PET scanning is a relatively old (late 1950s) technology. What has brought this once obscure diagnostic tool to usefulness since the 1950s is powerful computers and improved detectors. PET is an acronym for Positron Emission Tomography, which are probably all arcane terms for a non-tech geek.
A Positron is the antimatter version of an electron - it is an electron, but with a positive charge. Only a handful of radioactive isotopes decay by emitting positrons, and only a few of these are short-lived enough to justify injecting into a patient.
So we have described the first two terms: Positron and Emission.
Tomography means building a three dimensional image, for diagnostic purposes. So there we have it: Positron Emission Tomography.
An interesting thing about positrons; because they are antimatter, they don't last very long in a world that is made from normal matter. When a positron is created as a result of a nuclear decay, it very rapidly collides with a normal electron. The two particles are destroyed in a process called "mutual annihilation" and the destruction of their mass is converted (E = MC^2) to two gamma rays at energy levels of 511 KeV each. The gamma rays normally leave the point of destruction at a path 180 degrees from each other. This is useful to imaging, as will be seen later.
The positron-emitting isotopes chosen for imaging are Carbon-11 (half life 20 min), Nitrogen 13 (half life 10 min), Oxygen 15 (half life 2 min), Fluorine-18 (half life 110 min), and Rubidium-82 (half life 1.27 min). A short half-life is desirable for reduced patient exposure.
Unfortunately, these nuclides cannot simply be injected into the patient and expect imaging results. The positron emitter would simply diffuse throughout the body and the image would have no contrast. Instead the positron emitter is attached to a bio-active molecule that the body requires for metabolism. In this way the positron emitter accumulates in the body where it is desirable to generate an image.
The positron-emitter is therefore attached to a molecule that mimics blood sugar. As cells metabolize this blood sugar (and aggressively growing cancer cells will use a great deal more sugar than normal cells), the positron-emitting nuclide will accumulate there, emitting positrons and generating mutual annihilation gamma rays. This is how more gamma rays will be produced from tissue that is undergoing malignant growth than normal tissue.
After injection of the positron emitting tracer, the patient is placed inside a ring consisting of many, many gamma-ray detectors. Based on how the pairs of gamma rays are detected, an image can be constructed of where the positron emitter has accumulated. The pairing of the gamma rays helps to pinpoint exactly where the gamma rays originated. This is the purpose of the Coincidence Processing Unit in the diagram below - to ensure only matched pairs of 180 degree apart gamma rays are used to build the image.
Below, a diagram of how a PET scanner works (courtesy Wikipedia):
Below, an image created by a PET scan, showing where the pairs of positrons and electrons are giving off their opposed gamma rays.
Because PET scanning requires the use of such short-lived radio-isotopes, using this diagnostic tool presents some logistical problems. The positron-emitting nuclides are created in cyclotrons; large complex machines that hurl beams of relativistic electrons at target materials inside vacuum chambers. Most of these very short-lived isotopes (except for the Fluorine 18, half life 110 minutes) would decay before they could be transported to a PET scanner. For this reason, PET scanners are often located at universities where advanced nuclear and medical technologies exist side by side.
I find it fascinating how we have combined our understanding of nuclear decay, and the body's use of glucose for fuel to create an internal image that was undreamt of when the positron was first theoretically considered. We had to wait for someone to invent a bio-active chemical that would act as a carrier for the positron emitter, high efficiency detectors with coincidence counting, and powerful computers that could take all that data and generate an image from it. This confluence of diverse areas of science has led to an absolutely brilliant diagnostic tool!!!