A small field of view camera for hybrid gamma and optical imaging

The development of compact low profile gamma-ray detectors has allowed the production of small field of view, hand held imaging devices for use at the patient bedside and in operating theatres. The combination of an optical and a gamma camera, in a co-aligned configuration, offers high spatial resolution multi-modal imaging giving a superimposed scintigraphic and optical image. This innovative introduction of hybrid imaging offers new possibilities for assisting surgeons in localising the site of uptake in procedures such as sentinel node detection. Recent improvements to the camera system along with results of phantom and clinical imaging are reported.


Introduction
Nuclear medical images are typically acquired using large field of view (LFOV) gamma cameras designed for whole body scanning and SPECT imaging. However, a new generation of small field of view (SFOV) gamma cameras, now in development, has been designed to provide higher resolution capabilities for specific procedures, such as sentinel node localisation.
A number of researchers and manufacturers have been developing a range of small hand-held or portable systems for medical gamma imaging [1][2][3][4][5][6][7][8][9][10]. These smaller systems have all concentrated on a single imaging modality. This communication describes a compact dual-modality camera that combines optical and gamma ray imaging into a single system. We also report on a number of studies using thyroid and lymphatic phantoms and a recent clinical image showing the benefit of a combined optical and gamma image.

Compact Gamma Camera
The Compact Gamma Camera (CGC) combines an optical and a gamma camera in a configuration that offers high spatial resolution multi-modality imaging with the option for a fused image output. The basic design is flexible with a range of collimators that can be selected to balance the requirements of spatial resolution and sensitivity to match to a wide variety of applications.
The CGC consists of a charge-coupled device (CCD), a CsI(Tl) scintillator coupled to the CCD (e2v CCD97 BI [11]) and a tungsten pinhole collimator. The CCD to pinhole distance is fixed at 10 mm. The distance from the pinhole to the object under examination determines the magnification of the system. Gamma-ray photons arriving from the source will interact with the -1 -scintillator and generate optical photons, the number being related to the absorbed photon energy, which are subsequently detected by the CCD -allowing the position and the energy of the incident gamma photon to be recorded [2]. The hybrid camera head is surrounded by tungsten shielding (3 mm thickness) and sealed in a non-toxic plastic enclosure to provide both thermal and electrical isolation. Figure 1 shows the CGC being operated in a hand held mode.  h.s.) shows a simplified schematic of the hybrid system; the hybrid camera head comprises of the standard CGC, a thin (3 mm thickness) first surface mirror at 45 degrees and an optical camera outside the direct line of sight of the pinhole. A photograph of the optical mount is also shown (r.h.s.).
In an early version of the hybrid system [2] the optical camera was positioned directly in front of the gamma camera. A low cost webcam with a precision pinhole (100 µm diameter) was used to obtain the optical image. Although this design was simple to implement, it had a number of disadvantages -most significantly the absorption of a portion of the gamma rays by the electronics located on the back of the camera. In the updated configuration (see figure 2), gamma photons from the source will pass through the mirror with minimal absorption (< 3%) and scatter, whereas the optical photons are reflected towards the optical camera. Both the optical and gamma cameras are connected to a readout and control electronics system managed by a PC.
The CGC has been designed to be sensitive over the energy range 30-140 keV. At present, the CGC uses a Hamamatsu [12] CsI(Tl) columnar scintillator (600 µm thick on an amorphous carbon substrate) coupled directly to the CCD with Dow Corning optical grease. This thickness of scintillator has an absorption ranging from ∼ 100% to 20% over the energy range of interest [2].
The CGC uses a single pinhole collimator, partly to simplify the construction but also to extend the imaging field of view which, for a parallel-hole collimator, would be limited to the image area of the CCD (∼ 8 mm × 8 mm [2]). Two collimators were manufactured [13] from tungsten discs, 6 mm thick and 45 mm in diameter with pinholes of 0.5 mm or 1.0 mm diameter, both having an acceptance angle of 60 degrees. The transmission through 6 mm thick tungsten is < 4 × 10 −9 at 140 keV, thereby limiting detected gamma rays to those passing through the pinhole.
-2 - There are a number of factors that determine the sensitivity of gamma cameras that use pinhole collimators including photon energy, collimator material, pinhole diameter and acceptance angle. For a point-like object at 25 mm, the on-axis geometric sensitivities were estimated to be 4.3 × 10 −5 and 1.4 × 10 −4 (with 0.5 mm and 1.0 mm diameter pinholes respectively) for 140 keV gamma photons.

