Comparative Study of Imaging Characteristics of I-125 Imaging Using the Siemens Inveon Scanner and Siemens Symbia TruePoint ()
1. Introduction
Iodine-125 (125I) has been widely used for in vitro research because of its relatively long half-life (60.1 days). Various 125I-labeled nucleic acids, antibodies, and ligands are commercially available or easily made in the laboratory using commercially available reagents and kits [1] [2] . Recently, in vivo 125I imaging methods were developed, and both in vivo and in vitro experiments were performed using the same 125I-labeled tracers [3] - [7] . However, in vivo imaging in animals using 125I is limited because of its low energy (27 - 35 keV) compared to Technetium-99m (99mTc) or Iodine-123 (123I) [8] .
Although the clinical scanner was designed for imaging human patients, the clinical scanner has been used for animal imaging. Small animal imaging using a clinical scanner with a pinhole collimator and 99mTc or 123I has been reported [6] [9] - [12] . Although a small animal-dedicated scanner provides high resolution (submillimeter range), the clinical scanner is ready to use in clinics. Another advantage of the clinical scanner is the large field of view.
In this study, we investigated the feasibility for 125I imaging using a clinical scanner without a pinhole collimator and compared its performance characteristics to a small animal scanner.
2. Materials and Methods
2.1. System Description
The Symbia TruePoint (Siemens Medical Solutions USA, Inc., Knoxville, TN, USA) clinical single-photon emission computed tomography (SPECT)/computed tomography (CT) imaging system designed for human patients was used to assess 125I characteristics. Low energy high-resolution (LEHR) and low energy all-purpose (LEAP) collimators were used. Each collimator consisted of 148 holes with a 1.11-mm diameter and 90 holes with a 1.45-mm diameter across the flats. Each of the two detector heads of the Symbia TruePoint SPECT system consisted of a 15.8-mm crystal thickness and a SPECT field of view of 40 cm (axial) by 50 cm (diameter).
The Siemens Inveon small animal positron emission tomography (PET)/SPECT/CT imaging system (Siemens Medical Solutions USA, Inc.) was used to assess the performance of 125I. Mouse high resolution (MHR), mouse general purpose (MGP), and mouse high sensitivity (MHS) collimators with pinhole diameters of 0.5 mm, 1.0 mm, and 2.0 mm across the flats, respectively, were used. The Inveon SPECT/CT system consisted of two detector heads, an active area of 15 cm × 15 cm, [NaI(Tl)] scintillator crystal material, and 4624 crystal elements.
2.2. Spatial Resolution and Sensitivity
For the measurement of spatial resolution and sensitivity, a 1-mm point source was made using capillary tubes. The activities of 125I and 99mTc were 2.2 MBq and 3.3 MBq, respectively. The data acquisition time was 600 s. To determine the spatial resolution, profiles through peaks in count distributions were drawn in two orthogonal directions in middle slices. For each profile, full width at half maximum (FWHM) was measured by linearly interpolating the adjacent pixels at half or tenth of their maxima in the profile.
2.3. Image Quality
For the assessment of image quality, the National Electrical Manufacturers Association (NEMA) NU 4 Image Quality (IQ) Phantom was used [13] - [15] . The NU 4 IQ phantom is a cylindrical chamber that consists of three parts: five fillable rods with diameters of 1, 2, 3, 4, and 5 mm arranged in the top part; a uniform region in the middle part; and two cold components with a 10-mm diameter attached to the bottom part. The top and bottom regions are filled with nonradioactive water and air, respectively. The nonuniformity, recovery coefficient, and spillover ratio (SOR) were assessed using the NU 4 IQ phantom [14] . The initial activities for 99mTc and 125I were 11.87 MBq and 11.36 MBq, respectively.
2.4. Data Acquisition
The step-and-shoot mode was used for data acquisition. The acquisition time for each view was 100 s, total imaging time was 3200 s, rotation degree was 180˚, degree of configuration was 180˚, and magnification factor was 3.2.
2.5. Reconstruction
The ordered subset expectation maximization three-dimensional (OSEM-3D) algorithm was used in the small animal scanner, and the Flash 3D algorithm was used in the clinical scanner. The number of iterations was 8, and the number of sets was 4. The windows for 125I and 99mTc were 35 ± 11 keV and 140 ± 14 keV, respectively.
2.6. Animal Studies
Animal study was carried out in accordance with guideline provided by the Institutional Animal Care and Use Committee (IACUC) and Institutional review board (IRB) in Korea Institute of Radiological and Medical Sciences. In addition, all experimental protocols were approved by IACUC and IRB in Korea Institute of Radiological and Medical Sciences.
