications 2014, 35:405–415 Keywords: collimator, detective quantum efficiency, gamma camera, Monte Carlo technique a Department of Medical Physics, Karlstad Hospital, Karlstad, bDepartment of Radiation Physics, Institute of Clinical Sciences, the Sahlgrenska Academy at University of Gothenburg, cDepartment of Medical Physics and Biomedical Engineering, Sahlgrenska University Hospital, Gothenburg, dDepartment of Natural Science and Biomedicine, School of Health Sciences, Jo¨nko¨ping University and eHospital Physics, Department of Oncology, County Hospital Ryhov, Jo¨nko¨ping, Sweden
Correspondence to Ida Eriksson, MSc, Department of Medical Physics, Karlstad Hospital, SE-651 85 Karlstad, Sweden Tel: + 46 54 619 908; fax: + 46 54 616 370; e-mail: [email protected] Received 19 June 2013 Revised 4 October 2013 Accepted 25 October 2013
diameters increase the resolution of the system. In contrast, large hole diameters and small hole lengths increase the sensitivity of the system, allowing more photons to pass through the collimator and interact within the crystal, leading to lower noise. Historically, the parameters resolution and noise have mainly been analyzed separately in nuclear medicine [4–7]. However, in 2005 Starck et al. [8] proposed the use of detective quantum efficiency (DQE) as a relevant measure of the imaging properties of a gamma camera system. The DQE describes to what extent a system utilizes the information given as input as a function of spatial frequency and is dependent both on system sensitivity and on spatial resolution. The DQE is a concept widely used in, for example, digital projection radiography, as a measure of the performance of X-ray detectors [9–16]. Despite its less frequent use in nuclear medicine [17–19], there are some aspects that indicate that DQE actually may be of higher relevance for gamma camera systems than for X-ray detectors. First of all, DQE is of highest relevance when stochastic noise sources, such as quantum noise, have a large impact on the image quality, which is the case in nuclear medicine images. In contrast, in radiography the observer mostly performs DOI: 10.1097/MNM.0000000000000053
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
406 Nuclear Medicine Communications 2014, Vol 35 No 4
tasks that are mainly limited by the large amount of anatomical background contained in the images [16,20–30]. Although a ‘generalized DQE’ has been proposed, which, to some extent, takes this into account [31], DQE in its original form does not account for anatomical background. Second, the influence of the user on the DQE of the detector is more limited in radiography than in nuclear medicine, as the user in the former case can affect the DQE indirectly by changing the exposure and/or the beam quality, whereas in the latter case the DQE of the detector can be directly altered by, for example, adjusting the energy window. In a previous work by Eriksson et al. [32], it was shown that it is possible to use the Monte Carlo software SIMIND [33] for the determination of DQE of gamma camera systems. The Monte Carlo technique enables the simulation of complex systems and a time-efficient exploration of the relationship between the design and configuration of a gamma camera system and the resulting DQE. Therefore, the aim of the present study was to perform an extensive evaluation of available gamma camera systems in terms of their DQE and determine their dependency on relevant parameters such as collimator type, imaging depth, and energy window using the Monte Carlo technique.
Materials and methods Monte Carlo simulations
In the present work, SIMIND was used for DQE determination. SIMIND is a software system that describes a gamma (Anger) camera single-photon emission computed tomography (SPECT) system [33]. The user is able to model a gamma camera system by defining important imaging parameters of the desired system, such as collimator type, crystal thickness, intrinsic resolution, energy resolution, etc. However, the realism of the modeled system is limited and the camera’s spatial uniformity, energy response uniformity, and linearity are assumed to be ideal. Different sources and phantom setups can be modeled and the user chooses the type of acquisition (i.e. planar or SPECT imaging). After the Monte Carlo simulation, SIMIND produces result files that include the obtained image itself as well as a file containing other relevant measures of the simulation of the gamma camera system, such as the energy spectrum, the detector efficiency (the fraction between the number of photons detected in the energy window and the number of photons impinging on the crystal), and the sensitivity (the count rate in the energy window divided by the activity defined for the particular simulation). Values used in the simulations for a number of relevant parameters in SIMIND are given in Table 1. In contrast to previous work by Eriksson et al. [32], septal penetration and collimator scatter were taken into account in the
Description of parameter settings in the Monte Carlo simulations
Table 1
Parameters Collimator Crystal thickness Energy window Energy resolution Intrinsic resolution Pixel size Matrix size Number of photon histories Number of maximum scatter orders Phantom
Settings Varying (Table 2) 3/8 inch (0.95 cm) Upper energy threshold: 200 keV Lower energy threshold: varying Varying, 9.5–9.9% (Table 2) Varying, 0.33–0.38 cm (Table 2) 0.6 mm 1024 1 108 5 Rectangular water phantoms, width and height 50 cm respectively, varying thickness (20, 30, and 40 cm).
