radionuclide imaging Flashcards
What are the main differences between Radionuclide imaging from radiography with x-rays?
It utilises gamma radiation from a radioactive substance
The radioactive substance (in the form of a radiopharmaceutical) is administered to the patient (i.e. it is introduced into the patient’s body) so that the patient becomes the source of radiation
Images are created using equipment known as a gamma camera
Image contrast is largely determined by organ uptake (which depends on physiological function) rather than physical factors such as density and average atomic number
The radiation dose to the patient is largely determined by the nature and quantity (activity) of the radiopharmaceutical
What does contrast in radionuclide imaging depend on?
Contrast in radionuclide images depends primarily on the difference in the amount of radioactivity (activity) that has localised in organs and tissues. Thus, it depends mainly on physiological function rather than physical factors, such as density and average atomic number, which determine contrast in x-radiography.
What is the most common interaction in soft tissue with gamma radiation?
In soft tissue, Compton scattering is by far the most likely interaction mechanism, although photoelectric absorption and coherent scattering are also possible.
How is HVL calculated?
ln2/LAC
How much energy dose a scattered photon of 99Tc have?
In the case of 99mTc, photons that have suffered only one Compton interaction have a range of energies from about 90 keV (corresponding to 180° or back-scattering) to nearly 140 keV (corresponding to low-angle forward scattering).
Photons that have been Compton scattered many times within the patient (multiply scattered) may have very low energies.
What are the main components of a gamma camera?
The main components of a gamma camera are:
A collimator and a large-area radiation detector (both contained within a shielded gamma camera head)
Electronics for radiation detection and signal processing
A computer (for control, further signal processing and image acquisition)
Facilities for image storage and display
What proportion of the radiation from a patient reches the detector?
typically 0.01-0.1%
What is the problem of scattered radiation in radionuclide imaging?
If scattered photons were allowed to contribute significantly to the radionuclide image, this would have an adverse effect on contrast and spatial resolution because the point at which a photon was last scattered is not the photon’s point of origin. The collimator cannot distinguish between scattered and unscattered photons. Fortunately, scattered photons have lower energy and thus generate smaller electrical signals from the radiation detector. These are electronically rejected by the gamma camera
How does noise and spatial resolution compare in radionuclide imaging and radiographic imaging?
Radionuclide images generally have more noise and poorer spatial resolution than radiographs. The high noise is due to the relatively small number of gamma photons that contribute to the image.
The poor spatial resolution is related to the way in which the points of origin of the photons are determined and the properties of the collimator.
How can organ position effect signal?
different depths in tissue, Thus the gamma radiation from more superficial structures is less attenuated and therefore gives a higher signal
What is SPECT?
By acquiring a series of projection images around the long axis of the patient, it is possible to reconstruct a set of transverse tomographic slices. This is most often done by rotating the gamma camera head around the patient. The technique is known as single-photon emission computed tomography (SPECT) or sometimes just single-photon emission tomography (SPET). Most modern gamma cameras have more than one head and this has the potential to reduce total image acquisition times in both planar and tomographic imaging
What is effective halflife?
the activity of a radiopharmaceutical in the body decreases at a rate described by the effective half-life Teff.
1/Teff = 1/physical half life + 1/biological halflife
This decrease is caused by two processes:
Radioactive decay - described by the physical half-life Tphys
Elimination from the body by physiological processes such as clearance by the kidneys - described by the biological half-life Tbiol
What is patient dose in radionuclide imaging dependent on?
It is determined by the nature of the radiopharmaceutical (including the properties of the radionuclide label), its activity, its route of administration and its effective half-life.
Energy is deposited in the body through ionisations caused by Compton recoil electrons and photoelectrons from gamma interactions with tissues.
There may also be contributions from beta particles, internal conversion electrons and Auger electrons.
With sufficient information, it is possible to estimate the absorbed and equivalent dose to organs and tissues and the effective dose to the whole body
A radiopharmaceutical:
A. Provides contrast in radionuclide images when it is uniformly distributed in body organs and tissues
B. May emit both beta and gamma radiation
C. Is usually administered to the patient by intravenous injection
D. May be described as a sealed source of radioactivity
E. Has a biological half-life which is always the same as the physical half-life of the radionuclide
A. False. Contrast arises from differences in the uptake of the radiopharmaceutical in different organs and tissues.
B. True. Beta and gamma radiation may be emitted by radiopharmaceuticals. The presence of a large amount of beta radiation is undesirable for in vivo studies because it is not sufficiently penetrating and it increases the radiation dose to the patient. However beta emitters may be used for some in vitro investigations.
C. True. Intravenous injection is the most common route of administration, although other routes may be used for particular investigations (e.g. inhalation for lung ventilation studies).
D. False. Radiopharmaceuticals are unsealed sources of radioactivity.
E. False. The biological half life refers to the rate of clearance of a radiopharmaceutical from the body. The physical half life refers to the rate of radioactive decay. In general, the two values will be different.
Regarding a gamma camera collimator:
A. It is usually made of tungsten
B. It must have holes of circular cross-section
C. It attenuates gamma radiation mainly by photoelectric interactions
D. It acts in the same way as a lens
E. It has no direct effect on the radiation dose to the patient
A. Incorrect. Gamma camera collimators are made of lead.
B. Incorrect. The holes in a collimator may have any one of a variety of cross-sections (e.g. round, square, hexagonal).
C. Correct. In lead, most photons interact by photoelectric absorption.
D. Incorrect. There is no lens for gamma radiation. A collimator acts by selecting gamma photons according to their direction of travel
E. Correct. The radiation dose to the patient is determined by the nature of the radiopharmaceutical, its activity, its route of administration and the rate at which it is eliminated from the body. The collimator has no direct effect on the dose.
What are three methods of radionuclide production?
Cyclotron - produces radionuclides by bombarding stable nuclei with high-energy charged particles.
Nuclear reactor - nuclear fission or neutron activation of a stable target material.
Radionuclide generator - Radionuclides with short half-lives can be produced from a radionuclide generator, by decay of the parent radionuclide.
The generator allows the parent and daughter product to be easily separated.
How does a cyclotron work?
ions travel in a vacuum inside two D-shaped electrodes. A magnetic field causes the ions to travel in circular paths.
Application of an alternating voltage of the correct frequency between the Ds causes ions to be accelerated across the gap, thus gaining velocity and kinetic energy. As ion velocity increases, so does the radius of the circular path.
Ions with the greatest velocity and kinetic energy travel near the outside of the Ds, and are deflected onto a suitable target.
Nuclear reactions occur in the target to produce the required radionuclide
How are radionuclides produced in a nuclear reactor?
A naturally occurring uranium-235 (235U) nucleus absorbs a neutron which causes fission (breaking apart of the nucleus) with the release of more neutrons.
The fission is controlled by control rods, which can be inserted or withdrawn from the reactor core. These are made of a material which absorbs neutrons without producing fission.
Suitable target materials can be lowered into holes (ports) in the reactor so that they are irradiated by the neutrons. Neutron-capture reactions create radioisotopes of the target element; an example is the creation of radioactive molybdenum-99 (99Mo) from stable molybdenum-98 (98Mo).
The fission of 235U nuclei creates a mixture of radionuclides of lower mass number and lower atomic number. Some of these products of fission, such as iodine-131 (131I), are useful in nuclear medicine. Molybdenum-99 is also a fission product, and so there are two ways in which this radionuclide may be obtained from a nuclear reactor.
How is 99mTc produced?
beta-minus (β-) decay of parent radionuclide nuclide 99Mo (half-life 67 hours) produces a daughter radionuclide 99mTc (half-life 6 hours). Technetium-99m then decays by isomeric transition (IT) to 99Tc with the emission of 140 kiloelectron volt (keV) gamma rays that are suitable for medical imaging
What is the half life of 99mMo?
2.8 days (β- decay).
how is 99mTc extracted from the generator?
Technetium-99m is extracted from the generator by a process called elution. Immediately after each elution, the activity of 99mTc is reduced to zero.
How is a 99mTc generator made?
Fission produced 99Mo is adsorbed on alumina in a sterilised glass column surrounded by a lead or depleted uranium shield.
Technetium-99m is eluted from the alumina column, by an ion exchange mechanism when sterile saline is passed through it.
The resulting eluate is sodium pertechnetate (NaTcO4).
What is the half life of 99m Tc?
6-hour half-life
Long enough to be useful for imaging, but not so long as to result in an unacceptable radiation dose to the patient.
How does 99mTc decay?
Decay mode-isomeric transition
Only gamma rays are emitted, thus limiting radiation dose, since alpha and beta particles give a high localised radiation dose.
What is the energy of 99mTc emissions and its HVL?
140 keV gamma photon energy
Half-value thickness
0.3 mm lead (Pb)
46 mm tissue
3 mm NaI
High enough energy to pass through the body of the patient, but low enough to be stopped by the detector of the imaging device. Gamma rays with energies 50-500 keV can be used for imaging, about 150 keV being ideal. Facilitates shielding and collimation. Good penetration, scatter not excessive. Good detection efficiency.
WHat is the half life of the decay product of 99mTC?
Half-life of decay product 2.5 x 105 years. Effectively stable. No significant radiation dose to the patient from the decay product.
What are three are basic approaches to the preparation of radiopharmaceuticals?
Substitution of a stable nuclide with a radioisotope
Addition of radionuclide to chemical compound - eg with a kit
Incorporation of radionuclide onto autologous blood cells
What are the advantages of kit preparation?
The product is rapidly and easily prepared
The radiation hazard to the operator is low
The product is stable, sterile and pyrogen free
The product is reproducible and reliable
Unlabelled material has a long shelf-life (whereas labelled product has a limited shelf-life of a few hours)
The product is readily available and economic
Examples include:
99mTc methylene diphosphonate (MDP), used for bone scanning
99mTc macro aggregates (MAA), used for lung perfusion imaging
99mTc mercaptoacetyltriglycine (MAG3), used for renography
How is radionuclide activity measured?
with an ionisation chamber, into which a syringe or vial can be lowered. Electric current is proportional to radioactivity of a given radionuclide. The activity is indicated on a digital display, usually in megabequerels (MBq).
How is radionuclide purity measured?
Radionuclide purity is the proportion of radioactivity present in the stated radionuclide form. A possible contaminant in 99mTc radiopharmaceuticals is 99Mo from the generator.
Testing should be carried out daily for 99Mo breakthrough by placing the freshly eluted NaTcO4 in a 6-mm-thick lead container and measuring the ionisation current in the calibrator. The lead will absorb virtually all of the 140 keV gamma rays from the 99mTc, but only 50% of the 740 keV gamma rays from the 99Mo. The ionisation current reading is compared with the current expected from the maximum permitted level of 99Mo in the sample
What is radiochemical purity and how is it measured?
Radiochemical purity is a proportion of the total activity of a radioactive material that is present in the stated chemical form. Possible contaminants in 99mTc labelled radiopharmaceuticals are free 99mTcO4- and reduced forms of 99mTc.
Radiochemical purity is ascertained using paper or thin layer chromatography. A drop of the radiopharmaceutical is applied near the bottom (origin) of the chromatogram (Fig 3). This is placed in a solvent ensuring that the origin is not immersed. As the solvent migrates along the chromatogram different chemical species in the radiopharmaceutical are separated. The chromatogram is then analysed to calculate the percentage of bound, free and reduced 99mTc in the preparation. For radiopharmaceuticals, radiochemical purity should be >95%.
What is chemical purity and how is it measured?
Chemical purity is the fraction of the total mass that is present in the stated chemical form.
Possible chemical contaminant of 99mTc is aluminium (Al3+) from the alumina column in the generator.
commercially available kit. This contains Al3+ indicator papers and a standard Al3+ solution of 10 mg/ml. A drop of the standard solution is placed on the indicator paper next to a sample drop from the eluate. The intensity of the colour produced by the eluate should be less than that produced by the standard.
Regarding facilities for the production of radiopharmaceuticals:
A. All radiopharmaceuticals must be prepared in a sterile, pyrogen-free environment
B. Aseptic operations can be carried out in a cleanroom with a laminar flow cabinet, or an isolator with a filtered air supply
C. Cleanrooms have a negative pressure difference with respect to adjacent areas
D. Cleanrooms require a three-stage changing room
A. False. Only radiopharmaceuticals for injection must be sterile. Radiopharmaceuticals that are administered orally or by inhalation should be prepared in hygienic, but not necessarily sterile, conditions.
B. True. Both options provide an environment that conforms to EC GMP Grade A.
C. False. Cleanrooms have a positive pressure difference with respect to adjacent areas to prevent particulate contamination entering the room.
