DOI: 10.1148/rg.262055063
RadioGraphics 2006;26:513-537
© RSNA, 2006
MR Pulse Sequences: What Every Radiologist Wants to Know but Is Afraid to Ask1
Richard Bitar, MD, MSc,
General Leung, MSc,
Richard Perng, MD,
Sameh Tadros, MD,
Alan R. Moody, FRCR, FRCP,
Josee Sarrazin, MD, FRCP,
Caitlin McGregor, MD, FRCP,
Monique Christakis, MD, FRCP,
Sean Symons, MD, FRCP,
Andrew Nelson, RTR, RTMR and
Timothy P. Roberts, PhD
1 From the Department of Medical Imaging, University of Toronto, Fitzgerald Building, 150 College St, Room 112, Toronto, Ontario, Canada M5S 3E2 (R.B., A.R.M., J.S., C.M., M.C., S.S., T.P.R.); and Department of Medical Imaging, Sunnybrook and Women’s College Health Sciences Centre, Toronto, Ontario, Canada (R.B., G.L., R.P., S.T., A.R.M., J.S., C.M., M.C., S.S., A.N.). Recipient of an Excellence in Design award for an education exhibit at the 2004 RSNA Annual Meeting. Received March 24, 2005; revision requested May 12 and received June 7; accepted July 6. All authors have no financial relationships to disclose.
Address correspondence to R.B. (e-mail: richard.bitar{at}utoronto.ca).
 |
Abstract
|
|---|
The use of magnetic resonance (MR) imaging is growing exponentially, in part because of the excellent anatomic and pathologic detail provided by the modality and because of recent technologic advances that have led to faster acquisition times. Radiology residents now are introduced in their 1st year of training to the MR pulse sequences routinely used in clinical imaging, including various spin-echo, gradient-echo, inversion-recovery, echo-planar imaging, and MR angiographic sequences. However, to make optimal use of these techniques, radiologists also need a basic knowledge of the physics of MR imaging, including T1 recovery, T2 and T2* decay, repetition time, echo time, and chemical shift effects. In addition, an understanding of contrast weighting is very helpful to obtain better depiction of specific tissues for the diagnosis of various pathologic processes.
© RSNA, 2006
|
Introduction
|
|---|
Not long ago, many radiology residents did not study magnetic resonance (MR) imaging until their senior year; however, with recent technologic advances, including faster acquisition times and better anatomic and pathologic depiction, MR imaging is used with increasing frequency. In fact, residents now receive training in MR imaging and routinely use it while on call as early as the 1st year of residency. However, when residents are presented with MR images to interpret, they may not know where to start or, more important, why a specific pulse sequence with a particular contrast weighting was used. They may know little or nothing about the physical principles on which the sequence is based.
In this article, we describe the physical basis of the most common MR pulse sequences routinely used in clinical imaging. This is a vast and complicated subject, and only the fundamentals could be presented here; however, many excellent general (1–4) and detailed (5–30) references about the subject are available. To enhance understanding, our explanation of the physical principles is simplified; to highlight their practical relevance, clinical examples are interspersed throughout the article.
|
Physics Overview
|
|---|
MR imaging is based on the electromagnetic activity of atomic nuclei. Nuclei are made up of protons and neutrons, both of which have spins. MR-active nuclei are those that have a net spin because they are odd-numbered and the spins of their protons and neutrons do not cancel each other out (5). In clinical MR imaging, hydrogen (1H) nuclei are used most often because of their abundance in the body, but other nuclei—for example, fluorine (19F) nuclei—also may be used (6).
Each nucleus rotates around its own axis. As the nucleus spins, its motion induces a magnetic field (Fig 1a). When the nuclei are exposed to an external magnetic field (B0), the interaction of the magnetic fields (ie, the fields of the spinning nuclei and the externally applied field) causes the nuclei to wobble, or precess. The frequency at which precession occurs is defined by the Larmor equation, 
0 = B0 x 
, where 0 is the precessional frequency, B0 is the external magnetic field strength (measured in teslas), and is the gyromagnetic ratio (measured in megahertz per tesla), which is a constant for every atom at a particular magnetic field strength (eg, for 1H, /2
= 42.57 MHz/T) (7).
