A General Concepts of Radionuclide Imaging

The purpose of radionuclide imaging is to obtain a picture of the distribution of a radioactively labeled substance within the body after it has been administered (e.g., by intravenous injection) to a patient. This is accomplished by recording the emissions from the radioactivity with external radiation detectors placed at different locations outside the patient. The preferred emissions for this application are γ rays in the approximate energy range of 80 to 500 keV (or annihilation photons, 511 keV). Gamma rays of these energies are sufficiently penetrating in body tissues to be detected from deep-lying organs, can be stopped efficiently by dense scintillators, and are shielded adequately with reasonable thicknesses of lead (see Fig. 6-17—soft tissue has attenuation properties similar to water). Alpha particles and electrons (β particles, Auger and conversion electrons) are of little use for imaging because they cannot penetrate more than a few millimeters of tissue. Therefore they cannot escape from within the body and reach an external radiation detector, except from very superficial tissues. Bremsstrahlung (see Fig. 6-1) generated by electron emissions is more penetrating, but the intensity of this radiation generally is very weak.

Imaging system detectors must therefore have good detection efficiency for γ rays. It is also desirable that they have energy discrimination capability, so that γ rays that have lost positional information by Compton scattering within the body can be rejected based on their reduced energy (see Chapter 6, Section C.3). A sodium iodide [NaI(Tl)] scintillation detector (see Chapter 7, Section C) provides both of these features at a reasonable cost; for this reason it is currently the detector of choice for radionuclides with γ-ray emissions in the range of 80-300 keV.

The first attempts at radionuclide “imaging” occurred in the late 1940s. An array of radiation detectors was positioned on a matrix of measuring points around the head. Alternatively, a single detector was positioned manually for separate measurements at each point in the matrix. These devices were tedious to use and provided only very crude mappings of the distribution of radioactivity in the head (e.g., left-side versus right-side asymmetries).

A significant advance occurred in the early 1950s with the introduction of the rectilinear scanner by Benedict Cassen (see Fig. 1-2). With this instrument, the detector was scanned mechanically in a raster-like pattern over the area of interest. The image was a pattern of dots imprinted on a sheet of paper by a mechanical printer that followed the scanning motion of the detector, printing the dots as the γ rays were detected.

The principal disadvantage of the rectilinear scanner was its long imaging time (typically many minutes) because the image was formed by sequential measurements at many individual points within the imaged area. The first gamma-ray “camera” capable of recording at all points in the image at one time was described by Hal Anger in 1953. He used a pinhole aperture in a sheet of lead to project a γ-ray image of the radionuclide distribution onto a radiation detector composed of a NaI(Tl) screen and a sheet of x-ray film. The film was exposed by the scintillation light flashes generated by the γ rays in the NaI(Tl) screen. Unfortunately, this detection system (especially the film component) was so inefficient that hour-long exposures and therapeutic levels of administered radioactivity were needed to obtain satisfactory images.

In the late 1950s, Anger replaced the film-screen combination with a single, large-area, NaI(Tl) crystal and a photomultiplier (PM) tube assembly to greatly increase the detection efficiency of his “camera” concept. This instrument, the Anger scintillation camera,¹ or gamma camera, has been substantially refined and improved since that time. Although other ideas for nuclear-imaging instruments have come along since then, none, with the exception of modern positron emission tomography systems (see Chapter 18), has matched the gamma camera for a balance of image quality, detection efficiency, and ease of use in a hospital environment. The gamma camera has thus become the most widely used nuclear-imaging instrument for clinical applications.

B Basic Principles of the Gamma Camera

1 System Components

Figure 13-1 illustrates the basic principles of image formation with the gamma camera. The major components are a collimator, a large-area NaI(Tl) scintillation crystal, a light guide, and an array of PM tubes. Two features that differ from the conventional NaI(Tl) counting detectors described in Chapter 12 are crucial to image formation. The first is that an imaging collimator is used to define the direction of the detected γ rays. The collimator most commonly consists of a lead plate containing a large number of holes. By controlling which γ rays are accepted, the collimator forms a projected image of the γ-ray distribution on the surface of the NaI(Tl) crystal (see Section B.3). The second is that the NaI(Tl) crystal is viewed by an array of PM tubes, rather than a single PM tube. Signals from the PM tubes are fed to electronic or digital position logic circuits, which determine the X-Y location of each scintillation event, as it occurs, by using the weighted average of the PM tube signals (see Section B.2).

FIGURE 13-1 Basic principles and components of a modern gamma camera. The outputs of each photomultiplier (PM) tube are amplified and digitized using an analog-to-digital converter (ADC). The X-Y locations for each gamma ray that interacts in the NaI(Tl) crystal are computed from the digitized signals. The energy deposited by the gamma ray, E, which is proportional to the total measured pulse amplitude, also is computed by summing the individual PM tube signals. If E falls within the selected energy window, the event is accepted and placed at the appropriate X-Y location in the image.

Individual events also are analyzed for energy, E, by summing the signals from all PM tubes. When the pulse amplitude of an event falls within the selected energy window, it is accepted and the X and Y values are binned into a discrete two-dimensional array of image elements, or pixels. An image is formed from a histogram of the number of events at each possible X-Y location. Large numbers of events are required to form an interpretable image because each pixel must have a sufficient number of counts to achieve an acceptable signal-to-noise level. Because images often are formed in 64- × 64-pixel or 128- × 128-pixel arrays, the counting requirements are some 10³ to 10⁴ times higher than for a simple counting detector.