D. True. Cleanrooms require a three-stage changing room, for staff to change from outer clothing into sterilised cleanroom clothing.
Regarding the characteristics of 99mTc:
A. It has a half-life of 67 hours that is ideal for medical imaging
B. It decays by beta decay
C. It emits gamma photons that have an energy of 140 keV
D. It decays to 99Tc, a radionuclide with a very long half-life
A. False. The half-life of 99mTc is 6 hours. This is long enough to be useful for imaging, but not too long resulting in an unacceptable radiation dose to the patient. The parent radionuclide, 99Mo, has a half-life of 67 hours.
B. False. Technetium-99m decays by isomeric transition.
C. True. Note that an energy of 140 keV is good for imaging, being high enough to pass through the body of the patient but low enough to be stopped by the detector of the imaging device.
D. True. Technetium-99 has a half-life of 2.5 x 105 years. This means it is effectively stable and does not cause a significant radiation dose.
What is the most common type of collimator in the gamma camera?
The most common type is the parallel hole collimator in which hole shapes may be round, square, triangular or more typically, hexagonal
What are gamma camera collimators made from and why?
Collimators are made of lead because of its higher linear attenuation coefficient; the lead between two adjacent holes is called the septum. A collimator for imaging technetium-99m (99mTc) radiopharmaceuticals may have tens of thousands of holes each of about 1-2 mm in diameter
How can the collimators be made and which is best?
Parallel hole collimators may be made using a number of different methods:
Crimped lead foil sheets
A drilled lead block
Casting from molten lead
Drilling or casting is a better mode of collimator construction because no gaps are left in the septa, thus giving better image contrast and spatial resolution. However, crimped lead foil collimators are less expensive.
Why must the septa thickness be just right?
If the septa are too thin, there is an increased probability of penetration by gamma photons that are not travelling parallel to the axes of the holes. These would degrade contrast and spatial resolution.
Too thick and sensitivity is reduced due to more photons being attenuated
What does a diverging collimator do?
gives a minified image and may be useful when the object is larger than the field of view of the gamma camera.
What does a converging collimator do?
gives a magnified image and this may be useful when the object is smaller than the field of view
When is a pin-hole collimator used?
The pinhole is a single-hole collimator that is used to produce magnified views of small objects such as the thyroid. It consists of a small (3-5 mm) hole in a piece of lead or tungsten, mounted at the apex of a leaded cone.
Magnification is achieved provided that the pinhole-object distance is less than the detector-pinhole distance. The use of a pinhole collimator for small objects means that gamma photons interact with a much larger area of the radiation detector than would be the case with a parallel hole collimator.
what is the gamma camera detector normally made from?
large-area thallium (Tl)-doped sodium iodide (NaI) scintillation crystal - a single crystal of NaI to which a small amount of Tl has been added during its manufacture. Sodium iodide is a fluorescent material which converts the energy absorbed from a gamma photon interaction into a scintillation (a small flash of visible and ultraviolet light). Doping the NaI with Tl improves its light output. The crystal is usually 6-13 mm thick (i.e. ¼ to ½ an inch).
Why is the NaI crystal sealed in aluminium?
Modern detectors tend to be rectangular with sizes up to 60 × 40 cm. Sodium iodide is hygroscopic and so the crystal is hermetically sealed in an aluminium can with a glass window adjacent to one face so that the light can escape.
What lies behind the crystal in a gamma camera?
Behind the crystal is an array of 30-100 photomultiplier tubes (PMTs) that detect the visible and ultraviolet photons from the scintillation. These are separated from the crystal by a light pipe in the form of a slab of polymethyl methacrylate (commonly known by trade names such as Perspex® and Lucite®). Silicone grease is used to maintain good optical contact between the light pipe and both the exit window of the detector and the entrance window of the PMTs. Associated with each PMT is a pre-amplifier whose output is an analogue electrical voltage pulse.
How does the PMT work in a gamma camera?
The photomultiplier tube (PMT) is a vacuum tube that detects a very small amount of light such as that produced in a scintillation.
The light energy releases electrons from a photocathode deposited as a thin layer on the inside of the entrance window. Inside the tube are a series of electrodes called dynodes that are held at increasing positive potential to each other by a high voltage supply.
The electrons are accelerated from the photocathode to the first dynode by the potential difference and gain kinetic energy. The kinetic energy of each electron releases more than one electron from the dynode surface and these electrons are accelerated to the next dynode.
Thus the PMT acts as an electron amplifier and a relatively large amount of negative electrical charge is collected at the anode (the last dynode). The arrival of the negative charge (i.e. electrons) at the anode over a brief period of time generates a current pulse which forms the signal that is received by the pre-amplifier.
How is gain (amplification) stabilised in the PMTs?
One method of gain stabilisation is to periodically feed a fixed amount of light into each PMT, monitor the amount of charge produced at the anode and adjust the high voltage as necessary. A light emitting diode (LED) acts as a stable source of light and this is conducted to the PMTs by fibre optic cables.
Why are PMTs covered in mumetal?
shields the dynodes from the earth’s magnetic field which would otherwise deflect the electrons.
What is the main type of photon interaction in the gamma camera?
At 140 keV, most photons that interact with the crystal do so through the photoelectric effect with the remainder undergoing Compton scattering. Gamma photon energy is absorbed through photoelectrons and Compton recoil electrons to create a scintillation (small flash of visible and ultraviolet light). Thus a scintillation follows each gamma interaction
What does the light pipe do in the gamma camera?
By introducing a physical separation between the crystal and the PMTs, the light pipe (through which the visible and ultraviolet (UV) photons pass) helps to spread the light over many tubes. The amount of light received by a particular PMT depends on the inverse square of the distance from the scintillation to that PMT. Photomultiplier tubes close to the scintillation receive a relatively large amount of light while those further from the scintillation receive a relatively small amount of light.
How is the position and energy of a scintillation calculated?
the distribution of pulse heights from the PMTs carries information about the position of the scintillation in the plane of the crystal (and therefore the line of origin of the gamma photon in the patient). Furthermore, the sum of the pulse heights is proportional to the total amount of light produced and hence the energy absorbed from the gamma photon by the crystal.
How is the analogue signal from the PMTs digitised?
In modern digital gamma cameras, analogue voltage pulses from PMTs are digitised by analogue-to-digital converter (ADC) circuits and the calculation of X, Y and Z is performed digitally.
What is energy discrimination?
Scattered radiation is rejected by testing whether the corrected Z value represents the full energy of the gamma photon – this is called energy discrimination
What does digitising the signal from the PMTs allow?
the calculation of X, Y and Z is performed digitally. Subsequently, these are refined by the application of spatial linearity and energy corrections and the corrected values are used to create an image in digital computer memory.
What is spatial non-linearity in a gamma camera caused by?
Spatial non-linearity is caused by systematic mispositioning of counts, i.e. errors in the calculation of X and Y values. In particular, the locations of individual counts tend to be shifted towards the centre of the nearest PMT. Thus a straight line source of radioactivity tends to give a wavy line image and there is an increased count density at the locations of the PMTs when a uniform source is used. The Z value may vary with position due to factors such as variation in light production, light transmission, light detection and residual PMT gain. This can result in the fraction of counts rejected by energy discrimination varying from one area of the crystal to another. Positional variation in sensitivity (count rate per unit activity) may be due to manufacturing defects in the crystal and/or collimator.
What is the photopeak?
The spectrum of Z values has a prominent peak centred on the value corresponding to full gamma energy absorption; this is called the photopeak
What is the width of the photopeak used to measure?
because of variations in factors such as light production, transmission and detection, it has a measurable width. This is usually expressed as the full width at half maximum
The energy resolution of the scintillation crystal is given as a percentage by: (FWHM/energy at peak)x100
What is the compton band comprised of?
Counts in the Compton band comprise both:
Potentially useful unscattered radiation from the patient that has been scattered in the crystal
Scattered radiation from the patient that has interacted with the crystal by either mechanism
Unfortunately they are indistinguishable.
How is energy discrimation introduced?
Scattered radiation from the patient has an adverse effect on image contrast and spatial resolution and so it should be rejected. This cannot be done by the collimator; it can only be done electronically by energy discrimination. This is achieved by only accepting counts within the photopeak. Most commonly, a 20% acceptance window is used centred on the photopeak energy
How is the image formed in the gamma camera?
Each scintillation generates a trio of corrected X, Y and Z values.
The digital image is usually acquired in computer memory organised as a matrix of locations which may be imagined to be superimposed on the crystal.
The X and Y values define the position of the scintillation in the plane of the crystal and hence the address of the corresponding memory location.
Provided that the Z value is within the acceptance window set by the user, the contents of the identified memory location are increased by 1 count
Acquisition usually terminates after a preset number of counts (e.g. 1 million) has been accepted. This may take several minutes. Alternatively, it is also possible to acquire the image for a preset time, e.g. 100 s.
How is the image displayed from the gamma camera?
Once acquired, the digital image may be displayed on a video monitor with each image pixel corresponding to a particular memory location. It is possible to use monochrome or colour display scales with the brightness and/or colour of a given pixel depending on the count in the associated memory location.
How can gamma camera images be manipulated?
Images may be smoothed to reduce noise
Images may be windowed to increase contrast and highlight regions of low count density
Interpolation may be used to increase the display matrix relative to the acquisition matrix in order to make the pixels less apparent
Images may be added or subtracted and analysed to extract quantitative information
Regarding collimator septal thickness:
A. Iodine-131 requires a greater septal thickness than 99mTc
B. Low energy collimators may be used with photon energies up to 250 keV
C. A radiopharmaceutical labelled with 99mTc may be imaged with a medium energy collimator
D. In low-energy collimators, the septal thickness is typically about 0.3 cm
E. High-energy collimators may be used to image positron emitting radionuclides
A. True. The gamma photon energy of 131I is 364 keV while that of 99mTc is 140 keV.
B. False. The upper limit of photon energy for low-energy collimators is approximately 150 keV, depending on the manufacturer.
C. True. Technetium-99m has a gamma photon energy of 140 keV and medium-energy collimators may be used with energies up to approximately 300 keV. However, low-energy collimators are more suitable for 99mTc.
D. False. It is about 0.3 mm.
E. False. Positron emitters give annihilation radiation at 511 keV. The typical upper limit of photon energy for a high-energy collimator is 400 keV. Some manufacturers supply ultra-high-energy collimators for imaging annihilation radiation.
What are the three major parameters of image quality in radionuclide imaging?
Contrast (the difference in signal intensity between different parts of the image)
Noise (mainly due to random variations in signal intensity within the image)
Spatial resolution (detail or sharpness)
What is subject contrast?
Subject contrast CS is defined as the difference in activity concentration between the lesion and healthy tissue divided by the value for healthy tissue.
contrast is largely determined by the differential uptake of the radiopharmaceutical in a lesion and surrounding healthy tissue. Thus it depends primarily on physiological function rather than physical factors (such as the density and average atomic number of tissue) as in x-radiography.
What is the nnlimit of negative contrast?
The limit of negative subject contrast is -1 (when AL = 0)
What is the limit of positive contrast?
there is no theoretical limit to positive subject contrast. For this reason, radiopharmaceuticals that produce positive subject contrast are generally preferred to those that produce negative subject contrast.
What is image contrast?
difference between count density in the lesion and the healthy tissue
Count density is sometimes known as information density and can also be expressed as counts per pixel.
What is image contrast dependent on?
The subject contrast - greater the subject contrast, the greater the image contrast
Background gamma radiation. The greater the background, the lower the image contrast; this is usually an important contribution
Differential attenuation of gamma radiation from the lesion compared with healthy tissue. Attenuation of gamma radiation from a lesion deep within the body (i.e. far from the gamma camera collimator) is often greater than the average attenuation experienced by gamma radiation from surrounding healthy tissue and this decreases image contrast. It is less of a problem for superficial lesions
Patient movement. This will decrease image contrast (and spatial resolution) and is more likely with long image acquisition times
The spatial resolution of the gamma camera system. The poorer the spatial resolution, the lower the image contrast.
How does image contrast compare to subject contrast?
Image contrast is invariably lower than subject contrast.
What contributes to background radiation in radionuclide imaging?
Radioactive sources (other than the patient) in the vicinity of the gamma camera
Natural radioactivity in the environment
Radioactivity in tissue above and below the lesion
Septal penetration in the collimator
Scattered radiation from the patient
What is applied to help reduce background radiation effect?
Background subtraction is often applied to radionuclide images. This improves contrast, but increases noise.
What is structured noise?
