Volume 27, Issue 1 , Pages 11-20, January 2011
Evaluation of an uncollimated printed paper transmission source used under scatter limiting conditions
Article Outline
- Abstract
- Introduction
- Materials and methods
- Results
- Discussion
- Conclusions
- Acknowledgements
- Conflict of interest
- References
- Copyright
Abstract
Transmission sources used for image attenuation correction, allowing image quantification, are collimated to reduce scatter. We propose the same effect can be achieved for an uncollimated source by increasing source to patient distance. The aim was to compare planar image performance characteristics and absorbed doses of uncollimated and collimated radioactive printed paper transmission sources.
The scatter contribution to the uncollimated 99mTc source data was evaluated for different combinations of detector phantom distance, detector source distance and phantom source distance. Measurements were performed by increasing the Lucite phantom thickness in 1
cm steps to 20 cm. Spatial resolution, detection efficiency and entrance absorbed dose rate were measured for the uncollimated and collimated transmission source images.
Results derived from the energy spectra, obtained with the uncollimated transmission source indicate that scatter contribution increases with decreasing detector source distance. The scatter component in the uncollimated transmission images (detector source distances≥60cm; phantom source distances≥40cm) was comparable to that obtained with collimated transmission images. Attenuation coefficients obtained compared well (0.168cm−1 vs. 0.171cm−1). The full widths at half maxima differed by less than 0.9mm. The detection efficiency of the uncollimated source was 2.5 times higher than obtained with the collimated source. The entrance absorbed dose obtained from an uncollimated source was 3.75 times larger than that obtained from the collimated source.
An uncollimated transmission source (detector source distance≥60cm) results in acceptable image characteristics and presents a low cost, low dose, high efficiency option for transmission imaging.
Keywords: Transmission imaging, Attenuation correction, Printed source
Introduction
Several factors influence the quality of Nuclear Medicine images. The most important of these are the effect of attenuation in tissue and Compton scatter. Attenuation may lead to artefacts and inaccuracies in planar and reconstructed SPECT images, which are particularly disturbing in clinical studies. The importance of accurate compensation for attenuation effects in Nuclear Medicine studies is already widely recognized. Accurate image quantification plays an important role in radionuclide dosimetry [1], [2]. Several studies have demonstrated image quantification accuracy and quality can be considerably improved by applying an attenuation correction to planar images [3], [4], [5], [6].
In order to apply the attenuation correction accurately to planar images, pixel-by-pixel attenuation correction factors are required. Attenuation correction factors can be obtained of the person or object being imaged by using transmission imaging. Transmission images can be obtained with an external radionuclide source [3], [4]. Minarik et al. [1] proposed the use of a CT scout image of the patient to obtain whole-body transmission images for planar image quantification.
The practical advantages of CT over transmission imaging using an external radionuclide source include a high photon fluence rate allowing for fast data acquisition with a high statistical accuracy. Due to the higher fluence rate the transmission data can be acquired after the patien's emission data without a cross-contamination correction. In addition the X-ray source does not decay; therefore source replacement is not required. However the disadvantages of using CT scout view transmission are that CT equipment is very expensive and, due to the high fluence rate, the radiation dose should be monitored although Minarik et al. indicated the dose is not significant [1]. It is important to bear in mind that the X-rays emitted in CT are polychromatic and of a different energy than the photon energy used for emission imaging. The CT numbers obtained in the scout image have to be scaled to represent the attenuation of photons emitted by the radionuclide of interest. In addition CT generally has a higher spatial resolution and is reconstructed using a finer image matrix than SPECT images. Therefore the CT images require down-sampling to the same image matrix size as the emission image. Furthermore the CT unit produces a divergent beam of X-rays which will geometrically distort the image which requires correction [1].
