Fig. 9.1
X-ray imaging. X-rays are generated by a tube, travel through the patient and are detected by a detector placed behind the patient. a Projection radiography; a 2D image is taken, e.g. a chest X-ray. b Fluoroscopy; this is a real-time technique in which low-level X-rays are detected by an image intensifier placed underneath the patient couch. c CT scanning; X-rays are taken continuously while the tube/detector travels around the patient. d Rotational angiography; a fluoroscopy system rotates around the patient in a similar manner to a CT system
Fluoroscopy. X-rays are produced and detected continuously forming a 2D real-time image sequence (Fig. 9.1b). The X-ray beam intensity is much less than for projection radiography resulting in images which are noisier, but of sufficient quality to allow real-time visualisation. Fluoroscopy is widely used in examination of the arterial and venous systems and the chambers of the heart, where it is called ‘angiography’. As noted above the contrast between soft tissues is low and blood vessels would not show up without the use of an iodine based contrast agent which is injected into a vein or (less commonly) an artery. Fluoroscopy may be used purely for diagnosis, to investigate the location of any diseased region. More commonly fluoroscopy is combined with intervention in which the diseased area is identified then treated. In balloon angioplasty a catheter is inserted into the artery and directed to the area of stenosis then a balloon is inflated under high pressure (several atmospheres) resulting in increase in local lumen diameter with increase in blood flow. A stent may also be placed to help prevent restenosis.
Computed tomography (CT). In CT, X-ray data is collected at different angles around the patient (Fig. 9.1c). This is achieved by rotation of the tube and detector with typically 1000 datasets collected for a full 360° rotation. The individual projections are combined in the computer using a method called ‘back projection’ to form a 2D cross-sectional image in which the displayed data is related to the attenuation coefficient of the tissues. Early CT scanners collected 1 slice at a time and a 3D dataset was built up by sequential series of slices with the patient couch moved between each slice. In modern CT scanners data is collected continuously with constant rotation of the tube-detector around the patient and continuous movement of the patient through the CT system; so-called helical scanning. Further decrease in acquisition time is enabled by multi-slice detectors in which several slices can be collected simultaneously. Typically 64 slices are collected continuously. With helical scanning and multi-slice detectors a full thorax and abdominal scan takes a few seconds. This is a sufficiently short time to enable the patient to hold their breath during scanning which helps reduce registration artefacts associated with breathing in by different amounts for each slice. Modern CT scanners can acquire up to 320 slices simultaneously (up to 16 cm coverage in one rotation), or combine very fast imaging techniques with dual-X-ray source technology to significantly improve temporal resolution. These techniques allow collection of a time-gated heart scan in one cardiac cycle and total body angiography with single contrast injection protocol.
Rotational angiography. This is a variant of angiography in which there is rotation of the tube and detector around the patient through 180° (Fig. 9.1d). The data sequence can be used to view the arteries from a series of different angles. Alternatively the data can be reconstructed to produce a 3D image of the vascular system.
9.1.2 Magnetic Resonance Imaging (MRI)
MRI is based on control of the magnetisation of the nuclei of atoms within tissues. The nuclei in some atoms act as small magnets. This arises from the spin of the protons and neutrons which make up the nucleus. Pairs of spins tend to cancel. However where the atom has an odd number of neutrons or protons (or both) there is a net spin. In the human body the main atoms (by relative number) are: hydrogen (62 %), oxygen (24 %), carbon (12 %) and nitrogen (1.1 %). The most common isotopes of both oxygen and carbon have even numbers of protons and neutrons and hence no spin. Hydrogen has a single proton and has a spin of 1/2. Nitrogen has 7 neutrons and 7 protons and a spin of 1. Due to its abundance in the body (62 %) it is the hydrogen nucleus which gives rise to the signal in an MRI scanner; or in other words the vast majority of MRI is hydrogen (proton) imaging, and is associated with detection of water (H2O).
