Introduction

Nuclear magnetic resonance (NMR), the principle on which magnetic resonance (MR) imaging is based, was discovered independently in 1946 by Felix Bloch and Edward Purcell; they were jointly awarded the Nobel Prize for physics in 1952.¹ The first actual image was produced by Paul Lauterbur; although it was grainy and fuzzy, it nevertheless showed the difference between ordinary and ‘heavy’ water (D₂O). The importance of MR imaging has been such that he too went on to win a Nobel Prize of his own in 2003.² The first human image was captured in 1977.³

The name ‘nuclear magnetic resonance’ was altered to ‘magnetic resonance imaging’ in the 1980s. The official reason given was the public’s negative connotation with the word ‘nuclear’, particularly after the accident at the Three Mile Island power station in 1979. The unofficial explanation was that hospital workers feared that NMR might stand for ‘No More Radiographers’, whereas MRI could only mean ‘More Radiographers Immediately’.⁴

Magnetic resonance imaging uses powerful computer algorithms to generate multiplanar imaging slices through the body.⁵ Because it does not utilize ionizing radiation and provides a greater degree of contrast between the different soft tissues of the body than does computed tomography (CT), it is particularly useful in cardiovascular, musculoskeletal, neurological and oncological imaging.⁶ The relative lack of attendant safety issues also means that it is a modality that can be used to monitor the course of a condition or the response to treatment by repeat imaging.⁶ Although MR imaging is generally safe and non-invasive, there are circumstances in which it is contraindicated; these are detailed in Box 1.01.⁷

In place of x-rays, MR imaging uses a powerful magnetic field to align the magnetic moments of certain atoms within the body. Radiofrequency fields are then used to alter the alignment of this magnetization, which causes the atomic nuclei to produce a rotating magnetic field, specific to each tissue type, which is detectable by a receptor inside the scanner. This signal can then be manipulated to build sufficient information to construct an image of the body.^4,8

Magnetic resonance

Subatomic particles, such as protons, neutrons, positrons and electrons, have the property of spin, causing the particle to act as a magnetic dipole; that is, having north and south poles.^4,9 If two such particles pair up, the laws of magnetic attraction and repulsion mean that they will point in opposite directions and cancel each other out, so that the resultant particle has no overall magnetism. Certain nuclei, such as ¹H, which consists of a single proton, ³He, ¹³C, ²³Na and ³¹P, have an uneven number of protons and neutrons, and therefore have an unpaired particle, giving the nucleus an overall magnetism, known as the magnetic moment.^3,4

This effect is particularly strong in hydrogen, which does not have other particles to ‘dilute’ the relative strength of its magnetic moment; it is this nucleus, therefore, that is of primary importance in MR imaging.^4,10 Hydrogen has two spin states, sometimes referred to as ‘up’ and ‘down’.⁹ When these spins are placed in a strong external magnetic field, such as that found in an MRI scanner, they precess around an axis along the direction of the field (Figure 1.01).¹¹ Most protons have low energies and will align with the magnetic field; however, a small number have sufficient energy to have an anti-parallel alignment.¹¹

Figure 1.01 • The magnetic moment of the proton precesses about the flux lines of the applied magnetic field.

The frequency with which the protons precess (ω₀) is directly proportional to the strength of the magnetic field (B). In a magnetic field of 1 tesla, of the order found in MR scanners, hydrogen atoms will precess at precisely 42.6 MHz, which is in the radio wave band of the electromagnetic spectrum. This value, which varies from nucleus to nucleus as well as with magnetic field strength, is known as the Larmor frequency.^4,11

Contrast

Molecules, unless at the hypothetical limit of absolute zero, are constantly in motion. The greater the motion, the greater the inertia and the more difficult it is for the molecule to release its energy to its surroundings.⁴

Most molecules in the body contain hydrogen; amongst the commonest are fat and water. Water molecules are small and therefore have low inertia; however, they have a high inherent energy, which makes it more difficult for them to absorb more energy efficiently. By contrast, fat molecules consist of long chains of hydrocarbons – a central chain of interlinked carbon atoms with hydrogen attached to the sides. They therefore have high inertia and also possess a low inherent energy, meaning that they readily absorb additional energy.¹²

These characteristics mean that the different tissues in the body have different relaxation times and this is used to produce contrast: areas of high signal (which appears white), medium signal (grey) and low signal (black) which reflect the strength – or, more properly, prolongation – of the MR signal recovered from different locations within the sample.⁶

In addition to relaxation times, contrast is also affected by a number of other factors, including the relative density of the excited hydrogen and the extrinsic factors controlled by the radiographer. The two most important of these are the repetition time (TR) and the echo time (TE).¹⁴

Echo time (TE)

This is the length of time between the start of a radiofrequency pulse and the collection of the MR signal; it is also measured in milliseconds. The TE affects the relaxation times after the removal of the radiofrequency pulses and also the peak of the signal induced in the receiver coil.⁸ The differentiation of these factors is shown in Figure 1.02.

