In recent years hand-held mini γ-cameras have been increasingly used to perform scintigraphic imaging during surgery 1–4. The main applications of these small-sized imaging probes (IPs) concern the detection of positive sentinel lymph nodes in breast cancer and melanoma and the functional imaging of the thyroid and parathyroid glands. The use of an IP permits the acquisition of intraoperative imaging, thus replacing both the preoperative scintigraphy performed with large-sized γ-cameras and the acoustic localization during surgery through a portable nonimaging γ-probe. A few data have been published on specific phantom studies investigating the performances of these new imaging systems 2.
The aim of the present work was to evaluate the physical characteristics of a new hand-held γ-camera through specific phantom measurements relevant for its clinical use.
Materials and methods
The hand-held mini γ-camera used to obtain the measurements in the present study is the IP Guardian2 (Li-Tech, Roma, Italy), which is shown in Fig. 1. It consists of a CsI(Tl) scintillator crystal array composed of 18×18 elements with a sensitive area of 2.25 mm×2.25 mm, thickness of 5.00 mm, and distance of 2.45 mm from each other, matched with a tungsten square-holed collimator, with a 200-μm-thick and 24-mm-long septa. The system is coupled to a position sensitive photomultiplier tube, which has been described already by Trotta et al. 5. The full digital field of view (FOV) is 44.1 cm×44.1 mm. The final pixel size is 2.45 mm. The system is also enclosed in a lead shielding. The whole weight of the mini γ-camera is 1.2 kg. The portable control unit is integrated into a tablet personal computer.
IP Guardian2 is thus able to detect γ-emitters with energies ranging from 50 to 250 keV.
To characterize its performances, with particular attention to its clinical use, we took some phantom measurements normally proposed for nonimaging γ-probes used in surgery, in particular, depth transmission and count-rate linearity 6,7, together with some measurements usually ascertained with large-sized γ-cameras, such as spatial resolution, spatial linearity, uniformity, and image quality 8.
For determining the sensitivity and depth transmission measurements we used a 8.5 MBq 99mTc point source (1 mm diameter) positioned as shown in Fig. 2a.
The spatial resolution images were acquired with a capillary tube filled with 1.4 MBq of 99mTc (1 mm internal diameter), extracting the line-spread function profile.
Spatial linearity data were obtained by acquiring five images, with a 99mTc point source positioned at different locations within the FOV, and comparing the actual relative distances with the values deduced from the IP images. The linearity of response to the incident count rate was estimated through the decay method (activity ranged from 24.8 to 1.7 MBq in a volume of 2.5 ml) imaged at 1-h time intervals at the center of the FOV. The activity range was chosen to simulate values obtainable in clinical settings. In fact, in our center, sentinel lymph node patients underwent an intradermal injection of 60–70 MBq of 99mTc-nanocolloids so that during surgery (18–20 h after tracer injection) the local activity would be about 4 MBq.
For the spatial uniformity measurement, a Petri dish (9 cm diameter) filled with a uniform solution of 115 MBq/18 ml of 99mTc was used, acquiring 3240 kcounts overall (to have about 10 kcounts in each pixel). The integral and differential uniformity values were calculated, removing the edge pixels and smoothing the image with a 3×3 pixel filter, according to National Electrical Manufacturers Association (NEMA) standards 8. The integral uniformity is defined as the difference between the maximum and minimum pixel counts in the whole image divided by their sum, as per the following formula:
where Ci is the pixel count and i ranges within the whole image.
The differential uniformity as the difference between the maximum and minimum pixel counts within a set of five contiguous pixels in each row and column independently was calculated using the following formula:
where i ranges within five contiguous pixels.
Finally, to quantify the whole image quality we used four small hollow spheres, with internal diameters of 4.9, 6.2, 7.8, and 9.8 mm (corresponding to internal volumes of 0.063, 0.125, 0.250, and 0.500 ml), filled with 99mTc, with an activity concentration of 26 MBq/ml, and positioned as shown in Fig. 2b (acquisition time 30 s). The quantitative evaluation was carried out by calculating the contrast-to-noise ratio of each hot sphere using the following formula:
where Ci is the maximum count in the hot sphere and Cbkg and σbkg are the mean counts and SD in the background obtained by averaging six regions-of-interest of 2×2 pixels, respectively.
The main results obtained with the IP Guardian2 system are summarized in Table 1. The sensitivity of IP was calculated according to the depth transmission curve shown in Fig. 3 as the y-intercept obtained through the exponential fit of the data (Pearson’s coefficient, R2=0.999). As shown in Fig. 3 we also calculated the Plexiglas half-thickness, equal to 4.4 cm – that is, the thickness at which the counts decreased to 50%.