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Basic physics of the MR signal. (a) As 1H nuclei spin, they induce their own magnetic field (tan), the direction (magnetic axis) of which is depicted by an arrow (yellow). The 1H nuclei initially precess with a wobble at various angles (1–6), but when they are exposed to an external magnetic field (B0), they align with it. The sum of all magnetic moments is called the net magnetization vector (NMV). (b) When an RF pulse is applied, the net magnetization vector is flipped at an angle ( ), which produces two magnetization components: longitudinal magnetization (Mz ) and transverse magnetization (Mxy ). As the transverse magnetization precesses around a receiver coil, it induces a current (i). When the RF generator is turned off, T1 recovery and T2 and T2* decay occur.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Basic physics of the MR signal. (a) As 1H nuclei spin, they induce their own magnetic field (tan), the direction (magnetic axis) of which is depicted by an arrow (yellow). The 1H nuclei initially precess with a wobble at various angles (1–6), but when they are exposed to an external magnetic field (B0), they align with it. The sum of all magnetic moments is called the net magnetization vector (NMV). (b) When an RF pulse is applied, the net magnetization vector is flipped at an angle (), which produces two magnetization components: longitudinal magnetization (Mz ) and transverse magnetization (Mxy ). As the transverse magnetization precesses around a receiver coil, it induces a current (i). When the RF generator is turned off, T1 recovery and T2 and T2* decay occur.
|
|
Until the 1H nuclei are exposed to B0 magnetization, their axes are randomly aligned. However, when B0 magnetization is applied, the magnetic axes of the nuclei align with the magnetic axis of B0, some in parallel and others in opposition to it (5) (Fig 1a). The cumulative effect of all the magnetic moments of the nuclei is the net magnetization vector. When a radiofrequency (RF) pulse is applied, the RF excitation causes the net magnetization vector to flip by a certain angle, and this produces two magnetization vector components, longitudinal magnetization and transverse magnetization (Fig 1b). As the transverse magnetization precesses around a receiver coil, it induces a current in that coil, in accordance with the Faraday law of induction (8). This current becomes the MR signal.
When the RF energy source is turned off, the net magnetization vector realigns with the axis of B0 through the process of T1 recovery, during which the longitudinal magnetization increases in magnitude, or recovers. At the same time, the transverse magnetization decreases (decays) through additional mechanisms known as T2* decay and T2 decay. Different tissues have different T1, T2, and T2* values. Furthermore, T2* is dependent on the magnetic environment (the spatial uniformity of the external field). Fat has a shorter T1 (ie, recovers faster) and a shorter T2 (ie, decays faster) than water, which has a relatively long T1 and T2. T2* decay occurs very quickly in both fat and water (Fig 1b) (9).
During T1 (spin-lattice) relaxation, the longitudinal magnetization recovers as the spinning nuclei release energy into the environment (Fig 2a). During T2 (spin-spin) relaxation, the transverse magnetization is dephased because of interaction between the spinning nuclei and their magnetic fields (Fig 2b). In T2* signal decay, the transverse magnetization is dephased because of magnetic field inhomogeneities. The magnetic field is not exactly the same everywhere; in some places it is a bit stronger (B0 + )—for example, 1.505 T—and in others it is a bit weaker (B0 – )—for example, 1.495 T. Such differences may occur because of the presence of metallic objects, air, dental implants, or calcium, or they may be due to the limitations of magnet construction (Fig 2c) (6).
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Magnetization relaxation and decay. (a) T1 recovery (spin-lattice relaxation) involves recovery of the longitudinal magnetization (yellow) because of the release of energy (green) into the environment. The lattice is indicated in tan. (b) T2 decay (spin-spin relaxation) is decay of the transverse magnetization because of the interaction of the individual magnetic fields of spinning nuclei. Note that all nuclei initially spin in phase (as indicated by the similar position of the red bands at the bottom of each circle), then move out of phase (with red bands in different positions). (c) T2* decay is decay of the transverse magnetization because of magnetic field inhomogeneities (Fi). = flip angle, B0 = external magnetic field.
|
|
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Magnetization relaxation and decay. (a) T1 recovery (spin-lattice relaxation) involves recovery of the longitudinal magnetization (yellow) because of the release of energy (green) into the environment. The lattice is indicated in tan. (b) T2 decay (spin-spin relaxation) is decay of the transverse magnetization because of the interaction of the individual magnetic fields of spinning nuclei. Note that all nuclei initially spin in phase (as indicated by the similar position of the red bands at the bottom of each circle), then move out of phase (with red bands in different positions). (c) T2* decay is decay of the transverse magnetization because of magnetic field inhomogeneities (Fi). = flip angle, B0 = external magnetic field.
|
|
View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2c. Magnetization relaxation and decay. (a) T1 recovery (spin-lattice relaxation) involves recovery of the longitudinal magnetization (yellow) because of the release of energy (green) into the environment. The lattice is indicated in tan. (b) T2 decay (spin-spin relaxation) is decay of the transverse magnetization because of the interaction of the individual magnetic fields of spinning nuclei. Note that all nuclei initially spin in phase (as indicated by the similar position of the red bands at the bottom of each circle), then move out of phase (with red bands in different positions). (c) T2* decay is decay of the transverse magnetization because of magnetic field inhomogeneities (Fi). = flip angle, B0 = external magnetic field.
|
|
As mentioned earlier, the transverse component of the net magnetization vector induces a current in the receiver coil. For the induction of current, the nuclei must be spinning in phase; as the nuclei gradually spin out of phase, the signal induced in the coil decreases. This process is called free induction decay: Free refers to the fact that the system is no longer being forced out of equilibrium by the RF excitation; induction describes the mechanism through which the signal is detected; and decay refers to the decrease in signal amplitude over time.
