Chapter 37 Nuclear medicine imaging

David Wyn Jones, Julian MacDonald, Peter Hogg

Introduction

Nuclear medicine has three distinct practice areas: in-vitro laboratory-based diagnostics; unsealed source radionuclide therapy; and diagnostic radionuclide imaging. In all three areas the power of nuclear medicine is its ability to diagnose and/or treat disease at a physiological or molecular level.

In-vitro nuclear medicine is often performed remote from the imaging unit, in laboratories. Radioactive substances are used on human tissue and/or fluid samples to diagnose a wide range of pathologies. Examples of diagnostic tests include: analysis of renal function (glomerular filtration rate) using radiolabelled chromium (⁵¹Cr); assessment of vitamin B₁₂ absorption using radiolabelled cobalt (⁵⁷Co); and evaluation of thyroid function using radio-labelled iodine (¹²⁵I). Surprisingly, in-vitro nuclear medicine is performed at many more hospitals than is nuclear medicine imaging.

Unsealed source therapy is used to treat and/or palliate benign and malignant disease. The intention is to deliver an appropriate radiation dose to the offending tissue in order to cause cell death. Consequently, radioactive substances that emit particles (notably beta) are commonly used. Radionuclide therapy for malignant disease is generally performed in oncology units, not least because the radiation protection restrictions are stringent and expensive to implement. A wide range of malignant diseases can be treated in this fashion, an example being thyroid cancer using iodine (¹³¹I). The most common benign disease to be treated is thyrotoxicosis, again using ¹³¹I, but at a much lower dosage.

When radioactive substances are administered to patients, whether for diagnostic or therapeutic purposes, they are collectively referred to as radiopharmaceuticals. For diagnostic imaging these radiopharmaceuticals provide a way of visualising patterns of growth and biological activity in the organs of interest. This is achieved by imaging the distribution of radiopharmaceuticals which are selected based on their ability to be taken up in the area of interest. Abnormalities, trauma, or the effects of pathogenic invasion can be identified. The great advantage of nuclear medicine imaging is that, except in the case of trauma, physiological changes usually precede anatomical changes.¹

The modality is highly reliant on the skills of a multidisciplinary team; a suspected clinical condition needs to be matched to an appropriate nuclear medicine investigation, which usually involves complex medical and scientific decisions. The most obvious decision is whether to proceed, in terms of net benefit from the radiation dose received, and if so, which radiopharmaceutical and which imaging technique to use.

Equipment chronology

1896 Henri Becquerel discovers radioactivity.

1930s Cyclotron invented: providing means to produce usable quantities of radionuclides. Technetium-99m (^99mTc) first produced in the late 1930s.

1940s Radionuclides become available for medical use.

Early 1950s Cassen et al. produced a scintillation detector mounted in an automatic scanning gantry, which was probably the first incarnation of the rectilinear scanner.²

1953 First study involving the imaging positron emitters published.³ As cyclotrons became more available, development accelerated due to the availability of positron-emitting radionuclides that could be labelled as clinically useful molecules. Even so, positron emission tomography (PET) remained only a research tool until the late 1970s.

Late 1950s Commercial machines available: these devices allowed the acquisition of an image by tracing a collimated scintillation detector in a rectilinear pattern over the area of interest. Rectilinear scanners, however, were very slow and could not produce images of dynamic processes.

Hal Anger developed a scintillation detector, which has since become known as the gamma camera.⁴ This device is kept stationary and collects gamma rays over the field of view, resulting in much more rapid image acquisition than the rectilinear scanner and allowing dynamic imaging.

1960 Developed at the Brookhaven National laboratory the ⁹⁹Mo/^99mTc generator became commercially available. One of the earliest reported uses of ^99mTc was for brain scanning.⁵

1963 First single photon emission tomography study published.⁶

This technique acquires data at a series of angular positions around the patient allowing the production of multiplanar images.

By 1964 Commercial Anger gamma camera systems available.

