Adapting clinical gamma cameras for body monitoring in the
event of a large-scale radiological incident
JW Scuffham1,2, M Yip-Braidley2*, AL Shutt3, PJ Hinton1, A Nisbet1,2, DA Bradley2
1
Department of Medical Physics, Royal Surrey County Hospital NHS Foundation Trust, Guildford,
UK
2
Faculty of Engineering and Physical Sciences, University of Surrey, Guildford, UK
3
Centre for Radiation, Chemical and Environmental Hazards, Public Health England (PHE), Chilton,
UK
Email: [email protected]
Abstract
After a release of radionuclides, accidental or otherwise, there will be an urgent need to identify
members of the general public who have received a significant intake of radioactive material,
sufficient to require medical treatment or further investigation. A large number of people could be
contaminated in such an incident. For gamma-ray emitting radionuclides this screening could be
carried out using gamma camera medical imaging systems, such as those that are present in many
large UK hospital sites. By making a number of simple reversible changes such as removal of
collimators, these cameras could be employed as useful additional screening instruments as well as an
aid in contamination control. A study was carried out to investigate which systems were present in
sufficient number to offer wide scale coverage of UK population centres. Nine gamma cameras (eight
dual head and one single head) were assessed using point source and bottle mannequin (BOMAB)
phantom measurements so that a mathematical model could be developed for use with the MCNPX
Monte Carlo radiation transport code. The gamma camera models were assessed for practical seated
and supine geometries to give calibration factors for a list of target radionuclides that could be
released in a radiological incident. The Minimum Detectable Activities (MDAs) that were achieved
for a five minute measurement demonstrated that these systems are sufficiently sensitive to be used
for screening of the general public and are comparable to other body monitoring facilities. While
gamma cameras have on-board software that are designed for imaging and provide for a gamma-ray
energy range suitable for radionuclides for diagnostic imaging (such as 99mTc), they are not as
versatile as custom-built body monitoring systems.
* M Yip-Braidley is now at the The Institute of Cancer Research, 15 Cotswold Road, Sutton, SM2
5NG.
1. Introduction
In a radiological incident involving the accidental or deliberate release of radioactive material into the
environment, there may be a need to assess a large number of affected individuals for internal
radioactive contamination. International guidance exists for the triage, monitoring and treatment of
people exposed to ionizing radiation in a radiological incident (1), which includes recommendations
for establishing Radiation Monitoring Units (RMUs) at the site of the incident. In the UK, additional
guidance for the setting up and operation of RMUs has been published by Public Health England
(PHE) (2), who are also responsible for maintaining transportable body monitoring systems which can
be rapidly deployed in such an incident.
A very large-scale radiological incident may result in the need for additional capacity for body
monitoring beyond current provision. Many hospitals in the UK have Nuclear Medicine departments,
equipped with radiation monitoring devices and gamma cameras. These departments are staffed by
trained and experienced personnel, who are used to dealing with unsealed sources of radioactivity.
The availability of suitable equipment and expertise, together with the geographical spread of
hospitals across the country makes them ideally placed to facilitate additional body monitoring
capacity in large-scale radiological incidents as part of a centrally-controlled national contingency
plan.
Gamma cameras typically consist of large-area sodium iodide scintillation detectors that are highly
sensitive for the detection of low levels of radioactivity in the body. The suitability of gamma cameras
as mass screening devices in a radiological incident has been previously assessed in the USA (3)(4)
and Sweden (5). The use of gamma cameras for early radiological screening is of interest to several
European organizations involved in radiation protection. However, a study of the UK’s capability in
this area has not yet been reported.
In this paper, we have determined efficiency calibration factors of a number of current gamma
camera systems for a set of radionuclides that may be encountered following a radiation incident and
a range of possible subject ages. Physical calibration measurements were used to validate Monte
Carlo models of each gamma camera system, from which minimum detectable activities (MDAs)
were derived in realistic anthropomorphic digital phantoms. These were compared with MDAs for a
body monitoring facility currently used by PHE. In addition, we assessed the ease with which each
gamma camera system could be adapted to allow it to be used for screening through simple, reversible
changes to hardware and software settings.
2. Methods
2.1. Gamma camera survey
Comprehensive data on the types and geographic distribution of gamma camera systems at NHS sites
in the UK was a fundamental requirement for this project, and was determined using two sources of
data. The first was a selective extract of gamma camera types and locations from records held by the
Administration of Radioactive Substances Advisory Committee (ARSAC), under an agreement with
the Department of Health. When applying for or renewing an ARSAC certificate, clinicians must
submit information on the equipment in their department, including the makes and models of gamma
cameras. However, since ARSAC certificates must only be renewed every five years, and there is no
requirement to inform ARSAC of any changes in equipment, these data may contain out of date
information (6). Despite this shortcoming, the data currently represents the best available snapshot of
gamma camera provision in the UK.
The ARSAC data was incorporated into a Geographic Information System (7), which included
population data from the 2001 census and the locations of nuclear sites, supplied by PHE. This was
used to generate maps of the geographical spread of gamma cameras, and to identify the types of
camera closest to nuclear sites or areas of population densities. Based on this information, together
with the overall popularity of different cameras, a shortlist of gamma cameras was drawn up for
further investigation and characterisation as body monitoring devices. The aim was to select the
cameras that would optimise the national capability to respond quickly to a radiological emergency.