Repetition Time and Echo Time
At MR imaging, differences in T1, T2, and proton density (ie, the number of 1H nuclei) in various tissues create differences in tissue contrast on images (9). Two key parameters—repetition time (TR) and echo time (TE)—are key to the creation of image contrast. Figure 3 shows the symbols that are most commonly used to diagram pulse sequences (1–3) as well as the echoes detected, including the Hahn echo (with use of a spin-echo [SE] pulse sequence) and the gradient echo (GRE) (10). It is important to recognize these symbols, because they are invariably used to represent TR and TE.
TR is the time (usually measured in milliseconds) between the application of an RF excitation pulse and the start of the next RF pulse (1,2). TE (also usually measured in milliseconds) is the time between the application of the RF pulse and the peak of the echo detected (Fig 4a) (1,2). Both parameters affect contrast on MR images because they provide varying levels of sensitivity to differences in relaxation time between various tissues. At short TRs, the difference in relaxation time between fat and water can be detected (longitudinal magnetization recovers more quickly in fat than in water); at long TRs, it cannot be detected. Therefore, TR relates to T1 (Fig 4b) and affects contrast on T1-weighted images. At short TEs, differences in the T2 signal decay in fat and water cannot be detected; at long TEs, they can be detected. Therefore, TE relates to T2 (Fig 4b) and affects contrast on T2-weighted images.
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. (a) Schematic representation of TR and TE. (b) Graphs show the effects of short and long TR (left) and short and long TE (right) on T1 recovery and T2 decay in fat and water: TR relates to T1 and affects T1 weighting, whereas TE relates to T2 and affects T2 weighting. msc = milliseconds, Mxy = transverse magnetization, Mz = longitudinal magnetization.
|
|
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. (a) Schematic representation of TR and TE. (b) Graphs show the effects of short and long TR (left) and short and long TE (right) on T1 recovery and T2 decay in fat and water: TR relates to T1 and affects T1 weighting, whereas TE relates to T2 and affects T2 weighting. msc = milliseconds, Mxy = transverse magnetization, Mz = longitudinal magnetization.
|
|
When the TR is long and the TE is short, the differences in magnetization recovery and in signal decay between fat and water are not distinguishable (Fig 4b); therefore, the contrast observed on the resultant MR images is predominantly due to the difference in proton density between the two tissue types. Tissues with more protons have higher signal intensity, and those with fewer protons have lower signal intensity.
Tissue Contrast
All MR images are to some degree affected by each of the parameters that determine tissue contrast (ie, T1, T2, and proton density), but the TR and TE can be adjusted to emphasize a particular type of contrast. This may be done, for example, with T1 weighting. In T1-weighted MR imaging, while the images show all types of contrast, T1 contrast is accentuated. Table 1 describes the parameters used to obtain images with T1, T2, and proton-density weighting (1,3). T1-weighted images best depict the anatomy, and, if contrast material is used, they also may show pathologic entities; however, T2-weighted images provide the best depiction of disease, because most tissues that are involved in a pathologic process have a higher water content than is normal, and the fluid causes the affected areas to appear bright on T2-weighted images (1,3,9). Proton-density weighted MR images usually depict both the anatomy and the disease entity. Table 2 shows typical TR and TE values that may be used to achieve different weighting with SE and GRE sequences (1,3). The levels of signal intensity that characterize various tissues on T1- and T2-weighted images are shown in Figure 5 (1,3,4).
View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5. Diagram shows the signal intensity of various tissues at T1- and T2-weighted imaging. However, note that the signal characteristics of proteinaceous tissues vary according to the amount of protein content: Tissues with high concentrations of protein may have high signal intensity on T1-weighted images (T1WI) and low signal intensity on T2-weighted images (T2WI). CSF = cerebrospinal fluid. (Sources: References 1, 3, and 4.)
|
|
MR Signal Localization
How does the MR imaging system detect which tissue the signal is coming from? For this purpose, gradients are employed. Gradients are linear variations of the magnetic field strength in a selected region (11). Along with the magnetic field strength, the precessional frequency of 1H nuclei is also changed in the specific region (recall the Larmor equation). Three types of gradients are applied, according to the axis of imaging (x-, y-, or z-axis) (Fig 6). The section-selective gradient selects the section to be imaged (ie, the target of the RF excitation pulse). The phase-encoding gradient causes a phase shift in the spinning protons so that the MR imaging system computer can detect and encode the phase of the spin (eg, the location of the red bands in the nuclei shown in Fig 2). The frequency-encoding gradient also causes a shift—one of frequency rather than phase—that helps the MR system to detect the location of the spinning nuclei. Because this shift of frequency usually occurs when the echo is read, it is also called the readout gradient. Once the MR system processor has all of that information (ie, the frequency and phase of each spin), it can compute the exact location and amplitude of the signal. That information is then stored in k-space.