1967 Hounsfield develops computer algorithms for image production. These algorithms accounted for attenuation and scatter and converted the emission tomography technique to single photon emission computed tomography (SPECT). At this time reconstruction of data took several hours; however, owing to advances in computing, the same processes today take a few seconds.

1970s Radiopharmaceuticals developed allowing imaging of most organs in the body.

mid 1970s Rotating gantries developed to allow automatic SPECT acquisitions.

Late 1970s Clinical PET systems started to become commercially available.

1980s Cardiac radiopharmaceuticals became available.

Since the 1980s, systems for planar, dynamic and SPECT acquisition, have been commercially available and have been further developed and refined.

1990s Rectangular camera heads replaced circular ones to allow imaging of greater areas.

Late 1990s PET started to become routinely used as a clinical tool in the USA. A proliferation of literature started to appear to indicate that PET had a value in the diagnosis and management of certain malignant conditions, and it was not long before it was realised that PET imaging was an essential component in the management of certain cancers. The American healthcare economy then drove the PET market, and as a consequence PET scanning systems and cyclotrons became more available and at a lower cost. The increased clinical use of PET encouraged more research to be conducted into its potential applications and presently a large number of dedicated PET centres exist purely for research purposes.

Science and instrumentation

Radioactivity

The atoms of some substances are unstable owing to an imbalance in the number of protons and neutrons in the nucleus. Such substances emit radiation spontaneously and are said to be radioactive. Radiation is emitted from a radioactive atom when it undergoes disintegration, i.e. a transformation or decay to another atom. The radioactivity of a substance is defined as the number of disintegrations per second. The unit of radioactivity is the Becquerel (Bq): 1 Bq = 1 disintegration per second.

As atoms decay over time, the amount of radioactivity of a substance reduces. The time taken for the radioactivity to reduce to a half of its original value is called the half-life of the radionuclide. The radiation emitted by a radioactive substance can be of several types, e.g. alpha particles, beta particles, gamma rays, positrons. For imaging purposes, only penetrating radiation, such as gamma radiation, is of use.

Radionuclides

Table 37.1 gives the ideal radionuclide requirements for use in nuclear medicine imaging.

Table 37.1 Ideal radionuclide requirements for use in nuclear medicine imaging

Property	Ideal requirements
Radiation emitted	Detection relies on radiation being emitted from the body and thus requires a penetrating form of radiation, i.e. gamma rays
Energy	The gamma rays must possess sufficient energy to escape the body but, conversely, their energy must be low enough to allow them to be efficiently stopped within the detector
Half-life	The radioactivity must be sufficient to allow good image quality throughout the duration of the imaging period. The half-life must therefore be long enough to allow this. Conversely, if the half-life is much longer than the period of imaging, this may result in a higher exposure to the patient than is necessary
Cost and availability	The ideal radiopharmaceutical will be cheap and readily available

The physical characteristics of ^99mTc (half-life and gamma-ray emission) are ideal for gamma camera imaging in humans, not only because the gamma-ray energy is well suited to gamma camera detection, but also there is no particulate emission (reducing potential patient dose). Although its short half-life (approximately 6 hours) could be considered self-limiting in terms of geographical availability, the invention of the molybdenum (⁹⁹Mo) generator allowed for a ready supply of ^99mTc in most hospital locations. ⁹⁹Mo decays to ^99mTc, and using the generator principle this process can be capitalised on through daily or twice-daily elution (elution is a method of removing ^99mTc from ⁹⁹Mo in a sterile solution).

^99mTc can be chemically bound to an extensive range of non-radioactive chemical compounds, which can remain chemically stable for quite long times after introduction into the patient, allowing imaging to take place. Examples of uses of ^99mTc include phosphate labelled to ^99mTc, which permits bone imaging, and ^99mTc labelled to a chelate (e.g. diethylenediaminetetra-acetic acid) for renal imaging. Other commonly used radionuclides, and their uses in nuclear medicine imaging, are given in Table 37.2.

