Diagnostic Imaging II Student Project Compton Camera Ting-Tung Chang Introduction The Compton camera operates by exploiting the Compton Effect. It uses the kinematics of Compton scattering to contract a source image without the use of collimators or masks. Compton cameras have been built and tested for diverse applications including Compton telescopes for gamma-ray astronomy, as replacements for lead-collimated medical imagers, for industrial imaging, and for radioactive waste management. The Compton camera normally consists of two detectors. The first detector is designed to maximize the likelihood of Compton scattering and the second is designed to absorb the energy remaining after the Compton scatter. The useful range extends from kev to MeV. In contrast, the conventional camera is limited to at most a few hundred kev. Furthermore, it could obtain three dimensional images from a fixed position on one side of the source without the need for tomography. Compton Effect Compton Camera The Compton camera principle As illustrated in fig., a gamma photon emitted from point A with a known energy E, Compton scattering within the first detector at point B and undergoing some further interaction in the second detector at point C. If we assume that the coordinates of points B and C and the energy deposited in the first and second interaction are E and E. Then we can get E = E +E, and the angle θ between the incident direction and the direction of the scattered gamma ray is given by: cos( θ ) = + mc [], E E
where m c is the rest mass of electron. The direction of the incident gamma ray is confined to the cone shown in Fig. with its axis along the direction of the scattered gamma ray and its apex at the position of the interaction with the front plane detector. A large number of scattering events from a point source of gamma rays will define multiple cones which intersect at the location of the source. Generally the incident gamma ray energy E, is known or can be determined from peaks in the singles energy spectrum from each detector or in the coincident summed spectrum. A window is then set on the sum peak E + E = E. From events which fall within the window, the image is reconstructed on a plane or on the surface of a sphere centered on the front detector. The intersection of the cone with the image plane, as shown in fig., produces a curve which traces a conic section, typically an ellipse but parabolic and hyperbolic sections are possible for large angle scattering. A large number of events will produce multiple traces which intersect at the source. The simplest reconstruction is just a superposition of the traces which will build up a peak at the source position. The angular resolution of the reconstruction is limited by both the energy resolution and the position resolution of the front and back plane detectors and by the separation between the two detector planes. The Compton camera design
The first imaging gamma-ray detector was the Anger camera consisting of a pin-hole collimator in front of a position sensitive scintillator. The first Compton telescope was essentially a collimator rather than an imager, with an angular resolution of. It consisted of a single scintillator cell for the front plane and an array of nine cells in the back plane. The more recent UCR gamma-ray telescope consists of two arrays of 6 one meter long scintillator rods. The rods are viewed at each end by separate photomultiplier (PM) tubes for position sensitivity, achieving an angular resolution of. The basic concept of Compton camera comprises a detector system made of two parts, a scatter and an absorber. The direction of incident gamma-rays, Compton-scattering from the scatterer to the absorber, is reconstructed by analyzing the interaction positions and deposited energies in both detectors, i.e. by extracting the scattering angle, and applying the Compton formula. Hence the most relevant design criteria are:. Maximize the probability that the incident gamma-rays scatter by just a single scattering interaction from the scatterer to the absorber and that they are absorbed in the later. This determines the total detection efficiency and the sensitivity of the Compton camera.. Minimize the uncertainty with which the positions and energies of the gamma-ray interactions are determined and maximize the distance between scatterer and absorber. This maximizes the accuracy of determining the scattering angle and hence optimizes the resolution of the Compton camera. The scatter detector should have a high Compton scattering cross section. There are a number of options for absorbing the energy of the Compton scattered photon, from high density detectors such as sodium iodide, NaI(Tl), or bismuth germinate scintillators to multiple layers of silicon or germanium strip detectors. The desired events consist of a Compton scatter in the first detector followed by a full energy absorption in the absorption detector. Time coincidences are used to associate events in one detector with the other. In addition, since the energy of the photons emitted by the radionuclide is known, the use of germanium detectors means that the excellent energy resolution can also be exploited. Assuming that E is known, the uncertainty in the angle θ due to the error in the measured energy E is given below: Δθ mc = sin( θ )( E E ) ΔE The sin(θ) term in the denominator results in a large uncertainty for small angle scattering. The E terms in the denominator indicate that the uncertainty decreases rapidly with gamma ray energy. The finite width and depth of the detectors in each plane contribute to the geometrical uncertainty Δθ g, which represents an uncertainty in the vector connecting the positions of interactions in the two detector planes. The errors add in quadrature to give a total contribution to the angular uncertainty as follows: tan ( Δθ ) = tan ( Δθ ) + tan ( Δ θ g )