View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Schematic and table show the x-, y-, and z-axis gradients (Gx, Gy, and Gz, respectively) that are used for section selection and for phase and frequency encoding during acquisitions in the most common imaging planes.
|
|
k-Space and the Image Matrix
k-Space (named for k, the symbol for wave number) is a matrix of voxels within which the raw imaging data are stored in the MR imaging system (12). The horizontal axis (x-axis) of the matrix usually corresponds to the frequency, and the vertical axis (y-axis) usually corresponds to the phase (although the axes of frequency and phase may be easily interchanged) (Fig 7). The center of k-space contains information about gross form and tissue contrast, whereas the edges (periphery) of k-space contain information about spatial resolution (details and fine structures). The raw imaging data in k-space must be Fourier transformed to obtain the final image.
View larger version (48K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7. Schematic and corresponding MR images show the characteristics determined by data at the periphery of k-space (ie, spatial resolution, or definition of edges) and those determined by data at the center of k-space (ie, gross form and image contrast).
|
|
|
Diagrams and Clinical Applications of MR Sequences
|
|---|
Pulse sequences are wave forms of the gradients and RF pulses applied in MR image acquisition (2). The diagrams may be composed of several parallel lines if each parameter is diagrammed separately, or they may consist of only one or two lines if the parameters are superimposed. If the parameters are diagrammed separately, at least four lines are required: one for the RF pulse, and one each for the x-, y-, and z-axis gradients (2). The pulse sequence diagram is a schema of the timing of instructions sent to the RF generator and gradient amplifiers.
There are only two fundamental types of MR pulse sequences: SE and GRE. All other MR sequences are variations of these, with different parameters added on. MR pulse sequences can be either two-dimensional (2D), with one section acquired at a time, or three-dimensional (3D), with a volume of multiple sections obtained in a single acquisition.
SE Sequences
In SE sequences, a 90° pulse flips the net magnetization vector into the transverse plane (10). As the spinning nuclei go through T1, T2, and T2* relaxation, the transverse magnetization is gradually dephased. A 180° pulse is applied at a time equal to one-half of TE to rephase the spinning nuclei. When the nuclei are again spinning in phase (at total TE), an echo is produced and read (Fig 8a). Most conventional SE sequences are very long and therefore are not used frequently. However, advances in MR imaging technology have enabled a reduction in acquisition time with the use of fast SE sequences. Table 3 shows the names of the various SE sequences used by the major MR imaging system vendors.
View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. Application of an SE pulse sequence. (a) Diagram shows the application of an initial pulse at a 90° flip angle to redirect the net magnetization vector into the transverse plane; a subsequent interval of T1, T2, and T2* relaxation, accompanied by the gradual dephasing of the transverse magnetization; and a second pulse applied at a flip angle of 180° to bring the spinning nuclei again into phase so that an echo is produced. Note the locations of the section-selective (Slice) and phase- and frequency-encoding (Readout) gradients (G). (b, c) Coronal T1-weighted (b) and axial T2-weighted (c) SE images of the brain. (d) Sagittal proton-density weighted SE image of the knee.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. Application of an SE pulse sequence. (a) Diagram shows the application of an initial pulse at a 90° flip angle to redirect the net magnetization vector into the transverse plane; a subsequent interval of T1, T2, and T2* relaxation, accompanied by the gradual dephasing of the transverse magnetization; and a second pulse applied at a flip angle of 180° to bring the spinning nuclei again into phase so that an echo is produced. Note the locations of the section-selective (Slice) and phase- and frequency-encoding (Readout) gradients (G). (b, c) Coronal T1-weighted (b) and axial T2-weighted (c) SE images of the brain. (d) Sagittal proton-density weighted SE image of the knee.
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8c. Application of an SE pulse sequence. (a) Diagram shows the application of an initial pulse at a 90° flip angle to redirect the net magnetization vector into the transverse plane; a subsequent interval of T1, T2, and T2* relaxation, accompanied by the gradual dephasing of the transverse magnetization; and a second pulse applied at a flip angle of 180° to bring the spinning nuclei again into phase so that an echo is produced. Note the locations of the section-selective (Slice) and phase- and frequency-encoding (Readout) gradients (G). (b, c) Coronal T1-weighted (b) and axial T2-weighted (c) SE images of the brain. (d) Sagittal proton-density weighted SE image of the knee.
|
|
View larger version (30K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8d. Application of an SE pulse sequence. (a) Diagram shows the application of an initial pulse at a 90° flip angle to redirect the net magnetization vector into the transverse plane; a subsequent interval of T1, T2, and T2* relaxation, accompanied by the gradual dephasing of the transverse magnetization; and a second pulse applied at a flip angle of 180° to bring the spinning nuclei again into phase so that an echo is produced. Note the locations of the section-selective (Slice) and phase- and frequency-encoding (Readout) gradients (G). (b, c) Coronal T1-weighted (b) and axial T2-weighted (c) SE images of the brain. (d) Sagittal proton-density weighted SE image of the knee.