As shown in Fig. 4, the spatial resolution of the system at contact with the collimator was 2.5 mm (almost equal to the pixel size), and it worsened at increasing depth of the Plexiglas, as expected (Table 2). The linearity measurements of counts have been plotted in Fig. 5, together with the linear best fit (Pearson’s coefficient, R2=0.999). The integral and differential uniformity values, summarized in Table 1, were calculated on the flood field image as shown in Fig. 6. The useful FOV (UFOV) was obtained by removing the edge pixels, whereas the central FOV (CFOV) was obtained keeping only the central 12×12 pixels.
The whole image quality was estimated imaging four hot spheres in the IP standard setting. The display software can process the pixel data through a spatial interpolation. We used the default option (exponential) to save Fig. 7a and b. All the small lesions appeared well detectable when positioned at contact with the collimator (Fig. 7a), whereas the contrast decreased when a 5-cm-thick slab of Plexiglas was interposed (Fig. 7b), simulating a clinical situation with increased scatter. The contrast-to-noise ratios calculated in the images without application of any interpolation, normalized to the value of the biggest sphere imaged at contact with the IP collimator, are shown in Table 3.
The commercially available hand-held mini γ-camera systems are useful for performing small-area preoperative and intraoperative scintigraphy. They are being increasingly used, substituting in some specific applications the large-area γ-cameras normally used for scintigraphy. The IP Guardian2 system used in the present report was the new model commercialized by Li-Tech. Our data focussed on the IP Guardian2 extrinsic (i.e. with collimator) performances with specific attention to its clinical use. The high spatial resolution, good count-rate linearity, and the overall image quality confidently allow its use in the clinical setting. In addition, the probe is equipped with a portable touch-screen personal computer that facilitates its portability. The sensitivity and spatial resolution agreed well with manufacturer specifications. The extrinsic integral uniformity was 12.0% in the UFOV and 8.8% in the CFOV, whereas the differential uniformity was 5.7% in the UFOV and 4.0% in the CFOV, resulting in improved, or at least on a par with previously described, portable γ-cameras 2.
Specific mini phantoms equipped with smaller holes (diameters ranging from 4 to 1 mm) were reported in the literature and used to characterize the image quality of a similar mini γ-camera 3. In the present work we used four hollow spheres with slightly larger diameters (ranging from 5 to 10 mm) developed as an optional equipment of the NEMA cylindrical phantom but unlikely detectable by usual large-FOV γ-cameras: the hand-held mini γ-camera was able to detect and distinguish (i.e. spatially resolve) each hot sphere, even when a 5-cm-thick slab of plexiglass was interposed.
In contrast, the IP Guardian2 system appears a less flexible system compared with usual large-FOV γ-cameras, as no data on the energy spectra can be exported or analyzed and the user cannot perform any calibration, not even for improved centering of the energy window.
The IP Guardian2 imaging system demonstrated good performances, allowing its confident use for specific applications in which its portability and high resolution are very useful, in particular for radio-guided surgery.
This study was partially funded by Veneto Region, Italy (Ricerca Sanitaria Finalizzata no. 308/2009 – DGRV. 4273 del 29.12.2009).
Conflicts of interest
There are no conflicts of interest.
1. Soluri A, Massari R, Trotta C, Tofani A, di Santo G, di Pietro B, et al. Small field of view, high-resolution, portable γ-camera for axillary sentinel node detection. Nucl Instrum Methods A. 2006;569:273–276
2. Sanchez F, Benlloch JM, Escat B, Pavón N, Porras E, Kadi-Hanifi D, et al. Design and tests of a portable mini gamma camera. Med Phys. 2004;31:1384–1397
3. Lees JE, Bassford DJ, Blackshawn PE, Perkins AC. Design and use of a mini-phantom for high resolution planar gamma cameras. Appl Radiat Isot. 2010;68:2448–2451
4. Vermeeren L, Valdés Olmos RA, Klop WMC, Balm AJM, van den Brekel MWM. A portable γ-camera for intraoperative detection of sentinel nodes in the head and neck region. J Nucl Med. 2010;54:700–703
5. Trotta C, Massari R, Palermo N, Scopinaro F, Soluri A. New high spatial resolution portable camera in medical imaging. Nucl Instrum Methods A. 2007;577:604–610
6. Guidelines on the quality assurance of intraoperative gamma probes – Version 1. 2005 London UK Gamma Probe Working Group
7. Intraoperative probes for radio-guided surgery. Protocol for the quality control. 2001 Milan
8. NEMA Standard Publication NU-1 2001. Performance measurements of scintillation cameras. 2001 USA NEMA
Keywords:© 2013 Lippincott Williams & Wilkins, Inc.
mini imaging γ-camera; phantom study; radio-guided surgery