2.2. Gamma camera physical calibrations
Physical measurements were carried out on the shortlisted gamma camera models, for the purpose of
calibration and to provide a dataset that could be used to validate later Monte Carlo models (see
section 2.3). The systems were evaluated with their collimators removed, eliminating the capability to
perform spatial localisation, but improving the sensitivity of detection for low activities. Three
calibrated point sources (Amersham, UK) containing 241Am, 137Cs and 60Co were used to characterize
the response profile of the detector head. These radionuclides were chosen as they have a wide range
of gamma energy emissions (59.5, 661 and 1332.5 keV) to test the spectral capability of each gamma
camera. Each source was placed in turn on a polyurethane source stand (dimensions 13.5 x 18 x 5.5
cm), which was positioned directly on the imaging couch in the centre of the field of view. The
detector heads were moved to be equidistant from the point source at 15cm. Each source was
measured at different positions along the axial direction (-40, -20, -15, -10, -5, 0, +5, +10, +15, +20
and +40 cm relative to the initial position), using the imaging couch to translate the source. Static
images of 60 s duration were acquired at each position. For all acquisitions, the energy window was
set to the widest allowable by the camera acquisition software in order to maximize sensitivity. A
background measurement lasting 600 s was also acquired with no sources present. The background
corrected count rates were used to calculate detector efficiency factors for each of the three
radionuclides at each axial position.
To characterize the gamma camera in realistic measurement geometries, a BOttle Mannequin
ABsorber (BOMAB) anthropomorphic body phantom (7) was used. BOMAB phantoms consist of a
number of polyethylene containers that can be filled with radioactive solutions and assembled to
resemble people of different ages, Figure 1. To allow cross-comparison between different gamma
cameras, a completely sealed, leak-free phantom filled with a long-lived radionuclide that could be
transported between different hospital sites and easily assembled into different measurement
geometries is required. To facilitate this, standard BOMAB phantom containers were filled with a
radioactive gel mixture to minimise the chance of leakage during transport. The gel consisted of
water mixed with 1.1% agarose, 0.1 M hydrochloric acid, stable caesium chloride and 1050kBq of
137
Cs radioactive salt. The acidification maintains the chemical stability of the mixture, while the
addition of non-radioactive salts ensures that if any plating out onto the phantom surfaces occurs, the
non-radioactive salt is plated in greater quantity than the radioactive salt. These measures ensured
that the phantom activity distribution remained constant for the duration of the project.
Figure 1: BOMAB phantom assembled into adult configuration.
Gel BOMAB
phantom part name
Activity in gel
phantom part (Bq)
Fraction of total gel
phantom (%)
Fraction of
standard adult
BOMAB phantom
± 2σ
Head
74141 ± 534
7.03
5.79
Neck
11958 ± 214
1.13
1.49
Chest
295356 ± 1065
28.02
27.92
Pelvis
180010 ± 832
17.08
17.13
Arm 1
104177 ± 632
9.88
6.41
Arm 2
104032 ± 632
9.87
6.41
Thigh 1
71642 ± 524
6.80
11.14
Thigh 2
71554 ± 524
6.79
11.14
Shin 1
70552 ± 520
6.69
6.285
Shin 2
70755 ± 521
6.71
6.285
Table 1 - Description of the gel BOMAB phantom used in the study. The rightmost columns show the activity fractions in
the gel phantom compared to the standard 137Cs body monitoring phantom used at PHE.
The gel phantom was measured on the PHE body monitor and found that certain sections of the
phantom had more activity than anticipated which was attributed to the reuse of the 5 litre beakers
used to make up the gel. Ideally it would have been better to make up all 60 litres in a single
container, but due to the risk of burns and contamination from dispensing such large quantities of hot
gel, this was not possible. However, the difference between the arm, thigh and shin pairs was not
significant, so the phantom could still be used.
The following phantom ages were selected for the measurements: 3 months old, 8 and 12 years old
and adult (8). This range of ages not only represents the likely casualties in a radiological incident,
but also enabled the optimal measurement geometry on the gamma camera for each age group to be
determined. For the 3 month old configuration, the phantom was placed directly on the uncollimated
detector head, supported by a pillow; the other phantom configurations were placed in a seated
geometry with the detector heads positioned laterally at the maximum possible distance (Figure 2).
The seated geometry was chosen to allow maximal throughput of patients, whilst ensuring that the
detector heads are as close as possible to the torso, which is the most likely site of concentration for
inhaled (lungs) or ingested (stomach) radioactivity. This is also a reasonable geometry for detecting
radionuclides widely distributed in muscle (such as 134Cs or 137Cs).
Static images of each phantom were acquired for 60s with an open energy window. The
background corrected count rates recorded by the camera were used to calculate detector efficiency
factors in each of the measurement geometries.
Figure 2: BOMAB Phantom measurements on a Siemens Symbia gamma camera. Left: 3-month-old phantom placed
directly on the gamma camera head. Right: adult phantom in seated position.