|
|
As mentioned earlier, sequences that have a short TR and short TE are used to obtain T1 weighting. Those with a long TR and short TE result in proton-density weighting. When the TR is long and the TE is long, T2 weighting is achieved. Sequentially increasing the TE of a sequence weights it more heavily toward T2: This technique is used at MR cholangiopancreatography to obtain a detailed depiction of the bile ducts and pancreatic ducts (Fig 9). Increasing the TE also is useful for MR imaging of hemangiomas and cysts (Fig 10). Clinical illustrations of the contrast weightings of tissues and the advantages of choosing a particular contrast weighting at SE or fast SE MR imaging are shown in Figures 8–11.
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9. MR cholangiopancreatography. Sagittal fast SE image obtained with a heavily T2-weighted sequence (TE = 650) shows the common hepatic duct (arrowhead) and common bile duct (arrow).
|
|
View larger version (103K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Clinical examples of SE and fast SE sequences. (a, b) Liver hemangioma. (a) Axial T2-weighted fast SE image (TE = 82.9) shows a high-signal-intensity lesion (arrow) in the right lobe of the liver. (b) Axial T2-weighted fast SE image (TE = 180), obtained with heavier T2 weighting than a, shows retention of high signal intensity in the lesion (arrow), a feature that indicates a cyst or hemangioma. (c) Polycystic kidney disease. Axial T2-weighted fast SE image provides excellent depiction of cysts, which appear as areas of high signal intensity in the liver and kidney. Differences in signal intensity among the cysts are due to different protein concentrations.
|
|
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Clinical examples of SE and fast SE sequences. (a, b) Liver hemangioma. (a) Axial T2-weighted fast SE image (TE = 82.9) shows a high-signal-intensity lesion (arrow) in the right lobe of the liver. (b) Axial T2-weighted fast SE image (TE = 180), obtained with heavier T2 weighting than a, shows retention of high signal intensity in the lesion (arrow), a feature that indicates a cyst or hemangioma. (c) Polycystic kidney disease. Axial T2-weighted fast SE image provides excellent depiction of cysts, which appear as areas of high signal intensity in the liver and kidney. Differences in signal intensity among the cysts are due to different protein concentrations.
|
|
View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10c. Clinical examples of SE and fast SE sequences. (a, b) Liver hemangioma. (a) Axial T2-weighted fast SE image (TE = 82.9) shows a high-signal-intensity lesion (arrow) in the right lobe of the liver. (b) Axial T2-weighted fast SE image (TE = 180), obtained with heavier T2 weighting than a, shows retention of high signal intensity in the lesion (arrow), a feature that indicates a cyst or hemangioma. (c) Polycystic kidney disease. Axial T2-weighted fast SE image provides excellent depiction of cysts, which appear as areas of high signal intensity in the liver and kidney. Differences in signal intensity among the cysts are due to different protein concentrations.
|
|
View larger version (95K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Axial T1-weighted (a) and T2-weighted (b) fast SE images show a low-grade glioma. Because of hypercellularity, the tumor appears with hypointense signal in a and hyperintense signal in b. The cystic components and edema are better depicted in b than in a.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Axial T1-weighted (a) and T2-weighted (b) fast SE images show a low-grade glioma. Because of hypercellularity, the tumor appears with hypointense signal in a and hyperintense signal in b. The cystic components and edema are better depicted in b than in a.
|
|
SE-based Sequences
Fast SE Variants.—
In a fast or turbo SE sequence, a single 90° pulse is applied to flip the net magnetization vector, after which multiple 180° rephasing pulses are applied (13), each of which creates a Hahn echo (Fig 12a). All the echoes together are called an echo train, and the total number of 180° RF pulses and echoes is referred to as the echo train length. The acquisition time is greatly reduced with use of a fast SE sequence as opposed to a conventional SE sequence (Figs 12b, 12c). It is approximately proportional to 1/ETL, where ETL is the echo train length, for imaging of a single section or a small number of sections. However, at imaging of larger volumes, the reduction of acquisition time is highly dependent on the spatial coverage.