present work to increase the validity of the simulations. The photons were emitted in a wide solid angle of 901 to ensure that the whole camera surface was covered. The effect of changing the size of the energy window was studied by changing the lower energy threshold stepwise. The upper energy threshold was set to 200 keV for all simulations. With regard to the DQE determination, Eriksson et al. [32] showed that the complete frequency-dependent DQE can be determined from a single simulation of a point source, assuming that only uncorrelated quantum noise is present in the system. A 99mTc point source was therefore modeled and placed in a rectangular water phantom, which was positioned with a separation of 1 mm from the collimator. To study the effect of imaging depth on the DQE, the source was positioned in the middle of phantoms of different thicknesses (20, 30, and 40 cm). In addition, the source was placed superficially at 2 cm depth in the 20 cm phantom. At this depth, the collimator was randomly shifted in the imaging plane during the simulations to average out any effects on the DQE due to the lateral position of the source relative to the septal walls. Simulations were performed for 13 low-energy parallel hole collimators with different properties from four different manufacturers. The collimators were combined with the specific intrinsic properties of the gamma camera systems, as provided by the manufacturers. Table 2 describes the relevant parameters of the modeled systems. To investigate the effect of the intrinsic resolution on the DQE, additional simulations were performed in which the intrinsic resolution was altered for a given system. To estimate the size of the stochastic error in the DQE determination from the Monte Carlo simulation, repeated simulations of selected system setups were conducted. Detective quantum efficiency determination
As described above, the DQE describes to what extent an imaging system utilizes the incoming information and takes into consideration both system sensitivity and resolution,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 407
Table 2 Description of relevant parameters of simulated low-energy parallel hole collimators of types high sensitivity, general purpose, all purpose, high resolution, and ultrahigh resolution
Company and model
Intrinsic resolution (mm)a
Intrinsic energy resolution FWHM (%)a
GE Discovery
3.8
Mediso Anyscan
Type of collimator
Hole diameter (mm)
Hole length (mm)
Septal thickness (mm)
System sensitivity for 99mTc (cps/MBq)a
System resolution FWHM (mm)a
9.5
ELEGPb LEGP LEHR
2.50 1.90 1.50
40.0 35.0 35.0
0.40 0.20 0.20
144 121 72
10.3 9.0 7.4
3.6
9.7
LEHS LEGP LEHR LEUHR
2.50 1.90 1.50 1.40
26.0 35.0 35.0 40.0
0.30 0.20 0.16 0.16
378 117 72 45
14.1 8.9 7.4 6.5
Philips Brightview
3.3
9.6
LEGP LEHR
1.40 1.22
24.7 27.0
0.18 0.15
125 76
8.9 7.4
Siemens Symbia
3.8
9.9
LEHS LEAP LEHR LEUHR
2.54 1.45 1.11 1.16
24.1 24.1 24.1 35.8
0.36 0.20 0.16 0.13
459 149 91 45
15.6 9.4 7.5 6.0
FWHM, full-width at half-maximum. a Data obtained from the manufacturers. b Extended LEGP collimator.
which may be understood by the following expression [34]: MTF2 ðu; vÞ ; NNPSðu; vÞSNR2in
2.5
where MTF is the modulation transfer function, NNPS is the normalized noise power spectrum, and SNRin is the signal-to-noise ratio given as input to the system. In the present work, the DQE was determined using the relation [32]: DQEðu; vÞ ¼ MTF2 ðu; vÞDQEð0; 0Þ:
DQE × 10−7
1
2
The MTF was determined as the normalized absolute value of the Fourier transform of the point spread function obtained from the simulated point source image. The DQE(0,0) was determined as the ratio between the sensitivity (in counts per second per Bq) and the photon yield of the nuclide (89% for 99mTc). This relation is true only when the source is of small size; otherwise, the limited detector area will influence the result and lead to an underestimation of the DQE [32]. Hence, the use of the point source image is ideal also for this purpose. The sensitivities of the simulated systems were obtained from the SIMIND result files. The DQE was determined according to Eq. (2) for all simulated configurations described above. For presentation, a one-dimensional representation of the DQE was obtained by radial averaging of the two-dimensional DQE.