2.3. Monte Carlo model development and validation
Performing physical calibrations of each gamma camera in realistic measurement geometries for a
range of different radionuclides would have been an unfeasible task. Monte Carlo models of each
gamma camera systems were therefore developed using MCNPX version 2.6 (10), and validated
against the physical calibration measurements. This enabled the use of more realistic digital
phantoms, incorporating a range of different radionuclides without the need for further physical
measurements.
Initial models of the gamma camera heads were constructed using information from service
manuals where available, together with physical measurements and photographs of key internal
components (Figure 3). The gamma camera models, together with digital versions of the BOMAB and
point source phantoms, were used to calculate the number of events registered in the detector for each
of the measurements performed during the physical calibrations. This was done using a pulse height
detector tally and accepting only events occurring within the actual energy windows used on each
gamma camera system. The simulations were carried out on one stand-alone eight-node server and a
64-node Beowulf cluster computer at PHE. Simulations were run for 1×108 particles, with a typical
simulation time of 60 mins. Detector efficiency factors were derived for each of the physical
measurement geometries (point source and BOMAB phantoms). Minor manual adjustments were then
made to the gamma camera model designs in order to minimize the difference between the
simulations and the physical measurements.
Figure 3: MCNPX model components of a gamma camera using a seated adult BOMAB phantom. Left image is a slice
through the body in the XZ plane, the right image the same model position but with a YZ view showing the camera heads.
2.4. Simulations with voxel phantoms
Anthropomorphic digital phantoms have a more realistic internal organ structure than the parametric
BOMAB phantoms used to validate the simulations. These digital phantoms are typically created
from medical imaging (MRI or CT) of a human subject and consist of a matrix of voxels, with each
voxel representing a certain material in the body.
Voxel phantoms for adult and paediatric subjects were used to determine calibration factors and
minimum detectable activity values. These phantoms were the PHE NORMAN adult male phantom
(11) and the University of Florida child phantoms (12) with ages of 9 months, 8 years and 14 years
old. These were chosen as representative of the spread of subject sizes found in the UK general public
and were closest to the BOMAB phantoms. The voxel phantoms were originally supplied in a supine
geometry. To accommodate simulations in a seated geometry, the phantoms were partitioned into
sections and reconfigured by moving the leg sections as shown in Figure 4. It should be noted that
deformable anatomical phantoms are under development (13) that will more accurately represent a
seated geometry. However, our geometry was deemed adequate for this application.
Figure 4: MCNPX screen capture showing the PHE NORMAN voxel phantom arranged into a seated position
Radioactivity can be incorporated into these digital phantoms at the voxel level, allowing realistic
bio-distributions to be simulated. However, for the majority of the radionuclides in this study, we
assumed a uniform distribution of radioactivity, representing early-phase generalised uptake in the
respiratory and gastrointestinal tracts and the blood pool. A similar assumption was used previously in
the assessment of handheld detectors for rapid screening (14). We simulated uniform distributions for
60
Co, 75Se, 90Sr, 137Cs, 192Ir, 226Ra, 237Np and 241Am, all of which have one or more gamma energies or
Bremsstrahlung x-rays that could be measured using the detector technology present in gamma
cameras. We also simulated more realistic 131I uptake in the thyroid compartments of the voxel
phantoms.
The voxel phantoms and Monte Carlo models were used to calculate the MDAs for each
radionuclide and gamma camera studied. The MDAs were calculated using the following equations
(15):
𝐿
𝑀𝐷𝐴 =
(1)
𝜀 𝑃𝛾 𝑡
𝐿 = 2.71 + 4.65√𝐵
(2)
Where L is the detection limit of the system based on the background count B. The MDA is also
dependent on the detector efficiency ε, the gamma emission probability Pγ, and the counting period t.
The detector efficiency is defined as the ratio of the number of counts per second detected (Nd) and
the number of gamma rays per second emitted by the source (N):
𝜀=
𝑁𝑑
𝑁𝛾
(3)
3. Results
3.1. Gamma camera survey
Figure 5 shows the geographical spread of gamma cameras in the UK, as determined from the survey
of data held by ARSAC. As expected, the systems tend to be clustered around major conurbations,
with rural and remote areas less well served. Table 2 shows the list of gamma cameras identified in
the survey, sorted by the number of machines installed. Also shown in Table 2 for each camera is the
distance to the nearest known nuclear site within the UK, calculated using data provided by PHE.
These data was used to draw up a shortlist of camera makes and models to assess for adaptation as
body monitors. In drawing up the shortlist, we considered not only the popularity and distance to
nuclear sites, but also the relative ages of the models listed. As already noted, there is no requirement
to inform ARSAC when a gamma camera is replaced, and some of the older systems listed in our
survey are likely to have been replaced. For this reason, some newer gamma camera models with
fewer machines installed were chosen, as they would be expected to replace the older systems
identified in the survey. Furthermore, a single-headed camera was included in the study for
comparison with the now-conventional dual-headed models. The systems highlighted in bold type in
Table 2 indicate those shortlisted.