125I was injected into BALB/c mice (weight: 20 g, female, SLC JAPAN) and the mice were scanned for 1 hour after injection. Planar mode imaging and SPECT mode imaging were performed. The 1.0-MGP collimator was used in the small animal scanner, and the LEAP collimator was used in the clinical scanner. The activity concentrations for the animal scanner and clinical scanner were 18.9 MBq and 14.4 MBq, respectively.
In vivo radiolabeled antibody imaging was assessed using a human breast cancer mouse model. KPL-4 and MDA-MB-231 xenografts were established in female nude mice (6 weeks) began by subcutaneously implanting approximately 5 × 106 KPL-4 cells in the right hind leg and 5 × 106 MDA-MB-231 cells in the left hind leg. The size of the HER2-positive KPL-4 tumors and the HER2-negative MDA-MB-231 tumors were approximately 300 mm3. 125I-labeled HER2-targeting antibody was injected, and its activity concentration was 8.81 MBq.
3. Results
3.1. Spatial Resolution
Figure 1(a) shows the mean resolution as a function of source-to-collimator distance. The resolution measurements using 125I point sources are summarized in Figure 1(a). At source-to-collimator distances of 25, 30, and 35 mm, the spatial resolutions of 125I measured using the Inveon scanner were 3.32, 3.98, and 4.20 mm full width at half maximum (FWHM), respectively, using the MHR 0.5-mm pinhole (MHR-0.5) collimator; 3.98, 4.42, and 4.86 mm FWHM, respectively, using the MGP 1.0-mm pinhole (MGP-1) collimator; and 4.35, 4.88, and 5.43 mm FWHM, respectively, using the MHS 2.0-mm pinhole (MHS-2.0) collimator.
At source to collimator distances of 25, 30, and 35 mm, the resolutions of 125I measured using the Symbia TruePoint system were 8.28, 8.72, 8.89 mm FWHM, respectively, using the LEHR collimator and 9.06, 9.19, and 9.25 mm FWHM, respectively, using the LEAP collimator.
Figure 1(b) shows the spatial resolution of 99mTc point sources. The results were similar to a previous study and specification sheets provided by the manufacturer [15] - [17] .
3.2. Sensitivity
Figure 2 shows the sensitivity of 125I point sources as a function of source-to-collimator distance. At source- to-collimator distances of 25, 30, and 35 mm, the sensitivities for the Inveon scanner using the MHR-0.5 collimator were 29.49, 21.87, and 17.12 cps/MBq, respectively; the MGP-1.0 collimator were 64.41, 44.37, and 32.69 cps/MBq, respectively; and the MHS-2.0 collimator were 163.42, 116.88, and 87.18 cps/MBq, respectively.
![]()
(a)![]()
(b)
Figure 1. The resolution of the clinical scanner compared to the small animal scanner presented as a function of the source-to-collimator distance with each collimator. (a) Idoine-125 (125I) point source and (b) Technetium-99m (99mTc) point source.
![]()
(a)![]()
(b)
Figure 2. The sensitivity of the clinical scanner compared to the small animal scanner presented as a function of the source-to-collimator distance with each collimator. (a) Iodine-125 (125I) point source and (b) Technetium-99m (99mTc) point source.
At source-to-collimator distances of 25, 30, and 35 mm, the sensitivities for the Symbia TruePoint were 30.83, 30.55, and 30.35 cps/MBq, respectively, using the LEHR collimator and 68.89, 68.13, and 67.68 cps/MBq, respectively, using the LEAP collimator.
Figure 2(b) shows the sensitivity using 99mTc point sources. These results were similar to a previous study [15] - [17] .
3.3. Nonuniformity
Figure 3(a) shows the nonuniformity (%STDunif) of 125I measured by the NU 4 IQ phantom. The nonuniformities measured using the Inveon scanner were 87.97%, 29.74%, and 14.72% with the MHR-0.5, MGP-1.0, and MHS-2.0 collimators, respectively. The nonuniformities measured using the Symbia TruePoint scanner were 46.32% and 17.86% with the LEHR and LEAP collimators, respectively. These results indicated that uniformity enhanced by acquired counts.
The nonuniformity of 99mTc measured by the NU 4 IQ phantom was similar to 125I (Figure 3(b)).
3.4. Recovery Coefficient
Figure 4 shows the recovery coefficient of each hot rod in a cold background as a function of rod diameter. However, there are two problems, image noise and resolution of the clinical scanner, for evaluating recovery coefficients. Generally, the theoretical maximum value of RCrod is 1.0, but it can be more than 1.0 because of image noise. The resolution of the clinical scanner was 8.89 mm FWHM in the LEHR collimator and 9.25 mm FWHM in the LEAP collimator for 125I point sources. Therefore, detecting hot rods (1.0 - 5.0 mm) was impossible. For these reasons, calculation of this parameter on the clinical scanner was incomplete and was predicted by the position of the hot rod in the object based on the physical dimensions of the phantom.