Fig. 9.2
Behaviour of a magnetised nucleus (e.g. hydrogen atom) in a magnetic field. The magnetic moment will precess around the magnetic field
Underpinning Physics. If placed in a magnetic field the proton will precess; that is the magnetic axis of rotation sweeps out a cone around the magnetic field (Fig. 9.2). This is similar to the way a spinning top will behave where the axis of rotation rotates around the gravitational field. The rate of rotation is called the Larmor frequency and is equal to about 43 MHz per tesla (T); tesla is the SI unit of the magnetic field strength. For 1.5 and 3 T fields (the most common in clinical practice) the Larmor frequency is 64 and 128 MHz, respectively, which is in the radiofrequency range (i.e. the frequency range in the electromagnetic spectrum which consist of radio waves).
The signal used to form the MRI image arises due to the hydrogen nuclei absorbing energy from the MRI scanner then reemitting energy which is detected. A large magnet is used to align the magnetic moments of the hydrogen atoms (Fig. 9.3a). A coil transmitting in the radiofrequency range is used to change the direction of magnetisation of the protons (Fig. 9.3b). The degree of tilt of the magnetic moment depends on the duration and amplitude of the RF pulse. Only those nuclei precessing at the same frequency as the RF frequency are affected (hence the term ‘resonance’ in MRI). Once the RF pulse is switched off the nuclei will gradually return to their original magnetisation (Fig. 9.3c–f). In energy terms the protons are pushed into a higher energy state by the RF pulse. Once the RF pulse is switched off they transition (‘relax’) to a lower energy state and in the process emit energy. This energy is detected by the MR system, and this forms the MR signal; which is used to produce the MR image.
Fig. 9.3
RF pulse and subsequent relaxation. a The magnetic moment M o is aligned with the main magnetic field. b an RF pulse cause a 90° shift in alignment and the magnetic moment precesses; this is taken into account by the use of a rotating (x′, y′) frame of reference. c–f The longitudinal magnetization M z gradually recovers to its full strength while the transverse magnetization M y gradually dephases and reduces to zero. g–h Change in longitudinal and transverse magnetization with time
The detected signal is the composite of many individual protons relaxing. During relaxation, the magnetisation recovers in the direction of the magnetic field (T1 or spin-lattice relaxation) (Fig. 9.3g) and decays transverse to the field (T2 or spin-spin relaxation) (Fig. 9.3h). The T1 relaxation time is the time for the longitudinal magnetization to achieve 63 % of its equilibrium value. T1 is a measure of the rate of energy exchange between the protons and the neighbouring molecules. The T2 relaxation time is the time at which there is 37 % loss of the original transverse magnetization. T2 is a measure of the rate of decay of the transverse magnetization. T2 relaxation is associated with gradual loss of phase coherence of the protons. After the initial RF pulse the magnetic moments of the protons are all in phase. Gradually these become out of phase as a result of T2 relaxation. T1 and T2 have different values for different tissues and this provides the principle image contrast in MRI, along with the MRI signal strength which is related to the number of protons imaged (proton density). Values of T1 and T2 in tissues are provided by Selwyn (2014); see Table 12.3. Generally T1 values are longer than T2 values (for the same tissue) by a factor of 10–20.
Operating principles. The MRI scanner usually consists of a large magnet in the shape of a tube capable of generating field strengths of 1.5 or 3 T. The patient is positioned within the magnet. The magnetic field affects the alignment of the hydrogen nuclei so as to produce a net magnetisation along the direction of the field. A transmitting RF coil is used to tilt the magnetic moment of the hydrogen nuclei. When the alternating field is switched off the proton magnetic moments relax to their equilibrium orientation, emitting energy. This energy is detected by receiving coils positioned around the patient. Positional information is obtained by the use of gradient coils. These produce a smooth change in magnetic field strength and hence in Larmor frequency along different directions within the magnet bore. Gradient magnetic fields are switched on and off as needed during the image acquisition process. Information on spatial location is therefore encoded into the received signal because of these differences in Larmor frequency. There are three gradient coils; one for each direction; x, y and z. A full account of spatial encoding in MRI is beyond the scope of this book. However, the process allows acquisition of image slices in any orientation within the magnet bore, and of 3D as well as 2D data sets. These are very useful capabilities for cardiovascular imaging, as they allow for example acquisition of images aligned with the long and short axes of the left ventricle.
