Contents

PART ONE Physical Principles of MRI

1 Basic Principles

The hydrogen proton

Transmission of the MR signal

Image weighting and contrast

Spin echo and scan time

Spatial localisation

Fourier transformation

Pulse sequences: the quest for speed

2 MR System Hardware

Magnets

Shielding

Shimming

Gradients

Radiofrequency system

Faraday cage

Image processing and archiving

3 Imaging Parameters

Signal and noise

Contrast

Artefacts

4 Image Artefacts: Recognition and Reduction

Movement artefact

Phase wrap or aliasing

Frequency wrap

Chemical shift

Chemical misregistration

Susceptibility artefact

System-induced artefacts

Partial volume averaging

PART TWO MRI in Veterinary Practice

5 MRI Safety

Magnets and magnetic fields

Safety issues

Further reading

6 Using MRI in Clinical Veterinary Practice

Introduction

Anaesthesia

Patient positioning

Imaging planes

Contrast agents

Clinical indications

Head

Spine

Degenerative disease

Congenital abnormalities

Musculoskeletal system

Thorax, abdomen and pelvis

Foreign bodies

Anatomy Atlas for MRI Interpretation

Further Reading

Glossary

Index

PART ONE
Physical Principles of MRI

1

Basic Principles

The hydrogen proton
Transmission of the MR signal
Image weighting and contrast
Spin echo and scan time
Spatial localisation
Fourier transformation
Pulse sequences: the quest for speed

At the time of writing, most veterinary professionals, whether they be surgeons, nurses or students, would probably agree that their knowledge of magnetic resonance imaging (MRI) physics borders on non-existent. Indeed, many may be filled with a deep dread at the very thought of the subject. On the other hand most will have a working knowledge of radiography at least sufficient to know that a radiograph represents a record of the different densities of body tissues through which the x-ray beam has passed. In this chapter the nature of magnetic resonance (MR) will be examined and the measurement parameters involved in constructing a MR image will be discussed.

It is worth beginning by recapping briefly on some radiation physics. In conventional radiography and computed tomography (CT), image contrast, or greyscale, is dependent on density or, more specifically, electron density of tissues in the patient. The more electrons an atom has in its shell the more it will attenuate the x-ray beam. Dense tissues, such as cortical bone, will appear as white in the image whilst air, being least dense, appears black. Since electron density is the only measurement parameter, radiographic and CT appearances are consistent, predictable and, therefore, reproducible. In MRI, however, there are a number of measurement parameters which affect signal intensity and, subsequently, image contrast. This means that the operator can manipulate image contrast to the extent of turning the appearance of water, for example, from black to white. This may appear confusing until the principles are understood. In fact, it is the ability to manipulate contrast in this way that gives MRI its superior soft tissue differentiation.

Later, consideration will be given to how the operator can alter scan parameters in order to produce these changes in image contrast but first of all we should explore the hydrogen proton, how MRI uses radiofrequency (RF) energy to produce resonance and what happens as the proton relaxes when the RF pulse is turned off.

The hydrogen proton

There are several atoms that possess the ability to resonate and can be used to produce images. In fact any atom with an odd mass number such as carbon (13), sodium (23) and phosphorous (31) would be suitable, but in clinical use only hydrogen, with a mass number of one, is used. This is because a single hydrogen atom produces a relatively large magnetic moment and resonates very well; it is said to have a high gyromagnetic ratio (γ) and it is abundant within the body.

Hydrogen is the simplest of atoms, having a nucleus composed of a single proton (no neutrons) and has no orbiting electrons; hence it is often referred to simply as a proton.

The proton carries a positive electrical charge and spins on its own axis. This moving electrical charge, according to the laws of electromagnetic induction, creates a corresponding magnetic field around the proton so that it behaves like a tiny bar magnet having north and south poles (Figure 1.1). Such magnetic fields are described in physics as magnetic moments. Each magnetic moment possesses the properties of size and direction. Where two or more magnetic moments exist together, their size and direction (or vectors) can be combined to give their net magnetisation. Thus if two magnetic moments exist both having the same size and direction their net magnetisation will be double that of each individual. Conversely if they have the same size but opposite direction the two will cancel each other out and their net magnetisation will be zero. In the normal course of events the body’s many billions of microscopic magnetic moments are completely randomly orientated (Figure 1.2) and cancel each other out such that their macroscopic or net magnetic field is zero.

Figure 1.1 The hydrogen proton.

Figure 1.2 In the normal state of affairs magnetic moments are randomly orientated and cancel each other out.

The effects of an external magnetic field B0

When an animal is placed into the MRI scanner, the external magnetic field (referred to as B0) causes the protons to abandon their random orientation and ‘line up’ with the main magnetic field. Current knowledge of magnets and magnetic fields would suggest that the tiny magnetic fields of each proton would adopt an orientation parallel to the main field B0 with their north and south poles matching those of the main magnet. However the laws of quantum mechanics dictate that certain protons have sufficient thermal energy at room temperature to adopt an opposing, anti-parallel state. Indeed the two populations are almost identical. Moreover the protons are continually oscillating between the two states but at any given point in time, the ratio of anti-parallel to parallel states is one million to one million and six at a B0 field strength of 1 Tesla (1 T). This excess population of six in one million means that our patient’s total hydrogen content has a net magnetisation vector (NMV) in the parallel direction (Figure 1.3). With only six in two million protons contributing to the image it seems doubtful that the process will work at all. However, at 1.5 T 0.01 ml of water contains around 3 million billion such excess protons, so things begin to seem feasible.