View larger version (21K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12a. Fast SE pulse sequence. (a) Pulse sequence diagram shows the fast SE sequence used to obtain the image in b. G = gradient, n = number of repetitions. (b, c) Axial T2-weighted fast SE image (b) and conventional SE image (c) provide comparable depiction of a brain tumor. The acquisition time for conventional SE imaging was 7 minutes 17 seconds, whereas that for fast SE imaging with an echo train length of 16 was 34 seconds.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12b. Fast SE pulse sequence. (a) Pulse sequence diagram shows the fast SE sequence used to obtain the image in b. G = gradient, n = number of repetitions. (b, c) Axial T2-weighted fast SE image (b) and conventional SE image (c) provide comparable depiction of a brain tumor. The acquisition time for conventional SE imaging was 7 minutes 17 seconds, whereas that for fast SE imaging with an echo train length of 16 was 34 seconds.
|
|
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12c. Fast SE pulse sequence. (a) Pulse sequence diagram shows the fast SE sequence used to obtain the image in b. G = gradient, n = number of repetitions. (b, c) Axial T2-weighted fast SE image (b) and conventional SE image (c) provide comparable depiction of a brain tumor. The acquisition time for conventional SE imaging was 7 minutes 17 seconds, whereas that for fast SE imaging with an echo train length of 16 was 34 seconds.
|
|
Conventional Inversion Recovery.—
This is an SE sequence in which a 180° preparatory pulse is applied to flip the net magnetization vector 180° and null the signal from a particular entity (eg, water in tissue). When the RF pulse ceases, the spinning nuclei begin to relax. When the net magnetization vector for water passes the transverse plane (the null point for that tissue), the conventional 90° pulse is applied, and the SE sequence then continues as before (Fig 13). The interval between the 180° pulse and the 90° pulse is the TI.
View larger version (28K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13a. (a) Conventional inversion-recovery sequence diagram shows a 180° preparatory pulse applied to null the signal from either fat or water. At a predetermined inversion time (TI), a 90° pulse is applied, and the SE sequence is continued. G = gradient. (b) Coronal STIR image shows an insufficiency fracture of the distal tibia, with an extensive area of high signal intensity in bone marrow near the site of the fracture (arrow).
|
|
View larger version (52K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13b. (a) Conventional inversion-recovery sequence diagram shows a 180° preparatory pulse applied to null the signal from either fat or water. At a predetermined inversion time (TI), a 90° pulse is applied, and the SE sequence is continued. G = gradient. (b) Coronal STIR image shows an insufficiency fracture of the distal tibia, with an extensive area of high signal intensity in bone marrow near the site of the fracture (arrow).
|
|
At TI, the net magnetization vector of water is very weak, whereas that for body tissues is strong (Fig 14). When the net magnetization vectors are flipped by the 90° pulse, there is little or no transverse magnetization in water, so no signal is generated (fluid appears dark), whereas signal intensity ranges from low to high in tissues with a stronger net magnetization vector. Two important clinical implementations of the inversion-recovery concept are the short TI inversion-recovery (STIR) sequence and the fluid-attenuated inversion-recovery (FLAIR) sequence. Table 4 shows the names of the various inversion-recovery sequences used by the major MR imaging system vendors.
View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14. Diagrams show T1 recovery in water and in tissue with use of a conventional inversion-recovery sequence. Nulling of the water signal is seen at TI, when there is virtually no net magnetization vector (NMV) in water. When the 90° pulse flips the net magnetization vector into the transverse plane, little or no transverse magnetization (Tm) is present, and, therefore, no signal is detected in water. Lm = longitudinal magnetization.
|
|
STIR.—
In STIR sequences, an inversion-recovery pulse is used to null the signal from fat (14). When the net magnetization vector of fat passes its null point (at approximately 140 msec), the conventional 90° RF pulse is applied. Little or no longitudinal magnetization is present in fat, and the transverse magnetization in fat is insignificant. It is transverse magnetization that induces an electric current in the receiver coil, and, because the insignificant transverse magnetization of fat produces an insignificant current, no signal is generated from fat. STIR sequences provide excellent depiction of bone marrow edema (Fig 15), which may be the only indication of an occult fracture. Unlike conventional fat-saturation sequences (discussed later), STIR sequences are not affected by magnetic field inhomogeneities, so they are more efficient for nulling the signal from fat (Fig 16).
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15a. Comparison of fast SE and STIR sequences for depiction of bone marrow edema. (a) Diagram of the STIR sequence (TI = 100–180 msec for fat). (b, c) Coronal T1-weighted fast SE image (b) and coronal STIR image (c) both show pancarpal rheumatoid arthritis; however, the extent of bone marrow edema throughout the carpal bones, distal radius, and ulna is better depicted in c than in b.
|
|
View larger version (192K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15b. Comparison of fast SE and STIR sequences for depiction of bone marrow edema. (a) Diagram of the STIR sequence (TI = 100–180 msec for fat). (b, c) Coronal T1-weighted fast SE image (b) and coronal STIR image (c) both show pancarpal rheumatoid arthritis; however, the extent of bone marrow edema throughout the carpal bones, distal radius, and ulna is better depicted in c than in b.