Table 37.2 Other commonly used radionuclides

Radiopharmaceuticals for PET comprise radionuclides that emit positrons. The positrons lose energy in a short distance in the body and annihilate with atomic electrons to produce two 511 keV gamma-ray photons (180° apart) that allow coincidental detection of the tracer. PET radionuclides have to be produced by cyclotron. If the cyclotron is offsite, the half-life of the radionuclide must be sufficiently long to allow it to be transported to the imaging centre while enough activity remains. The most commonly used PET radionuclide is fluorine-18 (¹⁸F), which has a half-life of 1.8 hours.

Chemical component

The chemical component attached to the radionuclide determines where the radiopharmaceutical travels in the body. There are several ways in which a desirable distribution can be achieved, including:

• Using a chemical found physiologically in the organ of interest, e.g. iodine for imaging the thyroid.

• Using an analogue. This is a chemical that simulates one found physiologically, for instance thallium is a potassium analogue and thus can be used to image muscle. Similarly, fluorodeoxyglucose (FDG) is a glucose analogue and can be labelled with ¹⁸F for PET imaging to illustrate areas of high glucose metabolism, which has several clinical uses.

• Labelling cells that fight disease, thereby targeting the areas of disease, e.g. white blood cells or antibodies.

Radiopharmaceuticals are normally administered intravenously but are occasionally given subcutaneously, orally or via inhalation. Once incorporated, the radiopharmaceutical remains in the body for a period determined by the chemical form, the half-life of the radionuclide and the physiology of the patient. The patient will receive a radiation dose that will depend on the radioactivity administered and the residence time (i.e. the time during which the radionuclide is present in the body). The effective dose, which allows comparison with other imaging modalities using ionising radiation, is determined from the weighted sum of the absorbed doses to each organ. The weighting factors are organ dependent owing to their different radiosensitivities.

The gamma camera

Until the introduction of the gamma camera, imaging was performed on rectilinear scanners using a limited range of radiopharmaceuticals. These scanners tended to produce poor-quality low-resolution images. The gamma camera changed this, resulting in massively improved image quality, thereby increasing the diagnostic value of this modality. The basic principles of operation of the gamma camera have remained largely unchanged from its inception until today, and it continues to be used extensively.

The most fundamental part of any imaging system is the detector. In the case of a gamma camera the detector is a large crystal, normally rectangular, of a scintillation material that produces a weak flash of light when radiation is absorbed, due to excitation. Flashes of light are produced when gamma rays emitted by a patient, previously administered with a radiopharmaceutical, fall on the detector (Fig. 37.1).

Figure 37.1 Schematic diagram of a gamma camera showing the gamma rays emitted by the patient, the collimator allowing only those aligned with the collimator holes to pass through to the scintillation crystal. Light rays produced by the crystal are detected and quantified to obtain energy and positional information used to assign a count to the correct pixel location in the image matrix.

The light formed in the crystal is detected by photomultiplier tubes (PMTs) which convert the light to electronic signals whose magnitude is determined by the intensity of light reaching the PMT. There are, in fact, many PMTs packed into the space of the scintillator crystal, and those around the point of light emission will detect some amount of light, depending on their distance from that point. Those closest will detect more light and, in turn, produce a greater electronic pulse, whereas those further away will produce proportionally smaller pulses. The relative magnitude of these pulses can then be used to determine the point of light emission. The pixel count value in a corresponding location in a digital matrix can then be allocated, allowing the accumulation of an image (Fig. 37.2).

Figure 37.2 The intensity of light detected depends on the distance of the PMT from the position within the crystal where the gamma ray was absorbed and the light flash occurred. These relative intensities are used to reduce a weighted average of the PMT positions (Xa, Ya), (Xb, Yb) etc. to determine the origin of the light flash.

However, the system so far described does not provide a method of tracing the point of light emission in the detector back to the point of origin of the gamma ray within the patient, which is critical to producing a meaningful image. This is the function of the collimator.