Since the energy level required to achieve the anti-parallel state increases with the field strength of B0, and the patient’s thermal energy remains fairly constant, it follows that the magnitude of the NMV increases with the field strength of the MRI system we are using. This is an important relationship, since it is the NMV that contributes the useful MRI signal. Hence systems with high field strength magnets generate more signal from the same volume of tissue than lower field systems.

Figure 1.3 The influence of an external magnetic field is to align protons in the parallel and anti-parallel states.

A second influence of B0 is to cause spinning protons to precess. Just as a child’s spinning top begins to wobble under the influence of gravity, so protons are made to wobble or precess by B0. The exact frequency of this precession is given by the Larmor equation:

where ω0 can be referred to as the Larmor, precessional or resonant frequency and γ is the gyromagnetic ratio referred to earlier in this chapter and is a constant unique to each atom. Since γ is constant for hydrogen, it can be seen from this equation that precessional frequency is directly linked to field strength B0 thus:

The exact equation does not have to be remembered, but this is an important relationship to grasp as it will help the understanding of a number of other concepts which follow.

The major effect of this precessional motion is to introduce a transverse component to the magnetic field of each proton since each is now spinning at a slight tilt to B0 (Figure 1.4). Because the north/south poles of each proton are pointing in random directions at any one time (Figure 1.5), they still cancel each other out in the transverse plane so that the NMV is still in the parallel or longitudinal direction.

The effects of an RF pulse at the Larmor frequency: resonance

If a pulse of radiofrequency (RF) energy is now applied to protons in the system it can cause the hydrogen spins to react to it provided two important conditions are fulfilled. These are that the RF pulse must be applied at right angles to B0 and that it must be at the Larmor frequency; any other frequency at this field strength will have no effect on hydrogen.

Figure 1.4 Precession.

Figure 1.5 Out of phase in the transverse plane.

This reaction to the RF pulse is resonance and, essentially, two things happen. One is that the RF pulse imparts sufficient energy to allow more protons to adopt the anti-parallel state The six excess protons discussed earlier provide an illustration of what happens if enough RF energy is transmitted to allow three of these to flip into the anti-parallel position. They will then cancel out the other three in the parallel state and the NMV in the longitudinal plane will now be zero. The other effect, which takes place in the transverse plane, is to bring all our hydrogen spins into phase with each other. Now, instead of all the spins cancelling each other out, each microscopic magnetic field is in unison with its neighbours; they are said to be ‘in phase’ (Figure 1.6).

Figure 1.6 In phase in the transverse plane.

Figure 1.7 Net magnetisation passes through 90° from longitudinal to transverse planes.

Consequently their individual magnetic fields all add together so that the NMV is now at a maximum in the transverse plane. The NMV has shifted through 90° from longitudinal to transverse. If the RF transmission is terminated at this point it is said to be a 90° RF pulse (Figure 1.7). Note that the angle through which the NMV tilts or the ‘flip angle’ (α), in this case 90°, is a function of the strength and duration of the RF pulse. Other values for α will be encountered later.

And when the RF transmission is turned off …

Three things begin to happen simultaneously but independently of each other as soon as the RF transmission is turned off. Each will be considered in some detail but briefly what happens is this:

1. Because the NMV is now in the transverse plane and no longer overwhelmed by B0 it can be detected by a receiver coil. The absorbed RF energy is retransmitted as the useable MR signal. How much signal is transmitted will depend on how much hydrogen there is in a particular tissue; its proton density (PD).
2. The spins that were in phase with each other in the transverse plane are affected to varying degrees by other atoms locally and some begin to slow down relative to others; they begin to dephase. This is referred to as T2 relaxation, also called transverse or spin spin relaxation.
3. The extra protons that were able to use RF energy to adopt the anti-parallel state are now reliant again on thermal energy alone and begin to return to their usual state, thermal equilibrium. This, surprisingly enough, is called T1 relaxation. This process is also referred to as longitudinal recovery or spin lattice relaxation.

Transmission of the MR signal

The concept of electromagnetic induction teaches that a moving magnetic field will induce an electrical current in an adjacent conductor. That is exactly the situation in the spin system; the rotating magnetic field in the transverse plane will produce an electromagnetic radiation at the Larmor frequency. It is this RF emission that gives the useable MR signal that goes to make up the final image. The amount of signal generated by various tissues within the body is determined by the amount of hydrogen each contains, as well as their T1 and T2 relaxation times. Tissues containing lots of hydrogen such as fat and cerebrospinal fluid (CSF) will generate lots of signal. Conversely tissues like cortical bone and lung, which contain little or no hydrogen, will generate very low signal or even a signal void.

By placing a suitable receiver coil (discussed later) close to the patient the signals being emitted can be collected for conversion into shades of grey in the MR image.

T1 relaxation (longitudinal recovery)

Once the RF pulse is turned off, it no longer contributes energy to the spin system. In the absence of any external influence the hydrogen spins will return to their thermal equilibrium. Their acquired energy is given off partly as emitted RF radiation but mostly as heat to the surrounding tissues, or lattice. Hence T1 relaxation is sometimes referred to as spin lattice relaxation. This results in an exponential regrowth in longitudinal magnetisation (Figure 1.8). T1 relaxation time itself is defined as the time taken for 63% of magnetisation to realign with B0. This relaxation process is called recovery since it represents a return to maximum from zero.

Figure 1.8 Recovery of longitudinal magnetisation.

Figure 1.9 Decay of transverse magnetisation.