Several types of systems with transmission hardware modifications and external radionuclide sources have emerged for clinical implementation. Radionuclides used as transmission sources are (energy and half-life given in brackets): 241Am (60keV, 432.2yr), 153Gd (100keV, 241.6 days), 57Co (122keV; 270.9 days), 99mTc (140keV; 6.02 hrs), 123mTe (159keV; 119.7 days), 139Ce (166keV; 137.7 days) and 133Ba (360keV; 10.5yr) [7], [8]. A collimated sheet or scanning line source is proposed for planar transmission imaging [3], [4]. Limitations of the above mentioned transmission systems include: high source activity needed for collimated sources, complicated electronic equipment needed to separate transmission and emission images and insufficient counting statistics in the transmission image in an acceptably short imaging time.
In this study the use of a novel printed uncollimated external radionuclide flood source are examined. The flood source is obtained by printing an ink solution containing 99mTc onto a sheet of paper using an ink-jet printer as proposed by Van Staden et al. [9]. Printed radioactive sources were used by Larsson et al. in order to obtain scatter and attenuation free SPECT images [10].
Disadvantages associated with the use of an uncollimated transmission source are firstly the additional radiation dose to the patient compared to using a collimated source. Secondly, in order to apply an appropriate attenuation correction it is important to calculate the attenuation coefficient values for narrow beam geometry. When using an uncollimated transmission source scattered events may be added resulting in broad beam attenuation coefficient values. However a scatter correction can be applied to the transmission data as proposed by Kojima et al. [11]. The effect of the distance between the uncollimated transmission source and the detector on the amount of scatter being observed has not been investigated in literature. The reduction in absorbed radiation dose when the distance between the patient and the uncollimated transmission source is increased has also not been reported.
This study hypothesises that the scatter will decrease if the transmission source is placed further from the patient. By increasing the distance between the transmission source and the detector with short detector phantom/patient distance the source geometry will approach narrow beam geometry. When an uncollimated flood source is used, photons reaching the patient from different angles will penetrate the patient and reach the collimator (photons (i) and (ii) in Fig. 1(a)). Photons reaching the detector at angles larger than the collimator resolution will be absorbed by the collimator and will not be detected. However, the photons that will be absorbed by the collimator if no interaction in the patient occurs can pass the collimator and be detected if they are scattered in the patient so that their direction is changed to be within the resolution angle of the collimator (Fig. 1(a) photons (i) and (ii); and Fig. 1(b) photon (ii)). If the source-patient distance is increased, the number of photons emerging from the source at angles larger than the collimator resolution, which can interact with the patient, will decrease (Fig. 1(a) and (b), photon (i)). Therefore by increasing the source-patient distance or decreasing the detector phantom/patient distance the scatter contribution in the image is decreased. In addition, by increasing the source-patient distance the radiation dose to the patient will also decrease.
Figure 1
Explanation of the hypothesis that the scatter will decrease if the transmission source is placed further from the patient. In (a) the transmission source is placed closer than in (b). Gamma rays that can undergo scatter are illustrated and the difference in the interactions undergone by photons, i and ii, are indicated for setup (a) and (b).
The aim of the study was to evaluate the planar performance characteristics of an uncollimated radioactive paper transmission source by comparing its performance characteristics to those of a collimated radioactive paper transmission source. The hypotheses that the scatter and absorbed radiation dose will decrease if the uncollimated transmission source is placed further from the patient or the patient placed closer to the detector, were also tested.
Materials and methods
The scatter contribution to the data obtained with the uncollimated transmission source was evaluated by reviewing the following characteristics at different combinations of detector phantom distances, detector source distances and phantom source distances as defined in Fig. 2: the energy spectra obtained from the transmission data, the calculated attenuation coefficients and the scatter percentages. The image resolution, detection efficiency and absorbed radiation dose were measured for the uncollimated as well as for the collimated transmission source images.
Figure 2
The experimental setup used is illustrated. A source holder equipped with a printed transmission source supported on the bed parallel to the detector fitted with a low energy all purpose collimator. The Lucite attenuation sheets are shown on the patient support.
Data were acquired using a GE 400 AT gamma camera fitted with a low energy all purpose (LEAP) collimator. The camera was equipped with an Alfanuclear acquisition and processing station (Alfanuclear SAIYC, Buenos Aires). The LEAP collimator was selected for the preliminary evaluation due to the higher counting efficiency.