Structured noise is due to non-random variations in count density that are superimposed upon and interfere with the objects of interest. In radionuclide imaging, structured noise may arise from the distribution of the radiopharmaceutical for example bowel uptake of gallium-67 citrate in abdominal scintigraphy to detect inflammation or abscesses, or imaging system artefacts such as gamma camera non-uniformity.
What is unstructured noise?
is random noise, also called statistical noise or quantum mottle. This is due to random variations in count density which result from the random nature of radioactive decay.
How is randow noise related to counts?
the random noise in that region is the standard deviation (σ) of N (counts)
How is SNR related to counts?
square root of N
How is noise contrast related to counts?
1/ square root N
What factors effect relative random noise?
Image acquisition time - longer is better
Activity of radiopharmaceutical - higher is better
Sensitivity of the gamma camera system
How is Sensitivity of the gamma camera system defined?
count rate (counts per second (cps)) per unit activity (in units of cps per becquerel). The greater the system sensitivity, the lower the relative noise for a given acquisition time and administered activity. Unfortunately, the factors that increase sensitivity are also those that degrade contrast and/or spatial resolution: a wide absorbed energy (Z) acceptance window, a thick sodium iodide (NaI) crystal and a collimator with short holes of large diameter.
What is CNR?
contrast to noise ratio which is defined as the ratio of the modulus of the image contrast to the noise contrast in the surrounding tissue.
What is the rose criterion?
In order for the lesion to be detectable, the CNR must exceed a particular value known as the Rose criterion. This size of this value depends on factors such as lesion shape, spatial resolution, viewing distance, observer experience etc. It is typically about 4.
What is spatial resolution?
Spatial resolution is a measure of the gamma camera’s ability to record variations in activity concentration and to distinguish between small radioactive sources that are close to each other.
How can spatial resolution be quantified?
It may be quantified as the full width at half maximum (FWHM) of the curve of counts or count rate (line spread function) vs. distance when a line or point source of radioactivity is imaged (point spread function)
In radionuclide imaging, it is more usual to use the LSF.
What is the modulation transfer function?
The Fourier transform of the LSF is the modulation transfer function (MTF). The MTF is a complete mathematical description of the resolution properties of an imaging system. It gives the fraction of an object’s contrast that is recorded by the imaging system as a function of the size (actually spatial frequency) of the object.
What factors effect spatial resolution?
Intrinsic spatial resolution of the gamma camera
Spatial resolution of the collimator
Patient motion
Pixelation effects
Scattered radiation
Imperfections in image display or recording devices
WHy is patient motion a problem in radionuclide imaging?
Patient motion will degrade spatial resolution and is a particular problem in radionuclide imaging because of the relatively long imaging times.
What size of pixel is needed to avoid loss of function?
at least two pixels per FWHM are needed to avoid loss of detail. Given that FWHM values are comparatively large in radionuclide imaging, this can be achieved with a relatively coarse pixel matrix.
What factors effect intrinsic spatial resolution?
Energy and linearity correction of the X, Y and Z values produced by each scintillation
Photoelectron range in the NaI crystal
Gamma photon energy
Collection of scintillation light
Detection of scintillation light
Number and size of photomultiplier tubes
Thresholding the photomultiplier tube voltage signal
Thickness of the NaI(Tl) detector crystal
What is the typical intrinsic spatial resolution at 140kv?
Values of intrinsic spatial resolution at 140 keV vary between 2.5 mm FWHM (corresponding to about 4 line pairs cm-1) and 4 mm FWHM (corresponding to about 2.5 line pairs cm-1).
How does increased gamma energy effect spatial resolution?
The greater the gamma energy, the greater the number of scintillation photons reaching a particular photomultiplier tube (PMT) and so the smaller their relative statistical variation; this gives better spatial resolution.
How is collection of light improved?
Optimising light collection gives better spatial resolution by reducing the relative statistical variation in the number of scintillation photons per PMT. Collection is improved by good optical coupling and PMT shape - square or hexagonal PMTs cover a greater area of crystal than circular ones. In some designs of gamma camera head, the uniformity of light collection may be improved by the presence of a light guide between the detector and the PMTs.
Why does a thicker crystal give worse spatial resolution?
There is greater possible variation in the depth within the detector at which scintillations occur and this leads to variations in the distribution of scintillation light among the PMTs
There is greater probability of multiple Compton scatter of a gamma photon in the crystal; this produces a total absorbed energy Z value within the acceptance window but X and Y values that represent the position of the average luminous intensity of the scintillation rather than the true position at which the photon entered the crystal
What collimator factors effect spatial resolution?
using a collimator with long holes of small diameter and ensuring that the patient is as close as possible to the collimator improve spatial resolution
How is whole system spatial resolution calculated?
system resolution is given by the square root of the sum of the squares of the components. Thus system spatial resolution is worse than that of any single component and the collimator is the dominant factor.
What is the trade off in crystal thickness in the gamma camera?
A thick crystal optimises sensitivity by increasing the attenuation Optimal spatial resolution, on the other hand, requires a thin crystal. The compromise used in most gamma cameras is to use a thickness of 9.5 mm (3/8 inch).
What is the trade off in the collimator in the gamma camera?
For optimal sensitivity, a parallel hole collimator should have short holes of large diameter whereas the opposite is required for optimal spatial resolution. It is usual to have a range of low energy collimators (suitable for technetium-99m (99mTc)) ranging from high sensitivity to general purpose (a compromise between sensitivity and resolution) to high resolution.
Why does the sensitivity of a collimator not change with distance?
The spatial resolution of a parallel hole collimator becomes poorer as the distance between the source of radioactivity and the collimator face increases. However, the sensitivity remains approximately constant. The number of gamma photons passing through a particular collimator hole decreases with the square of the distance but the number of holes through which the photons can pass increases as the square of the distance. These two effects cancel each other out.
How does spatial resolution change with depth in tissue?
rapid degradation of spatial resolution with depth in tissue. Deep organs are further from the collimator face than more superficial organs. This fact, together with the increase in attenuation with depth, is the reason why it is common to take more than one view (with different orientations of the gamma camera head) in planar radionuclide imaging
In planar radionuclide bone imaging, which one of the following may improve the contrast between bone and surrounding soft tissue?
A. Increasing the administered activity
B. Using a high sensitivity collimator
C. Encouraging the patient to drink fluids
D. Placing an air gap between the patient and the collimator
A. Incorrect. Increasing the administered activity would produce the same relative increase of count density in both lesions and healthy tissue so the contrast would not change. However, the CNR would increase.
B. Incorrect. Using a high sensitivity collimator would produce the same relative increase of count density in both lesions and healthy tissue so the contrast would not change. However, the CNR would increase.
C. Correct. Compounds such as 99mTc labelled methylene diphosphonate (MDP) are excreted by the kidneys as well as being taken up by bone. Drinking fluids increases the rate of excretion of MDP which is not bound to bone and this improves both subject and image contrast.
D. Incorrect. Placing an air gap between the patient and the collimator would increase the distance between the source and the collimator, causing a decrease in spatial resolution and thereby reducing contrast.
In planar radionuclide imaging, relative noise decreases with which of the following three?
A. Increasing image acquisition time
B. Decreasing administered activity
C. Decreasing collimator sensitivity
D. Increasing detector sensitivity
E. Increasing total number of image counts
A. Correct.
B. Incorrect. Decreasing the administered activity would decrease the count density and therefore increase the relative noise.
C. Incorrect. Decreasing the collimator sensitivity would decrease the count density and therefore increase the relative noise.
D. Correct.
E. Correct.
Only one statement is correct regarding the spatial resolution of a parallel hole collimator. Which one is it?
A. Is independent of the distance between the patient and the collimator
B. Improves as hole length increases
C. Does not depend on hole diameter
D. May be optimised at the same time as sensitivity
E. Depends on the thickness of the NaI crystal
A. Incorrect. The spatial resolution of a parallel hole collimator improves as the distance between the patient and the collimator decreases.
B. Correct.
C. Incorrect. The spatial resolution of a parallel hole collimator is proportional to the hole diameter.
D. Incorrect. For optimal sensitivity, a parallel hole collimator should have short holes of large diameter whereas the opposite is required for optimal spatial resolution.
E. Incorrect. The thickness of the NaI crystal affects intrinsic resolution, not collimator resolution.
What are the three types of planar radionuclide imaging?
Static imaging
Dynamic imaging
Gated imaging
When is static radionuclide imaging used?
this is used where the distribution of a radiopharmaceutical within the patient is changing very slowly with time (is effectively stable for the duration of the acquisition).
When is dynamic radionuclide imaging used?
this is used where the radioactivity within the area of interest is changing rapidly with time.
When is gated radionuclide imaging used?
this is used to image organs with regular physiological motion. The most important examples are cardiac and respiratory gating
What is the difference between high resolution collimators and high sensitivity collimators?
High resolution collimators achieve better spatial resolution images at the expense of a low count rate and hence longer imaging times. High sensitivity collimators result in a higher count rate, but yield lower resolution images.
In practice, high resolution collimators are used where small structures need to be resolved, e.g. bone scans. High sensitivity or general purpose collimators are used in dynamic images where count rate is more important than anatomical detail such as renography.
How can SNR in planar radionuclide imaging be increased?
Increasing imaging time. This is limited by the need to maintain patient comfort in order to avoid motion and the logistics of timing patients throughout the working day
Increasing the amount of administered radioactivity. This is limited by the legal requirement to ensure that doses arising from a medical exposure are kept as low as reasonably practicable consistent with the intended purpose
Ensuring acceptable gamma camera sensitivity. The sensitivity of the gamma camera, expressed in counts per second per megabecquerel (counts/s/MBq), should be checked as part of the routine quality control program for the camera
How does choice of matrix effect the image?
The choice of matrix alters the size of the pixels into which the image is acquired. Using a smaller matrix, images appear ‘coarser’. For high quality imaging, a larger matrix should be chosen. One disadvantage of using a larger matrix is the larger amount of computer memory required.
Where should the camera be in relation to the patient?
the gamma camera should be as close as possible to the patient during imaging to optimise spatial resolution.
How does zoom effect the radionuclide image?
reduces pixel size
Reducing the pixel size by a factor of N reduces the counts per pixel by a factor of N^2 so the SNR is decreased.
What is list mode and why can it be beneficial?
In list mode the X and Y values of each detected event are recorded together with regular timing information. The data can then be split into segments or frames post-acquisition.
The advantage of list mode is that the frame length does not need to be known in advance, so it can be used in any dynamic imaging where the physiological rate is unknown. It is also used in gated cardiac blood pool imaging where there are potential variations in the R-R interval.
The disadvantage of list mode acquisition is the large amount of computer memory required to store each detected event individually.
How is the pixel count and brightness displayed determined?
a look-up table
Most display screens are 8 bits deep, so the look-up table has 256 different intensities or display levels.
In many nuclear medicine images, most of the counts are concentrated in a small number of pixels, often not in the area of interest for example, the bladder in a bone image. Use of a linear look-up table would result in the majority of display levels being used to display this small number of pixels and differences between the pixels of interest would not be seen. This problem can be overcome using non-linear look-up tables
What is linear filtering?
One of the most common forms of linear filtering is convolution (a mathematical function).
Filtering enables smoothing and sharpening.
What is the relationship between convolution in the spatial domain and the frequency domain?
Convolution in the spatial domain is equal to multiplication in the frequency domain
What is the process to perform convolution in the frequency domain?
Take the Fourier transforms of i(x,y) and f(x,y), apply pointwise multiplication, and take the inverse Fourier transform of the result
What advantage does frequency domain convolution have over direct convolution?
It requires less computational power
In the frequency domain, what does high spatial frequency represent?
Fine spatial detail
In the frequency domain, what does low spatial frequency represent?
Larger structures
What happens to the modulation transfer function (MTF) of the gamma camera with increasing spatial frequency?
It decreases rapidly
How does image noise behave in relation to frequency?
It is independent of frequency
What is the effect on the signal-to-noise ratio (SNR) as spatial frequency increases?
The SNR decreases
How can the SNR be improved in the context of spatial frequencies?
By attenuating the higher spatial frequencies
What type of filter can be applied to improve SNR by attenuating noise?
Low-pass filter such as Hanning or Butterworth filter
What does ROI stand for in image analysis?
Region of Interest
What is the purpose of ROI analysis?
To generate the total number of counts in a region and its area in pixels.
How can the area in ROI analysis be defined?
By the user or automatically by the processing system.
What are the common options for drawing the area in ROI analysis?
- Rectangular
- Circular
- Freehand
What techniques are used in ROI analysis to define the region?