3.5. Spillover Ratio
Figure 5(a) shows the results of SOR using the 125I-filled NU 4 IQ phantom. The SORwat was 20.41%, 19.79%, 16.95% in the Inveon scanner with the MHR-0.5, MGP-1.0, and MHS-2.0 collimators, respectively. The SORair in the Inveon scanner was 14.71%, 12.65%, 8.96%, with the MHR-0.5, MGP-1.0, and MHS-2.0 collimators, respectively. The SORwat in the Symbia TruePoint scanner was 36.30% and 32.30% with the LEHR and LEAP collimators, respectively. The SORair in the Symbia TruePoint scanner was 28.14% and 23% with the LEHR and LEAP collimators, respectively.
Figure 5(b) shows the SOR results measured by 99mTc-filled NU 4 IQ phantom.
![]()
(a)![]()
(b)
Figure 3. Nonuniformity (%STDunif) of the clinical scanner compared to the small animal scanner. (a) NU 4 Image Quality (IQ) Phantom filled with Iodine-125 (125I) and (b) Technetium-99m (99mTc).
![]()
Figure 4. Recovery coefficient for Iodine-125 (125I)-filled rods of 1, 2, 3, 4, and 5 mm diameter.
![]()
(a)![]()
(b)
Figure 5. Spillover ratios (SORs) for cold-water and air compartments. (a) NU 4 Image Quality (IQ) Phantom filled with Iodine-125 (125I) and (b) Technetium-99m (99mTc).
3.6. Animal Studies
Small Animal Imaging
Figure 6(a) and Figure 6(b) show the 125I images of a BALB/c mouse acquired by the Inveon scanner using the planar and SPECT modes, respectively. Figure 6(c) and Figure 6(d) show the 125I images of a BALB/c mouse acquired by the Symbia TruePoint scanner using the planar and SPECT modes, respectively.
Figure 6(e) shows the 125I-affibody image of human breast cancer KPL-4 and MDA-MB-231 xenograft mice using the Symbia TruePoint scanner. Two mice were scanned simultaneously using the clinical scanner at 1 hour after 125I-affibody injection. The tumors were faintly visualized as accumulation regions on the hind leg.
4. Discussion
This comparative result of 125I showed that the sensitivity was comparable between the Inveon scanner and Symbia TruePoint scanner. An important limitation of the evaluation of the clinical scanner stemmed mostly from the source-to-collimator distance and small phantom object size. The recovery coefficient was indefinite, and the spillover ratio was overestimated in the clinical scanner. Nevertheless, the uniformity of the clinical scanner was acceptable because of the parallel collimator.
For the reasons mentioned above, attempting small animal imaging by a clinical scanner is more important than statistically analysis. Not surprisingly, the performance of the small animal scanner was undoubtedly better than that of the clinical scanner. Even small components could be distinguished with less blurring than clinical scanner. However, Figure 6(e) implies an unexpected possibility for small animal imaging beyond numeric data. Accumulation of 125I-labeled target agents in the hind legs of a xenograft mouse model of human cancer was confirmed sufficiently for the purpose in spite of planar mode imaging. These results are acquired by the daily routine methods for patient imaging in hospitals without special modifications, such as attaching pinhole collimators.
The small animal imaging system is an important tool with great potential for development of pharmaceutical agents aimed at diagnosis and therapy. Therefore, the research demand for small animal imaging has also increased over the last decade [4] - [7] . The resolution of small animal scanners is generally in the submillimeter range. This range is a considerable improvement compared to the 8 mm or higher range resolution of existing clinical scanners [18] [19] . Because a high-resolution imaging system is not always available in clinical environment, clinical scanner for small animal imaging would be feasible imaging modality for animal imaging and research.
5. Conclusion
This study collectively showed that 125I animal imaging using the Symbia TruePoint scanner was feasible for animal imaging. The advantages of using the clinical scanner include its large field of view, which would allow imaging of 2 - 3 mice, and availability, since it could be used where a dedicated small animal scanner system is not available. Small animal imaging using a clinical scanner is more accessible and readily available. This study shows that using clinical scanner can expand small animal imaging research opportunities. Our results also showed the possibility of 125I for clinical use although 125I imaging was not feasible due to low energy of 125I for clinical imaging.
Supported
This work was supported by Priority Research Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (50458-2011 (PI: Ilhan Lim)). This work was supported by the Korea Science and Engineering Foundation (KOSEF) (No. 2015001667 (PI: Joo Hyun Kang) and No. 1711026888 (PI: Kook-Hyun Yu)).
NOTES
*Corresponding authors. Authors with equal contributions.