The 99mTc flood source was supported on the gamma camera bed (Fig. 2). The flood source containing approximately 370MBq 99mTc was printed on a A4 paper sheet using an ink-jet printer. Details about the printing process and the determination of the activity were explained by Van Staden et al. [9]. These paper sources were restricted to A4 as all measurements were limited to the central 9×9cm of the image reducing the total amount of activity required. It is possible to extend the source to cover the full field of view of the camera. The sources were also laminated to prevent contamination. The IM512P acquisition software (Alfanuclear SAIYC, Buenos Aires) allows the acquisition of the energy spectrum for each pixel. Transmission data were acquired for the uncollimated as well as a collimated source. The latter incorporated a low energy high sensitivity collimator (LEHS), placed on top of the transmission source on the source holder. All measurements were repeated three times, except when stated differently.
Scatter contribution
The performance of the transmission source was evaluated for different phantom and source positions in relation to the detector distance by obtaining three data sets. In each data set one of the three distances (detector phantom distance, detector source distance or phantom source distance) was kept constant and the other two distances varied (Table 1, Fig. 2). For data set 1 the detector phantom distance was kept at 20cm and the detector source distance was increased in steps of 20cm. The phantom source distance changed accordingly. During the measurement of data set 2 the detector source distance was fixed at 80cm and the detector phantom distance was increased in steps of 20cm, thereby decreasing the phantom source distance. For data set 3 the phantom source distance was kept at 0cm. The detector phantom as well as the detector source distances were increased in steps of 20cm by moving the phantom and source simultaneously. Measurements for data set 1 and 2 were repeated three times, however measurements for data set 3 were only obtained once. Results are given as an average of the three repeated measurements for data sets 1 and 2.
Table 1. Distance measurements as obtained for each data set.
| Data set | Description of distance | Measurement distances (cm) | |||
|---|---|---|---|---|---|
| Data set 1 | Detector phantom distance | 20 | 20 | 20 | 20 |
| Detector source distance | 20 | 40 | 60 | 80 | |
| Phantom source distance | 0 | 20 | 40 | 60 | |
| Data set 2 | Detector phantom distance | 20 | 40 | 60 | 80 |
| Detector source distance | 80 | 80 | 80 | 80 | |
| Phantom source distance | 60 | 40 | 20 | 0 | |
| Data set 3 | Detector phantom distance | 20 | 40 | 60 | 80 |
| Detector source distance | 20 | 40 | 60 | 80 | |
| Phantom source distance | 0 | 0 | 0 | 0 | |
The IM512P acquisition software allows the acquisition of the energy spectrum for each pixel. Two dimensional energy spectral data matrices were acquired, resulting in an energy spectrum in each pixel of a 128×128 matrix. Energy spectral data matrices for 99mTc photons (140keV) were obtained for data sets 1, 2 and 3 with the uncollimated transmission source. Similar spectral data matrices were also obtained with the collimated transmission source for data set 1. Energy spectral data were obtained for the photons transmitted through various thicknesses (0–20cm) of Lucite plates. The energy spectrum data were acquired in ∼0.5keV channels for 0–500keV. Only information from 40keV to 180keV was considered. The attenuating material consisted of 19 Lucite slabs each with a dimension of 20×20cm2 and a thickness of 1.07cm. The energy spectra of the different pixels in that region were summed resulting in a single energy spectrum for each matrix obtained from a region with size 9×9cm2 in the central area of the Lucite slabs. The energy spectra for each data set were corrected for decay and acquisition time.
The spectra obtained as described in Section Energy spectra were used to determine the attenuation coefficients for Lucite from the three data sets. The total counts in a 20% energy window were plotted against the thickness of Lucite for each spectrum obtained. An exponential function was fitted to the counts at increasing thicknesses. Attenuation coefficient values were calculated from these exponential functions. Attenuation coefficient values were also obtained for the collimated transmission source with the setup as in data set 1. This latter value should correspond to the “narrow beam geometry” attenuation coefficient value for 99mTc through Lucite.