Figure 1.02 • The radiofrequency pulse (RF) is applied at regular intervals. The repetition time (TR) is the time between the applications of the pulse. The echo time (TE) is the time from the application of the pulse to the collection of the magnetic resonance signal.

The precise design of the imaging pulse sequences allows one contrast mechanism to be emphasized while the others are minimized. This ability to choose different contrast mechanisms is what gives MR imaging its tremendous flexibility.⁴ In the spine (Figure 1.03), the spinal canal structures have very different appearance on T1-weighting compared to T2-weighting, whilst the vertebral bodies and discs remain almost unchanged. The differences in signal intensities between the two types of images are summarized in Table 1.01.

Figure 1.03 • Two midsagittal MR images of the lumbar spine. The image in (A) is a T1-weighted image; note how the neurons of grey matter (black arrows), have a medium intensity (grey) signal whilst the cerebrospinal fluid (white arrows) appears dark. These contrasts are reversed in the T2-weighted image (B). Of incidental note is the focal higher-intensity signal in the body of the T11 vertebra in both T1-weighted and T2-weighted images; this represents a haemangioma. This finding is also responsible for the lighter appearance of the L1 and L2 vertebral bodies on both images.

Table 1.01 Signal intensities seen in T1-weighted and T2-weighted images

Structure	T1-weighting	T2-weighting
Air	No signal	No signal
Avascular necrosis	Low signal	High signal
Calcification	No/low signal	No/low signal
Cortical bone	No/low signal	No/low signal
Degenerative fatty deposition	High signal	Low signal
Fast-flowing blood	No signal	No signal
Fat	High signal	Low signal
Fluid cysts	Low signal	High signal
Haemangioma	High signal	High signal
Infarction	Low signal	High signal
Infection	Low signal	High signal
Lipoma	High signal	Low signal
Scar tissue	No signal	No signal
Sclerosis	Low signal	High signal
Slow-flowing blood	High signal	High signal
Tendons	No/low signal	No/low signal
Tumours	Low signal	High signal

There are conditions in which it is not possible to generate enough image contrast to adequately show the anatomy or pathology of clinical interest by adjusting the imaging parameters alone. In such cases, a contrast agent may be administered. This may be as simple as getting the patient to drink a glass of water to assist in imaging the stomach and small bowel; however, most of the contrast agents used in MR imaging are selected for their specific paramagnetic properties.⁴

The most common paramagnetic contrast agent used in musculoskeletal imaging is gadolinium. Gadolinium-enhanced tissues and fluids appear extremely bright on T1-weighted images and this provides high sensitivity for detection of neovascular tissue associated with fibrous tissue, such as that found in tumours, and permits assessment of brain perfusion in cases of cerebrovascular incidents. It also allows assessment of scar tissue following discal surgery. The addition of gadolinium adds a considerable degree of invasiveness and potential toxicity to the imaging procedure and is contraindicated in patients with impaired kidney function (nephrogenic fibrosis).⁷

Pulse sequences

There are a number of ways in which the operator can further manipulate the image. A pulse sequence is defined as a series of radiofrequency pulses, which may be further modified by the application of gradients, a change in magnetic flux density along the length of the scanner, and by changing the intervening time periods.¹⁵

The main purpose of these pulse sequences is to manipulate the TE and TR to produce different types of contrast and to rephase the spin of the hydrogen nuclei, which tend to gradually un-phase owing to inhomogeneities in the applied magnetic field causing signal decay.^4,14

Spin echo (SE)

Spin echo (SE) uses a radiofrequency pulse to rephase in the manner already described; however, in addition to the initial 90° pulse, a second pulse is added at 180° to rephase any nuclei that have decreased their precessional frequency and reform the net magnetization vector. This regenerates the induced MR signal, which had been decaying, and this regeneration can be measured. The regenerated signal is called an ‘echo’ and, because a radiofrequency pulse has been used to generate it, it is specifically termed spin echo. Spin echo is used to produce T1-weighted (short TR, short TE) and T2-weighted (long TR, long TE) images. By adding an additional 180° pulse (these pulses do not affect TR, which is still defined as the gap between the application of 90° pulses), it is possible to produce two images per pulse. Both will have a long TR; however, the first will have a short TE and produces a proton density weighted image. The second has a long TR and TE and is used to produce a standard T2-weighted image (Figure 1.04). This is known as a dual echo sequence.¹⁵

Figure 1.04 • Pulse sequences used to produce T1-weighted, T2-weighted and proton density weighted signals using spin echo and dual echo sequences.