Figure 5: Map of the locations of gamma cameras within the UK
Gamma Camera
GE Infinia
Siemens e.cam
Siemens Symbia
GE Millennium
Philips Skylight
GE DST
Philips/ADAC Forte
Philips Axis
Philips Brightview
Philips Argus
Siemens c-cam
Philips Cardo MD
Philips Vertex
Number of
Machines
69
60
43
28
26
22
20
17
12
6
6
5
5
Distance to nearest
nuclear site (km)
5
7
7
3
34
5
7
13
25
26
48
44
66
GE Discovery NM670
GE Ventri
Philips Irix
Nucline Mediso
Elscint SPX6
GE Starcam
Siemens Diacam
Siemens Orbiter
Discovery 530c
Nucline TH45
Elscint Apex
Elscint Helix
Elscint SP4M
Elscint SP6
Neurofocus SPECT
Nucline DHV
Philips Meridian
Philips Precedence
Philips Prism 2000
Siemens Digicam
Siemens Mobile
Scintronix GRC-1
Scintronix 480
4
3
3
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
15
61
47
7
34
31
53
79
57
47
57
71
48
34
37
7
69
28
58
65
38
53
64
Table 2: The gamma cameras models in the UK, ordered by the number of machines installed. For each camera, the shortest
distance to a nuclear site is also shown. The shortlisted gamma cameras are shown in bold type.
3.2. Gamma camera physical calibrations
All the gamma cameras studied were capable of being adapted for seated body monitoring by
removing the collimators, moving the imaging couch away from the gantry and positioning a chair
between the detector heads. Some camera models required Perspex (polymethlymethacrylate) covers
to be attached to the heads to protect the fragile scintillation crystal (Table 3). These covers may
adversely affect the sensitivity of the detector for radionuclides with low energy emissions or pure
beta emitters. Some of the cameras had safety interlocks that limited the movement of the gantry
when the collimators were removed. However, in all cases these interlocks could be disabled to allow
careful manual re-positioning of the gantry. Thus, on all systems it was possible to rotate the gantry
by 90 degrees to allow an infant to be placed on a pillow directly on the detector head. All changes
made during setup were reversible and easily carried out by an experienced operator.
For each gamma camera, an open energy window was set to cover the maximum possible spectral
range available. This was achieved using standard software tools available on the acquisitions
systems, which are readily accessible to the operator (although some cameras required the user to
login with enhanced privileges). Some systems required more than one energy window in order to
cover the entire energy range. Table 3 summarises the settings used on each of the systems. The
energy window lower limits were variable between systems, and reflect the slightly different pulse
height thresholds for event detection used by each gamma camera. Similarly, the systems studied had
a range of different upper energy window limits, determined by the camera multi-channel analyser.
These limits could not be altered using the acquisition software available. Some of the radionuclides
studied have full-energy peaks that are above the maximum energy range of the majority of clinical
gamma cameras (eg, 60Co, 1173.2 and 1332.5keV). Sensitivity for detection of these radionuclides
will therefore not be optimal. However, since the camera will register events that have been Compton
scattered, either in the patient or the detector itself, these radionuclides can still be measured.
Gamma Camera
GE Infinia
Siemens e.cam
Siemens Symbia
GE Millennium
Philips Skylight
Philips Forte
Philips Brightview
GE Discovery
Nucline Mediso
Number of energy
windows used
2
1
1
2
2
1
1
2
1
Energy window
range (keV)
49 – 511
50 – 511
50 – 511
40 – 1024
50 – 910
50 – 910
56 – 920
40 – 511
79 – 1400
Perspex
Covers?
Yes
No
No
Yes
No
No
No
Yes
No
Table 3: Energy window settings and Perspex protection covers for each gamma cameras evaluated in this study.
3.3. Monte Carlo model validation
The detector efficiencies calculated for the physical point source measurements were compared to the
equivalent figures derived from the Monte Carlo simulations. Figure 6 shows a comparison between
the physical and Monte Carlo point source results for the Siemens Symbia for 241Am, 137Cs, 60Co - one
of the gamma camera models evaluated in the study. The difference in the detector efficiencies
observed between the two gamma camera detector heads is due to attenuation of the imaging couch
upon which the point sources were placed. Consequently, the difference is most pronounced for the
lower energy of 241Am. The error bars represent a 5% error bound on the Monte Carlo simulation
result to account for the anticipated variation in source positioning and activity. Relatively close
agreement can be seen in the responses across the fields of view of both detector heads. The relative
differences between the simulation and physical measurements were calculated for each point source
position; Table 4 summarises these validation results for all camera models studied.
