With positron emission tomography (PET), the positron of a β
+ emitter radioisotope annihilates with a nearby electron
producing a pair of back-to-back 511 keV gamma rays that can be
detected in a scanner surrounding the point source. The position of
the point source is somewhere along the Line of Response (LOR) that
passes through the positions where the 511 keV gammas are detected.
In standard PET, an image reconstruction algorithm is used to
combine these LORs into a final image. This paper presents a new
tomographic imaging technique to locate the position of a β
+
emitting point source without using a standard PET image
reconstruction algorithm. The data were collected with a
Proof-of-Concept (PoC) PET scanner which has high spatial and energy
resolutions. The imaging technique presented in this paper uses
events where a gamma undergoes Compton scattering. The positions
and energies deposited by the Compton scattered gamma define the
surface of a Compton cone (CC) which is the locus of all possible
positions of the point source, allowed by the Compton kinematics.
The position of the same point source is also located somewhere on
the LOR. Therefore, the position of the point source is defined by
the 3 gammas and is given by the intersection point of the LOR and
the Compton cone inside the Field of View (FOV) of the scanner. We
refer to this method as CC×LOR. This new technique can locate
the point source with an uncertainty of about 1 mm, after
collecting a minimum of 200 CC×LOR events.