Submillimeter nuclear medical imaging with a Compton Camera using triple coincidences of collinear \beta^{+}-annihilation photons and \gamma rays

Submillimeter nuclear medical imaging with a Compton Camera using triple coincidences of collinear -annihilation photons and rays

C. Lang, D. Habs, P.G. Thirolf, A. Zoglauer

CHL-ICTR-PHE12-arXiv-2.dvi keepcomments

Fakultät für Physik, Ludwig-Maximilians Universität München, Garching, Germany

Space Science Laboratory, University of California, Berkeley, CA, USA

PET cameras had their breakthrough as an imaging instrument for clinical application studies due to the large variety of radioactively labeled tracer molecules [1]. These tracers carry an e emitting isotope to e.g. a tumor. Annihilation of the positron into two back-to-back 511 keV photons allows to restrict the source origin in 2 dimensions onto a line of response (LOR).

Superimposing the LORs of different decay events locates the source distribution of the emitter in 3D. Modern PET systems reach a spatial resolution of 3-10 mm. A disadvantage of this technique is the diffusion of the positron before its decay with a typical range of ca. 1 mm (depending on its energy). This motion and Compton scattering of the 511 keV photons within the patient limit the performance of PET. We present a nuclear medical imaging technique, able to reach submillimeter spatial resolution in 3 dimensions with a reduced activity application compared to conventional PET. This ’gamma-PET’ technique draws on specific e sources simultaneously emitting an additional photon with the decay. Exploiting the triple coincidence between the positron annihilation and the third photon, it is possible to separate the reconstructed ’true’ events from background [2,3]. Therefore the spatial uncertainty introduced by the motion of the e or by Compton scattering within the patient can be strongly reduced in the direction normal to the annihilation. Especially Sc is of interest, which decays into Ca, emitting an 1157 keV photon. With its halflife of 3.9 h, Sc has to be produced from a Ti generator (t = 60.4 a) [4]. Presently this cannot be performed in clinically relevant quantities, however, this may change with the soon expected availability of highly brilliant gamma beams [2]. Sc has already been applied in patient studies [5]. We combined the Compton camera technique, i.e. the measurement of photon energies and positions of Compton scattering interactions, with a PET camera defining the LOR. Due to the Compton kinematics and subsequent photon absorption, a Compton camera can reconstruct the origin of a primary photon on the surface of a ’Compton cone’ [6]. Superimposing such cones from different events reduces the reconstructed source distribution in 3D to the few-millimeter range. The gamma-PET technique is different, it will intersect the Compton cone with the LOR from the same annihilation/gamma coincidence event, thus allowing to reconstruct the source position in 3D from single events.

In order to test the feasibility of this technique, Monte-Carlo simulations and image reconstruction has been performed using the ’Medium Energy Gamma-Ray Astronomy’ library MEGAlib [7]. MEGAlib is a software framework designed to simulate and analyze data from Compton cameras. For the requirements of the gamma-PET technique we modified MEGAlib to realize an event reconstruction from the intersection of the Compton cone with the LOR. The simulated geometry consists of four Compton camera modules in a quadratic arrangement surrounding an HO sphere of 6 cm diameter in a distance of 3.5 cm from the sphere center. Two of them serve for detecting the annihilation photons defining the LOR, the others allow for detecting the third photon, enabling to reconstruct the Compton cone. Inside the sphere two Na point-like, non-medical test sources were placed at a distance of 0.4 mm. Each camera module consists of a LaBr scintillator crystal (50x50x30 mm) read out by a 2D-segmented photomultiplier with 64 pixels (6x6 mm each) and a double-sided silicon strip detector with 128 strips on each side, an active area of 50x50 mm and an optimized thickness of 2 mm.

Figure 1 shows the source distribution after 60 iterations of the reconstruction algorithm.

Figure 1: Reconstructed source distribution of two non-medical, long-lived Na point-like test sources placed in a water sphere (6 cm diameter) with a distance of 0.4 mm between them. The reconstruction was based on triple coincidences between the two annihilation photons and the third 1275 keV gamma ray after 60 iterations of the reconstruction algorithm. (Left) ungated projection on the x axis.

The two Na sources can clearly be resolved. The spatial resolution amounts to 0.2 mm (FWHM) in each direction, surpassing the performance of conventional PET by about an order of magnitude. The simulated detector geometry exhibits a coincidence detection efficiency of 1.92e-7 per Na decay. In 25 of these events the intersection between the LOR and the Compton cone could be identified. Starting with only 0.7 MBq of source activity (ca. 200-500 times less compared to conventional PET), about 20 intersections (sufficient for source reconstruction) can be identified after an exposure time of 450 s.

This work was supported by the DFG Cluster of Excellence MAP (Munich-Centre for Advanced photonics).

[1] P. Suetens, Fundamentals of Medical Imaging, Camb. Univ. Press (2002).
[2] D. Habs and U. Koester, Appl. Phys. B 103, 501 (2011).
[3] C. Grignon et al., Nucl. Instr. Meth. A 571, 142 (2007).
[4] M. Pruszynski et al., Appl. Rad. Isot. 68, 1636 (2010).
[5] F. Roesch and R.P. Baum, Dalton Trans. 40, 6104 (2011).
[6] G. Kanbach et al., Nucl. Instr. Meth. A 541, 310 (2005).
[7] A. Zoglauer et al., New Astron. Rev. 50, 629 (2006).

Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description