The series of radiofrequency pulses and gradients employed during MR image acquisition is known as a ‘pulse sequence’. Different sequences produce images with a variety of geometrical, temporal and contrast properties, and development of new pulse sequences remains a very active research area. It is possible, for example, to produce images of the heart in which blood appears dark (‘black blood imaging’) or bright (‘bright blood imaging’), and these have different clinical applications. It is also possible to image the beating heart in real time (‘MR fluoroscopy’) and to produce images that are sensitive primarily to flowing blood (‘MR angiography’). Contrast agents used in MRI are chosen for their magnetic properties, rather than for high atomic number as in the case of X-ray imaging. These contrast agents can be particularly useful in MR angiography, but also for assessment of myocardial perfusion and viability (see also Sect. 9.1.5). There are also a variety of quantitative MR methods, such as mapping of T1 and T2 relaxation times which has recently undergone something of resurgence particularly for the investigation of diffuse cardiac disease.
9.1.3 Ultrasound
Essential physics: Ultrasound for medical diagnostic use consists of high frequency sound waves in the frequency range 2–20 MHz. Ultrasound is generated by a transducer which is in contact with the patient and the waves pass into the patient along a thin beam. The waves are scattered by the tissues, with some of the energy returning back to the transducer where the waves are detected. The depth D from which the scattered ultrasound arises is calculated from the time T R between transmission and reception, assuming that the speed of propagation c is 1540 m s−1 (Eq. 9.1). This is called the pulse-echo technique (Fig. 9.4).
Fig. 9.4
The pulse-echo technique for ultrasound using a simplified model of 2 tissues. a Transmission. An ultrasound is transmitted by the transducer at time zero. The pulse travels along a well defined beam. b Some of the acoustic energy is reflected at the interface between the tissues and travels back along the beam. c The reflected ultrasound is detected by the transducer and the machine notes the time (T R ). The ultrasound machine calculates the depth D from which the echoes arose as the speed of sound (assumed to be 1540 m s−1 multiplied by T R /2)
(9.1)
The amplitude of scattered ultrasound is determined by local differences in acoustic impedance within the tissues. These differences are especially high at the boundaries of organs resulting in high amplitude scattering.
Image formation: The amplitude of the received ultrasound is displayed in grey scale on the image; called a B-mode (brightness mode) image. The transducer sweeps the beam through the tissues and the image is built up line by line (Fig. 9.5). A 2D image typically is collected in 15–50 ms which is equivalent to 20–70 frames/second. This frame rate is sufficiently high for the image display to appear in real time. Higher frame rates can be achieved by collecting several scan lines simultaneously.
Fig. 9.5
Building up a single ultrasound frame line by line
Most ultrasound in clinical practice is based on the acquisition of real-time 2D images as described above. It is also possible to collect 3D ultrasound data (Prager et al. 2010; Fenster et al. 2011). This may be achieved through collection of a series of 2D images which from a technical point of view is the easiest way. The existing transducers for 2D scanning are adapted to enable them to collect 3D data. Typically this involves mechanical oscillation of the transducer to and fro to build up the 3D image. Dedicated transducers designed specifically for 3D imaging are able to steer the ultrasound beam within a 3D volume enabling collection of 3D data without the need for mechanical components. Using a single-beam system the number of volumes collected is low at 1–2 per second. Applications where real-time visualisation of 3D movement is required include cardiac scanning. Commercial ultrasound systems for cardiac scanning can achieve over a hundred volumes per second by simultaneous collection of many scan lines simultaneously.