Thresholding techniques.
In ROI analysis, how is the region typically drawn?
To include pixels with count values above a certain percentage of the maximum count value in the image.
True or False: ROI analysis can only be performed manually by the user.
False
Which images are used in ROI analysis to calculate the geometric mean?
Anterior and posterior images
What does the geometric mean calculation in ROI analysis correct for?
Depth
What can regions of interest on dynamic studies generate?
Time-activity curves
Points on the curve represent the number of counts or the count-rate within the ROI in each frame of the study.
What does the term time-activity curve imply?
Activity is proportional to counts or count-rate.
What parameters can be derived from time-activity curves?
- Area under the curve
- Curve slope
- Time to reach a peak
What types of curves are used to define the rate of washout of the radiopharmaceutical from an organ?
Single exponential or bi-exponential curves
True or False: The curve slope is not a parameter that can be derived from time-activity curves.
False
What is the significance of the time to reach a peak in a time-activity curve?
It indicates the time taken for activity to reach its maximum level.
WHat do static images provide information about?
Organ size
Organ shape
Organ position
Regions of increased and/or decreased uptake
What is whole body scanning?
A method by which a single (whole body) image is obtained instead of lots of ‘spot’ views.
What are the two modes of scanning in whole body scanning?
- Continuous mode
- Step and shoot mode
Describe continuous mode in whole body scanning.
The scanning couch travels on rails at a constant speed, and the computer ‘reconstructs’ the image.
Describe step and shoot mode in whole body scanning.
Typically 4-6 static images are taken to cover the length of the body and then ‘zipped’ together.
What is the typical scan speed for whole body scanning?
12 cm/min
How long does a typical whole body scan take?
~15 minutes
What is an advantage of whole body scanning?
Guarantees that no area of the body will be missed between spot views.
How can anterior and posterior views be acquired in whole body scanning?
- Simultaneously in a dual headed system
- Sequentially in a single headed system
What is a disadvantage of whole body scanning?
The theoretical loss of resolution.
What is autocontouring used for in whole body scanning?
To ensure that the camera heads are as close as possible to the patient at all times.
What determines the choice of frame time in imaging studies?
The study being performed
What are the frame times needed for imaging rapid changes?
Short frame times (down to 0.1 seconds)
What frame times are suitable for slowly changing distributions?
Long frame times (up to 300 seconds or more)
Can multiple frame times be used within the same study?
Yes
What do dynamic images provide?
Variation of distribution of radioactivity with time
What type of images do dynamic images offer?
Low resolution ‘static’ images of frames grouped together
What is required for image processing in dynamic imaging?
Background subtraction, frame summation, region of interest analysis, generation of time activity curves
What is the gall bladder ejection fraction (EF)?
A measure assessed following an injection of radiopharmaceuticals like 99mTc-labelled HIDA.
EF is used to evaluate gall bladder function.
What radiopharmaceutical is commonly used to assess gall bladder ejection fraction?
99mTc-labelled hepatobiliary iminodiacetic acid (HIDA).
HIDA is a compound used in nuclear medicine for imaging the liver, gallbladder, and bile ducts.
What is the purpose of gated blood pool imaging?
To assess ventricular function both qualitatively and by calculation of ejection fraction.
Ejection fraction is a key measurement of heart function, indicating the percentage of blood pumped out of the ventricles with each heartbeat.
What radiolabel is used in gated blood pool imaging?
99mTc.
99mTc (technetium-99m) is a commonly used radioisotope in medical imaging due to its favorable properties.
At what orientation is imaging typically carried out in gated blood pool imaging?
LAO 40° orientation.
LAO stands for Left Anterior Oblique, which helps maximize the separation between the left and right ventricles during imaging.
What triggers the acquisition of images in gated blood pool imaging?
The R wave of the ECG.
The R wave represents the electrical activity of the ventricles during the heartbeat.
What happens when the next R wave is received during imaging?
Information from the first portion of the cardiac cycle is added to that of the previous cycle.
This process allows for continuous data accumulation for improved accuracy.
What is the typical total number of counts acquired in gated blood pool imaging?
Typically 3 million counts.
A higher count improves the quality of the imaging data.
What mode can be used for data acquisition when the R-R interval is irregular?
List mode.
List mode allows for more flexible data collection, accommodating variations in heart rhythm.
Why is SPECT better than planar gamma camera imaging?
In SPECT, the 3D distribution of a radiopharmaceutical is represented as a series of 2D slices reconstructed in a transverse plane, which can be further reconstructed in coronal, sagittal or oblique planes. These 2D slice images provide improved contrast by removing or reducing the effects of activity in overlying and underlying structures and improved depth information by better localisation of image structures within the three orthogonal slice planes.
Single photon emission CT imaging enables improved contrast to be obtained by removing or reducing the effects of activity in overlying and underlying structures.
What tracers can be imaged by SPECT?
Almost any radiopharmaceutical distribution can be imaged using SPECT, provided that:
The distribution of the radiopharmaceutical does not change significantly within the duration of the image acquisition time, typically 20-40 minutes
The acquisition time is long enough for sufficient numbers of gamma photons to be acquired to ensure the noise within the tomographic images is kept to an acceptable minimum
WHat are 3 key applications of SPECT?
Cerebral blood flow
Skeletal imaging
Myocardial perfusion
Why are single gamma camera heads not used for SPECT?
Whilst such systems are capable of performing SPECT, they are limited by long acquisition times and the relatively crude design of the gantry and the patient table. They have largely been superseded in clinical practice by the widespread introduction of multiple head systems.
What is the standard equipment for SPECT imaging?
The most common are dual head, large field-of-view (FOV) systems which enable both planar and SPECT imaging to be performed with shorter acquisition times whilst maintaining image quality
What are SPECT/CT systems?
the addition of x-ray CT capability to dual head gamma camera systems. These hybrid devices usually employ a fan beam x-ray source and detector array mounted on the gantry, between the gamma camera heads.
64-slice systems with full diagnostic CT capability are now available.
Single photon emission CT/CT systems enable the superimposition or fusion of the SPECT (functional) and CT (anatomical) images resulting in improved localisation of features within the images.
The CT image obtained from such systems can also be used to perform accurate attenuation correction of the SPECT images.
WHat is the energy resolution for modern gamma camera detectors?
The energy resolution of modern gamma camera detectors is less than 10% at low energies (<160 keV), which is more than adequate for performing SPECT. In general the lower the energy resolution, the better the system will be at minimising the effects of scattered photons and maximising image contrast.
How should a sensitivity correction map be acquired?
The sensitivity correction map should be acquired on a regular basis and derived from a high-count image (~10 000 counts/pixel, giving a statistical error of 1%), obtained by imaging a uniform flood source of the appropriate radionuclide.
Each head is a 2D detector and must be highly uniform with an integral uniformity typically <3%. This can be achieved by applying a sensitivity correction map to all the acquired images, in addition to the usual energy and linearity corrections
What sort of collimators are preffered in SPECT?
In SPECT, it is important that the collimator holes are as parallel as possible, and both hole size and septa size are as uniform as possible, across the entire detector FOV, in order to maximise system uniformity. Collimators are usually manufactured either using thin lead strips (foil collimator) or cast in lead (microcast collimator). The latter are preferred for SPECT as they result in much more uniform hole and septa size, although they are significantly more expensive than foil collimators.
Although a choice of collimator can give a choice between improving resolution or sensitivity, a high-resolution collimator should be used wherever possible. This will maximise the spatial resolution throughout the depth of the patient, reducing image distortion during the reconstruction. High-resolution collimators show a better preservation of resolution with distance from the collimator, compared with general purpose collimators.
How can accurate body contouring of the gamma camera head be achieved?
It is particularly important in SPECT to minimise the detector - patient distance, in order to maximise image resolution. This is known as ‘body contouring’ and can be achieved either by using an elliptical orbit or automatically by using an infrared detection device.
These devices, incorporating two rows of infrared beams in front of the collimator face, continuously maintain the detector-patient distance by ensuring that only the outer row beam is interrupted by the patient outline. Any rapid patient movement which interrupts both beams will immediately disable all gantry motion.
What is the typical length of the SPECT scan?
typically 20–40 minutes.
What are important factors of the table in SPECT?
Be manufactured from carbon fibre or similar material
Support 200 kg patient without significant sagging
Have a narrow width (~35 cm)
Have low attenuation of gamma photons (<10% at 140 keV)
Have restraint straps
Have a narrower head restraint for brain imaging
What extra QA does SPECT need vs planar imaging?
Single photon emission CT requires extra routine QC procedures in addition to those routinely performed for planar imaging. The two most important requirements are correction for sensitivity variation and COR offset. They should be performed separately for each gamma camera detector head.
How is QA for COR performed?
For accurate reconstruction, the location of the detector axis of rotation is assumed to be at pixel position 64.5, 64.5 (for a 128 x 128 matrix) - the COR. A perpendicular line drawn from a location on the detector equivalent to this pixel position should pass through the axis of rotation. Where these two lines do not intersect, a COR offset is defined as the separation between them.
Centre of rotation is usually measured from a series of images of a small point source (~20 megabecquerel (MBq) 99mTc in 0.1 ml) (Fig 1). These are taken from typically 120 different angles around the source, positioned in the centre of the FOV of the camera head, offset by ~15 cm from the axis of rotation (in the X direction).
Using data from the images, graphs of the X and Y co-ordinates (pixel positions) of the point source are plotted against gantry rotation angle.
How can the overall performance of a SPECT system be tested?
Overall tomographic performance can be assessed from SPECT images acquired using a Jaszczak test object
Very high-count images can be acquired by filling the tank with typically 500 MBq of 99mTc and using very long acquisition times. The resulting SPECT slice images can be used to assess uniformity and resolution in orthogonal slice directions
Which of the following are advantages of SPECT imaging?
A. Short frame times compared with planar imaging
B. Improved image contrast
C. Improved feature localisation at depth
D. Avoid the need for planar imaging
The correct answers are B and C.
The purpose of performing SPECT, which is a tomographic technique, is to give improved contrast and feature localisation at depth compared with planar radionuclide imaging. However, it takes about 30 minutes to acquire a complete set of projection images (at different angles around the patient) for SPECT reconstruction, and this is comparable to or longer than the time it takes to acquire a set of planar images. In many clinical situations, such as bone imaging, both SPECT and planar imaging are done. However, there are some situations, such as myocardial perfusion imaging, in which it is usual just to do SPECT.
What factors may influence imagaing parameters in SPECT?
Factors which may influence the choice of imaging parameters include the administered activity, the relative uptake of the radiopharmaceutical, the organ or body area of interest and the desired resolution and noise level within the SPECT images.
The final choice is also is often determined by the ability of the patient to keep still for the duration of the acquisition.
What is the realtive noise of SPECT vs XR CT?
In general, SPECT images contain much more noise than x-ray CT images, irrespective of the imaging parameters used, so the aim should be to maximise image resolution during acquisition and to control or minimise image noise during the reconstruction process.
How does the acquisition matrix in SPECT effect the image?
The acquisition matrix used determines the spatial sampling (maximum resolution), counts per pixel (image noise) and minimum slice thickness of the SPECT data.
Large matrices (and hence small pixel size) will produce the best resolution and thinnest slices but at the expense of the highest image noise
How can the impact of noise be reduced in slice reconstruction?
During the slice reconstruction process the impact of image noise can be reduced, at the expense of resolution, by either increasing the slice thickness (Fig 2), use of smoother reconstruction filters, displaying the slice data in a 64 × 64 pixel matrix, or a combination these.
What is the typical length of time for each projection frame in SPECT?
20-40 seconds. This will depend on the administered activity, collimator and the ability/comfort of the patient to keep still during the acquisition.
What should the angular step be in SPECT?
To ensure that angular sampling is similar to the spatial sampling, 3° steps should be used for a 128 × 128 matrix. If the angular range is 360° (i.e. a complete circle), the number of projection frames is 360/3 = 120, which is close to 128.
What collimators are used in SPECT of the brain?
Fan beam collimators, used for brain imaging, improve sensitivity whilst preserving resolution by utilising more of the detector FOV to collect image data (magnification). As these have non-parallel holes, they can only be used with circular orbits and require specific reconstruction techniques.
WHy are images acquired with arms above the head in SPECT?
It is usual to extend the patient’s arms above their heads for chest (cardiac) and abdominal imaging to remove the effect of radiation attenuation in the arms and minimising patient – detector distance.
How is the heart normally imaged in SPECT?