Scatter percentage determinationThe total counts in a 20% energy window were obtained from each spectrum mentioned in Section Energy spectra. The spectrum obtained with no attenuating medium was assumed to contain only primary counts (i.e. those originating directly as a result of unscattered photons from decaying atoms). The theoretical attenuated primary counts (i.e. excluding scattered events) at different thicknesses were calculated using the narrow beam geometry attenuation coefficient value for 99mTc through Lucite (0.1777cm−1) [12]. The scattered counts were obtained by subtracting the calculated primary counts from the total observed counts. The scatter percentage in the 20% energy window was calculated as the ratio of scattered to total counts and expressed as a percentage.
Spatial resolution
The image resolution was evaluated by acquiring transmission images of a 2mm thick lead strip as has been proposed by Cao and Tsui [13]. The lead strip was placed at depths of 1cm, 10cm and 18cm in the Lucite phantom when measuring the resolution in the scatter medium. These distances correspond with the lead detector distance, since the phantom was placed against the detector (detector phantom distance=20cm). These measurements were repeated with the collimated transmission source with the lead strip at a lead detector distance of 10cm. The phantom consisting of 20cm Lucite slabs was used as the scattering medium. From these images the derivative of the transmission edge response function was determined to calculate the full width at half maximum (FWHM) and the full width at tenth maximum (FWTM) for each configuration.
System detection efficiency
Images of the flood source were acquired with no scatter medium placed between the gamma camera (fitted with the LEAP collimator) and the source. Images were acquired in a 20% 99mTc window at 5, 10, 15, 20, 25 and 30cm distances from the gamma camera for 120s with the collimated as well as the uncollimated source. The relative detection efficiency was calculated as the ratio of the counts obtained with no source collimator to counts obtain with the source collimator.
Absorbed dose rate
An ionisation chamber was used to measure the entrance absorbed dose rate (EAD rate) from a printed collimated flood source containing 78MBq 99mTc. It was measured at detector source distances of 20, 40, 60 and 80cm in front of a 20cm Lucite phantom (data set 1). The measured dose rates were adjusted to a transmission source activity of 370MBq suggested for clinical use with a collimated transmission source [3].
The EAD rate was then converted to a total EAD taking decay into account and assuming that the total acquisition time for a clinical planar study will be limited to 5min. The total EAD for an acquisition period of 5min with a collimated transmission source (EADref) was utilised as a reference to compare the doses without a collimator.
In order to have comparable images for attenuation correction, the collected counts, from collimated and uncollimated flood sources, should have the same order of magnitude. As the efficiency of the uncollimated source is higher than the efficiency of the collimated source by a factor of 2.5 (see results Section System detection efficiency), the measured EAD was adjusted for the differences in efficiency, resulting in sensitivity adjusted EAD (SEAD). The EAD for the uncollimated source relative to the EADref is defined as the relative entrance absorbed dose (READ) and is utilised as a measure of the increase in dose relative to a collimated source.
Results
Scatter contribution
Energy spectraFig. 3(a) shows the average energy spectra obtained for the uncollimated transmission source through 20cm of Lucite at different detector source distances (20cm, 40cm, 60cm and 80cm) and a fixed detector phantom distance of 20cm (data set 1). The spectrum obtained for the collimated transmission source, 80cm from the detector, attenuated through 20cm of Lucite is also shown (C-80cm). The energy spectra were normalised to 100 at 150keV in order to demonstrate the scatter component in the spectra.
Figure 3
Energy spectra (normalised to 100 at 150
keV) obtained with 20cm of Lucite as attenuating medium. (a) The spectra for data set 1 at different detector source distances of the uncollimated source (20cm, 40cm, 60cm, 80cm) as well as the spectrum obtained with the collimated source at a detector source distance of 80cm (C-80cm) are shown for the detector phantom distance fixed at 20cm. (b) Energy spectra are shown for data set 2 at different detector phantom distances (20cm, 40cm…) with the detector source distance fixed at 80cm. The spectrum obtained from the collimated source at a distance of 80cm and the detector phantom distance at 20cm is also shown (C-20cm). (c) Energy spectra for data set 3 obtained with different detector phantom distances with the phantom source distance at 0cm are shown here.