Results With regard to the stochastic uncertainty of the results, the relative SD of the obtained DQE was proportional to the inverse square root of the DQE value, irrespective of system, imaging parameters, and spatial frequency. For the number of emitted photons used in the present work (Table 1), the
Above 1 2 4 8 16 32 64 128 256 512 1024
2.0 Spatial frequency (cycles/cm)
DQEðu; vÞ ¼
Fig. 1
1.5
1.0
0.5
0.0 0
50
100 200 150 Lower energy threshold (keV)
250
Contour plot of DQE as a function of spatial frequency and lower energy window threshold for a GE GP system at an imaging depth of 10 cm. The lightest gray part of the plot corresponds to a DQE between 10 – 7 and 2 10 – 7, and the second lightest gray part corresponds to a DQE between 2 10 – 7 and 4 10 – 7, etc. DQE, detective quantum efficiency; GE, General Electric; GP, general purpose.
uncertainty was B0.1% for DQE values in the order of 10 – 4, B1% for DQE values around 10 – 6, etc. In Fig. 1, an example of a contour plot of DQE as a function of spatial frequency and lower energy window threshold is shown for a General Electric (GE) system
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
408 Nuclear Medicine Communications 2014, Vol 35 No 4
with a general-purpose (GP) collimator at an imaging depth of 10 cm. Except for very low spatial frequencies, the highest DQE for a given frequency was obtained with a lower energy window threshold of around 130 keV. However, for the lowest frequencies (below B0.05/cm) the lower energy threshold resulting in the highest DQE for a given frequency rapidly decreased with frequency. A similar behavior was seen for all systems irrespective of collimator type and manufacturer. This is reflected in Fig. 2, in which the DQE is shown as a function of lower energy threshold for the specific spatial frequencies of 0.03/cm (a) and 0.5/cm (b) for all 13 simulated gamma camera systems. As can be seen, the dependency of the DQE on the energy threshold is similar for all systems.
between the manufacturers was reduced for the GP and all-purpose (AP) systems (Fig. 5). Specifically, the DQE curves of the GE-extended GP system (ELEGP) decreased more rapidly with frequency, having, together with Siemens AP, the highest values at low spatial frequencies (< 0.2/cm) but the lowest values of all manufacturers at higher frequencies.
The energy window resulting in the highest DQE for a given system showed a small dependency on the imaging depth, as exemplified in Fig. 3, in which DQE contour plots for a Siemens high-resolution (HR) collimator at imaging depths of 2, 10, 15, and 20 cm are shown. For all systems, it increased from slightly below 130 keV at 2 cm depth to slightly above 130 keV at 20 cm depth for most frequencies. For the Siemens HR collimator (Fig. 3), it ranged from 125 to 133 keV for any imaging depth and any frequency above 0.2/cm. For a given frequency, the range was even smaller.
For a given manufacturer, the high-sensitivity (HS) collimator resulted in the highest DQE at low spatial frequencies, whereas the HR and ultrahigh-resolution (UHR) collimators showed higher DQE at higher frequencies. This is exemplified in Fig. 6, in which the DQE results for the four Siemens collimators at 2 and 10 cm depth are shown for a lower energy threshold of 130 keV. At 10 cm depth (Fig. 6b), the DQE was highest for the HS collimator for frequencies less than 0.4/cm, but at frequencies greater than 0.5/cm all the other three collimators had higher DQE. The AP collimator had the highest DQE in the frequency interval between 0.4 and 0.6/cm and the HR collimator had the highest DQE in the interval between 0.6 and 0.8/cm. For frequencies greater than 0.8/cm the UHR collimator gave the highest DQE values. A similar relationship between the collimators was seen at 2 cm depth (Fig. 6a), but shifted to higher spatial frequencies.
