Professional Documents
Culture Documents
Ebook Diffusion Weighted MR Imaging of The Brain Head and Neck and Spine 3Rd Edition Toshio Moritani Editor Online PDF All Chapter
Ebook Diffusion Weighted MR Imaging of The Brain Head and Neck and Spine 3Rd Edition Toshio Moritani Editor Online PDF All Chapter
https://ebookmeta.com/product/imaging-anatomy-head-and-neck-1st-
edition-philip-r-chapman/
https://ebookmeta.com/product/magnetic-resonance-imaging-of-the-
brain-and-spine-fifth-edition-scott-w-atlas/
https://ebookmeta.com/product/diagnostic-imaging-head-and-
neck-4th-edition-bernadette-l-koch-md/
https://ebookmeta.com/product/head-and-neck-pathology-3rd-
edition-lester-d-r-thompson/
Atlas of head and neck pathology 3rd Edition Bruce M.
Wenig
https://ebookmeta.com/product/atlas-of-head-and-neck-
pathology-3rd-edition-bruce-m-wenig/
https://ebookmeta.com/product/head-and-neck-imaging-a-multi-
disciplinary-team-approach-1st-edition-taranjit-singh-tatla/
https://ebookmeta.com/product/textbook-of-ear-nose-throat-and-
head-neck-surgery-clinical-and-practical-3rd-edition-p-hazarika/
https://ebookmeta.com/product/pet-ct-in-head-and-neck-cancer-
clinicians-guides-to-radionuclide-hybrid-imaging-wai-lup-wong-
editor/
https://ebookmeta.com/product/oxford-handbook-of-head-and-neck-
anatomy-1st-edition-van-gijn/
Diffusion Weighted
MR Imaging
of the Brain, Head
and Neck, and Spine
Toshio Moritani
Aristides A. Capizzano
Editors
Third Edition
123
Diffusion Weighted MR Imaging of the 1
Editors 4
Diffusion Weighted MR 5
Spine 8
Third Edition 9
10 Editors
Toshio Moritani Aristides A. Capizzano
Department of Radiology Department of Radiology
University of Michigan Division of Neuroradiology
Director of Clinical Neuroradiology University of Michigan
Research, Division of Neuroradiology Ann Arbor, MI
Ann Arbor, MI USA
USA
27 This Springer imprint is published by the registered company Springer Nature Switzerland AG
28 The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Lila, my mom, for being always there, to my wife, Nicoll, for 29
her relentless patience and to Angela and Emma, source of my 30
inspiration. 31
Aristides A. Capizzano 32
Toshio Moritani 38
Preface to the 3rd edition 39
vii
viii Preface to the 3rd edition
We are sincerely grateful to all the authors and contributors to this book (see 125
the list in authors and contributors section). 126
We wish to thank our Faculty colleagues, residents, fellows, MRI tech- 127
nologists, and other coworkers in the Departments of Radiology and allied 128
Departments at the University of Michigan and the University of Iowa. 129
Our deepest gratitude goes to Ms Danielle Dobbs and Ms Sarah Abate, 130
Media Division, Department of Radiology at the University of Michigan, for 131
the beautiful illustrations. 132
We would like to especially thank Dr Ashok Srinivasan (Director, 133
Neuroradiology, Department of Radiology, University of Michigan), Dr 134
Thomas Chenevert (Professor, Department of Radiology, University of 135
Michigan), Dr Vikas Glani (Chairman, Department of Radiology, University 136
of Michigan), Dr N Reed Dunnick (Former Chairman, Department of 137
Radiology, University of Michigan), Dr Colin P Derdeyn (Chairman, 138
Department of Radiology, University of Iowa), Dr Joan E Maley (Director, 139
Neuroradiology, Department of Radiology, University of Iowa), Dr Yutaka 140
Sato (Director, Pediatric Radiology, Department of Radiology, University of 141
Iowa), and Dr Yuji Numaguchi (Department of Radiology, St. Luke’s 142
International Hospital, Tokyo, Japan), who gave us encouragement and 143
support. 144
ix
Contents 148
xi
180
179
178 xii Contents
181
182 13 Demyelinating Diseases ������������������������������������������������������������������ 313
183 Aristides A. Capizzano and Toshio Moritani
184 14 Degenerative Diseases of the CNS�������������������������������������������������� 353
185 Aristides A. Capizzano and Toshio Moritani