The heart is a small moving object, located off-centre from the body axis and is usually imaged over 180° from a left posterior oblique (LPO) to right anterior oblique (RAO) projection, with the heads in L mode.
This restricted angle acquisition is used because the data from the heart obtained from the other 180° projections, has reduced resolution (due to increased distance from the detector) and reduced counts (due to an increased photon attenuation in tissue) and thus contributes more noise than useful data to the final images.
To demonstrate cardiac wall motion electrocardiogram (ECG) gating can be used during the SPECT acquisition. Typically, eight frames per cardiac cycle are acquired into a 64 × 64 matrix, for each projection angle. This data can be reconstructed into a single SPECT dataset, or eight individual SPECT datasets which represent the eight frames within the cardiac cycle.
What are the three steps of reconstructio in SPECT to make a 3D image?
The process of reconstruction can be divided into three separate stages, namely creation of 1D profiles from each projection angle, filtering of these profiles and finally, ‘back projection’ of the filtered profiles to create the reconstructed slice image. This can all be done in real space using convolution or in frequency space using Fourier transformation.
WHat is convolution in image reconstruction of SPECT?
There is a need to filter each profile mathematically before back-projection in order to remove the background noise and star effects in the reconstructed image. The filter usually performs some form of smoothing. The filter is applied sequentially to each count value by a process known as convolution.
What does FBP remove from the image?
star artefact
What is a downside to convolution?
The mathematical process of convolution is performed in real space but it is complicated and time consuming. Here real space means ordinary space in which the profiles and the filter are expressed as a graph of counts (or amplitude) against distance (or pixel number). However, it is more elegant and efficient to do the filtering in frequency space. Convolution in real space becomes multiplication in frequency space and from a computation perspective, it is much easier and faster to do.
How does a fourier graph represent data?
The Fourier representation of a profile or a filter is a graph of amplitude against spatial frequency. If distance in real space is expressed in cm for example, the corresponding unit of spatial frequency would be cm-1 (per cm or cycles per cm).
What is the procedure for SPECT imaging with Fourier filtering?
- Acquire projections (profiles of counts against distance in a particular direction) at a large number of projection angles in real space as usual
- Calculate the one-dimensional (1D) Fourier transform (FT) of each projection – the result is an amplitude against spatial frequency
- Choose an appropriate frequency space filter - this is the 1D FT of the corresponding real space convolution filter
- At each spatial frequency, multiply the amplitude of the FT of the projection by the amplitude of the FT of the filter - this yields the FT of the filtered projection
- Calculate the inverse FT for each projection to obtain the filtered projections in real space and arrange as a sinogram
- Back project the filtered profiles in real space as usual
What is the compromise in the choice of reconstruction filter?
Choice of filter is a compromise between noise reduction (degree of smoothing) and preservation of image detail (resolution) in the final reconstructed images.
What is the amount of smoothing in a filter dictated by?
the cut-off or critical frequency of the filter, which is the maximum spatial frequency (minimum spatial dimension or resolution) which is present in the image.
WHat does Iterative reconstruction normally start with?
Iterative reconstruction techniques start with an initial estimate of the reconstruction - usually the FBP solution.
How is attenuation of gamma rays from depth in SPECT corrected for?
The effects of photon attenuation can be corrected either by using analytical techniques or by using direct measurement of the attenuation coefficients within the body:
Analytical
If the linear attenuation coefficient of tissue can be regarded as approximately uniform, an analytical technique can be used to correct for the effects of attenuation eg the Chang method in which attenuation is assumed to vary exponentially with distance in the patient and the edge of the patient can be accurately defined on the images.
Direct measurement
Direct measurement of the distribution of attenuation coefficients within the body, using either an external radioactive transmission source (e.g. gadolinium-153) or an x-ray CT image, enables more accurate attenuation correction of the emission image to be made.
What can the cine view in SPECT be used to check for?
Possible patient motion
Truncation (incomplete projection dataset)
Interference from injection site
Possible attenuation arising from metal objects, arms, breasts etc.
What are the 2 main methods used for creating 3D visualisations from a continuous stack of slices?
Surface rendering
Surface rendering (Fig 1) requires a threshold or segmentation value to be applied to the pixels in the data to define the surface for display. In SPECT this is usually done using an iso-contouring method.
Maximum intensity projection
Maximum intensity projection is the simplest method for 3D visualisation or rendering and produces a semi-transparent display. In maximum intensity projection, the pixel intensity in the 3D object is just the maximum intensity encountered along the projection line from the pixel to the viewing plane. It works best when the object of interest in the image dataset is the brightest.
Using a dual detector gamma camera in SPECT results in:
A. Improved image resolution
B. Increased image counts
C. Shorter acquisition times
D. Attenuation corrected images
The correct answers are B and C.
Compared with a single detector, using a dual detector gamma camera in SPECT results in:
Increased image counts if the total acquisition time and other factors are unchanged
Shorter acquisition times if the total number of counts and other factors are unchanged
Image artefacts can be caused by:
A. Patient movement
B. Lack of image counts
C. Injection site in the FOV
D. Incorrect alignment of the centre of rotation
Image artefacts can be caused by all of these factors.
Artefacts due to a low number of counts are caused by poor signal-to-noise ratio in the projection images. There is usually an accumulation of activity at the injection site and this may appear as a ‘hot’ lesion. Incorrect alignment of the centre of rotation causes ring-shaped artefacts.
Single photon emission CT image resolution is improved by:
A. Using a 128 × 128 matrix
B. Using a 64 × 64 matrix
C. Using a high resolution collimator
D. Using a smooth reconstruction filter
E. Using ‘step and shoot’ acquisition
F. Increasing the administered activity
The correct answers are A and C.
Single photon emission CT image resolution is improved by using a:
128 × 128 matrix because pixel size is smaller
High-resolution collimator because the collimator holes are narrower and longer
Which three of the following are necessary for performing SPECT?
A. A dual detector gamma camera
B. A radionuclide which emits a single gamma photon energy
C. A radionuclide which emits gamma photons
D. A gamma camera system capable of rotating around the patient
E. A radiopharmaceutical specifically for use with SPECT
F. A radiopharmaceutical whose distribution within the patient is fixed
The correct answers are C, D and F.
The following are necessary for performing SPECT:
A radionuclide which emits gamma photons but the photons may have more than one energy
A gamma camera system capable of rotating around the patient to enable the acquisition of projection images at multiple angles
A radiopharmaceutical whose distribution within the patient is fixed during the acquisition of the projection images in order to permit accurate reconstruction of SPECT images
Which three of the following improve SPECT image quality?
A. Integral uniformity <5%
B. High resolution collimators
C. Integral uniformity >10%
D. Dual detector gamma camera
The correct answers are A, B and D.
Poor gamma camera uniformity results in artefacts in the reconstructed SPECT images. The use of high-resolution collimators improves spatial resolution while dual detectors give increased counts (for a given acquisition time) and hence lower noise.
Regarding SPECT reconstruction:
A. For practical purposes, back projection requires the use of a filter
B. The filter removes background counts
C. Filter property affects image resolution
D. Filters can correct for photon attenuation
E. Iterative reconstruction can improve image quality
F. Reconstruction can be performed through any plane
A. True. Simple back projection produces blurring, which is removed by the use of a basic ramp filter.
B. False. A ramp filter removes much of the background due to blurring but it does not remove background counts due to the presence of radioactivity in tissues surrounding the target organ(s) or lesion(s).
C. True. A filter that gives good spatial resolution in the reconstructed images is associated with high noise and vice versa.
D. False. Correction of the emission images for the attenuation of gamma photons is not part of the function of the reconstruction filter. It can be done by assuming a uniform value of linear attenuation coefficient or by using attenuation coefficient maps (in the form of transmission CT images for example).
E. True. In general, iterative reconstruction gives better contrast and spatial resolution than filtered back projection.
F. False. In SPECT, transverse slices are reconstructed although a stack of contiguous transverse slices may be used to create slices in other planes.
Regarding SPECT:
A. It is a useful adjunct to planar imaging
B. Oblique slices are only used in cardiac applications
C. A 3D display is useful for detecting patient motion
D. It is possible to correct for attenuation with CT imaging
A. True. For greatest diagnostic sensitivity, SPECT is usually performed alongside planar radionuclide imaging. A notable exception is myocardial perfusion imaging, for which planar imaging is rarely done.
B. False. Oblique slices are the norm in myocardial perfusion imaging; the slices are oriented with respect to the long axis of the left ventricle rather than that of the body. However, oblique slices may be created from a stack of contiguous transverse slices for any types of SPECT investigation.
C. False. Patient motion during SPECT acquisition may be detected by inspecting the projection data before reconstruction. The projection images may be viewed as a cine loop or they may be transformed into a sinogram or a linogram.
D. True. A registered x-ray CT transmission image provides a map of linear attenuation coefficients, which may be used to correct gamma photon attenuation in the corresponding emission SPECT image.
How is PET like SPECT?
Like SPECT, PET is a tomographic radionuclide imaging technique (Fig 1). Both methods produce digital slice images in which the signal (pixel brightness and/or colour) is largely determined by activity concentration (i.e. the activity of a radiopharmaceutical in the corresponding voxel of tissue). This, in turn, depends on physiological function. These techniques can be used for the non-invasive visualisation of biological and biochemical events at the molecular level in living subjects, an approach that has become known as molecular imaging
How is PET different to SPECT?
Unlike SPECT, however, PET relies on the detection of two gamma photons, specifically the pair of photons that result from the annihilation of a positron and an electron.
What sort of decay do PET tracers undergo?
Of necessity, therefore, PET radiopharmaceuticals are labelled with radionuclides that decay by positron emission. The annihilation photons travel in opposite directions (at 180° to each other) and are detected by opposed radiation detectors. In most implementations of PET, near simultaneous detection of the two photons allows the localisation of their origin to a line joining the detectors without the need for absorptive collimation. This is called annihilation coincidence detection.
What is the product of an anihilation?
During annihilation, the mass of the two particles is converted into energy in the form of annihilation radiation - two photons of energy 511 kiloelectron volts (keV). This is the energy equivalent E of the mass m of a positron or electron according to Einstein’s famous equation E = mc2, where c is the speed of light.
When do radionuclides decay by PE?
Radionuclides tend to decay by positron emission when the nucleus contains a relatively large number of protons compared with the number of neutrons. Such radionuclides are said to be proton rich and are often cyclotron produced.
What is the subatomic process of PE?
Positron emission may be regarded as a proton changing into a neutron with the emission of a positron and a neutrino plus energy. Excess energy released during the decay is shared as kinetic energy of the positron and the neutrino.
What is an important example of PE decay used in PET?
F18 to O18
WHat process competes with PE?
EC as electron capture decay is also equivalent to the conversion of a proton to a neutron
What is the range of a positron?
individual positron paths are tortuous and their exact shape and length unpredictable. This means that the range is also variable from one positron to the next; this is the distance between the point of emission and the point of annihilation. The range also varies because successive positrons are emitted with different kinetic energy. The maximum range depends on the maximum positron kinetic energy but the mean range is significantly shorter (typically about 25% of maximum in soft tissue).
What leads to dose absoprtion in PET?
As they travel through matter, positrons lose most of their kinetic energy through collisions (i.e. interactions with bound electrons) although some energy is lost through bremsstrahlung. The collisions cause excitations and ionisations and constitute the main mechanism for the deposition of radiation dose in PET
How is F18 produced?
cyclotron
What is the half-life of F18?
just under 2 hours
WHat is the most commonly labelled tracer in PET?
Fluorodeoxyglucose is a tracer for glucose metabolism and may be used to image tumours, the brain and the heart after intravenous administration.
What radiopharmiceuticals are used in myocardial perfusion?
myocardial perfusion may be imaged with 13N ammonia and ionic 82Rb (which, like Thallium-201 (201Tl), is an analogue of potassium).
What is the dose to the patient proportional to in PET?
The radiation dose to the patient is proportional to the activity administered and depends on factors such as the physical properties of the radionuclide, the physiological properties of the radiopharmaceutical and its route of administration. Most of the dose to the patient is delivered by the positrons which are non-penetrating charged particles.
What is the effective dose from IV 18FDG?
For intravenously administered 18FDG, the whole-body effective dose is 0.02 millisieverts (mSv) per megabecquerel (MBq).
For tumour and myocardial imaging, the diagnostic reference level (DRL) is 400 MBq giving an effective dose of 8 mSv. For various kinds of brain imaging, the DRL is lower at 250 MBq and this gives an effective dose of 5 mSv. The activity administered to a child should be reduced in proportion to body weight or surface area.
WHat is the staff dose in PET due to?
