Investigation into a Compton Camera

The primary limitation in SPECT imaing is in the poor spatial resolution and high patient dosage. This is required so that enough counts can be detected by the camera in order to provide a statistically accurate image. The typical lead collimator allows only 1 in 10,000 photons emitted from the patient to be detected. Thus, without the collimator the performance of the system should improve substantially. This may be achieved by using a Compton camera.

The Compton Camera

The potential gains that the Compton camera offers over the Anger Camera are primarily due to the absence of the lead collimator. This is expected to (i) give great gains in system sensitivity, and (ii) allow the acquisition of data representing multiple angular views of the source distribution from a single position; reducing the necessary camera motion.
Both these effects may reduce the radiation dose delivered to the patient in order to produce a useful image. First, the systems increased sensitivity will allow the use of lower activity levels and/or isotopes with shorter half-lives. Secondly, the reduced angular motion of the camera will result in less wasted time between angular stops. The mechanical complexity of the system will also be reduced because of the abscence of the massive lead collimator.
Compton cameras are in current use in the nuclear energy industry for site and environmental surveys.

Principle of operation

The generic Compton camera design consists of 2 detectors. The function of the first detector is to promote and to detect Compton interactions. This is commonly referred to as the "scatter detector". Si, Ge and Ar have all been proposed and experimented with as scatter detector materials, primarily due to their high Compton scatter cross-sections in the 100-600 KeV range.
The second detector is referred to as the "absorbtion detector". As the name suggests, its function is to absorb scattered gamma rays, measuring both position and energy. Typical choices for this detector are CdZnTe, NaI, BGO and Xe, due to their high photo-absorbtion cross-sections in the energy range of interest.
Although NaI and BGO detectors have worse energy and position resolutions than their solid state and gas couterparts, they were the first absorbtion detector materials investigated, and continue to be considered. This is because they are already in wide use in SPECT cameras. This could allow the development of a "Compton camera add-on" to convert existing cameras at low cost.
Because the exact energy of the photons emitted by the radioisotope is known, the sum of the energies deposited in the scatter and absorption detectors is used to reject photons that do not offer accurate information about their source because of scattering within the patient or other effects. The probability of the photon scattering though a given angle is described by the Klein-Nishina formula. This formula also relates the energy deposited in the scatter detector to the angle through which the photon was scattered.

The positions of the interactions, coupled with the scatter angle, limits the photon's possible source location to a cone whose axis is in line with the positions of the two interactions, and whose opening angle is equal to the scatter angle. Uncertainties in the positions and energies of the interactions give the cone of possible source locations a non-zero thickness.
This type of information about the possible source distribution requires specialized reconstructions techniques. These techniques are under investigation by a number of research groups, including ours.

Solid State and Solid State / Scintillation Compton Cameras

Solid state detectors are being investigated for use in compton cameras because of their high energy and spatial resolution compared to scintillation crystals. As stated earlier, Si is the best solid-state choice for the scatter detector.
Solid state detectors require more supporting electronics than scintillation or gas detectors because each pixel in the detector uses it's own readout channel.
Si detectors are also very sensitive to thermal noise, and require cooling in order to operate at acceptable levels of energy resolution.
Solid state detectors usually have a smaller field of view (approx. 10 cm) than existing anger cameras (50 cm).
Si detectors are limited to approximately 5 mm thickness because of the technical challenges and costs involved in making them thicker. Thicker Ge detectors can be produced without much trouble, but it's less than ideal because of its high photoelectric cross-section.
The possibility of using Si strip detectors "edge-on" has been proposed by Beauville, Collins, Kipins, Nygren, Oltmann and Hauptman. This would present a greater apparent thickness to the gamma source, and allow much greater efficiency of detection. The number of read-out channels would be extremely high for this approach, complicating the electronics of the detector.

Another approach is to stack multiple layers of Si detectors. This type of "multi-layer" Compton camera would allow the possibility of the photon scattering more than once before it is absorbed. This kind of thing is referred to as a three(or more) - vertex event.
In three(or more)-vertex events, the possibility of using knowledge of the gamma photon's polarization is being investigated by Basko, Zeng, and Gullberg. They say that this could reduce to cone of possible source positions to a "V", because all scatters after the first will lie in the plane of the photon's polarization.