A collimator is essentially a block of attenuating material with a network of holes and is attached to the gamma camera between the detector crystal and the patient. The holes allow gamma rays travelling in a certain direction to pass through to the crystal and be detected. Gamma rays travelling in different directions are attenuated, and so the collimator effectively acts as a filter, allowing only gamma rays travelling in a known direction to contribute to the image. In this way there is a direct one-to-one mapping between the origin of the gamma ray and the position of the pixel within the image matrix. A flow diagram illustrating the process of nuclear medicine image formation is shown in Figure 37.3.

Figure 37.3 The imaging process.

The scintillation material used is sodium iodide, which contains a small amount of thallium impurity [NaI(Tl)]. The thallium impurity significantly increases the amount of light produced. The properties of the detector affect many aspects of the image, as detailed below:

• Thickness. The thicker the crystal, the more likely the radiation will be absorbed, thereby improving efficiency. However, if the crystal is too thick there will be an increase in the drift of the light photons, adding to the uncertainty in the calculated point of interaction and hence a worsening in the spatial resolution, i.e. blurring.

• Light output. Different scintillator materials produce different amounts of light per interaction and release the light at different rates. The latter affects the duration of the processing required to assign the event to a position within the image matrix, a period known as the dead-time. Ideally a large amount of light is required in a short time. It is also essential that the light intensity produced is proportional to the energy of the radiation absorbed. This allows energy discrimination and hence the ability to disregard gamma rays outside a certain energy range, which is used to reduce scatter.

• Transparency. This factor affects the amount of light lost as it passes through the crystal before being detected by the PMTs. The more light photons that pass through the crystal and are detected by the PMTs, the more accurate the calculated point of interaction will be.

Collimators vary with regard to their thickness and the number, direction and diameter of the holes. The most common type of collimator used is the parallel hole collimator, in which the holes, as well as being parallel with each other, are perpendicular to the camera face. The number of holes, and hence the thickness of the attenuating material between them, known as the septa, is altered to allow imaging for different energies. For example when imaging ¹³¹I gamma rays of 364 keV, a collimator with fewer holes and correspondingly thicker septa is essential to prevent penetration of the radiation through the septa. High-resolution high-sensitivity collimators are also generally available and feature variations in collimator thickness and hence hole length. There is a trade-off between resolution and sensitivity such that a high-resolution collimator will have lower sensitivity and therefore take longer to obtain the same number of counts than a high-sensitivity collimator, and vice versa.

Other geometric arrangements of holes are also available. The holes of a diverging hole collimator fan outwards and allow demagnification of the object, which is useful for large objects. Converging hole or fan beam collimators fan inwards, providing magnification of the object, and are commonly used in brain imaging. Pinhole collimators have a single aperture at the end of a lead shield that allows magnification of objects near the collimator and demagnification further away. The magnitude of the effect of all these collimators depends on the distance from the collimator, and hence distortions in the image will occur. This is particularly the case with the pinhole collimator, which can really only be used with thin objects. Nonetheless, the pinhole collimator is a very useful way of providing magnified images of, for example, the thyroid gland or small bone joints in children.

Multiheaded gamma cameras

Gamma cameras can be purchased with one, two or three detector heads, each consisting of scintillation crystal, collimator and associated electronics. In most common use are the dual-headed systems, and many of the available systems can offer flexibility in the position and orientation of the heads. The advantage of multihead gamma cameras is basically that of speed. Dual-headed cameras can be used for whole-body scanning systems: one head can acquire data anteriorly and the other posteriorly simultaneously. For SPECT imaging, multiheaded cameras allow each head to acquire data from part of the complete revolution, thereby speeding up acquisition by a factor equal to the number of detector heads. Three-headed systems are most commonly used for dedicated brain SPECT imaging.

Single photon emission computed tomography (SPECT)

The majority of nuclear medicine studies require the acquisition of static or dynamic planar images, which are acquired with the gamma camera in a fixed position against the patient for the duration of imaging. SPECT imaging, on the other hand, acquires a series of images as the gamma camera rotates around the patient. These images, or projections, can then be mathematically reconstructed to form a 3D dataset from which slices through the body or 3D visualisations can be formed, in a similar way to X-ray CT. Reconstruction techniques are beyond the scope of this chapter, but are based on back-projection or iterative algorithms.