In contrast to the patient, radiation dose to members of staff and members of the public is due to the highly penetrating annihilation radiation. All PET radiopharmacies and imaging rooms should be designated as controlled areas and fitted with ambient dose rate monitors. They require significantly increased shielding compared with general radionuclide imaging facilities and the same is true for containers used to store and transport radiopharmaceuticals.
Imaging staff should wear personal dosemeters and minimise the time spent in close contact with the patient once the radiopharmaceutical has been administered. The potential dose to a member of staff who declares herself pregnant should be carefully reviewed.
A lactating patient should not breast feed or cuddle her child for several hours after the administration of 18FDG. The decision to perform PET imaging on a pregnant patient should involve a careful assessment of the risks and benefits to both mother and child.
How different is the HVL of lead at 140keV and 511keV?
the HVT of lead at 511 keV is about 15 times greater than it is at 140 keV. 0.408 cm vs 0.027 cm
What is the design of PET detectors?
The annihilation radiation is detected by scintillation crystals. In most designs of PET scanner, the crystals are in the form of blocks that are laid side by side to form one or more rings mounted on a gantry with a horizontal axis along which the patient lies
The rings of detector blocks surround the patient and the external gantry looks very similar to that of an x-ray CT scannerImmediately beyond each end ring is a lead shield and the distance between the shields defines the axial field-of-view (FOV) of the scanner. Typically, this about 15 cm and so this length of the patient is imaged at any given time. The patient couch moves along the ring axis and it may take about 10 couch positions to cover the length of the patient’s body. Other designs of PET scanner are also possible.
What are the requirements for a PET scintillator material?
A high LAC at 511 keV (i.e. high density and effective atomic number)
A high ratio of photoelectric to Compton scatter interactions
A high number of scintillation photons per unit of gamma photon energy deposited (giving good energy resolution)
A short scintillation light decay time
WHy is NaI not used for PET scintillator crystals?
Sodium iodide (NaI) is the scintillator that is usually used in planar radionuclide imaging and SPECT where most of the work is done with 99mTc at 140 keV and gamma photon energies for other radionuclides are less than 400 keV.
However, NaI is not ideal for PET, because its relatively low value of LAC at 511 keV leads to correspondingly low intrinsic detection efficiency. It also has a relatively long scintillation light decay time, which is a disadvantage when the timing of gamma interactions is important.
What is the most common scintillator crystal for PET?
bismuth germanate (BGO) which has a higher LAC because of its higher density and effective atomic number. However, both the scintillation light output and decay time of BGO are inferior to those of NaI.
Newer materials, such as lutetium oxyorthosilicate (LSO) and gadolinium oxyorthosilicate (GSO) have a more suitable combination of properties.
How many PMTs serve each detector block?
A detector block is viewed by four photomultiplier tubes (PMTs) that receive the scintillation light.
How are the detector blocks modified to increase spatial resolution?
Each block is subdivided into an array of detector elements by making partial slits (cuts) with a fine saw (Fig 1). The slits between the elements are filled with a reflective material that serves to reduce the optical cross-talk between the elements.
Each element is a few mm square but each block is several cm thick to increase intrinsic efficiency.
How does annihilation coincidence detection work?
Positron-electron annihilation produces a pair of photons travelling in opposite directions and if these are detected simultaneously by two detector elements in the ring system, the point of origin of the photons must lie along a line joining the detector elements. For each scintillation, the digitised voltage pulses from the four PMTs are used to locate the detector element in which the photon interaction occurred, and they are also summed to generate a signal that represents the total energy deposited by the photon. In the scanner electronics, the time of arrival of each energy signal is recorded to a precision of 1 or 2 ns (where 1 ns = 10-9 s). Two signals are deemed to be coincident when they arrive within a specified coincidence timing window which is typically set at 10 ns (with a range of about 6-12 ns in different scanners). When a coincidence is detected in two elements, the scanner computer identifies the line of response and stores one event in a memory location corresponding to that line. Data acquisition is complete when a sufficient number of coincident events have been recorded (typically many tens of millions).
What is the difference between 2D and 3D data acquisition in PET?
For 2D imaging, coincidences are confined to each individual ring of detector elements and this is achieved through the use of thin annular collimators that are placed immediately in front of the detectors. Unlike collimators in planar radionuclide imaging, collimators in PET do not actually form the image. Instead, they ensure that that the coincidences used to reconstruct a particular transverse tomographic slice are due only to annihilation photons that originated in that slice. The slice thickness is determined by the size of the detector elements in the axial direction.
In 3D acquisition, such collimators are not used and coincidences from a much greater volume of tissue are recorded. This greatly increases the total count rate and the coincidence detection rate compared with 2D acquisition. This mode of acquisition is likely to be most useful in situations in which there is relatively little scattered radiation and/or the administered activity is relatively low (for example, brain imaging or investigations on children).
What are the collimators in 2D PET normally made from?
The collimators are usually made of tungsten which is highly attenuating.
A true coincidence is the simultaneous detection of two unscattered photons from a single annihilation. what are the 2 types of unwanted coincidence?
A scatter coincidence occurs when one or both of these photons are Compton scattered in the patient and both are detected. A random (or accidental or chance) coincidence occurs when photons from different annihilations happen to be detected simultaneously. Scatters and randoms generate coincident events that are assigned to lines of response that do not intersect the real positions of annihilations.
How is some scatter removed in PET?
The energy signals from each quartet of PMTs are subject to energy discrimination so that only signals that lie in the photopeak are accepted. This reduces the number of scatter coincidences but does not completely eliminate them; the photopeak window is wide because of the poor energy resolution of the scintillators at 511 keV. Further scatter correction may be done by estimating the scatter fraction for each line of response.
WHat increases the proportion of random scatter in PET?
The fraction of random coincidences is also greater in 3D compared with 2D acquisition. It increases with administered activity and duration of coincidence timing window. The number of random coincidences may be measured or calculated and subtracted from the total number of coincidences for each line of response.
Why can attenuation of annihilation photons be a problem in PET?
Annihilation photons are attenuated by body tissues (mainly through Compton scattering). Since attenuation is not the same for all lines of response (being greater for those that traverse larger thicknesses of tissue), it may cause non-uniformities and artefacts in reconstructed images
How can attenuation be corrected for in PET?
Attenuation may be corrected by assuming a cross-sectional shape (such as an ellipse) for the patient and a uniform value of μ within this shape. Correction may also be achieved by measuring the transmission of 511 keV photons through the patient for each line of response. This is usually done using a retractable rod source that is parallel to the scanner axis and rotates inside the ring between the detectors and the patient. The source is usually germanium-68 (half-life 288 days) which decays by electron capture to gallium-68 (half-life 68 minutes); the daughter is a positron emitter and hence gives rise to annihilation radiation. A blank or reference scan is made without the patient and a transmission scan with the patient in position; the ratio of blank to transmission gives the attenuation correction factors.
How is normalisation performed in PET?
A typical PET scanner may have 10 000 to 20 000 individual detector elements which may have small variations in factors such as dimensions and the fraction of scintillation light that reaches the PMTs. This means that when exposed to the same radiation source, each line of response does not record the same number of counts. Correction for this effect is called normalisation and it is done using the same rotating rod source as is used for attenuation correction with no objects in the FOV. For a given line of response, the normalisation (correction) factor is the ratio of the measured counts for that line divided by the average counts for all lines.
What is dead time in PET?
A radiation detector is insensitive for a period of time (the dead time) following the detection of a photon. Counts that occur during the dead time are lost.
when is dead time a particular problem in PET?
This is a particular problem in high count rate applications such as PET, in which dead time count losses are greater with 3D acquisitions compared with 2D. The dead time may be measured and observed counts corrected using a mathematical model of detector behaviour.
Why do whole body scans need to correct for decay?
When whole body scans are acquired in several segments (with the same acquisition time) by moving the patient couch, the counts should be corrected for radioactive decay. The longer the delay of a segment acquisition, the greater is the decay factor. Because of the short half-lives of the PET radionuclides, decay correction factors for a given delay are larger than those for 99mTc for example.
How can PET be reconstructed?
both FBP and IR
A contiguous stack of 2D transverse slices produces a 3D image volume that may be reformatted into coronal or sagittal views or, indeed, views at any arbitrary orientation. This is the case irrespective of whether the transverse slices were reconstructed by FBP or iteration
How does reconstruction of 3D PET vary from 2D PET?
In 2D PET, lines of response (LORs) are confined to planes at right angles to the detector axis. In 3D PET, on the other hand, better use is made of the emitted radiation, since LORs are possible between all detector elements within the axial FOV of the scanner.
3D data are also usually stored as sinograms but there are many more of them to accommodate oblique LORs. Several approximate methods are available to convert the data into a set of parallel transverse sinograms so that conventional 2D FBP can be used for reconstruction. However, a 3D reprojection algorithm may also be used to back project filtered data into a truly 3D image matrix from which desired slices may be extracted.
Alternatively, iterative approaches can be easily adapted to reconstructing 3D PET data although the computational complexity increases dramatically.
WHat is contrast determined by in PET?
the differential uptake of the radiopharmaceutical in a lesion compared with healthy surrounding tissue. In planar imaging, the presence of radioactivity in front of and behind the lesion means that image contrast is lower than subject contrast. Tomographic techniques overcome this problem and so, like SPECT, one of the main advantages of PET is a markedly improved image contrast which is particularly important for low contrast lesions.
Random and scatter coincidences add a background to the reconstructed images thus reducing image contrast.
is there structured noise in PET or SPECT?
In planar radionuclide imaging, the presence of radioactivity in front of and behind the lesion is a source of structured noise as well as a cause of contrast reduction. Tomographic techniques such as SPECT and PET largely remove structured noise.
What is random noise defined by in PET?
As in planar imaging and SPECT, random noise is equal to √N where N is total number of counts
What effects the sensitivity of the PET scanner?
count rate per unit activity
Intrinsic detection efficiency. A high intrinsic efficiency requires a high linear attenuation coefficient for the scintillator material and a suitably large detector thickness
Geometric detection efficiency. This is the fraction of emitted annihilation photons that reach the detectors and depends on the solid angle subtended by a particular voxel at the detectors; this is much greater for 3D acquisition compared with 2D
The width of the photopeak acceptance window. A wide window produces greater sensitivity, but it also increases the scatter coincidence rate
How does sensitivity of PET compare to SPECT?
Thicker detectors and the absence of absorptive imaging collimators mean that PET sensitivity is much greater than that of SPECT (by a factor of about 10 for 2D and about 100 for 3D) and so PET images are less noisy than SPECT images.
What determines spatial resolution in PET?
Positron range - the longer the range, the poorer is the spatial resolution. For 15O, the mean range is 2 mm while its value for 18F is 0.6 mm giving better resolution.
Non-colinearity of the annihilation photons - If the positron and the electron have any residual momentum at the time of annihilation, the angle between the directions of travel of the photons will not be exactly 180°. The effect increases with the diameter of the detector ring.
Size of detector elements
The smaller the size, the better is the spatial resolution; 5 mm is a typical element size
Depth of interaction in the detectors
The greater the thickness of the detector, the poorer is the spatial resolution. This effect means that resolution is best at the centre of the detector ring and degrades towards the periphery. This is the reverse of the situation for SPECT.
Reconstruction filter
Since PET has significantly greater count rate sensitivity than SPECT, more counts are acquired and thus noise is less of a problem. This means that PET images are usually reconstructed with a higher spatial frequency cut-off so that their final spatial resolution is superior to that of SPECT images.
What are typical spatial resolutions for PET?
Typical values of axial resolution are 4 mm and 6 mm at the centre and the periphery respectively. Typical transaxial values are 6 mm and 7 mm.
What do many problems with PET scanners give rise to?
linear artefacts in the sinogram of a blank scan obtained by rotating the transmission source
What artefacts can arise from PET detectors?
In SPECT, small non-uniformities in detector response can create noticeable ring artefacts but in PET they have a much smaller impact on reconstructed images. Complete detector failure produces dark linear artefacts in the sinogram but might still have relatively little impact on images, depending on the applied correction and reconstruction algorithms.
What artefacts can attenuation cause in PET?
Reconstruction of PET images without correction for tissue attenuation leads to apparent higher activity in superficial structures.
Cold areas may result from attenuating materials such as metallic joint implants and other objects carried by the patient, but these are usually readily recognisable.
What artefacts can reconstruction cause in PET?
Filtered back projection leads to streak artefacts that are not apparent in iterative reconstruction.
What artefact can patient movement cause in PET?