Doppler ultrasound: Information on motion of blood may be obtained by Doppler ultrasound. The Doppler effect is familiar within everyday living; it is the difference between the transmitted and perceived frequency of sound when there is motion of either the source or the observer. An everyday example is the change in pitch of the siren from an ambulance as it passes the observer. In ultrasound if the transmitted ultrasound has a frequency f t and the tissue is moving with a velocity v then the ultrasound received by the transducer will have a slightly different frequency f r . The difference in frequency (f r − f r ) is called the Doppler shift f d . The Doppler shift is related to the velocity and the direction of motion as illustrated in Fig. 9.6 and shown in Eq. 9.2.
where θ is the angle between the Doppler beam and the direction of motion.
Fig. 9.6
Schematic for Doppler ultrasound estimation of blood velocity. Blood of velocity v travels in a direction which makes an angle θ with respect to the Doppler beam
(9.2)
(9.3)
The Doppler shift can be used to measure velocity of the tissues (Eq. 9.3). This requires measurement of the Doppler frequency shift f d and of the angle θ. Commercial systems adopt 2 main display modes which use Doppler shift data. Spectral Doppler is a real-time display of Doppler frequencies versus time from blood flow; essentially this is the time–velocity waveform. Colour flow is a real-time display of the pattern of blood flow with the mean Doppler frequency displayed at each pixel.
Endoscopy: In clinical practice the vast majority of ultrasound examinations involve the use of transducers placed on the skin, so-called transcutaneous imaging. It is also possible to collect ultrasound images from inside the patient. Intravascular ultrasound involves a miniature ultrasound system at the distal end of a catheter inserted in the blood stream via an arterial puncture, and is described in the section on ‘catheter based imaging’ (9.1.7). An endoscope is a long flexible instrument which is inserted into a body orifice such as the oesophagus for examination of the tissues from the inside. An ultrasound transducer may be incorporated at the distal end of the endoscope allowing ultrasound imaging of the tissues. For cardiovascular imaging the most widespread endoscopic technique is TOE or transoesophageal echocardiography. The oesophagus passes very close to the heart, so that very high quality cardiac images can be obtained using TOE compared to transcutaneous imaging. Clinically TOE is most commonly used to help guide surgical procedures such as cardiac valve replacement and in haemodynamic monitoring during surgery.
9.1.4 Gamma Camera Imaging and Positron Emission Tomography
The term ‘nuclear medicine’ refers to techniques used to diagnose and treat clinical disorders using radioactive isotopes. The two main imaging techniques in nuclear medicine are gamma camera imaging and positron emission tomography (PET). Both of these imaging techniques involve the injection or inhalation of radioactive chemicals containing a radioactive atom. The radioactive atoms subsequently decay resulting in the production of gamma rays (high energy electromagnetic particles). It is the detection of the gamma rays which forms the basis for imaging in gamma camera imaging and PET.
Radioisotopes used in nuclear medicine have short half-lives so do not occur naturally. Some, particularly those used in PET, are produced in a cyclotron. If there is no on-site cyclotron then radioisotopes must be transported from the cyclotron to the nuclear medicine department where the imaging takes place. In some cases a parent radioisotope is produced with a relatively long half-life which decays into the isotope of interest; this is referred to as a ‘generator’. By far the most common radioisotope used in gamma camera imaging is technetium 99m (or Tc-99m) which has a half life of 6 h. In practice a molybdenum-99 generator is produced; Mo-99 is a fission product from nuclear reactors and has a half life of 66 h and decays to Tc-99m allowing Tc-99m to be drawn off over a period of several days. Tc-99m may be bound to several different molecules and used for imaging of different organs (e.g. lung, bone, heart, liver, thyroid). Other isotopes used in gamma camera imaging include iodine (I-123 or I-131) for thyroid imaging, indium (In-111) for white cell imaging and gallium (Ga-67) for imaging of inflammation. Typical radioisotopes used in PET are oxygen (O-15), carbon (C-11), nitrogen (N-13) and fluorine (F-18). These have short half-lives of 2–100 min. These radioisotopes may be used unaltered in PET or they may be incorporated into molecules for use in scanning; for example water (O-15), ammonia (N-13), acetate or carbon dioxide (C-11) and, the most commonly used PET tracer, fluorodeoxyglucose (FDG) (F-18).