Figure 6: (a) 241Am (b) 137Cs and (c) 60Co point source results for Siemens Symbia
Gamma Camera
GE Infinia
Siemens e.cam
Siemens Symbia
GE Millennium
Philips Skylight
Philips Forte
Philips Brightview
GE Discovery
Nucline Mediso
Min
0.7
0.4
0.0
2.2
0.7
0.0
0.9
5.2
0.1
Relative difference (%) between physical and
Monte Carlo detector efficiencies
Am-241
Cs-137
Co-60
Max Mean Min
Max Mean Min
Max
37.7
15.0
5.0
23.8
12.0
2.1
33.1
11.7
3.4
0.4
38.1
13.9
0.6
11.9
30.2
7.2
0.9
18.1
8.6
0.7
14.0
105
22.9
1.0
69.2
28.8
2.5
40.4
23.8
6.1
17.0
31.6
23.4
5.2
20.5
54.4
13.0
2.7
27.2
15.0
0.1
10.8
203
85.4
0.4
10.6
6.2
0.3
12.0
37.7
22.6
1.1
18.6
10.3
6.0
26.3
24.5
6.1
2.4
9.2
6.5
0.0
15.5
Mean
12.2
6.2
5.3
14.8
11.6
3.6
6.2
14.8
4.1
Table 4: Summary of point source validation results
As can be seen in Table 4, for the majority of gamma camera models, the Monte Carlo and
physical measurement results differ by less than 15%. This was considered acceptable given the
limited details available on the construction of the detector heads and the experimental errors
associated with the physical measurements. The largest contribution to the relative difference occur
from the source for the detector head below the bed, which is greatest for the Am-241 source for the
GE Millennium and Philips Brightview. The contribution to the relative difference due to the bed is
50% of the total value when considering the relative efficiency for the upper detector only. This
suggests the mathematical model for the bed may be attenuating the radiation transmitted through it
differently to the actual physical gamma camera for certain models of gamma camera. For the seated
geometry for person measurements, the bed is not used.
Physical and simulated detector efficiencies were also calculated for each of the BOMAB phantom
configurations studied. The BOMAB phantom activity fraction in each gel phantom section from
Table 1 was used in the simulated source in the mathematical phantom. The percentage differences
between the physical and simulated results are reported in Table 5. The mean and 95% (2σ) standard
deviation for each phantom is 27.2 ± 39.7 (3 months), 45.9 ± 27.5 (8 years), 33.0 ± 27.5 (12 years)
and 30.4 ± 35.2 (adult). The differences between the physical and simulated results are greater for the
BOMAB phantoms than for the point source measurements. This reflects the more complex
geometry, particularly the off-axis contribution, which was not addressed in our point source
validation measurements. The gamma cameras also have limitations with respect to upper energy
windows for some models, and the Monte Carlo simulation only simulates the radiation detection
processes in the crystal, so any processing of the results by the gamma camera operations software
may not be taken fully into account.
The measurements of an actual person will be subject to a number of measurement uncertainties
with respect to variations in body size and height. It is unlikely that the average UK adult or child will
reflect the idealized BOMAB phantom. The chair used for a seated position will probably be different
to the standard metal framed chair used for this evaluation. The person will be measured in their
outdoor clothes and the background count within the hospital nuclear medicine department might vary
depending on the radioisotopes in use. This system is also only intended for emergency screening, so
whereas a total measurement uncertainty of 20% would be acceptable for laboratory measurements,
for radiation screening purposes, a total measurement uncertainty of 50% would be tolerated during a
radiological incident.
Gamma Camera
GE Infinia
Siemens e.cam
Siemens Symbia
GE Millennium
Philips Skylight
Philips Forte
Philips Brightview
GE Discovery
Nucline Mediso
Relative difference (%) between physical and
Monte Carlo detector efficiencies for the BOMAB phantom
3 month old
8 year old
12 year old
Adult
(supine)
(seated)
(seated)
(seated)
-16.0
-2.7
-31.7
-22.6
-18.4
14.5
-17.9
11.2
31.7
-34.1
20.4
43.3
44.1
35.1
44.5
-40.5
-35.7
-42.6
-11.3
-38.7
-29.5
-32.5
-25.8
-33.2
-30.1
-24.8
23.5
-38.2
-23.9
-31.7
-16.4
-36.5
-6.3
9.3
Table 5: Summary of BOMAB phantom validation results
3.4. Voxel phantom simulations
The validated Monte Carlo models of each gamma camera were used in conjunction with the voxel
phantoms to calculate MDA values for each radionuclide. The simulations assumed a 5 minute
acquisition time per subject, as this is the standard time currently used by PHE for their transportable
body monitoring equipment. Background acquisitions were simulated based on the physical
background results obtained for each system. Table 6 summarises the MDAs calculated for each
gamma camera, radionuclide and voxel phantom subject age. A comparison was also carried out with
the PHE static bed monitoring system with a five minute measurement count time. This system has
five 6” x 4” NaI(Tl) low background detectors and is located in a steel room constructed from 6”
naval armour plate to reduce the background within the room.
As can be seen from Table 6, the MDA is generally lowest for the 9 month old phantom. This is
due to the higher geometric efficiency afforded by placing the subject on a pillow directly on the
detector head. In general, the MDAs are greater for the seated child subjects when compared to the
seated adult subject. This is most likely to be due to the greater volume of the adult subjects bringing
them closer to the detector heads. Whilst close proximity to the detector heads affords high detection
efficiency, it is also likely that the efficiency will have a greater dependence on the subject geometry.
Small inter- and intra-subject variations in positioning are therefore likely to have a greater influence
on the MDA than if the subject was located further away from the detectors, although we did not
investigate this effect.
