SPADnet improves on the state of the art in many respects. The core differentiating features of the proposed component are innovative packaging techniques based on through silicon vias, the concept of spatial oversampling, and an advanced fully digital network communication. Our project represents a timely unification of emerging technology in detectors, 3D-IC packaging and sensor networks towards the goal of PET imaging.
A recent state of the art overview is provided in the public version of SPADnet's Deliverable D1.2 SPADnet Design: Specifications & Planning. We will also comment in the following on the state-of-the-art related to the target application (PET-Positron Emission Tomography).
Pre-clinical PET imaging systems
The most recent commercially available pre-clinical PET systems are based on more heterogeneous concept (see table below) than the human scanners. All systems use lutetium based scintillator material, either LSO or LYSO, and the majority have pixellated crystals. The axial and tangential size of the pixels are in the range of 1,12 – 2 mm, while the thickness (radial dimension) varies between 9,8 – 10 mm. There is only one commercially available pre-clinical PET system that employs continuous crystals.
In small animal scanners the diameter of the PET ring is relatively small with respect to the length of the crystals, in comparison with human systems. This results in a parallax error caused by missing or imprecise information on the depth of interaction (DOI) in the crystal. There are several ways to correct for this effect, like optical DOI encoding, use of a phoswich detector structure, use of a monolithic scintillator, or the sandwich approach that we propose in SPADnet.
There is wide variety of inorganic scintillator materials suitable for PET design concepts. In the early systems [Casey86] NaI, BGO and GSO was used, but one of the major change in the clinical PET detectors was the introduction of the new scintillator Lutetium Oxyorthosilicate (LSO) in 1992. LSO and LYSO currently offer the best combination of luminosity, decay time and coincidence detection efficiency. Because the target application for the SPADnet sensor is PET, we have selected LYSO by the previously mentioned reasons. In order to use the SPADnet detector module for other imaging modalities different scintillator material has to be selected (eg. NaI for SPECT).
In order to achieve high spatial resolution pixellated crystals are used to reduce the well-known spread [Derenzo93],[Moses93] of the scintillation light cone emitted by the crystal. Each pixel is separated by a thin reflector layer to avoid the crosstalk between the neighboring crystal pins. However, in order to achieve further improvements in terms of spatial resolution in such systems, the crystal pins need to be made ever smaller, creating a number of technical problems and elevated cost.
Beside the scintillator the other major component of the detector is the sensor. Currently the most widely used sensors in the PET and SPECT technology are the photomultiplier tubes (PMTs). Comparing to the other solutions PMTs have very high gain, low dark current and relatively good temperature stability making them almost ideal sensors for most of the applications. One significant drawback of PMTs is their extreme sensitivity to magnetic fields, which renders them unsuitable for use in magnetic field or vicinity of an MRI scanner. In order to improve the spatial resolution multi anode photomultiplier tubes are introduced with 64, 144 or 256 channels with the same area as a standard 2 inch wide PMT. These position sensitive photomultipliers (PSPMTs) are applied to date in dedicated fields such as small animal imaging only and they have not fully translated into the world of clinical PET.
Solid-state photodetectors (APD, SiPM, SPAD) have also several advantages over PMTs: high quantum efficiency, compact and flexible shape that can be adapted to individual crystals, ruggedness and insensitivity to magnetic field.
In academic labs different groups [Raylman06][Mackewn05][Schlyer07][Judenhofer07] have made considerable progress with APD-based PET inserts for pre-clinical scanners. This approach is also employed in the first prototype of a clinical PET/MR scanner for brain imaging that is being developed by Siemens. The concept has proven to work and the first studies are ongoing in preclinical prototypes, and also on the brain scanner prototypes at a limited number of clinical sites. However, the systems still suffer significant difficulties in their operation and are far from a stable system that can be commercialized. In addition, these APD based systems suffer from a comparatively poor timing resolution that makes them incompatible with time-of-flight readout.
Multicell Geiger-mode Avalanche Photodiodes, also known as Silicon Photomultipliers (SiPMs), are a novel type of photodetectors that have undergone a rapid development [Buzhan01] [Herbert06] [Piemonte07]. They are small, light and insensitive to magnetic fields. In addition, their improved characteristics in comparison to other solid-state photodetectors, namely high gain and excellent signal-to-noise ratio enables the SiPMs to be employed as photodetectors in PET and PET/MR.
Fast scintillators and photodetectors also allow the implementation of time-of-flight (TOF) techniques in human systems. The difference in the arrival time between the two detectors in coincidence can be used to reduce the uncertainty in the annihilation point of the event to a limited region along the line of response, and thus diminish the noise in the reconstructed images. A timing resolution of 500ps results in a reduction of a factor of five in the variance in the reconstructed image compared to conventional PET [Moses03].
- [Buzhan01] Buzhan et al., "An advanced study of silicon photomultiplier," ICFA Instrumentation bulletin, 23, 28, (2001).
- [Casey86] Casey, M. , Nutt, R. “A Multicrystal Two Dimensional BGO Detector System for Positron Emission Tomography”, IEEE Trans. Nuc. Sci, 33(1) 460-463, (1986).
- [Derenzo93] Derenzo, S.E., Moses, W.W., et al., “Critical instrumentation issues for <2mm resolution, high sensitivity brain PET”, Quant. Brain Func. 25, (1993).
- [Herbert06] Herbert, D. et al. “First results of scintillator readout with Silicon Photomultipliers”, IEEE Trans. Nuc. Sci, 53(1), 389-394, (2006).
- [Judenhofer07] Judenhofer, M.S. et al. “PET/MR images acquired with a compact, MR compatible PET detector in a 7T magnet”, Radiology, 244, 807-814, (2007).
- [Mackewn05] Mackewn, J. et al. “Design and development of an MR compatible PET scanner for imaging small animals”, IEEE Trans. Nuc. Sci., 52(5), 1376-1380, (2005).
- [Moses03] Moses, W. W. et al. “Time of Flight in PET Revisited”, IEEE TNS 50(5), 1325-1330, (2005).
- [Moses93] Moses, W.W. and Derenzo, S.E. “Empirical observation of performance degradation in positron emission tomographs utilizing block detectors,” J. Nucl. Med., 34, 101, (1993).
- [Piemonte07] Piemonte, C. et al., "Characterization of the first prototypes of Silicon Photomultiplier at ITC-irst", IEEE Trans. Nucl. Sci. 54(1), 236-244, (2007).
- [Raylman06] Raylman, R. R. et al., “Combined MRI-PET scanner: A Monte Carlo evaluation of the improvements in PET resolution due to the effects of a static homogeneous magnetic field”, IEEE Trans. Nucl. Sci., 43, 2406-2412, (2006).
- [Schlyer07] Schlyer, D. “First images from the BNL simultaneous PET/MRI scanner”, J. Nucl Med., 48 (Supplement2), 89, (2007).