Positron Emission Tomography (PET) is a modern nuclear medicine imaging technique-using tracers that decay with the emission of positrons. PET is a safe noninvasive visualization technique. It provides serial quantitative imaging of the spatial distribution of any compound labeled with a positron emitting radionuclide in transverse sections of the body. PET scanner does not detect positrons directly but uses important features of positron annihilation to determine their spatial location by using a coincidence detection system. PET distinguishes from other medical tomographic imaging techniques in its ability to provide quantitative information about tissue function in vivo.

Positron & Coincidence detection of Annihilation Radiation: PET scanners are based on the principle that the annihilation of one positron generates two photons opposite to each other at the same time. Coincidence detection is then used to detect these annihilation events. The positron is an anti-electron that, after the emission by a radionuclide and traveling a short distance, will combine with an electron from the surroundings and annihilate. After loosing most of its kinetic energy, a positron interacts with an electron (forming positronium with a brief existence) in such a fashion that the masses of two particles are converted into electromagnetic radiation in the form of two photons, in order to conserve energy and linear momentum, each carrying an energy of approximately 511 Kev and travelling nearly co-linearly in opposite directions. This process is called annihilation, and the two photons are called annihilation radiation. In PET, the positrons are detected in vivo by the detection of the annihilation radiation externally coincidentally. Measurement of annihilation radiation is based on the fact that the photon to be detected transfers its energy by accelerating electrons within the detector. For most PET applications, the detection of the annihilation radiation is based on the use of scintillation counters.

Coincidence Events: Coincidence events are classified into several classes 1.True coincidences, 2. Random coincidences, 3. Scatter coincidences 4. Multiple coincidences.

True coincidences: Ideally, the only coincidence events recorded by a positron tomograph are those which arise from unique positron annihilation that occurred along the line between two activated detectors. These are referred to as true coincidences. They carry useful information about the location of the positron emitter.

Random Coincidences: Because of the finite width of the coincidence time window there is a possibility that two unrelated photons will be detected and registered as a coincidences. These unrelated events are referred to a random coincidences. They do not carry any spatial information about the activity distribution. In most modern examinations with PET, the contribution of random coincidences is a major component of noise.

1. Operation of the coincidence systems at relatively low counting rates
2. Reduction of the coincidence resolving time
3. Subtraction of the contribution of random coincidences obtained by the measurement of the single rates in each of the detectors operated in coincidence, or by direct measurement of the contribution of random counts to the detected signal.

Scatter coincidences: Scatter coincidences are those in which one or both of the annihilation photons are diverted from their original path before reaching the detector as a result of compton interactions with electrons in surrounding tissues.

Radiation Detectors in PET: PET images represent the distribution of positron-emitting radionuclides. A necessary component of any PET device is detectors sensitive to annihilation radiation. Some of the most important physical characteristics of PET detectors are 1) high efficiency for the detection of annihilation photons, 2) short coincidence resolving time, and 3) ability to provide high spatial resolution. Three types of detectors have been deemed, by various investigators, worthy of incorporation into PET devices:

1. Scintillation detectors - NaI (Tl), BGO, CsF.
2. Multiwire proportional counters
3. Semiconductor Photosensitive detectors optically coupled to scintillating crystals.

Depending on the arrangement of the detectors and the kind of detectors used, there are various scanner configurations. For example, there are PET scanners with rotational heads, with polygonal detector rings, circular rings, etc.

Configuration of a PET System: The configuration of a PET system consists of

1. Data acquisition system- radiation detectors & the gantry
2. The electronic components- that record and process the data
3. Image display system

Reconstruction of Image in PET: The transverse tomographic image is reconstructed from a series of measurements taken from different angles around the object to be imaged. The mathematical algorithms used for the reconstruction is Filtered Back Projection method.

The PET Image: The quality of PET image can be expressed by three physical characteristics: 1) spatial resolution, 2) contrast resolution, 3) temporal resolution. Good image resolution not only improves the visualization of small structures, it is also critically important for providing accurate quantitative measures of radioactivity concentrations in PET. Image resolution in PET is related to the positron range and angulation of the positron emitting isotope, the detector size, the reconstruction filter used, the amount of scattered radiation and the noise level of the measurements, are all important factors for PET scanner performance.

Operational Guidelines for PET: The data obtained from PET scanners range from the purely qualitative, where images are examined visually, to the quantitative, where it is desired that every pixel value in the image correspond numerically to the amount of radioactivity in the region of the body being examined. The accuracy and reliability of image data from a PET scanner in specific scanning situations can be predicted through measurements of its various performance characteristics. Aspects of PET scanning that affect the acquisition and analysis of image data include - spatial resolution, noise effects, count rate performance, scatter and its correction, attenuation correction and 3-D imaging.