Figure 4 shows DQE curves for HR systems from four different manufacturers (GE, Mediso, Philips, and Siemens) at imaging depths of 2, 10, 15, and 20 cm with an energy window of 130–200 keV. At a given depth, the DQE curves from the four manufacturers had similar shape, whereas the Siemens and the Philips system resulted in slightly higher DQE compared with Mediso and GE. The similarity in the shape of the DQE curves
As described above, the DQE reached its maximum for most spatial frequencies and for all systems at a lower energy threshold of B130 keV. In Fig. 7, the DQE curves at 10 cm depth for Philips GP and Philips HR collimators with lower energy thresholds of 126 and 130 keV can be seen. Except at very low frequencies, the smaller energy window starting at 130 keV gave slightly higher DQE for both collimator types. It can also be seen that the HR
Fig. 2 (a)
(b) 2.5E−04
9.0E−06
Philips GP Siemens AP Mediso GP Philips HR GE LEGP Siemens HR Mediso HR GE HR GE ELEGP Mediso UHR Siemens UHR Mediso HS Siemens HS
8.0E−06 2.0E−04
7.0E−06 6.0E−06 DQE
DQE
1.5E−04
1.0E−04
5.0E−06 4.0E−06 3.0E−06 2.0E−06
5.0E−05
1.0E−06 0.0E+00
0.0E+00 0
50
100
150
Lower energy window threshold (keV)
200
0
50
100
150
200
Lower energy window threshold (keV)
DQE as a function of lower energy window threshold for 13 LE parallel hole collimators at 10 cm imaging depth at spatial frequencies of 0.03/cm (a) and 0.5/cm (b). AP, all purpose; DQE, detective quantum efficiency; GP, general purpose; HR, high resolution; HS, high sensitivity; LE, low-energy; UHR, ultrahigh-resolution.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 409
Fig. 3
(a)
(b)
2.5
2.5
DQE × 10−7
DQE × 10−7 2.0 Spatial frequency (cycles/cm)
Spatial frequency (cycles/cm)
2.0
1.5
1.0
0.5
1.5
1.0
0.5
0.0
0.0 0
50
150 100 Lower energy threshold (keV)
200
(c) 2.5
50
0
250
250
(d) 2.5
DQE × 10−7
DQE × 10−7
2.0 Spatial frequency (cycles/cm)
2.0 Spatial frequency (cycles/cm)
200 100 150 Lower energy threshold (keV)
1.5
1.0
1.5
1.0
0.5
0.5
0.0
0.0 0
50
100 150 Lower energy threshold (keV)
200
250
50
0
Above 1 2 4 8 16 32
100 150 Lower energy threshold (keV)
200
250
64 128 256 512 1024
Contour plots of DQE as a function of spatial frequency and lower energy window threshold for a Siemens HR system at imaging depths of 2 cm (a) and 10 cm (b) in a 20-cm-thick phantom, 15 cm in a 30-cm-thick phantom (c), and 20 cm in a 40-cm-thick phantom (d). The lightest gray part of the plot corresponds to a DQE between 10 – 7 and 2 10 – 7 and the second lightest gray part to a DQE between 2 10 – 7 and 4 10 – 7, etc. DQE, detective quantum efficiency; HR, high-resolution.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
410 Nuclear Medicine Communications 2014, Vol 35 No 4
Fig. 4
9.0E−05
3.5E−05
8.0E−05
3.0E−05
7.0E−05
2.5E−05
5.0E−05
DQE
DQE
6.0E−05
4.0E−05
2.0E−05 1.5E−05
3.0E−05
1.0E−05
2.0E−05
5.0E−05
1.0E−05 0.0E+00
0.0E+05 0
1
0.5
1.5
2
0
Spatial frequency (cm−1)
2
1.5
Spatial frequency (cm−1)
1.8E−05
1.0E−05
1.6E−05
9.0E−06 8.0E−06
1.4E−05
7.0E−06
1.2E−05
6.0E−06
1.0E−05
DQE
DQE
1
0.5
8.0E−05
5.0E−06 4.0E−06
6.0E−05
3.0E−06
4.0E−05
2.0E−06
2.0E−05
1.0E−06
0.0E+00
0.0E+00 0
0.5
1
1.5
0
Spatial frequency (cm−1)
0.5
1
1.5
Spatial frequency (cm−1) Siemens HR
Mediso HR
Philips HR
GE HR
DQE as a function of spatial frequency for HR collimators from different manufacturers (GE, Mediso, Philips, and Siemens) with energy window 130–200 keV at (a) 2 cm depth in a 20-cm-thick water phantom, (b) 10 cm depth in a 20-cm-thick phantom, (c) 15 cm depth in a 30-cm-thick phantom, and (d) 20 cm depth in a 40-cm-thick phantom. DQE, detective quantum efficiency; HR, high-resolution.
collimator resulted in higher DQE compared with the GP collimator at frequencies greater than 0.6/cm.
DQE between the two systems, whereas at 10 cm the DQE difference was small (< 5 10 – 7 at all frequencies).