The scatter contribution in the spectra decreases with increase in detector source distance for the uncollimated transmission source (Fig. 3(a)). The spectra obtained with the detector source distances of 60cm and larger compare well with the spectrum of the collimated source. The relative counts of the 60cm and 80cm spectra were however slightly higher than those of the C-80cm spectrum. This can be explained due to the fact that small angle Compton scatter will contribute to the counts in the 20% energy window region (126–154keV).
Fig. 3(b) shows energy spectra obtained for the uncollimated transmission source through 20cm of Lucite at a fixed detector source distance of 80cm with different phantom detector distances (data set 2). The spectrum obtained for the collimated transmission source, 80cm from the detector, attenuated through 20cm of Lucite is also shown (C-20cm). The scatter contribution in the spectra increases with increase in detector phantom distance (this implies a decrease in the source phantom distance) for the uncollimated transmission source. Also, the uncollimated spectra at detector phantom distances of 20cm and 40cm compared well with the spectra for the collimated source at 20cm (C-20cm).
Fig. 3(c) shows energy spectra obtained for data set 3. The uncollimated transmission source was positioned at a fixed phantom source distance of 0cm with different phantom detector distances. It can be seen from Fig. 3(c) that the scatter contribution in the spectra was significant in all cases.
The results show that the phantom source distance is the main determinant of the scatter component.
Attenuation coefficient determinationThe attenuation coefficient value increases with an increase in detector source distance for the uncollimated source in data set 1 (Table 2). This is indicative of a decrease in scatter with an increase in the detector source distance. However, the “narrow beam geometry” attenuation coefficient value, obtained from the collimated transmission data (0.170±0.001cm−1), is also underestimated in comparison to the published narrow beam attenuation coefficient [12] of 0.1777cm−1. The attenuation coefficient values determined for data set 3 is also given in Table 2. The phantom source distance was fixed at 0cm. The average attenuation coefficient for the different detector source distances was 0.143±0.001cm−1. These values imply a large amount of scatter is present since the theoretical narrow beam attenuation coefficient for 99mTc through Lucite is 0.1777cm−1. Table 2 shows a decrease in the attenuation coefficient value with an increase in the detector phantom distance (0.171–0.144cm−1), indicating the influence of the increase in scatter contribution (data set 2).
Table 2. Attenuation coefficient values (cm−1) as a function of detector source distances for uncollimated transmission sources (UTSs) in data set 1 and data set 3 (source phantom distance=0cm) as well as for collimated transmission sources (CTSs) in data set 1 are given. Attenuation coefficient values for data set 2 at different detector phantom distances for the uncollimated transmission source (UTS) are also shown.
| Distances | Data set 1, UTS | Data set 1, CTS | Data set 2, UTS | Data set 3, UTS |
|---|---|---|---|---|
| 20cm | 0.141 | 0.170 | 0.171 | 0.142 |
| 40cm | 0.160 | 0.170 | 0.169 | 0.143 |
| 60cm | 0.167 | 0.170 | 0.164 | 0.143 |
| 80cm | 0.168 | 0.170 | 0.144 | 0.143 |
These results confirm scatter decreases with an increase in the detector phantom distance, and for detector phantom distances larger than 60cm, the attenuation coefficients for the collimated and uncollimated transmission sources were similar.
Fig. 4(a) shows larger detector source distances result in a smaller scatter percentage in the 20% energy window for the uncollimated transmission image. The uncollimated scatter percentage at 20cm Lucite is 16% for the source at a distance of 80cm and a detector phantom distance of 20cm (data set 1). The corresponding scatter percentage for the collimated source was 15% (Fig. 4(b)). When imaging the uncollimated transmission source at larger distances from the camera the experimental setup approaches the narrow beam geometry as we hypothesized and explained in Fig. 1. Fig. 4(c) and (d) confirms the results obtained in Sections , . There is an increase in the scatter percentage in the 20% energy window with an increase in the detector phantom distance or a decrease in the phantom source distance. Fig. 4(d) shows the scatter contribution stays constant with increasing detector source distance, while keeping the phantom close to the source (phantom source distance=0cm) (data set 3).