Images are displayed on a computer monitor, where image brightness and contrast may be manipulated and different color tables may be employed. More sophisticated digital image processing is discussed in Chapter 20.

Most modern gamma cameras are completely digital, in the sense that the output of each PM tube is directly digitized by an analog-to-digital converter (ADC). The calculation of X-Y position and pulse-height are performed in software based on the digitized PM tube signals, and errors in energy and positioning caused by noise and pulse distortions caused by the analog positioning circuitry are eliminated. This approach also permits improved handling of pulse pile-up at high counting rates, as described in Section B.2.

The gamma camera can be used for static imaging studies, in which an image of an unchanging radionuclide distribution can be recorded over an extended imaging time (e.g., minutes). Single contiguous images of the whole body can be obtained by scanning the gamma camera across the entire length of the patient. This can be achieved by moving either the bed or the gamma camera while adjusting the event positioning computation to account for this movement. Clinically important whole-body studies include bone scans of the skeleton, and the localization of tumors or their metastases in the body.

The gamma camera also can be used for dynamic imaging studies, in which changes in the radionuclide distribution can be observed, as rapidly as several images per second. This allows physiologic information to be obtained, such as the rate of tracer uptake or clearance from an organ of interest. Images also can be synchronized to electrocardiogram signals, permitting images of the heart in different phases of the cardiac cycle to be formed. These gated images can provide important information on cardiac function.

2 Detector System and Electronics

The gamma camera employs a single, large-area, rectangular NaI(Tl) detector crystal, usually 6- to 12.5-mm thick with sizes of up to 60 × 40 cm. Round crystals of 25 to 50 cm in diameter were used in many older systems. The NaI(Tl) crystal is surrounded by a highly reflective material such as TiO₂ to maximize light output and hermetically sealed inside a thin aluminum casing to protect it from moisture. An optical glass window on the back surface of the casing permits the scintillation light to reach the PM tubes. A cross section of a typical gamma camera crystal assembly is shown in Figure 13-2. The choice of thickness of the NaI(Tl) crystal is a trade-off between its detection efficiency (which increases with increasing thickness) and, as shown in Chapter 14, Section A.1, its intrinsic spatial resolution (which deteriorates with increasing thickness). Most general-purpose gamma cameras have crystal thicknesses of approximately 9.5 mm. For lower-energy γ emitters, such as ^99mTc and ²⁰¹Tl, however, detection efficiency is adequate even with 6-mm-thick detector crystals.

FIGURE 13-2 Schematic cross-section of a NaI(Tl) crystal assembly for a gamma camera.

An array of PM tubes is coupled optically to the back face of the crystal with a silicone-based adhesive or grease. Round PM tubes are arranged in a hexagonal pattern to maximize the area of the NaI(Tl) crystal that is covered. Some cameras use hexagonal (or rarely, square) cross-section PM tubes for better coverage of the NaI(Tl) crystal. Typical PM tube sizes are 5 cm in diameter. Most modern cameras employ between 30 and 100 PM tubes. Figure 13-3 shows a photograph of a 30-tube model. The PM tubes are encased in a thin magnetic shield (Chapter 7, Section C.2) to prevent changes in the gain caused by changes in the orientation of the gamma camera relative to the earth’s magnetic field. The ultrasensitivity of PM tubes to magnetic fields also makes gamma cameras susceptible to the stray fields from magnetic resonance imaging systems.

FIGURE 13-3 A rectangular gamma camera detector with the cover removed showing the photomultiplier (PM) tubes mounted on the NaI(Tl) crystal. In this example, the gamma camera detector measures 50 × 15 cm and is read out by 30 PM tubes 5 cm in diameter. This is a digital camera in which each of the PM tube outputs is individually digitized.

(Courtesy Dr. Joel Karp, University of Pennsylvania, Philadelphia, PA.)

Many manufacturers employ plastic light guides between the detector crystal and PM tubes, whereas others couple the PM tubes directly to the crystal. The functions of the light guide are to increase the light collection efficiency, by channeling scintillation light away from the gaps between the PM tubes, and to improve the uniformity of light collection as a function of position. The latter effect is achieved by painting or etching a carefully designed pattern onto the entrance face of the light guide. The use of the PM tubes with hexagonal or square cross-sections that can be tiled without gaps on the NaI(Tl) crystal may in some cases allow elimination of the light guide, assuming there is sufficient spreading of the scintillation light in the glass entrance window of the PM tube for accurate positioning.

The detector crystal and PM tube array are enclosed in a light-tight, lead-lined protective housing. In most modern cameras, most of the electronics (such as preamplifiers, pulse-height analyzers, automatic gain control, pulse pile-up rejection circuits and ADCs) are mounted directly on the individual PM tube bases within the detector housing to minimize signal distortions that can occur in long cable runs between the detector head and control console.

The amount of light detected by a particular PM tube is inversely related to the lateral distance between the interaction site and the center of that PM tube. This is illustrated in one dimension in Figure 13-4. Ideally, the relationship between signal amplitude and location with respect to the center of a PM tube would be linear. This would enable the position of an event to be determined by taking a weighted average or centroid of the PM tube signals using the simple relationships shown in Figure 13-4. In practice, however, the response is more complex, with a plateau directly beneath the PM tube (because the PM tube is not a “point” detector) and long, flat tails caused by reflections of light from the back and side surfaces of the NaI(Tl) crystal. Therefore a calibration for spatial nonlinearity is required (see Chapter 14, Section B).