In order to obtain spatial information about the gamma-ray emissions from an imaging subject (e.g. a person's heart muscle cells which have absorbed an intravenous injected radioactive, usually thallium-201 or
technetium-99m, medicinal imaging agent) a method of correlating the detected photons with their point of origin is required. The conventional method is to place a
collimator over the detection crystal/PMT array. The collimator consists of a thick sheet of
lead, typically thick, with thousands of adjacent holes through it. There are three types of collimators: low energy, medium energy, and high energy collimators. As the collimators transitioned from low energy to high energy, the hole sizes, thickness, and septations between the holes also increased. Given a fixed septal thickness, the collimator resolution decreases with increased efficiency and also increasing distance of the source from the collimator. The individual holes limit photons which can be detected by the crystal to a cone shape; the point of the cone is at the midline center of any given hole and extends from the collimator surface outward. However, the collimator is also one of the sources of blurring within the image; lead does not totally attenuate incident gamma photons, there can be some
crosstalk between holes. Unlike a lens, as used in visible light cameras, the collimator attenuates most (>99%) of incident photons and thus greatly limits the sensitivity of the camera system. Large amounts of radiation must be present so as to provide enough exposure for the camera system to detect sufficient scintillation dots to form a picture. however, none have entered widespread routine clinical use. The best current camera system designs can differentiate two separate point sources of gamma photons located at 6 to 12 mm depending on distance from the collimator, the type of collimator and radio-nucleide. Spatial resolution decreases rapidly at increasing distances from the camera face. This limits the spatial accuracy of the computer image: it is a fuzzy image made up of many dots of detected but not precisely located scintillation. This is a major limitation for heart muscle imaging systems; the thickest normal heart muscle in the left ventricle is about 1.2 cm and most of the left ventricle muscle is about 0.8 cm, always moving and much of it beyond 5 cm from the collimator face. To help compensate, better imaging systems limit scintillation counting to a portion of the heart contraction cycle, called gating, however this further limits system sensitivity. ==See also==