|
|
View larger version (155K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15c. Comparison of fast SE and STIR sequences for depiction of bone marrow edema. (a) Diagram of the STIR sequence (TI = 100–180 msec for fat). (b, c) Coronal T1-weighted fast SE image (b) and coronal STIR image (c) both show pancarpal rheumatoid arthritis; however, the extent of bone marrow edema throughout the carpal bones, distal radius, and ulna is better depicted in c than in b.
|
|
View larger version (85K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16a. Comparison of magnetic field inhomogeneities with fast SE versus STIR sequences. (a) Sagittal T2-weighted fast SE image obtained with spectral fat suppression, which requires a uniform magnetic field, shows incomplete fat saturation in regions where there is field inhomogeneity, such as at the irregular air–soft tissue interfaces of the toes (arrow). (b) Sagittal image obtained with STIR, which is less susceptible than fast SE sequences to magnetic field inhomogeneities, provides more uniform and more complete fat saturation (arrow). Bone infarcts in the distal tibia and heel appear as areas of high signal intensity on both images.
|
|
View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16b. Comparison of magnetic field inhomogeneities with fast SE versus STIR sequences. (a) Sagittal T2-weighted fast SE image obtained with spectral fat suppression, which requires a uniform magnetic field, shows incomplete fat saturation in regions where there is field inhomogeneity, such as at the irregular air–soft tissue interfaces of the toes (arrow). (b) Sagittal image obtained with STIR, which is less susceptible than fast SE sequences to magnetic field inhomogeneities, provides more uniform and more complete fat saturation (arrow). Bone infarcts in the distal tibia and heel appear as areas of high signal intensity on both images.
|
|
FLAIR.—
In FLAIR sequences, an inversion-recovery pulse is used to null the signal from cerebrospinal fluid (15). When the net magnetization vector of cerebrospinal fluid passes its null point, little or no longitudinal magnetization is present in the fluid. The transverse magnetization of cerebrospinal fluid is insignificant, and therefore no signal is generated (Fig 17). Elimination of the signal from cerebrospinal fluid is useful for detecting lesions that otherwise are not easily distinguishable or for delineating hyperintense lesions that border fluid-containing spaces such as sulci or ventricles in the brain (Fig 17b, 17c).
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17a. Comparison of fast SE and FLAIR sequences for depiction of lung cancer metastases to brain. (a) Diagram of the FLAIR sequence shows a TI of 1700–2200 msec for cerebrospinal fluid. (b) Axial T2-weighted fast SE image shows white matter abnormalities in the left temporal lobe. (c) Axial T2-weighted FLAIR image obtained with nulling of the signal from cerebrospinal fluid shows the metastatic lesions more clearly.
|
|
View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17b. Comparison of fast SE and FLAIR sequences for depiction of lung cancer metastases to brain. (a) Diagram of the FLAIR sequence shows a TI of 1700–2200 msec for cerebrospinal fluid. (b) Axial T2-weighted fast SE image shows white matter abnormalities in the left temporal lobe. (c) Axial T2-weighted FLAIR image obtained with nulling of the signal from cerebrospinal fluid shows the metastatic lesions more clearly.
|
|
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 17c. Comparison of fast SE and FLAIR sequences for depiction of lung cancer metastases to brain. (a) Diagram of the FLAIR sequence shows a TI of 1700–2200 msec for cerebrospinal fluid. (b) Axial T2-weighted fast SE image shows white matter abnormalities in the left temporal lobe. (c) Axial T2-weighted FLAIR image obtained with nulling of the signal from cerebrospinal fluid shows the metastatic lesions more clearly.
|
|
GRE Sequences
In a GRE sequence, an RF pulse is applied that partly flips the net magnetization vector into the transverse plane (variable flip angle) (16). Gradients, as opposed to RF pulses, are used to dephase (negative gradient) and rephase (positive gradients) transverse magnetization (Figs 3, 18). Because gradients do not refocus field inhomogeneities, GRE sequences with long TEs are T2* weighted (because of magnetic susceptibility) rather than T2 weighted like SE sequences. Table 5 lists important differences between SE and GRE sequences. The names of various GRE sequences used by the major MR imaging system vendors are listed in Table 6.
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18a. Comparison of T1-weighted and T2-weighted GRE pulse sequences. (a) GRE pulse sequence diagram shows a variable flip angle and the gradients used to dephase (negative) and rephase (positive) transverse magnetization. Note the locations of the section-selective (Gslice ), phase-encoding (Gphase ), and frequency-encoding (Gfreq/read ) gradients. (b–d) Axial T1-weighted GRE images of the brain (b) and upper abdomen (c), and axial T2*-weighted GRE image of the brain (d), obtained with the pulse sequence in a.
|
|
View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18b. Comparison of T1-weighted and T2-weighted GRE pulse sequences. (a) GRE pulse sequence diagram shows a variable flip angle and the gradients used to dephase (negative) and rephase (positive) transverse magnetization. Note the locations of the section-selective (Gslice ), phase-encoding (Gphase ), and frequency-encoding (Gfreq/read ) gradients. (b–d) Axial T1-weighted GRE images of the brain (b) and upper abdomen (c), and axial T2*-weighted GRE image of the brain (d), obtained with the pulse sequence in a.