In gamma camera imaging gamma rays from the patient are detected one gamma ray at a time using a 2D detector (Fig. 9.7a). Over time, typically a few minutes, an image is built corresponding to the distribution of radioactivity within the patient. If the camera is rotated around the patient then the data acquired may be used to generate a 3D image in a similar manner to CT scanning; this is called SPECT (single-photon emission computed tomography) (Fig. 9.7b). The article by Rahmima and Zaidib (2008) compares PET and SPECT.
Fig. 9.7
a Gamma camera imaging; radionuclides in the patient decay producing gamma rays which are detected by the camera. b SPECT; the gamma camera is rotated around the patient and from the acquired data a 3D reconstructed image is produced
In PET, the decay of each radioactive atom produces a positron. The positron is the anti-particle of the electron so has a positive charge. The positron travels for a short distance from its point of creation gradually slowing down until it collides with an electron. The mass of the positron and electron is completely converted into energy (‘annihilation’) resulting in the formation of two gamma rays which travel in opposite directions (Fig. 9.8a). In PET the gamma ray pair is detected by sensors which surround the subject (Fig. 9.8b). A line may then be drawn between the two detection points and the gamma rays must have originated somewhere along this line. Data is collected over many millions of annihilation events and is reconstructed in the computer in a similar manner to CT imaging to obtain a 2D cross-sectional image of the number of events occurring at each location within the image. In practice a 3D dataset is collected. A whole body dataset may be collected if the patient/volunteer is passed slowly through the detectors. Typical scan times are 20–60 min. The distance the positron travels before annihilation is 1–3 mm. This places an inherent limitation on the accuracy of localisation of the origin of the positron, and hence on the achievable spatial resolution which is typically 4–6 mm in the human. In practice a low radiation dose PET scan is followed immediately by a CT scan (delivered within the same imaging system) for the purposes of attenuation correction and localisation of PET uptake to local anatomy.
Fig. 9.8
Principles of PET imaging. a A radioactive nucleus decays releasing a positron. This travels a short distance then collides with an electron undergoing annihilation and two gamma rays are produced which travel in opposite directions. b In a patient a radioactive nucleus decays and the gamma rays are detected by the detector
9.1.5 Contrast Agents for Ultrasound and MRI
The term ‘contrast agent’ is used to refer to a material which when injected into the patient provides a significant change (increase or decrease) in the imaged quantity enabling improved visualisation of particular structures; or in other words an agent which improves image contrast. Often a contrast agent is simply associated with an increase or decrease in received signal strength. Contrast agents for X-ray imaging were noted above where an iodinated compound is injected into the vascular system to enable visualisation of arteries and veins, and where barium is swallowed by mouth or pumped into the rectum to enable visualisation of the gastrointestinal system.
Ultrasound contrast agents. For ultrasound, contrast agents are based on microbubbles (Fig. 9.9) (Sboros 2008). These are spherical consisting of a gas encapsulated by a thin shell. These are usually injected into a vein in the arm; they pass through the lung capillary bed; enter the systemic arterial system and travel to the organ of interest where they are imaged using ultrasound. Microbubbles are commonly manufactured with diameters in the range 2–6 μm in order that they can pass through the capillaries in the lung. The behaviour of the microbubble in the acoustic field depends on the peak acoustic pressure (Fig. 9.10). At low pressure the bubble behaves in a linear manner with small oscillation. At higher peak pressure the bubble will oscillate with non-linear motion including resonance. This non-linear motion induces additional (harmonic) frequencies in the scattered ultrasound. At very high peak pressure the bubble shell will burst. A number of display modes are designed to take advantage of these different behaviours; examples are as follows.
Fig. 9.9
An ultrasound contrast agent
Fig. 9.10
Behaviour of an ultrasound contrast agent with increasing acoustic power
Non–linear imaging. Intermediate acoustic pressure is used which induces bubble resonance and non-linear behaviour. The scattered ultrasound contains harmonic frequencies which are detected and displayed.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