The simulations of the PHE static body monitor utilises the same measurement conditions of a five
minute screening measurement on a human subject as was applied when using a gamma camera. The
Monte Carlo simulation of the body monitor simulates a full spectrum measured from the
radionuclide and uses the full energy spectrum to determine MDAs for each of the radionuclides. The
results for the PHE body monitor are substantially lower than most of the gamma cameras for the 8
and 14 year old children and for adults. This is largely due to the use of a shielded room to reduce
intrinsic background and the five large volume NaI(Tl) detectors equally spaced around the subject
designed to measure gamma rays with energies higher than 300 keV.
The MDA for the 3 month infant measured in the PHE body monitor is higher than for most of the
gamma cameras. The standard supine measurement for infants on the PHE body monitor is to position
them at the head location for the adult which is further away from the detectors (approximately 30 cm
away) compared to a 3 cm infant to detector distance when directly placed on the camera head. Three
of the five detectors are at a substantial distance from the subject for the standard PHE subject
measurement geometry.
Gamma Camera
GE Infinia
9 month old
8 year old
14 year old
Adult
Siemens e.cam
9 month old
8 year old
14 year old
Adult
Siemens Symbia
9 month old
8 year old
14 year old
Adult
GE Millennium
9 month old
8 year old
14 year old
Adult
Philips
Skylight
9 month old
8 year old
14 year old
Adult
Philips Forte
9 month old
8 year old
14 year old
Adult
Philips
Brightview
9 month old
8 year old
14 year old
Adult
GE Discovery
9 month old
8 year old
14 year old
Adult
Nucline Mediso
9 month old
8 year old
14 year old
Adult
PHE Body
monitor
9 month old
8 year old
14 year old
Adult
241
60
217
686
827
968
99
232
229
239
66
147
141
145
89
217
217
224
68
126
232
123
31
72
74
78
288
847
976
1110
67
167
180
196
79
212
225
240
29
73
78
85
44016
122923
151010
165630
2727
7256
8256
9797
230
647
865
854
99
207
223
198
75
146
152
135
99
207
226
200
61
238
453
262
29
65
75
67
295
767
980
942
63
150
181
167
84
195
227
207
27
64
77
71
46150
114856
138105
128627
2635
6468
7955
7867
252
823
1088
1085
126
298
315
283
86
192
200
178
111
271
292
264
81
343
533
383
33
85
96
88
330
991
1257
1219
71
195
233
217
93
252
290
268
31
84
100
93
52331
147804
177024
168577
2829
8153
9923
9885
93
246
321
317
53
106
115
102
35
65
71
62
49
99
109
97
34
127
233
140
16
35
40
36
149
369
465
446
34
78
93
86
46
102
119
108
15
34
41
37
18583
46514
57525
55767
1308
3172
3772
3691
744
1601
2107
2069
168
258
294
267
84
130
147
134
184
281
321
290
180
466
1011
559
84
125
147
134
1077
2119
2695
2593
215
352
424
389
259
402
476
435
103
156
187
172
142791
234562
299335
300206
8041
15480
18453
18360
2579
5996
8202
8393
78
180
203
185
43
98
109
100
76
178
205
187
71
249
522
295
30
71
85
78
844
2038
2670
2634
90
217
263
248
101
239
286
265
40
89
108
101
127513
287105
370503
322957
4874
12150
13313
13685
585
1649
2261
2270
113
281
307
276
65
159
174
155
108
270
303
271
98
363
702
422
39
98
115
104
492
1355
1748
1694
92
241
292
270
118
303
359
327
43
103
124
114
86189
218924
251525
254702
4334
11885
14042
13556
208
592
787
776
101
229
246
218
70
147
155
137
92
210
231
203
61
247
504
284
27
67
77
69
284
767
988
943
58
150
182
167
76
192
227
204
28
65
78
72
38488
100874
119812
125495
2535
6471
7923
7867
2111
9869
14422
25892
130
581
744
1095
76
340
431
624
124
547
711
1061
101
176
313
436
43
192
257
390
789
3619
5315
8989
110
491
681
1069
139
617
839
1277
43
193
269
424
123531
566042
874290
1527241
4222
19478
27904
44564
702
884
1016
1250
86
93
96
102
28
31
32
33
64
71
73
78
47
67
80
82
31
35
36
39
715
854
948
1101
95
108
115
127
104
117
123
133
42
48
112
124
109547
115293
134199
14708
5377
6114
6506
7226
Bi
Co
137
Minimum Detectable Activity (Bq)
131
192
237
233
214
I
Ir
Np
Pa
Pb
214
Am
Cs
75
Se
Table 6: MDAs calculated for each camera, radionuclide and subject age
90
Sr
90
Y
Figure 7 compares the distribution of MDAs for each subject size for 241Am, 137Cs, 60Co and 90Sr.
Figure 7: Box and whisker plots illustrating the distribution of MDAs for each subject age for four example radionuclides.
The boxes represent the interquartile range and the whiskers represent the range of MDAs observed for all gamma cameras.
Overall, the lowest MDA was achieved by the GE Millennium, with 15 Bq for 75Se. This was the
only gamma camera we evaluated with a 12.7 mm thick scintillation crystal; all other cameras had 9.5
mm thick crystals. The highest MDA was 1.5 MBq for 90Sr on the Nucline Mediso. This was the only
single-head gamma camera evaluated, and the geometric efficiency is therefore greatly reduced.