Figure 4
Scatter percentage as a function of thickness of attenuating medium. Values are given for data set 1 at various detector source distances (20
cm, 40cm, 60cm, 80cm) for the uncollimated (a) as well as for the collimated transmission source (b). Similar values are shown for data set 2 at a fixed detector source distance of 80cm with different phantom–detector distances (c). The scatter percentage is also shown for data set 3 at different detector source distances with a fixed source phantom distance of 0cm (d).
Spatial resolution
The resolution expressed as FWHM obtained with the uncollimated as well as for the collimated transmission source is shown in Fig. 5. The error bars on the graphs indicate twice the standard deviation of the three measurements obtained at each detector source distance. No error bars are shown for the values obtained with the collimated transmission source since these measurements were not repeated. The FWTM values ranged from 15.9mm to 29.3mm when no scatter was present and from 16.1mm to 31.1mm with scatter present. The FWTM values showed a similar trend as the FWHM values displayed in Fig. 5.
Figure 5
Resolution results obtained for transmission images acquired using a 2
mm thick lead strip. The FWHM values are shown for different detector source distances (DSD) with no scatter medium as well as with scatter medium placed between the detector and the transmission source. The 2mm thick lead strip was positioned at lead detector distances of 1cm, 10cm and 18cm with an uncollimated transmission source (UTS). Error bars are displayed as twice the standard deviation of the three repeated measurements. Resolution images were also obtained with a collimated transmission source (CTS) at a lead detector distance of 10cm.
At a lead edge detector distance of 1cm there was no difference in the resolution values with an increase in detector source distance. The additional scatter medium had no influence at this distance on the resolution values.
At a lead detector position of 10cm no difference was seen in the resolution values with an increase in detector source distance when no scatter medium was present and the uncollimated transmission source was used. The FWHM shows a minor deterioration at a detector source distance of 20cm when scatter medium was added (12.6mm vs. 12.1mm when no scatter medium was added).
The resolution values obtained with the uncollimated transmission source and a lead detector distance of 10cm, at all source detector distances shows an increase in comparison to the values obtained with the collimated transmission source. This average difference obtained for the FWHM values was 1.2mm without scatter medium. When scatter was added this average difference was 0.9mm. These results correspond to similar findings obtained by Cao and Tsui [13]. However the influence of different detector source distances was not evaluated by Cao and Tsui [13].
At the lead position of 18cm there was a significant difference between the resolution values with and without the scatter medium present. There was a marked decrease in the difference with an increase in the detector source distance.
From the results mentioned it can be concluded that the resolution of the transmission images obtained with the uncollimated transmission source with scatter medium present, improved with an increase in the detector source distance. The resolution values obtained with the uncollimated transmission source were also comparable to the resolution values obtained with the collimated transmission source. This was caused by a marked decrease in the resolution values with an increase in the detector source distance, while the values without the scatter medium remained constant.
Detection efficiency
The detection efficiency for the uncollimated source was 2.5±0.01 times higher than for the collimated source measured at 5, 10, 15, 20, 25 and 30cm distances from the gamma camera. This will result in shorter transmission imaging times for the same image quality.
Absorbed dose rate
The entrance absorbed dose rate for the collimated source remained constant at all detector source distances at a value of 2.857μGy/h, and this resulted in an EADref of 0.237μGy for a 370MBq 99mTc collimated source acquired for 5min.
The EAD rate and sensitivity adjusted EAD (SEAD) for a 2min acquisition time using the uncollimated source is given in Table 3. The dose rate from the uncollimated source at a DSD of 80cm is only 3.75 times higher than that of the collimated transmission source. These doses are significantly lower than currently accepted values for CT transmission imaging used for SPECT attenuation correction [14], [15].