This can seriously degrade spatial resolution. When performing whole body scans, unusual appearances may result if the patient moves between couch positions.
How can TOF be used in PET?
This is a method of creating a set of 2D tomographic images or a 3D image volume without the need for tomographic reconstruction, which means that each annihilation is directly localised to a voxel in the patient. It relies on measuring the difference in the time Δt at which the annihilation photons reach a pair of opposed detector elements.
Thus a 1 cm depth resolution requires a time resolution of 67 picoseconds (ps) (= 0.067 ns). There exist electronic circuits that are capable of measuring this magnitude of time difference, but available scintillators do not have ideal properties.
With the fastest available scintillators and careful design of electronic components, it is possible to achieve a time resolution of a few hundred ps, corresponding to a spatial resolution of a few centimetres. However, SNRs are improved compared with reconstructed PET images, because there is less noise propagation than in FBP, for example.
When is PET-CT useful?
Image contrast in PET relies heavily on physiological and biochemical function and it is often the case that areas of abnormal uptake are easily identified but difficult to locate within organs and tissues. In these situations, registration and fusion of the PET images with those from a modality that provides better anatomical information (such as x-ray CT) is very helpful. However, if the dual modality images are acquired on different scanners and at different times, it can be difficult to register them with sufficient accuracy.
To address this problem, systems have been developed that incorporate a PET scanner and an x-ray CT scanner in the same physical gantry. Since the patient does not move relative to the couch, accurate image registration is much easier. Furthermore, the CT scans may be used to correct for attenuation in the PET scans, although the average x-ray photon energy is much less than the energy of annihilation photons.
In PET, radiation dose to the patient:
A. Is mainly due to positrons
B. May be reduced by thick lead shielding
C. Is of the same order of magnitude as that in SPECT
D. Is typically 80 mSv for tumour imaging with 18F-FDG
E. Results from the attenuation of annihilation photons by body tissues
A. True. Most of the dose to the patient is delivered by the positrons which are non-penetrating charged particles.
B. False. The radiopharmaceutical is inside the patient so it is not possible to reduce the dose to the patient by shielding. Both PET radiopharmacies and imaging rooms require thick lead shielding, but this is primarily to protect staff rather than the patient.
C. True. Typical effective doses from PET imaging are in the range 5 to 8 mSv.
D. False. Whole tumour imaging with 18F-FDG typically gives an effective dose of 8 mSv.
E. True. Although most of the dose to the patient is delivered by positrons, attenuation of annihilation photons by body tissues also contributes to the dose.
A scintillation detector material for PET should have high values of:
A. Linear attenuation coefficient and light output
B. Light output and light decay time
C. Physical density and light decay time
D. Light output and effective atomic number
E. Effective atomic number and physical density
A. True.
B. False.
C. False.
D. True.
E. True.
The desirable features of a scintillation detector material for PET are high LAC at 511 keV (that is, high density and effective atomic number) and high light output. A high value of light decay time is a disadvantage because the timing of gamma interactions is important in PET.
A designer of a PET scanner can improve its spatial resolution by reducing:
A. The length of the detector elements in the axial direction
B. The range of the positrons in tissue
C. The thickness of the detector elements in the radial direction
D. The non-colinearity of the annihilation photons
E. The length of the detector elements in the transaxial direction
A. True. Reducing the size of the detector elements will improve the spatial resolution.
B. False. Spatial resolution is affected by the range of the positrons in tissue, but this is a physical property and cannot be reduced by the designer.
C. True. Reducing the thickness of the detector elements will improve the spatial resolution.
D. False. Spatial resolution is affected by the non-colinearity of the annihilation photons, but this is determined by the laws of physics and cannot be reduced by the designer.
E. True. Reducing the size of the detector elements will improve the spatial resolution.
What forms of ionizing radiation can radionuclides cause and how are they different?
The radiation may be in the form of particles (e.g. electrons or positrons) and high-energy photons (X or gamma radiation)
Particulate radiation is non-penetrating (i.e. absorbed within a relatively small thickness of material), whereas high-energy photons may penetrate many centimetres
What are the two classifications of hazard in radionuclide imaging?
External radiation hazard - from a source of radiation outside the body, for example gamma rays emitted from a radiopharmaceutical in a syringe. Radiation risks to staff and the public mainly arise from external hazards.
Internal radiation hazard - from a radioactive substance inside the body following injection, absorption through the skin, ingestion or inhalation. This causes the main radiation risk to patients. Internal radiation can also be a hazard to staff and members of the public if radioactive material is allowed to spread into the environment in an uncontrolled manner; this is known as contamination.
How is the hazard different in Xray and radionuclide?
In x-ray imaging…
The hazard is only from external radiation.
In radionuclide imaging…The hazard is from both external and internal radiation.
How are the sources different in xray and radionuclide?
In x-ray imaging… The source of radiation is an x-ray tube in a well-defined location.
In radionuclide imaging… The source of radiation is radioactive material which can spread to different locations, intentionally or unintentionally.
How are the radiation dose different in xray and radionuclide?
In xray - The radiation dose to the patient is confined to the region of diagnostic interest by collimation.
In radionuclide - The radiation dose to the patient is generally not confined to the region of diagnostic interest - it depends on the distribution of the radiopharmaceutical in the body.
How are the radiation source different in xray and radionuclide?
In xray - There is a single main source of radiation, the x-ray tube (although there will also be some sources of scattered radiation such as the patient).
In radionuclide - There are multiple sources of radiation including the radiopharmaceutical (before administration), the patient (after administration), waste material and possibly contamination.
How are the dose rate different in xray and radionuclide?
In Xray - The dose rate is relatively high.
In radionuclide - The dose rate is relatively low and decreases as the activity decays.
How are the radiation production different in xray and radionuclide?
In xray - An x-ray tube only produces radiation while the exposure button is pressed and exposure times are usually short.
In radionuclide - Radioactive material produces radiation continuously so the duration of exposure is potentially much longer, particularly if the material enters the body.
How is the radiation energy different in xray and radionuclide?
In Xray - The effective energy of the radiation is typically around 50 keV.
In radionuclide - The energy of the radiation is higher, for example 140 keV for technetium-99m (99mTc), so shielding may be more problematic. This is particularly so in positron emission tomography (PET) imaging, with photon energies of 511 keV.
What are the 3 principles of dose reduction for external hazards?
Time - minimising the time that a person is exposed to the radiation source as dos = dose rate x time
Distance - The dose rate can be reduced by increasing the distance from the source. For small sources, the inverse square law applies, so doubling the distance will reduce the dose rate by a factor of four. Close to larger sources such as patients, the dose rate does not decrease so rapidly with distance but the general principle still applies.
Shielding - Dense materials of high atomic number such as lead or tungsten are used for shielding. Emitters of higher energy radiation require greater thicknesses of shielding.
What is TVL?
The TVL is the thickness required to reduce the dose rate by a factor of 10, so 3 mm of lead will reduce the dose rate from a 99mTc source by a factor of about 1000.
How does the principles of dose reduction change for internal sources?
The internal radiation dose arising from radioactivity that has entered the body depends on the properties of the radionuclide, its activity, its chemical form, and the rate of uptake and clearance in different organs. Once inside the body, there is usually little that can be done to reduce the radiation dose. For a particular substance, the radiation dose is proportional to the activity. Therefore the main factor in dose reduction is control of the amount of radioactive material entering the body.
the radiation dose to the patient arises almost entirely from internal radiation.
How does dose topatients relate to number and type of images acquired?
The radiation dose does not depend directly on the number or type of images acquired. However, single-photon emission computed tomography (SPECT) studies generally result in higher radiation doses than their planar counterparts because a higher activity is administered to achieve the required image quality.
How is absorbed dose calculated in radionuclide imaging?
If the distribution of the radiopharmaceutical in the body as a function of time is known, the cumulated activity (area under the activity-time curve) in each source organ can be found. Mathematical modelling can then be used to estimate the absorbed dose to each target organ arising from all the source organs. Absorbed dose is expressed in grays (Gy)
How is absorbed dose related to equivalent dose in radionuclide imaging?
Absorbed dose is expressed in grays (Gy) and is numerically the same as the equivalent dose in sieverts (Sv) in radionuclide imaging because the radiation weighting factor is 1.
How is effective dose calculated from equivalent dose in radionuclide imaging?
Each organ’s equivalent dose is multiplied by a tissue weighting factor based on that organ’s relative sensitivity to radiation. The weighted equivalent doses to all organs are summed to give the effective dose from the examination. The unit of effective dose is also the sievert; the millisievert (mSv) and microsievert (μSv) are more useful in practice.
Who specifies DRLs for radionuclide imaging?
the Administration of Radioactive Substances Advisory Committee (ARSAC) of Public Health England, issues guidance on the administration of radiopharmaceuticals to patients. This specifies diagnostic reference levels (DRLs), which are the administered activities that should not normally be exceeded for particular investigations. Adhering to the DRL(read a full definition of this term) ensures that the effective dose to the patient is controlled.
What does IRMER 2017 specify regarding radionuclid imaging?
The Ionising Radiation (Medical Exposure) Regulations 2017 require employers and practitioners to hold a licence for the administration of radioactive substances for a specified purpose at any given medical radiological installation. Licences are only granted to those with appropriate training, facilities and support staff.
How can patients reduce their effective dose?
After administration of radiopharmaceuticals that are excreted in the urine, patients should be encouraged to drink plenty of fluids to increase the rate of clearance and thus reduce the radiation dose. They should also be advised to empty their bladder frequently in order to reduce the radiation dose to the bladder, gonads and pelvic bone marrow.
The radiation dose to the thyroid may be minimised by administration of thyroid-blocking agents when using radioiodide-labelled pharmaceuticals (unless an image of the thyroid gland is required).H
ow can administered activity be scaled for paediatric use?
The activity to be administered may be calculated by scaling down the activity for an adult in proportion to the child’s weight as a fraction of the ‘standard’ adult weight of 70 kg. This will result in an effective dose similar to that for an adult. An alternative method is to scale down the activity in proportion to the child’s body surface area. The ARSAC(read a full definition of this term) Notes for Guidance include a table of factors for this purpose based on the surface area estimated from the weight of the child. Use of these factors will produce an image quality and an imaging time comparable with those expected for adults, but with a higher effective dose.
what is the minimum dose for a paediatric examination?
The ARSAC also recommends that the administered activity for any examination should not be less than 10% of the adult activity. For some investigations, higher minimum activities are specified in order to ensure that adequate image quality is obtained.
What are the recommendations around radionuclide imaging for pregnant patients?
The ARSAC(read a full definition of this term) recommends that a radionuclide procedure resulting in an absorbed dose to the foetus of more than 1 mGy requires particular justification. Many common procedures fall into this category. In cases where radionuclide imaging of a pregnant patient is justified, the foetus will receive an external radiation dose from radioactivity in adjacent maternal organs and may receive an internal radiation dose if the radiopharmaceutical can cross the placenta.
How do doses to staff mainly occur in radionuclide imaging?
Radiation doses to staff mainly arise from the external radiation hazard. Much of the radiation dose is associated with patient handling after the radiopharmaceutical has been administered. The weighted average value for a radionuclide imaging technologist is about 1.5 μSv per procedure, or around 2 to 3 mSv per year depending on the department’s workload and case mix.
How is staff hand dose reduced in radionuclide imaging?
When injecting a radiopharmaceutical, the external radiation dose to the skin of the hand and fingers can be reduced by a factor of 10 or more by using a syringe shield
How is internal staff dose minimised?
Staff will receive an internal radiation dose if radioactive material enters the body by absorption through the skin, ingestion or inhalation. It is therefore important to minimise the spread of radioactive materials into the working environment by taking the following precautions:
Restrict access to areas where unsealed sources are handled
Wear appropriate protective clothing
Use careful working practices
Dispose of radioactive waste properly
Decontaminate promptly after a spill
WHy can skin exposure cause such high localised skin dose?
The exposure can continue for a long time, particularly if the contamination is not recognised and/or the substance is difficult to remove
The distance from the source to the skin is very small
There is no shielding between the source and the skin
What is the dose limit for members of the public?
1 mSv
What should breast feeding mothers do regarding radionuclide imaging?
If a patient is breast feeding, consideration should be given to postponing the examination or using an alternative technique. If it is clinically necessary to proceed with the examination, then the following precautions should be taken:
Express milk and ‘bank’ at least one feed prior to the test
Feed the baby just before the test
Express breast milk 3-4 hours after the test and discard
After some examinations, interrupt breast feeding for a further specified period, between 12 and 48 hours depending on the radiopharmaceutical (Table 1)
After administration of 131I, stop breast feeding altogether
What precautions should be undertaken during radionuclide administration?