The intrinsic spatial resolution affected the DQE mainly at superficial depths, which is illustrated in Fig. 8. In the figure, a standard Siemens HR system with intrinsic resolution of 3.8 mm is compared with a modified system for which the intrinsic resolution was set to 3.3 mm (the highest intrinsic resolution of any of the included systems, cf. Table 2). All other parameter values for the modified systems were equal to those of the standard Siemens HR system. At 2 cm, there was an evident difference in the
Discussion In the present work, a large number of gamma camera systems have been evaluated in terms of their DQE using the Monte Carlo technique. Using only a Monte Carlo simulation of the response of a system to a point source, the entire DQE of the system could be determined. This enabled numerous systems and parameter settings to be evaluated in an efficient way, leading to high precision in the results at modest computational times.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 411
Fig. 5
(a)
(b)
1.6E−04
6.0E−05
1.4E−04 5.0E-05 1.2E−04 4.0E−05 DQE
DQE
1.0E−04 8.0E−05
3.0E−05
6.0E−04 2.0E−05 4.0E−05 1.0E−05
2.0E−05 0.0E+05
0.0E+00 0
(c)
0.5 1 1.5 Spatial frequency (cm−1)
2
0
(d)
3.0E−05
0.5 1 Spatial frequency (cm−1)
1.5
0.4
0.8
1.6E−05 1.4E−05
2.5E−05 1.2E−05 1.0E−04 DQE
DQE
2.0E−05
1.5E−05
8.0E−06 6.0E−06
1.0E−05
4.0E−05 5.0E−05
2.0E−05 0.0E+00
0.0E+00 0
0.2
0.4
0.6
0.8
1
0
Spatial frequency (cm−1)
0.2
0.6
Spatial frequency (cm−1) Siemens AP
GE LEGP
GE ELEGP
Mediso GP
Philips GP DQE as a function of spatial frequency for extended low energy GP (ELEGP), GP and AP collimators from different manufacturers (GE, Mediso, Philips, and Siemens) with energy window 130–200 keV at (a) 2 cm depth in a 20-cm-thick water phantom, (b) 10 cm depth in a 20-cm-thick phantom, (c) 15 cm depth in a 30-cm-thick phantom, and (d) 20 cm depth in a 40-cm-thick phantom. AP, all-purpose; DQE, detective quantum efficiency; GP, general purpose.
The height and shape of the DQE curve was clearly affected by the collimator design. As illustrated in Figs 4 and 5, the more sensitive collimators resulted in higher DQE at low frequencies but the DQE decreased more rapidly as the frequency increased. This effect could also be seen for the ELEGP collimator, mainly intended for
use with 123I [35], when compared with regular, less sensitive, GP and AP collimators (Fig. 5). The resolution properties of the system became more important as the imaging depth increased, and at 10 cm depth the DQE curves for the HR and UHR collimators crossed the HS and GP curves at lower frequencies than at 2 cm imaging
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
412 Nuclear Medicine Communications 2014, Vol 35 No 4
Fig. 6
1.8E−04
5.0E−05
4.0E−04
4.5E−05
1.6E−04
3.5E−04
4.0E−05
1.4E−04
3.5E−05
3.0E−04 DQE
(b)
4.5E−04
3.0E−05
2.5E−04
2.5E−05 2.0E−04
7.0E−06 6.0E−06 5.0E−06
1.2E−04 DQE
(a)
1.0E−04
4.0E−06
8.0E−05
3.0E−06
2.0E−05
1.5E−04
1.5E−05
6.0E−05 2.0E−06
1.0E−04
1.0E−05
4.0E−05
5.0E−05
5.0E−06
2.0E−05
0.0E+06
0.0E+00
0.0E+00 0
2
0.5 1 1.5 Spatial frequency (cm−1)
1.0E−06 0.0E+06 0
Siemens HS
Siemens HR
Siemens AP
Siemens UHR
0.5 1 Spatial frequency (cm−1)
DQE as a function of spatial frequency for Siemens high-sensitivity (HS), all-purpose (AP), high-resolution (HR), and ultrahigh-resolution (UHR) collimators at (a) 2 cm depth and (b) 10 cm depth in a 20-cm-thick water phantom, energy window 130–200 keV. The figure uses two y-axes; the left axis corresponds to DQE values at frequencies below (a) 0.8/cm and (b) 0.5/cm and the right axis corresponds to DQE values at higher frequencies. DQE, detective quantum efficiency.
Fig. 7
6.0E−05
GP 126-200 keV GP 130-200 keV
5.0E−05
HR 126-200 keV
8.0E−06 7.0E−06
Buvat et al. [36] that despite its inferior spatial resolution the LEHR collimator is better than the LEUHR collimator for detection of small tumors in planar scintimammography.
DQE as a function of spatial frequency for Philips general-purpose (GP) and high-resolution (HR) collimators at 10 cm depth in a 20-cm-thick water phantom with energy windows 126–200 and 130–200 keV. The figure uses two y-axes; the left axis corresponds to DQE values at frequencies below 0.5/cm and the right axis corresponds to DQE values at higher frequencies. DQE, detective quantum efficiency.