The main problem with methods of image acquisition is the length of time taken; this can be reduced by fast or turbo spin echo (FSE/TSE) whereby a chain of 180° rephasing pulses are applied in between the 90° pulses. This can reduce the imaging time by factors of as much as 16; however, there is an inevitable trade-off with image quality and artifacts from blood flow, which usually is not imaged but can be ‘frozen’ if the image acquisition is fast enough. This technique depends on altering the gradient between each pulse so that the phase is slightly shifted and can be differentiated by the computer algorithm.¹⁶

Clinical aspects of MRI

Although the physics behind NMR and the engineering that allows us to ‘see’ MR images are complicated, the principles are worth understanding; without them, the clinician can have no real comprehension of what they are seeing on the computer screen or view box – or why. It is therefore worth taking the time to get your head around these terms and their impact on and consequences for the images you are going to learn to interpret; it will also help you to answer the questions that will come from patients about their films and what they mean for them.

Patients will also have questions for you from the moment you decide that imaging is required – why MR imaging and not x-ray? If you cannot readily answer this question, you should revisit your clinical and diagnostic thinking. They will also, as a rule, want to know how much it is going to cost, particularly where the cost of the procedure is not covered by insurance schemes or public health. As a clinician, it is your job to screen the patient for contraindications (Box 1.01); there is nothing guaranteed to infuriate both the imaging centre and patient more than for them to have their time and effort wasted by an inappropriate referral. On occasion, it may be necessary to prescribe tranquillizers (and arrange a chauffeur) for a nervous, claustrophobic patient who cannot gain access to imaging by an open scanner; it may also be necessary to organize a skull x-ray to screen for metal fragments in and around the orbit in those patients who have worked with lathes or sheet metal or who have seen active service in the armed forces.⁷

Most imaging centres will have their own bespoke forms; a typical such form (for the author’s own ‘local’ MR scanner) is shown in Figure 1.05. It is important to give the radiologist as much information as possible so that the appropriate protocols can be used to answer the clinical question that you want answered: the radiologist can’t refine your differential diagnosis if you don’t have one! Regional protocols are discussed on a chapter by chapter basis; however, as a rule, you can expect to receive back a CD or hard copy films containing T1- and T2-weighted images in at least two planes. These should be correlated with the radiologist’s report (a good way to practise) and with your own clinical findings and thought processes – remember, the radiologist has not seen the patient, you have.

Figure 1.05 • A typical referral form for an MR imaging centre. The clinician has room to give a clinical history as well as the patient’s demographic details. Contraindications are clearly listed and have to be signed off before sending.

(Courtesy of Gloucestershire Magnetic Resonance Imaging Service.)

The patient may also be concerned about the process and what will happen to them, particularly if they have not had a scan before. When they arrive at the centre, they will be asked to complete a consent form and will be rescreened for contraindications. They will then be asked to change into a gown and to remove all jewellery and eyeglasses. Depending on the region to be imaged, surface coils may be applied to the area; these act as a local receiver that can be placed immediately adjacent to smaller or off-centre structures such as the distal extremities to improve the image quality.

A traditional MR imaging scanner has, from the patient’s perspective, the appearance of a long tunnel, around 1.60 m in length. The opening is fairly small; indeed, most centres have an upper limit of 225–280 lb beyond which it is physically impossible to fit the patient in the scanner. In this instance, the open scanner again provides an alternative (Figure 1.06).

Figure 1.06 • Recent improvements in receptor analysis have allowed the development of open-sided MRI scanners (A), which ameliorate the effects of claustrophobia – a condition that affects up to 10% of people and can preclude the use of a conventional scanner (B).

A horizontal table ‘feeds’ the patient into the scanner, where they will remain for the duration of the scan, usually 25–40 minutes. Communication with the control booth is maintained using speakers or headphones, which can also be used to help reduce the significant banging noises from the radiofrequency pulses turning on and off. Music can help in this regard; patients are often allowed to bring a CD of their choice; this can also help make them feel more comfortable. Instructions usually consist of no more than getting the patient to lie as still as possible while the scan takes place.

Most scanning units are happy to organize visits from referring physicians, and personal knowledge and familiarity with the centre and its personnel can be a significant aid in helping to assuage a patient’s concerns or anxieties.