Imaging software
The system can operate in either single-or dual-modality mode. Each image format can be displayed independently or in a fused co-aligned image with adjustable contrast and colour tables to help interpretation.
A full gamma image consists of many individual frames acquired over a period of time, typically ranging from 100-20,000 frames. The image is analysed on a frame by frame basis using a "blob-detection" algorithm with automatic scale selection [2]. This analysis approach gives information on the number and intensity of all gamma interactions within the scintillator, which in turn gives information on the energy distribution of the gamma events. The optical images are captured directly from the optical camera electronics using bespoke interface software which allows for the capture of images from multiple cameras. Each image is imported directly into the main acquisition software for display and analysis. The current optical cameras are simple webcams chosen for their low profile, low cost and availability.

Modelling
A new Monte Carlo model, independently developed in IDL, 1 has been specifically designed to simulate the response of a pinhole-collimated scintillator-based gamma camera to incident photons. This model was developed to allow greater flexibility in both input and output information than a proprietary software package. The aim of its development was to fully understand the responses of components of the CGC.

Scope of model
The model simulates gamma photon attenuation from photoelectric absorption, Compton and Rayleigh scattering along with fluorescence and Auger transitions. Within the scintillator, energy deposition and transport (including reflection effects) are modelled. A simulated image is formed based upon the detection of scintillation photons in the CCD, along with the modelled effects of noise from the CCD and associated electronics. Final results may be viewed in real time or recorded for further analysis.

Modelling results
The model allows for the investigation of response from individual camera components along with the device as a whole. The model has been used to investigate sensitivity, spatial resolution and energy resolution and shows a good agreement with theoretical and experimental values. Figure 3 shows an example simulated image and an experimental image acquired with the CGC under similar conditions. A full description of the model will be reported in a future paper. We anticipate that the simulation will be used to guide future developments of the CGC.

Imaging phantoms
A number of different phantoms were used to provide a variety of shapes and configurations to explore the imaging response of the CGC system.
-4 -Thyroid phantom. For thyroid simulations a Picker Nuclear thyroid phantom (Part # 3602) was used. This phantom has a total volume of 35 ml; the left hand lobe has a depth of 18.4 mm and the right hand a depth of 9.2 mm. The phantom also contains nodules, three of which were cold spots where no activity was present, and one of which was a hot spot with a depth of 18.4 mm in the 9.2 mm lobe. The phantom is sealed with a 5 mm Perspex cap. The area of the entire phantom is 60 mm 2 , approximately 50% larger than a typical patient thyroid.
The phantom was filled with 99m Tc from a stock solution to an activity of 100 MBq. The phantom was placed 100 mm below the camera (fitted with the 0.5 mm diameter pinhole) and a 60000 frame image accumulated (∼ 30 minutes acquisition time). Figure 4 shows a schematic of the Picker phantom (l.h.s.) and an example gamma-ray image (r.h.s.). The 12 mm nodes, both hot and cold spots, are clearly distinguishable, as is the 9 mm cold spot on the low activity side of the phantom. The smallest cold spot (top left hand side) can be identified although with a much lower contrast-to-noise ratio. The long exposure image, shown in figure 4, allowed us to explore the stability and contrastto-noise ratio of the CGC. To gauge the response and detectability of the system, a sequence of shorter exposure images was accumulated. Figure 5 shows the time sequence with exposure times ranging from 200-2000 s. As expected, the clarity of the phantom image improves with exposure time. Ultimately the choice of exposure time will rest with the clinician and will vary depending on the application and activity of the site under investigation.
Lymph node phantoms. A lymphoscintigraphic phantom has been constructed to assess the capability of the CGC for lymphatic vessel drainage imaging. In this phantom, lymphatic vessels were simulated with a 0.55 mm internal diameter cannula (0.2 MBq/mm of 99m Tc) and an injection site simulated using a 10 mm maximal diameter Eppendorf tube (60 MBq of 99m Tc). Varying thicknesses of Perspex, placed between the phantom and camera, were used to simulate differing -5 - depths of vessel within the body. The same phantom was also imaged using a conventional Large Field of View (LFOV) camera (Mediso Nucline X-Ring-C).
Fused optical and gamma images of the simulated lymphatic vessel produced by the CGC (figure 6) show good alignment of the two modalities -allowing accurate localisation of activity within the field of view.
In figure 7, the simulated injection site is within the FOV. Even in the presence of the highactivity injection site, the targeted lymphatic vessel can still be seen after an acquisition time of ∼ 2 minutes. Increasing the exposure time to 10 minutes improves the contrast of the image (figure 7A).   Figure 8 shows the relationship between the measured full width half maximum (FWHM) of the lymphatic vessel image and the distance from the injection site for both the CGC and the LFOV camera. The measured FWHM was significantly smaller for the CGC than the LFOV camera regardless of separation or depth of Perspex. When the distance from the centre of the modelled injection site was greater than approximately 15 mm the FWHM of the lymphatic vessel reached a constant level (variation < 5%), which reflects the fundamental spatial resolution of each camera (measured at 100 mm from the source).