A narrow energy window (corresponding to an B15% energy window) was preferable for all gamma camera systems because of the reduction in scattered radiation independent of imaging depth (Figs 2b and 3). A narrow energy window setting requires a good intrinsic energy resolution of the system. Although the intrinsic energy resolution varied from 9.5 to 9.9% for the different manufacturers, the variation in lower energy threshold resulting in the highest DQE was small, and all systems had sufficient energy resolution for a 15% energy window. The intrinsic energy resolution and the intrinsic spatial resolution are imaging parameters that are determined by the construction of the gamma camera system and cannot be changed by the user of the system. The effect of intrinsic resolution could mainly be seen at superficial depths (Fig. 8). Deeper in the phantom the resolution properties of the collimator became more important because of the larger distance between the source and the detector and also because of the increased amount of scattered radiation in the phantom.
depth (Fig. 6). However, the UHR collimator required relatively high frequencies before it resulted in the highest DQE. This corresponds well with the finding by
While designing a collimator, a compromise between sensitivity and spatial resolution has to be made. It can be argued that the optimal compromise is reached for a collimator for which the DQE is maximized. As such, the
HR 130-200 keV
4.0E−05
6.0E−06
DQE
5.0E−06
3.0E−05
4.0E−06 3.0E−06
2.0E−05
2.0E−06 1.0E−05
1.0E−06
0.0E+00 0
0.2
0.4
0.6
0.8
1
0.0E+06 1.2
Spatial frequency (cm−1)
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 413
Fig. 8
(a)
(b)
9.0E−05 8.0E−05
3.5E−05
3.0E−05
7.0E−05 2.5E−05
2.0E−05
5.0E−05
DQE
DQE
6.0E−05
4.0E−05
1.5E−05
3.0E−05 1.0E−05 2.0E−05 5.0E−06
1.0E−05
0.0E+00
0.0E+00 0
0.5 1 1.5 Spatial frequency (cm−1)
2
0
Intrinsic resolution 3.3 mm
0.2
0.4 0.6 Spatial frequency (cm−1)
0.8
1
Intrinsic resolution 3.8 mm
DQE as a function of spatial frequency for Siemens HR systems with intrinsic spatial resolution of 3.8 and 3.3 mm with energy window 130–200 keV at (a) 2 cm depth and (b) 10 cm depth in a 20-cm-thick water phantom. DQE, detective quantum efficiency; HR, high-resolution.
success of the compromise is dependent on, for example, imaging depth and imaging task (as different spatial frequencies are of different relevance for different imaging tasks). Therefore, it may be difficult to compare systems resulting in DQE curves with different shapes that cross each other. However, if the DQE curves differ only in magnitude, comparison is simplified and can be used to assess the design of the system. For example, Fig. 4 indicates that especially Philips, and to some degree also Siemens, has succeeded slightly better than GE and Mediso in designing an efficient HR collimator, whereas the differently shaped DQE curves in Fig. 5 complicate the assessment of the GP collimators. Although the higher DQE at superficial depths is partly due to the higher intrinsic resolution, the difference at larger depths is mainly due to differences in collimator design (cf. Fig. 8). In the simulations, the point source was located at different depths in uniform rectangular water phantoms of different thicknesses. In the clinical situation, the reviewer is disturbed by background noise in the region of interest, originating from uptake of activity in other organs, blood, and soft tissue. Any effects from the background noise are not taken into account by the DQE concept. As previously described, the importance of the anatomical background compared with the importance of the quantum noise is lower in nuclear medicine than in
radiography. Nevertheless, the extent to which the background noise affects the clinical interpretation of the results obtained in the present work should be investigated. A further limitation in the present work is the fact that the Monte Carlo simulations account only for the quantum noise. Other noise sources or sources of nonhomogeneity in the images, such as nonuniformity of the gamma camera, are not included. For example, at spatial frequencies below 0.5/cm, the noise power spectrum may be dominated by nonstochastic noise [37]. From this it can be understood that the DQE of the systems in the present study is overestimated at low spatial frequencies. Alternatively, the results can be seen to represent the ideal DQE for perfectly calibrated gamma camera systems for which sources to nonstochastic noise have been removed. The extent to which the excluded noise sources affect the final results is not investigated here, but the influence can be expected to differ depending on the energy window and collimator type being used. For example, the nonuniformity of the gamma camera has a larger effect on an image acquired with a narrow energy window [38]. However, in the previous work by Eriksson et al. [32] a comparison between Monte Carlo-generated DQE curves and the experimental determinations of DQE performed by Starck et al. [8] was made, and a good agreement was found irrespective of energy window, collimator, and spatial frequency. This indicates that, for a
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
414 Nuclear Medicine Communications 2014, Vol 35 No 4
well-calibrated gamma camera system, the quantum noise is the dominating noise source affecting the DQE, in turn indicating the potential of the Monte Carlo technique for determination of the DQE. In addition to the potential systematic errors discussed here, a Monte Carlo simulation is always associated with stochastic errors. The large amount of photons included in the present simulations resulted in small stochastic errors for relevant spatial frequencies. Although the DQE takes both sensitivity and resolution into account, it may be difficult to compare different systems if the DQE curves have different shapes and cross each other, as noted above. Despite these difficulties, the DQE concept has recently, to some extent, been used for optimization in nuclear medicine imaging. For example, Robert et al. [18,19] used the DQE concept to find the most suitable collimator to improve the high frequency content in CdZnTe gamma camera scintimammography. Furthermore, Jeon and Gyuseong [17] compared the performance of a coded aperture collimator and pinhole collimators with the DQE for a small gamma camera system. Taking into account the arguably higher relevance of DQE in nuclear medicine than in radiography and the already wide use of the concept in radiography, there is potential for an increase in the use of DQE for evaluating the performance of gamma camera systems. In summary, the present paper gives an extensive evaluation of the DQE of gamma camera systems, which was performed using the Monte Carlo technique. The dependency of the DQE on imaging properties such as imaging depth, energy window, and collimator type has been investigated. The results indicate that the manufacturers have succeeded differently in their attempt at designing a system constituting an optimal compromise between sensitivity and spatial resolution.