Spatial Resolution: The spatial resolution can be defined in terms of the amount by which a system smears out the image of a point source of radioactivity. The spatial resolution of the scanner determines how small a structure one can visualize. Structures that are smaller than the spatial resolution of the scanner in any direction can not be perfectly distinguished from adjacent structures and will be blurred by the activity from nearby areas. Spatial resolution is considered in transverse and axial directions. The transverse spatial resolution is characterized by the spread of a point or line source in the directions along a line from the center of the gantry to the source position (radial) and perpendicular to that line (tangential). A profile of the counts measured along a line through the point source image is called the Point spread function (PSF). On a line perpendicular to the line source it is called the Line Spread Function (LSF). Passing a narrow line source between a coincidence detectors will give a count profile as a function of position (LSF). The LSF of the individual coincidence detector pairs is a common measure of spatial accuracy and also defines the limit of spatial resolution for the final PET image. Resolution can be expressed as FWHM of the point or line spread function measurements. Therefore, resolution is the minimum distance by which two point sources must be separated in order to be distinguished as separate sources in an image. A count profile through the source is approximately Gaussian in shape. The width of this curve at one-half its maximum value (FWHM) is most commonly used to characterize the resolution in the radial and tangential directions.

LSF measurements taken over the central third of the region between coincidence detectors are remarkably constant for both cylindrical & rectangular detectors shapes, and that region constitutes the most useful FOV for accurate measurements in PET. The pixel size in the reconstructed image has no effect on the spatial resolution of the scanners. However, the pixel size should be no more than one third of the FWHM. The ability of annihilation coincidence detection (ACD) to localize is restricted to the direction orthogonal to the line connecting the two detectors, and there is no localization along the line connecting the detectors.

Axial Resolution: The terms axial resolution, slice thickness, slice profile width, and slice separation are used to characterize the ability of the scanners to discriminate between structures in the out-of-line (across plane) direction. Axial resolution refers to the width in the axial direction of a stationary point source and is analogous to transverse resolution. Slice thickness and slice profile width refer to the width of the axial response function obtained when a point source of activity is moved through the scanner axially in fine steps. Slice separation is the distance centers of adjacent slices. Some investigators have proposed looking at the FWHM. Fourier transform of axial response function or modulation transfer or Equivalent Width (EW) in addition to citing the more familiar FWHM.

Sampling: Sampling refers to the spacing of the coincidence line across the fields of view.

Energy Resolution: Energy resolution is defined as the FWHM of the detector response to photons of the same energy.

Linearity: Linearity is the ability of the camera to produce a linear image of a linear object.

Sensitivity: The sensitivity of a scanner is defined as the count rate measured per unit activity concentration in a given distribution of radioactivity. Sensitivity can be also been expressed as the ratio of counts observed divided by nuclear disintegration. Sensitivity is influenced by a number of parameters including (1) detector material, (2) detector ring diameter, (3) axial acceptance window, (4) the inter-ring collimators.

Scatter and Attenuation: Scatter and attenuation of gamma rays in the body of a patient are important sources of error in quantification and of image distortion in both PET & SPECT. Scatter can amount to between 10 % and 60 % of the total number of events being detected in a given study. Attenuation reduces the number of counts coming from the center of the body by a factor of 10 for SPECT and 20 for PET (4-5 in case of brain also). Therefore, the images do not represent the true distribution of radiotracers in the patient unless they are corrected for attenuation and scatter. If the energy lost by scattering is greater than the resolution width of the detection system, the scatter event can be rejected by energy discrimination by Pulse Height Analysis. The amount of energy lost in the process of scattering depends on the angle of scatter, and is also a function of the gamma ray energy. A number of methods can be used to compensate for scatter coincidences in PET (e.g. Convolution & De-convolution method of scatter correction).

1. Calculated attenuation correction method.
2. Measured attenuation correction method
3. Hybrid attenuation correction method.

1. Preposition
2. Postposition
3. Iterative reconstruction
4. Chang's Multiplicative method
5. Chang's Reprojection method.

Partial-Volume Effect: The loss of accuracy in quantification of the activity in small structures as a result of limitations in spatial resolution is called the partial-volume effect. PVE is greater for small objects. Small objects near the resolution limits of the device appear to contain smaller concentrations of radioactivity than they actually do. The ratio of the apparent concentration to true concentration is called the recovery coefficient (RC). The RC for a three-dimensional object is a product of the RCs in each dimension. In principle, if an ECT system has a known and uniform spatial resolution and if the size of the object is known, a "recovery coefficient correction factor" can be applied to correct for the partial-volume effect for small objects.

Time-of-Flight PET: The PET system designed with fast electronic circuitry capable of measuring the time of arrival of annihilation photons between two coincidence detectors is referred to as TOF-PET.