Moreover, this model had the smallest field of view at 49 x 39cm compared with Siemens Symbia
with a field of view of 53 x 39 cm. This accounts for all the MDAs for this system being much higher
than for the other systems.
The MDAs for pure - emitting radionuclides are much greater than for radionuclides with x-ray or
gamma emissions in all of the gamma cameras studied, as illustrated by 90Sr in Figure 7. This is
because the majority of the events detected by the gamma camera are bremsstrahlung x-rays, which
have a broad spectrum that will be attenuated within the subject. We observed that the MDA is highly
dependent on the - emission energy, as can be seen by comparing the results for 90Y and 90Sr
(maximum emission energies 2.28 MeV and 0.55 MeV respectively). We believe this is because
higher energy bremsstrahlung emissions result in a larger component of Compton scattered photons
reaching the detector within the energy window range.
The influence of the low-energy threshold on MDA is illustrated by comparing the results for the
different radionuclides on the Nucline Mediso. This is the only system we evaluated with a lowenergy threshold above the principal emission energy of 241Am (59.5 keV). The MDA for 241Am for
this camera was therefore much greater relative to other radionuclides than for the other systems we
evaluated.
Figure 8: A screenshot of the spectrum collected on the Philips Forte gamma camera from a 60Co source. The x-axis is in
units of channels, and the y-axis is in units of counts. Despite the camera upper energy window of 926 keV (at around 2250
channels), a substantial number of counts are collected due to Compton scattering in the energy region 50 keV to 910 keV.
It was anticipated that the upper energy window limit would adversely affect the MDA for
radionuclides with principal photopeaks above this limit, such as 60Co. However, we did not observe
significantly greater MDA values for systems with lower energy window limits. For example, the
60
Co adult MDAs for the Siemens Symbia and Philips Skylight were 178 Bq and 134 Bq, despite their
energy window limits being 511 keV and 910 keV respectively. A screenshot of a Philips Forte
gamma camera in Figure 8 depicts a 60Co source spectrum with an upper cut off of 910 keV. The
Compton spectrum demonstrates that a large number of counts are recorded despite the energy cut-off
being too low to measure the 60Co full energy peaks at 1173 and 1332 keV directly.
We also anticipated that the presence of Perspex protective covers on some cameras may adversely
affect the MDAs for low-energy or beta-emitting radionuclides. However, we did not observe a
significant difference between the cameras with Perspex covers and those without. For example, the
MDAs for the Siemens Symbia for 90Sr and 241Am were 168.6 and 1.1 kBq respectively, whereas the
corresponding MDAs for the GE Discovery, which has Perspex covers, were actually lower (125.5
and 0.8 kBq respectively).
4. Discussion
We have shown that by making simple, reversible changes, clinical gamma cameras can be used as
efficient whole-body contamination monitoring devices. The wide geographical spread of gamma
cameras and their close correlation to population density shows that they could be a valuable source
of additional monitoring capacity in the event of a large-scale radiological incident. By measurement
and using Monte Carlo modelling techniques, we have carried out the most comprehensive study to
date of the sensitivity of gamma cameras to a range of key radionuclides. We have shown that the
limitations of clinical systems in terms of spectroscopic range do not significantly hinder the ability to
detect radionuclides with higher-energy gamma emissions. We also concluded that dual-headed
gamma cameras with thicker scintillation crystals are best suited to the task of body monitoring, but
that the more common thinner crystals still provide an acceptable level of sensitivity, with MDAs
comparable to body monitoring devices currently used by PHE. It should be noted, however, that the
restricted spectral information typically available from clinical gamma cameras means that there will
be limited ability to discriminate between different radionuclides if a subject were to be exposed to a
composite source. For this reason, the use of hospital gamma cameras is primarily useful in initial
screening only.
In this study, the collection of broad spectrum energy data from a wide range of gamma camera
systems has required varying levels of adaptation of acquisition software and hardware between
manufacturers and even between systems from the same manufacturer. For obvious reasons this
equipment is optimised for imaging 140 keV photons emitted by Technetium-99m, however, the
ability to perform list mode acquisitions and the existence of simple tools to manage data acquired in
this way would have greatly facilitated this work and the ability to use gamma cameras for this
purpose. Given the frequent developments in radiopharmaceuticals used for diagnosis and therapy in
nuclear medicine, the equipment manufacturers should be encouraged to build flexibility into their
data acquisition systems so that other applications can be investigated and developed. On a practical
level, one of the major challenges associated with using clinical gamma cameras in a radiological
emergency would be the control of contamination. Robust measures to prevent the contamination of
facilities and equipment with potentially long-lived radionuclides would have to be put in place before
the systems could be used. Nuclear Medicine staff are trained in handling unsealed radionuclides, and
all departments will maintain spill kits. However, these are likely to be insufficient for dealing with a
major incident. Items such as plastic sheeting to protect waiting areas, walkways and the gamma
camera heads themselves may not be readily available.