The choice of detector material is affected by the ratio of the photoelectric absorption to the Compton scattering cross section, which increases rapidly with atomic number Z. The former dominates below about 4 kev for Ge while the Compton Effect remains dominant down to about 6 kev for Si. For high Z semiconductors such as CdTe and HgI, the cross-over occurs at 3-4 kev. From Eq. [], the energy deposited in the first plane, E = E E = E E + α( cos( θ )) where α = E / m o c is the ratio of the incident photon energy to the electron rest mass. When this ratio is small, the energy deposited is also small and can fall below the noise threshold of the detector. All of these factors combine to make Ge marginal for 99m Te gamma rays at 4 kev where a 5 kev energy threshold eliminates all events with less than 3. The optimum geometry of the detector system is influenced by the above factors which suggest an arrangement that favors scattering angles of 3 or greater. The Klein-Nishina equations for Compton scattering show that the scattering probability per unit solid angle is largest for small angle scattering. However, the available solid angle increases with the scattering angle so that the probability per unit scattering angle is peaked around 3 for 66 kev gamma rays as shown in figure below. While only a single detector can be placed at, there is room for several detectors placed in a ring at 3. This also impacts on the question of open versus closepacked arrays of detectors. Close-packed arrays tend to favor small angle scattering whereas open arrays can be arranged at larger scattering angles. In addition, there is the increased probability of multiple scatterings into adjacent detectors in a closepacked array, which limits the useful detector thickness. Further, the ellipses generated by back-projections from adjacent detectors will tend to be similar and thus give poorer point spread functions than for detectors with larger separations. For these reasons, systems using open arrays generally have the advantage. Detectors Type
The Monte Carlo simulations suggest that both Si and Ge are feasible scattering detectors. The excellent efficiency that results from using Ge as the first detector has been demonstrated by Singh et al. and is the detector of choice for their ECGC system. Moreover, it is necessary to stack individual silicon detectors to achieve competitive efficiency. However, the need for cryogenic cooling of Ge, the expense of Ge detectors and the interesting advances recently made in silicon detector manufacture have led us to start investigations using silicon detectors. A crude prototype consisting of a single 5 X 5 cm microstrip detector operated in coincidence with a Pho gamma camera is being developed in collaboration with members of the high-energy physics group at Imperial College. Having established a basic working system, additional planes of Si will be added thus increasing the sensitivity. An interesting possibility that will also be examined is the use of two-dimensionally position-sensitive silicon drift detectors as the first detector. The potentially excellent energy resolution afforded by the low capacitance of these devices and their outstanding position resolution make them attractive. Conclusion Compton Cameras have advantages over mechanically collimated systems of having a wider field of view and of using a much larger fraction of the gamma rays emitted by the source in order to produce an image. This gives them increased sensitivity allowing improved signal-to-noise ratio, shorter counting time, or when used for medical imaging, it could reduce the radiation dosage to the patient. Compare to conventional detector, it can image high energy photons ( kev~ MeV), and it has the ability to obtain 3D image from a fixed position on one side of the source without the need for tomography. With the improvement of reconstruction algorithms and hardware, the Compton camera could apply to modern SPECT and PET scanners. Reference [] G.W. Philips, Gamma-ray imaging with Compton cameras, Nuclear Instruments & Methods in Physics Research B 99 (995) 674-677 [] W. Gast et al., Gamma-ray imaging with segmented tracking detectors. [3] G. J. Royle et al., Design of a Compton camera for imaging 66 kev radionuclide distributions, Nuclear Instruments & Methods in Physics Research A 348 (994) 63-66 [4] C. Solomon et al., Gamma Ray Imaging with Silicon Detectors A Compton Camera for Radionuclide Imaging in Medicine, Nuclear Instruments & Methods in Physics Research A 73 (988) 787-79 [5] Singh M, Brechner RR. Experimental test-object study of electronically collimated SPECT. J Nucl Med 99;3:78-86.