This imaging technique allows much greater contrast owing to the effective removal of overlying structures present in planar imaging. Some common applications are in assessing myocardial perfusion, brain functionality and bone lesions.

SPECT-CT systems

It is becoming increasingly common for gamma cameras and CT scanners to be housed on the same system to give what are termed SPECT-CT systems. The CT scanners for this purpose range from low-dose non-diagnostic CT to fully fledged multislice diagnostic CT systems, depending on their intended use.

The CT component provides two advantages:

1 Attenuation correction. Gamma rays emitted from within the patient are attenuated by various anatomical structures before they leave the patient and are detected by the gamma camera. The amount of attenuation varies depending on the path the gamma ray travels along from its point of origin, i.e. which anatomical structures the rays have to pass through, and this will vary with the orientation of the camera during a SPECT acquisition.

The number of gamma rays detected may not actually represent the distribution of the radiopharmaceutical in the body: for example, during myocardial perfusion imaging on large-breasted female patients there is more attenuation from the front than from the side, and this gives rise to artificially low counts in the anterior wall of the heart when the data are reconstructed.

To obtain an accurate image of the actual distribution, it is necessary to know the attenuation of the various anatomical structures so that the attenuation differences at different angles can be corrected for. This is achieved by acquiring a transmission image via CT. If X-rays of similar energy as the gamma rays are used, the resulting CT image will effectively be an attenuation map and can be used as a correction in the reconstruction process. The resolution of the CT images for this purpose need not be particularly high, and it is therefore possible to use a low-dose CT protocol. Some systems acquire the CT over a relatively long time period compared to conventional CT examinations, which are normally performed during a breath-hold. This has the advantage that the CT images are more consistent with the relatively long SPECT acquisition, making the attenuation map a better match than one obtained from a breath-hold.

2 Image fusion. The limited spatial resolution of nuclear medicine imaging, together with the efficient targeting of some radiopharmaceuticals, can result in specific uptake which is difficult to localise. On the other hand, a CT image provides good anatomical detail without the functional information. By overlaying the nuclear medicine SPECT images onto the corresponding CT image, the best of both modalities can be obtained. It is possible, for example, to see exactly which bone is affected in infection or trauma of complex areas such as the hand or foot. Hybrid systems are calibrated such that the CT and SPECT images can be accurately and consistently co-registered, and this must be checked as part of routine quality control.

Dedicated specialised systems

As well as three-headed gamma cameras for dedicated brain SPECT imaging, other dedicated systems have been developed, particularly for myocardial perfusion imaging. These include systems based on traditional gamma camera technology but with fixed head positions and smaller fields of view, with patient chairs designed to maximise comfort and minimise movement. Other systems are based on different detector technology, i.e. solid-state detectors such as cadmium zinc telluride (CZT); a number of individual CZT detectors may be used to scan the heart volume, producing similar resolution images in a significantly reduced time compared to conventional SPECT imaging.

Positron emission tomography (PET)

PET scanning is gaining popularity in the UK and is becoming part of a routine nuclear medicine service. PET imaging is desirable because positron emitting radionuclides are relatively simple to label to biologically active organic molecules. By far the most common PET radiopharmaceutical is ¹⁸F-FDG, which provides an image of glucose metabolism that is useful in oncology, cardiology and neurology.

The basic principle of PET imaging is shown in Figure 37.4A,B. PET tracers emit positrons that annihilate with electrons to form two 511 keV photons which are emitted in opposite directions. PET scanners consist of a ring of detectors in which the two photons are detected coincidentally. The point of emission in the patient must then be somewhere along a line between the two detection events. When sufficient coincident events have been accumulated, the distribution in the body is indicated by a superimposition of these lines. Reconstruction of the data produces a 3D dataset, which can be used to obtain slices through the area of interest.