Table 3. Entrance absorbed dose (EAD) rate, sensitivity adjusted EAD (SEAD) and relative EAD (READ) for the uncollimated transmission source. Acquisition time was 5min for the collimated and 2min for the uncollimated transmission source.
| DSD (cm) | EAD rate (μGy/h) | SEAD (μGy) | READ (SEAD/EADref) |
|---|---|---|---|
| 20 | 919.55 | 30.30 | 131.12 |
| 40 | 305.99 | 10.08 | 43.63 |
| 60 | 62.31 | 2.05 | 8.88 |
| 80 | 26.32 | 0.87 | 3.75 |
Discussion
In this study, the optimum detector source and detector phantom distances using a novel uncollimated radioactive paper transmission source were evaluated for application during planar image acquisition. The printed transmission source could easily be attached to an existing gamma camera since no additional heavy collimator or modification was required.
The results obtained from the acquired energy spectra with the collimated and uncollimated transmission source and the scatter percentage indicate that for the uncollimated transmission source and detector phantom distance of 20cm, the scatter contribution increases with a decrease in detector source distances. This suggests that the uncollimated transmission source would be most effective at distances 60cm and further from the detector in order to minimise the detection of scattered photons. The scatter can be kept to a minimum if the phantom or patient is close to the detector. When an uncollimated transmission source is used, the scatter present would result in an underestimation of attenuation and therefore overestimation of activity during quantification during emission imaging. The presence of scatter also reduces image spatial resolution. By using the uncollimated source at detector source distances and detector phantom distances of larger than 60cm and 40cm respectively scatter is reduced and image quantification and spatial resolution should approach that obtained with the collimated transmission source.
It was shown when the source detector and phantom source distances are increased to 60cm and 40cm respectively the attenuation coefficient for the uncollimated transmission source increased to 0.168cm−1 in comparison to the value of 0.171±0.001cm−1 for the collimated transmission source. Therefore, accuracy of image quantification using the uncollimated transmission source could be achieved.
An increase in the detector source distance resulted in an improvement in the resolution values when measured in a scatter medium at the lead detector distances of 10 and 18cm. Therefore, at detector source distances larger than 60cm the resolution using an uncollimated transmission source compares well with the resolution using a collimated transmission source. The FWHM difference is approximately 0.9mm. This is considered as acceptable for transmission imaging with the purpose of applying an attenuation correction to emission. The influence of the detector source distance on the resolution, as obtained with the uncollimated transmission source, was not studied by Cao and Tsui [13].
It was shown that the detection efficiency of the uncollimated transmission source was nearly constant when the detector source distance (data set 1) was moved from 0cm to 30cm. The detection efficiency of the uncollimated transmission source was 2.5 times higher than obtained with the collimated transmission source. The higher detection efficiency of the uncollimated transmission source allows a 2.5 times shorter transmission imaging acquisition time to acquire the same number of image counts as for a collimated transmission source. The reduction in the transmission acquisition time will also proportionately reduce the absorbed entrance dose. The entrance absorbed dose resulting from using an uncollimated transmission source will be 3.75 times larger than obtained from a collimated source acquired for the same number of counts.
Although the absorbed entrance dose obtained from the uncollimated transmission source is larger than obtained from the collimated transmission source, it is still low. To be clinically relevant the acquisition time for the collimated transmission source and for the uncollimated source was assumed 5 and 2 min respectively. The EAD for the uncollimated transmission source to acquire the same number of counts as with a collimated scan will be 0.877μGy at a detector source distance of 80cm. The published CT dose from SPECT/CT imaging is more than 4.6mGy [14] or is in the range of 6–14mSv [15].
Potential advantages of a printed paper transmission source are firstly that the intensity distribution of the radioactive source can be modified to be suitable for a specific patient. Cellar et al. [16] proposed the use of a series of collimated line sources with different intensities in order to create a non-uniform transmission profile. This will result in a more uniform transmission image of the patient. An important advantage of the use of the printed paper transmission source configuration is that the distribution of activity in the source system is tailored to the attenuation in the human body, minimising the problem encountered when too few counts are recorded in some pixels of the transmission scan [16]. Secondly the printed radioactive source can be manufactured at low cost to evaluate the use of different activity distributions in the transmission source. An optimal activity distribution can then be identified. In this study 99mTc was used for the preliminary evaluation of a uniform transmission source; however, a long life radionuclide may be used once the required non-uniform configuration has been determined. Thirdly the advantage of an uncollimated transmission source is, older existing gamma cameras can be easily modified to allow attachment of the printed transmission source, since the source construction is light. Fourthly the photon-flux detected with an uncollimated transmission source is much higher than when collimation is used.