Imaging patients
It is anticipated that the combination of a portable gamma camera with an optical camera will have practical benefits for a number of procedures. The most obvious benefit is due to the registration of the gamma image onto the anatomical optical image, allowing the clinician to make a better assessment of radiopharmaceutical localisation during surgery leading to a more accurate diagnosis. As part of the development of the CGC, patients who were undergoing clinical investigations at the nuclear medicine clinic at Queen's Medical Centre, Nottingham were imaged with the hybrid camera following their standard clinical test as permitted following ethical approval.
Lacrimal drainage. A nuclear lacrimal drainage scan is a non-invasive study that uses a low radiation dose introduced to the eye to assess the performance of the nasolacrimal drainage system (tear duct). This procedure can be important for pre-operative planning -to determine the position of an obstruction -and post-operatively for surgical evaluation [1,14]. Figure 9 shows a recent CGC image of a patient undergoing a lacrimal drainage investigation. 1 MBq of 99m Tc-pertechnetate in a saline solution was introduced to each eye.
The gamma image is a summation of 2000 frames overlaid onto an optical image of the patient using the fusion software described in section 2. Figure 9 clearly shows good agreement between the expected location of the activity and that determined by the anatomical image.

Conclusions
Intra-operative imaging is used widely in medicine, and the use of gamma camera imaging has the potential to further improve surgical outcomes. The concept of sentinel lymph node biopsy has been applied to assist in the identification of patients for complete lymph node dissection [15,16] and has become an important technique for staging cancer and, in particular, determining whether the disease has spread from the primary tumour to the lymphatic system. This procedure is becoming standard, and is carried out over 2 million times annually. For this procedure to be successful it is necessary to identify the sentinel node pre-operatively.
The clinical advantages of fusing the gamma image to the optical image are clear and should be especially beneficial for use during surgical procedures. The CGC may have many practical benefits for a number of clinical procedures including diagnosis, surgical investigation and the visualisation of drug delivery.
We anticipate that by providing the surgeon with an overlaid image during an operation confidence in localisation of sites of uptake can be increased, reducing the procedure time and providing both a better diagnosis and outcome.
The CGC has been shown to be capable of high resolution gamma imaging combined with optical imaging to produce fused images offering improvements in a number of nuclear medicine applications. The camera has been used to produce dual-modality images from phantom simulations and patients undergoing clinical examinations.
The CGC offers combined optical and gamma imaging that will facilitate new techniques to address both existing and emerging healthcare needs, especially point-of-care and intraoperative imaging.