5. Conclusion
If a national policy is to be put in place for the use of gamma cameras in the event of a radiological
incident, this would need to be with the agreement of the individual departments, and clear guidance
to Nuclear Medicine staff would be required. This guidance should give specific instructions not just
on the technical details of adapting gamma cameras for the task of body monitoring, but also on
managing a large-scale triage unit and the control of contamination. In collaboration with the
Department of Health, future work on this project will include drafting such guidance, as well as
maintaining and updating the database of clinical gamma cameras in the UK, and deriving calibrations
for new gamma cameras as they enter the market. The major difficulty for the early stages of
radiological screening is getting information about the release: what radionuclides may be present and
in what areas as environmental monitoring may not have been carried out. Even in these situations,
the gamma cameras are more sensitive than many of the handheld radiation detectors that might be
available at the hospital. Even with a measurement of total counts of gamma radiation over the
background count, these cameras can be used effectively to identify those people who require
decontamination and follow up measurements when specialist body monitoring facilities are deployed
later in the incident.
This study was originally carried out in 2012 and although some of the gamma camera models
have been superseded by updated systems that have been installed at NHS sites, the methodology and
results should be indicative of the results that can be attained by these newer systems.
Acknowledgments
This project was funded by the Department of Health and the Home Office under the Policy
Research Programme (PRP) project “0470103 - Adapting Medical Imaging Systems in the Event of a
Radiological Incident”. The project team would like to thank the Department of Health and
Administration of Radioactive Substances Advisory Committee (ARSAC) for technical assistance in
obtaining data on the number of installed gamma cameras at NHS sites in the UK.
JWS was supported by an NIHR Chief Scientific Officer post-doctoral Fellowship during this
work.
The project team would also like to acknowledge the support and assistance of the staff at the
various hospitals who allowed us access to their clinical systems: Royal Surrey County Hospital, St
Richard’s Hospital, St Peter’s Hospital, Royal Berkshire Hospital, St Bartholomew’s Hospital,
Kingston Hospital and Maidstone Hospital.
References
1. Rojas-Palma, C., et al., et al. Triage, Monitoring and Treatment of people exposed to ionising
radiation following a malevolent act. s.l. : NRPA, 2009. 978-82-90362-27-5.
2. Thompson, N.J., et al., et al. Radiation Monitoring Units: Planning and Operational Guidance.
Health Protection Agency. Didcot, Oxfordshire, UK : s.n., 2011. HPA-CRCE-017.
3. Suitability of nuclear medicine gamma cameras as gamma spectrometers in the event of a
radiological emergency. Engdahl, J.C. and Bharwani, K. 2005, Nuclear Instruments and Methods in
Physics Research A, Vol. 533, pp. 569-577. doi:10.1016/j.nima.2005.07.051.
4. Anigstein, R., Olsher, R. H. and Loomis, D. A. Using Gamma Cameras to Assess Internal
Contamination from Intakes of Radioisotopes. Centers for Disease Control and Prevention. 2010.
[Accessed June 2012]. http://www.bt.cdc.gov/radiation/clinicians/evaluation/pdf/Instructions.pdf.
5. A gamma camera for measurements of internal contamination after a radiological accident.
Wallstrom, E., Alpsten, M. and Mattsson, S. 2, 1999, Journal of Radiological Protection, Vol. 19, pp.
143-154.
6. Administration of Radioactive Substances Advisory Committee. Notes for Guidance on the
Clinical Administration of Radiopharmaceuticals and Use of Sealed Radioactive Sources. 2006.
http://www.arsac.org.uk/notes_for_guidance/documents/ARSACNFG2006Corrected2011.pdf.
7. ArcGIS product webpage. [Online] [Cited: August 20, 2015.] https://www.arcgis.com/features/.
8. The BRMD BOMAB phantom family. Kramer, G.H., Burns, L. and Noel, L. 6, 1991, Health Physics,
Vol. 61, pp. 895-902.
9. Youngman, M. Transportable in vivo monitoring system for accident monitoring of internal
contamination. Didcot : National Radiological Protection Board, 2002. ISBN 0859514838.
10. Hendricks, J.S., et al., et al. MCNPX 2.6.0 Extensions. s.l. : Los Alamos Laboratory, 2008. LA-UR08-2216.
11. Induced current densities from low-frequency magnetic fields in a 2 mm resolution, anatomically
realistic model of the body. Dimbylow, P J. 1998, Physics in Medicine and Biology, Vol. 43, pp. 221230.
12. The UF Family of Reference Hybrid Phantoms for Computation Radiation Dosimetry . Lee, C., et
al., et al. 2, January 2010, Physics in Medicine and Biology, Vol. 55, pp. 339–363.
13. Na, Y.H, et al., et al. Mesh-Based and Anatomically Adjustable Adult Phantoms and a Case Study
in Virtual Calibration of a Lung Counter for Female Workers. [book auth.] X.G Xu and K.F. Eckerman.
Handbook of Anatomical Models for Radiation Dosimetry. Boca Raton, Florida : CRC Press, 2009.
14. Youngman, M.J., et al., et al. Guidance on Screening People for Internal Radioactive
Contamination. Health Protection Agency. Didcot, Oxfordshire, UK : s.n., 2011. HPA-CRCE-014. ISBN:
978-0-85951-687-7.
15. Gilmore, G. and Hemingway, J.D. Practical gamma-ray spectroscopy. 2nd Edition. Chichester,
UK : John Wiley & Sons, Ltd, 2008. ISBN:9780470861967.