The use of an uncollimated transmission source, due to the low cost, ease of implementation, the higher detection efficiency and low absorbed entrance radiation dose, is an appealing alternative source to be used during transmission imaging. The findings of this study for planar imaging should also apply to SPECT, since SPECT reconstruction use planar acquired images as input data.
Conclusions
In this study the optimum detector source and detector phantom distances using a novel uncollimated radioactive paper transmission source were evaluated for application during planar image acquisition. The imaging characteristics of the uncollimated paper transmission source at a source detector distance of 60cm or more compared well with the collimated transmission source. The scatter component in the uncollimated transmission images was acceptable when the transmission source was placed at a distance of 60cm and further from the detector. The reduction in scatter at these distances resulted in acceptable values for attenuation coefficients, image quantification, spatial resolution and entrance absorbed radiation dose. Similar information is not available from previously published data [13].
Acknowledgements
The authors would like to thank Prof AC Otto of the Nuclear Medicine Department, Universitas Hospital for use of the equipment in the department.
Conflict of interest
None.
References
- . A new method to obtain transmission images for planar whole-body activity quantification. Cancer Biother Radiopharmaceut. 2005;20:72–76
- . Attenuation correction: a better cardiac SPECT. ACC Curr J Rev. 2000;9(Suppl. 1):25S–31S
- . Impact of scatter and attenuation corrections for iodine-131 two-dimensional quantitative imaging in patients. Cancer Biother Radiopharmaceut. 2003;18:191–199
- . A comparison of planar scintigraphy and SPECT measurement of total lung deposition of inhaled aerosol. J Aerosol Med. 2003;16:9–19
- . Comparison of conventional, model-based quantitative planar, and quantitative SPECT image processing methods for organ activity estimation using In-111 agents. Phys Med Biol. 2006;51:3967–3981
- Evaluation of quantitative imaging methods for organ activity and residence time estimation using a population of phantoms having realistic variations in anatomy and uptake. Med Phys. 2009;36:612–619
- . Determination of the attenuation map in emission tomography. J Nucl Med. 2003;44:291–315
- . Implementation of a Tc-99m and Ce-139 scanning line source for attenuation correction in SPECT using a dual opposing scintillation camera. Med Phys. 2000;27:1523–1534
- . Production of radioactive quality assurance phantoms using a standard inkjet printer. Phys Med Biol. 2007;52:N329–N337
- . A novel phantom design for emission tomography enabling scatter- and attenuation-“free” single-photon emission tomography imaging. Eur J Nucl Med. 2000;27:131–139
- . Accurate scatter correction for transmission computed tomography using an uncollimated line array source. Ann Nucl Med. 2004;18:45–50
- . Tables of X-ray mass attenuation coefficients and mass energy-absorption coefficients. National Institute of Standards and Technology; 1996;http://physics.nist.gov/PhysRefData/XrayMassCoef/cover.html[accessed 22.09.09]
- . Performance characteristics of transmission imaging using a uniform sheet source with parallel-hole collimation. Med Phys. 1992;19:1205–1212
- . Evaluation of a low-dose/slow rotating SPECT-CT system. Phys Med Biol. 2008;53:2495–2508
- SPECT/CT. J Nucl Med. 2008;49:1305–1319
- . Multiple line source array for SPECT transmission scans: simulation, phantom and patient studies. J Nucl Med. 1998;39:2183–2189
PII: S1120-1797(10)00006-2
doi:10.1016/j.ejmp.2010.01.004
© 2010 Associazione Italiana di Fisica Medica. Published by Elsevier Inc. All rights reserved.
Volume 27, Issue 1 , Pages 11-20, January 2011