|
|
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18c. Comparison of T1-weighted and T2-weighted GRE pulse sequences. (a) GRE pulse sequence diagram shows a variable flip angle and the gradients used to dephase (negative) and rephase (positive) transverse magnetization. Note the locations of the section-selective (Gslice ), phase-encoding (Gphase ), and frequency-encoding (Gfreq/read ) gradients. (b–d) Axial T1-weighted GRE images of the brain (b) and upper abdomen (c), and axial T2*-weighted GRE image of the brain (d), obtained with the pulse sequence in a.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18d. Comparison of T1-weighted and T2-weighted GRE pulse sequences. (a) GRE pulse sequence diagram shows a variable flip angle and the gradients used to dephase (negative) and rephase (positive) transverse magnetization. Note the locations of the section-selective (Gslice ), phase-encoding (Gphase ), and frequency-encoding (Gfreq/read ) gradients. (b–d) Axial T1-weighted GRE images of the brain (b) and upper abdomen (c), and axial T2*-weighted GRE image of the brain (d), obtained with the pulse sequence in a.
|
|
GRE sequences are sensitive to magnetic field inhomogeneity secondary to magnetic susceptibility differences between tissues. Magnetic susceptibility–related signal loss, or susceptibility artifact, is caused by magnetic field (B0) inhomogeneity (17) and can be described in terms of T2* signal decay. This inhomogeneity (local variation in B0) usually occurs at the interface between entities (eg, tissue and air) that have different magnetic susceptibilities. Because magnetic fields vary locally, some spinning nuclei precess faster (recall the Larmor equation, o = B0 x ) than others; so when the individual vectors are added to obtain the net magnetization vector, there is a progressive decrease in the magnitude of the net magnetization vector over time. This decrease results in a progressive decrease in signal intensity, which eventually leads to a signal void. This feature of GRE sequences is exploited for the detection of hemorrhage, as the iron in hemoglobin becomes magnetized locally (produces its own local magnetic field) and thus dephases the spinning nuclei. The technique is particularly helpful for diagnosing hemorrhagic contusions such as those in the brain (Fig 19) and in pigmented villonodular synovitis (Fig 20). SE sequences, on the other hand, while they are relatively immune from magnetic susceptibility artifacts, are also less sensitive in depicting hemorrhage and calcification.
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19a. Comparison of FLAIR and GRE sequences for the depiction of hemorrhage. (a) Axial T2-weighted FLAIR image shows an area of high signal intensity in the right parietal lobe, a finding indicative of hemorrhage. (b) Axial T2-weighted GRE image shows signal loss due to the magnetic susceptibility of hemoglobin in the area of hemorrhage.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19b. Comparison of FLAIR and GRE sequences for the depiction of hemorrhage. (a) Axial T2-weighted FLAIR image shows an area of high signal intensity in the right parietal lobe, a finding indicative of hemorrhage. (b) Axial T2-weighted GRE image shows signal loss due to the magnetic susceptibility of hemoglobin in the area of hemorrhage.
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 20a. Pigmented villonodular synovitis. (a) Coronal protondensity weighted fast SE image obtained with fat saturation shows a large parameniscal cyst that contains punctate low-signal-intensity foci (arrow). (b) Coronal T2*-weighted GRE image depicts the punctate foci (arrow) much more prominently, with a visual effect (blooming artifact) that is related to the magnetic susceptibility of hemosiderin in the area affected by pigmented villonodular synovitis.
|
|
View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 20b. Pigmented villonodular synovitis. (a) Coronal protondensity weighted fast SE image obtained with fat saturation shows a large parameniscal cyst that contains punctate low-signal-intensity foci (arrow). (b) Coronal T2*-weighted GRE image depicts the punctate foci (arrow) much more prominently, with a visual effect (blooming artifact) that is related to the magnetic susceptibility of hemosiderin in the area affected by pigmented villonodular synovitis.
|
|
Magnetic susceptibility imaging is the basis of cerebral perfusion studies, in which the T2* effects (ie, signal decrease) created by gadolinium (a metal injected intravenously as a chelated ion in aqueous solution, typically in the form of gadopentetate dimeglumine) are sensitively depicted by GRE sequences (18) (Fig 21). Magnetic susceptibility is also used in blood oxygenation level–dependent (BOLD) imaging, in which the relative amount of deoxyhemoglobin in the cerebral vasculature is measured as a reflection of neuronal activity. BOLD MR imaging is widely used for mapping of human brain function.
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 21a. MR cerebral perfusion study with normal findings. (a) Axial T1-weighted SE image. (b, c) Corresponding perfusion images show negative enhancement (b) and the maximum enhancement slope (c).
|
|
Top
Abstract
Introduction