2 3 4
5
6
7
8
9
10
11
12
13
14 15
16
17
18
Acknowledgements
19
The authors thank Michael Ljungberg for answering questions and concerns regarding the simulations in SIMIND.
20
¨ rebro Regional This work was supported by the Uppsala-O Research Council (RFR-314501), (RFR-229021), (RFR153131), the Centre for Clinical Research Va¨rmland (LIVFOU-299291), (LIVFOU-217791), (LIVFOU-110761, LIVFOU-149561), and the Swedish Federal Government under the LUA/ALF agreement (ALFGBG-147491).
21
22
23
24
Conflicts of interest
There are no conflicts of interest.
25
References
26
1
Anger HO. A new instrument for mapping gamma-ray emitters. Biol Med Quart Rep 1957; UCRL-3653:8.
Anger HO. Scintillation camera. Rec Sci Instrum 1958; 29:27. Petersen TE, Furenlid LR. SPECT detectors: the Anger camera and beyond. Phys Med Biol 2011; 56:R145–R182. Imbert L, Poussier S, Franken PR, Songy B, Verger A, Morel O, et al. Compared performance of high-sensitivity cameras dedicated to myocardial perfusion SPECT: a comprehensive analysis of phantom and human images. J Nucl Med 2012; 53:1897–1903. Rasouli M, Takavar A, Ay MR, Saber S, Loudos G. Effects of crystal pixel size and collimator geometry on the performance of a pixelated crystal g-camera using Monte Carlo simulation. J Nucl Med Technol 2010; 38: 199–204. A¨rlig A˚, Jacobsson L, Larsson A, Ljungberg M, Wikkelso¨ C. Selection of collimator for rCBF studies and evaluation of triple-headed SPET using noise-resolution plots. Nucl Med Commun 1997; 18:655–661. Weinmann AL, Hruska CB, O’Connor MK. Design of optimal collimation for dedicated molecular breast imaging systems. Med Phys 2009; 36: 846–856. Starck S-A˚, Ba˚th M, Carlsson S. The use of detective quantum efficiency (DQE) in evaluating the performance of gamma camera systems. Phys Med Biol 2005; 50:1601–1609. Dobbins JT III, Ergun DL, Rutz L, Hinshaw DA. DQE(f) of four generations of computed radiography acquisition devices. Med Phys 1995; 22: 1581–1593. Kengyelics SM, Davies AG, Cowen AR. A comparison of the physical imaging properties of Fuji ST-V, ST-VA and ST-VN computed radiography imaging plates. Med Phys 1998; 25:2163–2169. Flynn MJ, Samei E. Experimental comparison of noise and resolution for 2k and 4k storage phosphor radiography systems. Med Phys 1999; 26: 1612–1623. Moy JP. Signal-to-noise ratio and spatial resolution in X-ray electronic imagers: is the MTF a relevant parameter. Med Phys 2000; 27: 86–93. Granfors PR, Aufrichtig R. Performance of a 41 41-cm2 amorphous silicon flat panel X-ray detector for radiographic imaging applications. Med Phys 2000; 27:1324–1331. Samei E, Flynn MJ. An experimental comparison of detector performance for computed radiography systems. Med Phys 2002; 29:447–459. Ba˚th M, Sund P, Ma˚nsson LG. Evaluation of the imaging properties of two generations of a CCD-based system for digital chest radiography. Med Phys 2002; 29:2286–2297. Sund P, Ba˚th M, Kheddache S, Ma˚nsson LG. Comparison of visual grading analysis and determination of detective quantum efficiency for evaluating system performance in digital chest radiography. Eur Radiol 2004; 14: 48–58. Jeon H, Gyuseong C. The detective quantum efficiency (DQE) for evaluating the performance of a small gamma camera system with uniformly redundant array (URA) collimator. Nucl Instrum Methods Phys Res A 2008; 591: 279–281. Robert C, Monte´mont G, Rebuffel V, Buvat I, Gue´rin L, Verger L. Simulation-based evaluation and optimization of a new CdZnTe gammacamera architecture (HiSens). Phys Med Biol 2010; 55:2709–2726. Robert C, Monte´mont G, Rebuffel V, Verger L, Buvat I. Optimization of parallel hole collimator/CdZnTe gamma-camera architecture for scintimammography. Med Phys 2011; 38:1806–1819. Samei E, Flynn MJ, Eyler WR. Detection of subtle lung nodules: relative influence of quantum and anatomic noise on chest radiographs. Radiology 1999; 213:727–734. Bochud FO, Valley J-F, Verdun FR, Hessler C, Schnyder P. Estimation of the noisy component of anatomical backgrounds. Med Phys 1999; 26:1365–1370. Burgess AE, Jacobson FL, Judy PF. Human observer detection experiments with mammograms and power-law noise. Med Phys 2001; 28:419–437. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Kheddache S, Grahn A, Ruschin M, et al. Nodule detection in digital chest radiography: introduction to the radius chest trial. Radiat Prot Dosimetry 2005; 114:85–91. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Kheddache S, Grahn A, Bochud FO, et al. Nodule detection in digital chest radiography: part of image background acting as pure noise. Radiat Prot Dosimetry 2005; 114: 102–108. Ba˚th M, Ha˚kansson M, Bo¨rjesson S, Hoeschen C, Tischenko O, Kheddache S, et al. Nodule detection in digital chest radiography: effect of anatomical noise. Radiat Prot Dosimetry 2005; 114:109–113. Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Flinck A, Ullman G, Ma˚nsson LG. Nodule detection in digital chest radiography: effect of nodule location. Radiat Prot Dosimetry 2005; 114:92–96.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
DQE comparison of gamma camera systems Eriksson et al. 415
27 Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Johnsson A˚A, Ma˚nsson LG. Nodule detection in digital chest radiography: effect of system noise. Radiat Prot Dosimetry 2005; 114:97–101. 28 Ha˚kansson M, Ba˚th M, Bo¨rjesson S, Kheddache S, Grahn A, Ruschin M, et al. Nodule detection in digital chest radiography: summary of the radius chest trial. Radiat Prot Dosimetry 2005; 114:114–120. 29 Tingberg A, Ba˚th M, Ha˚kansson M, Medin J, Besjakov J, Sandborg M, et al. Evaluation of image quality of lumbar spine images: a comparison between FFE and VGA. Radiat Prot Dosimetry 2005; 114:53–61. 30 Ruschin M, Timberg P, Ba˚th M, Hemdal T, Svahn T, Saunders R, et al. Dose dependence of mass and microcalcification detection in digital mammography: free response human observer studies. Med Phys 2007; 34:400–407. 31 Richard S, Siewerdsen JH, Jaffray DA, Moseley DJ, Bakhtiar B. Generalized DQE analysis of radiographic and dual-energy imaging using flat-panel detectors. Med Phys 2005; 32:1397–1414. 32 Eriksson I, Starck S-A˚, Ba˚th M. Determination of the detective quantum efficiency of gamma camera systems: a Monte Carlo study. Radiat Prot Dosimetry 2010; 139:219–227.
33
34
35
36
37
38
Ljungberg M, Strand SEA. Monte Carlo program for the simulation of scintillation camera characteristics. Comput Methods Programs Biomed 1989; 29:257–272. Dobbins JT III. Image quality metrics for digital systems. In: Beutel J, Kundel HL, Van Metter RL, editors. Handbook of medical imaging vol 1 physics and psychophysics. Bellingham, WA: SPIE Press; 2000. pp. 161–222. Larsson A, Jakobson MoS, Ljungberg M, Riklund K. Dopamine D2 receptor SPECT with 123I-IBZM: evaluation of collimator and post-filtering when using model-based compensation – a Monte Carlo study. Phys Med Biol 2010; 55:1971–1988. Buvat I, Laffont S, Le Cloirec J, Bourguet P, Di Paola R. Importance of the choice of the collimator for the detection of small lesions in scintmammography: a phantom study. Phys Med Biol 2001; 46:1343–1355. Grossman LW, Anderson MP, Jennings RJ, Kruger JB, Lukes SJ, Wagner RF, Warr CP. Noise analysis of scintillation camera images: stochastic and non-stochastic effects. Phys Med Biol 1986; 31:941–953. International Atomic Energy Agency (IAEA). IAEA quality control atlas for scintillation camera systems. Vienna: IAEA; 2003.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