Wearing disposable impervious gloves
Using a syringe shield
Placing disposable absorbent material to catch potential spills
Verifying venous access first (e.g. with a cannula or butterfly needle), then injecting the radiopharmaceutical and flushing with saline; this will minimise the time that the syringe needs to be handled and will also reduce the risk of extravasation of the radiopharmaceutical
What precautions should be undertaken after radionuclide administration?
Use a contamination monitor to check the area and yourself for activity (Fig 1)
Decontaminate if any spills have occurred
Dispose of the ‘empty’ syringe appropriately; note that it will contain residual activity
Dispose of waste such as gloves and swabs appropriately
Wash your hands
Ensure that a record of the administration is made in the patient notes
What is the effective dose range for most adult radionuclide studies?
Most radionuclide imaging examinations of adults involve effective doses in the range 1 mSv to 12 mSv
Regarding a syringe shield used when injecting a radiopharmaceutical:
A. It may be made of tungsten
B. It reduces the external dose rate from a syringe by a factor of 10 or more
C. It reduces the risk of contamination
D. It allows the clinician to inject safely without wearing gloves
E. It reduces the internal radiation dose to the patient
A. True. A syringe shield is typically made of a heavy metal such as tungsten, with a lead glass window.
B. True. A syringe shield reduces the external dose rate from a syringe by a factor of 10 or more.
C. False. Contamination may occur if the radiopharmaceutical is spilt during the injection. The syringe shield does not prevent this.
D. False. Impervious gloves must always be worn when injecting a radiopharmaceutical. They protect the clinician against the risk of contamination.
E. False. The use of a syringe shield will reduce the external radiation dose to the patient but has no effect on the much larger internal radiation dose.
What is the difference between quality assurance and quality control?
QA = All those planned and systematic actions necessary to provide confidence that a system or process will perform satisfactorily in service.
QC = The set of operations (programming, coordinating, carrying out) intended to maintain or improve quality.
What are the performance parameters for gamma cameras?
Uniformity
Spatial resolution
Linearity
Sensitivity
Count rate capability
Energy resolution
Centre of rotation
What does uniformity of a gamma camera assess?
the ability of the gamma camera to produce an image where count values are equal in every pixel, when irradiated by a uniform source.
What are the two measures of uniformity of a gamma camera?
There are two measures of uniformity:
Intrinsic uniformity - measured with the collimator removed. measured with a point source positioned on the central axis of the gamma camera detector at a distance equal to 5.5 times the diameter of the detector. This assesses the response of the detector with no contribution from irregularities in the collimator
System uniformity - measured with the collimator in place. measured with a large-area uniform radioactive source known as a flood source. This may be either a liquid-filled phantom to which 99mTc can be added, or a solid source consisting of cobalt-57 (57Co) embedded in plastic. This assesses the response of the imaging system as used clinically
Gamma camera uniformity performance will depend on the spatial linearity and energy response of the system.
Daily/weekly measurements should be made qualitatively and quantitatively to detect both gross abnormalities (e.g. failure of a PM tube) and degradation of performance over time.
What is made from uniformity tests on gamma cameras?
The calibrations and correction ‘maps’ necessary are produced either at manufacture, or as part of the commissioning process (sensitivity maps may be re-made on a more regular basis, if uniformity measurements become unsatisfactory).
What makes up system spatial resolution of a gamma camera?
The system spatial resolution (Rs) of the gamma camera is a combination of the collimator resolution (Rc) with that of the detector assembly itself (the intrinsic resolution (Ri)).
Rs^2 = Ri^2 + Rc^2
How is spatial resolution of the gamma camera measured qualitatively?
Qualitative measurements are made using a suitable transmission phantom in combination with a flood source for system measurements or a point source at a distance of five times the diameter of the crystal for intrinsic measurements. Two such phantoms are the:
Anger pie phantom
Quadrant bar phantom
A phantom image is inspected to find the minimum resolvable detail diameter or bar spacing.
How is spatial resolution of a gamma camera measured quantitavely?
imaging a line or point source. A profile taken through the image gives the line spread function (LSF) or point spread function (PSF). The full width at half maximum (FWHM) of the LSF (or PSF) is quoted as a measure of spatial resolution.
The FWHM is equal to the minimum separation required between two line sources if they are just to be resolved (the Rayleigh criterion).
The full width at tenth maximum (FWTM) may also be quoted.
How does spatial resoluition change with depth?
it gets worse with increased depth
What is spatial linearity?
Spatial linearity measures the spatial distortion of an image.
The quantitative assessment of spatial linearity requires the use of phantoms with parallel equally spaced lines or a regular array of holes.
What is sensitivity of a gamma camera and how is it measured?
Sensitivity is a measure of the proportion of gamma rays emitted from a radionuclide source which is detected within the photopeak of the collimated gamma camera.
It is measured by imaging a small phantom containing a known amount of radioactivity (measured in MBq), for a known time (measured in seconds) at a distance of 10 cm from the camera face.
The background corrected number of counts in the image is found and the sensitivity then expressed as counts per second per MBq (cps/MBq).
WHat is count rate capability of a gamma camera?
the ability of the gamma camera to register count rate linearly in response to incident count rate, even as count rates increase.
The measure of count rate capability is the 20% count rate loss - that is the count rate at which the recorded count rate is 20% lower than the expected count rate. Counts are lost due to the dead time of the detector.
Why cannot all mono-energetic photons be recorded at the same energy in the crystal?
due to Compton scattering interactions within the crystal, the spectrum of pulse heights (Z) has the shape shown in Fig 1, with a photopeak of a finite width and a Compton band to the lower energy side of the photopeak. The photopeak has finite width because a range of Z values (proportional to deposited energy) is produced even when all of the photon energy is deposited in the detector.
What is energy resolution?
Energy resolution is a measure of the width of the photopeak at half of the maximum amplitude, expressed as a percentage of the photopeak energy.
For example, for 99mTc:
FWHM = 11.5 keV
Energy = 140 keV
Energy resolution = 8.2%
How is COR measured?
only for SPECT
The COR is usually measured from a series of images of a small point source (~20 megabecquerel (MBq) 99mTc in 0.1 ml). These are taken from typically 120 different angles around the source, positioned in the centre of the FOV of the camera head, offset by ~15 cm from the axis of rotation (in the X direction).
Using data from the images, graphs of the X and Y co-ordinates (pixel positions) of the point source are plotted against gantry rotation angle.
ideally the offset should be less than the equivalent of ±1 mm.
What are the typical values for energy resolution?
<10%
What are the typical values for sensitivity?
System sensitivity (low-energy high-resolution collimator) > 60 cps/MBq
System sensitivity (low-energy general purpose collimator) ~ 150 cps/MBq
What are the typical values for system spatial resolution?
System spatial resolution (FWHM at 10 cm, low-energy high-resolution collimator) ~ 7 mm
System spatial resolution (FWHM at 10 cm, low-energy general purpose collimator) ~ 8 mm
A solid plastic flood source, suitable for uniformity measurements, contains principally which isotope?
he long half-life of 57Co makes it suitable for incorporating into a solid flood source. It emits photons with an energy of 122 keV and can therefore be used as a convenient alternative to 99mTc.
Gamma camera sensitivity is:
A. A measure of the sharpness of the image produced by the gamma camera
B. The ability of the gamma camera to register count rate linearly in response to incident count rate
C. The ability of the gamma camera to produce an image where count values are equal in every pixel, when irradiated by a uniform source
D. A measure of the proportion of gamma rays emitted from a radionuclide source which is detected in the photopeak of the collimated gamma camera
A. Incorrect. Spatial resolution is a measure of the sharpness of the image.
B. Incorrect. Count rate capability is the ability of the gamma camera to register count rate linearly in response to incident count rate.
C. Incorrect . Uniformity is the ability of the gamma camera to produce an image where count values are equal in every pixel, when irradiated by a uniform source.
D. Correct. Sensitivity is a measure of the proportion of gamma rays emitted from a radionuclide source which is detected within the photopeak of the collimated gamma camera.
The following aspect of performance should be checked daily:
A. Uniformity
B. Energy resolution
C. Spatial linearity
D. SPECT COR
A. Correct. A qualitative check of uniformity should be made daily.
B. Incorrect. Energy resolution usually only needs to be checked annually unless a fault is suspected.
C. Incorrect. Spatial resolution usually only needs to be checked at acceptance testing, or if other tests show unusual results.
D. Incorrect. Checking of the SPECT COR usually only needs to take place monthly unless a fault is suspected.
What can cause major non-uniformities in the gamma camera image?
Photomultiplier (PM) tube failure
Cracked or broken scintillation crystal
What can cause minor non-unifority in the gamma camera?
Minor non-uniformities can result in either photon deficient or photon enhanced areas within an image, which are not necessarily obvious on clinical images. For example:
Energy and/or linearity calibration incorrect or not applied
Collimator damage (e.g. a few damaged septa)
Incorrect user sensitivity map applied
What acquisition and/or processing parameters can result in artefact?
Incorrect energy window (e.g. cobalt-57 (57Co) used for imaging 99mTc)
Incorrect collimator (e.g. low energy collimator used to image a medium or high energy radionuclide)
Insufficient counts due to incorrect image acquisition time (e.g. 10 seconds instead of 100 seconds)
Count loss in final frames of a gated cardiac study due to incorrect R-R acceptance window
Inappropriate use of a smoothing filter on low count images, including SPECT, resulting in ‘blob’ type artefacts
Incorrect application of image display windowing in cases of areas with high count density (e.g. full bladder in a bone scan)
Regarding radionuclide image artefacts:
A. All image artefacts are avoidable if correct equipment and technique are used
B. Image artefacts always show as ‘hot spots’ on the image
C. Image artefacts may be indistinguishable from true clinical/pathological appearances of an image
D. Image artefacts may be related to equipment malfunction
E. Image artefacts may be caused by patient movement
A. False.
B. False.
C. True.
D. True.
E. True.
Radionuclide image artefacts are not caused by:
A. Equipment malfunction
B. Poor radiopharmaceutical production
C. External radiopharmaceutical or urine contamination
D. Environmental conditions
E. Incorrect patient identification
The correct answers are D and E.
Image artefacts are not caused by environmental conditions, such as temperature or atmospheric pressure, unless changes in temperature are rapid enough to cause damage to the scintillation crystal of the gamma camera. Manufacturers recommend that temperatures should not be allowed to change by more than 2°C per hour, so gamma camera rooms should be temperature controlled.
Incorrect patient identification will not lead to artefacts on the image, although of course the clinical relevance of the images will be lost, leading to the possibility of incorrect diagnosis.
Incorrect gantry centre of rotation calibration for SPECT imaging will result in:
A. ‘Hot spots’ on the reconstructed slices
B. Photon deficient areas on the reconstructed slices
C. ‘Star’ artefacts on the reconstructed slices
D. Photon deficient or photon enhanced ‘ring’ artefacts on the reconstructed slices
E. ‘Hot spots’ on the raw projection data
The correct answers is D.
In SPECT systems, incorrect gantry centre of rotation calibration, or detector head mis-alignment will result in non-uniformities within the reconstructed SPECT slices. These generally appear as either photon deficient or photon enhanced ‘ring’ artefacts.
Photon attenuation artefacts may be caused by:
A. External radiopharmaceutical contamination
B. Coins in the patient’s pocket
C. Extravasation of the radiopharmaceutical injection
D. Barium-based contrast medium from previous x-ray examinations
E. Patient pacemaker
The correct answers are B, D and E.
Photon attenuation artefacts may be caused by any high density objects such as coins, belt buckles, jewellery and mobile phones, which should be removed from the patient prior to imaging, and also implants, such as pacemakers and aneurysm clips, which cannot be removed.
High density contrast media, such as those based on barium, will also cause attenuation artefacts.
Extravasation of the radiopharmaceutical injection can lead to image artefacts due to:
A. Interference from intense photon count at the injection site
B. Injection site mimicking uptake
C. Insufficient counts in the organ/target of interest
D. Increased attenuation
E. Extra activity in the bladder
The correct answers are A, B and C.
Extravasation of the radiopharmaceutical injection leads to a large amount of the radiopharmaceutical being left around the injection site, and a reduced amount entering the blood stream.
This means that in severe cases there is intense photon count at the injection site, which can obscure parts of the image and lead to problems with windowing the image correctly.
In less severe cases the injection site could mimic an area of uptake. The reduction of injection entering the circulation means that there will ultimately be less uptake and therefore insufficient counts in the area/organ of interest.