186 15 Toxic and Metabolic Diseases���������������������������������������������������������� 391
187 Aristides A. Capizzano and Toshio Moritani
188 16 Infectious Diseases �������������������������������������������������������������������������� 429
189 Toshio Moritani, Yoshiaki Ota, and Patricia Kirby
190 17 Trauma���������������������������������������������������������������������������������������������� 487
191 Vikas Jain and Toshio Moritani
192 18 Brain Neoplasm�������������������������������������������������������������������������������� 521
193 Jayapalli Rajiv Bapuraj, Toshio Moritani, Shotaro Naganawa,
194 and Akio Hiwatashi
195 19 Pediatrics������������������������������������������������������������������������������������������ 627
196 Lillian Lai and Toshio Moritani
197 20 Head and Neck �������������������������������������������������������������������������������� 715
198 Jerry Kovoor, Jack Kademian, and Toshio Moritani
199 21 Spine Infection���������������������������������������������������������������������������������� 777
200 John Kim, Eiyu Matsumoto, and Toshio Moritani
201 22 Primary and Metastatic Spine Tumors������������������������������������������ 803
202 Patrick W. Hitchon, Shotaro Naganawa, John Kim,
203 Royce W. Woodroffe, Logan C. Helland, Mark C. Smith,
204 and Toshio Moritani
205 23 Spinal Cord Lesions������������������������������������������������������������������������ 839
206 John Kim, Duy Q. Bui, and Toshio Moritani
Aaron Berg Department of Radiology, Sanford Health, Sioux Falls, SD, 221
USA 222
xiii
xiv Contributors
247
284
Myles Horton Section of Neurology, Department of Medicine, University
248 of Manitoba, Winnipeg, MB, Canada
249 Masahiro Ida Department of Radiology, Mito Medical Center, Ibaraki,
250 Japan
251 Karra A. Jones Department of Pathology, University of Iowa Hospitals &
252 Clinics, Iowa City, IA, USA
253 Jinsuh Kim Department of Radiology, University of Alabama at
254 Birmingham, Birmingham, AL, USA
255 Toshibumi Kinoshita Department of Radiology, Akita Cerebrospinal and
256 Cardiovascular Center, Akita, Japan
257 Takashi Kitanosono Northfield International, INC., Honolulu, HI, USA
258 Nobuo Kobayashi Department of Radiology, St. Luke’s International
259 Hospital, Tokyo, Japan
260 Ryo Kurokawa Department of Radiology, University of Tokyo, Tokyo,
261 Japan
262 Andrew Lee Department of Ophthalmology, Blanton Eye Institute, Houston
263 Methodist Hospital, Houston, TX, USA
264 Ho Kyu Lee Department of Radiology, Jeju National University Hospital,
265 Jeju, Republic of Korea
266 Michael Lee Department of Radiology, University of Michigan, Ann Arbor,
267 MI, USA
268 Vincet A. Magnotta Department of Radiology, University of Iowa Hospitals
269 & Clinics, Iowa City, IA, USA
270 Joy Matsui Department of Radiology, UT Southwestern Medical Center,
271 Dallas, TX, USA
272 Mitsuru Matsuki Department of Radiology, Kindai University, Osaka,
273 Japan
274 Harushi Mori Department of Radiology, Jichi Medical University, Tochigi,
275 Japan
276 Minoru Morikawa Department of Radiology, Nagasaki University,
277 Nagasaki, Japan
278 Kristine M. Mosier Department of Radiology & Imaging Sciences, Indiana
279 University, Indianapolis, IN, USA
280 Hisao Nakamura Department of Radiology, St. Marianna University School
281 of Medicine, Kawasaki, Japan
282 Jared W. Nelson Department of Radiology, University of Iowa Hospitals &
283 Clinics, Iowa City, IA, USA
Contributors xv
Min D. Tang-Schomer Penn Center for Brain Injury and Repair and 315
Department of Neurosurgery, University of Pennsylvania, Philadelphia, PA, 316
USA 317
Principles of DW Imaging 2
1
2
Basics of Diffusion Measurements
by MRI
1
3 Arun Venkataraman and Jianhui Zhong
A. Venkataraman · J. Zhong (*) netic field for a certain duration to “tag” different 49
Department of Physics and Astronomy, Center for locations in the tissue of interest. This equates to 50
Advanced Brain Imaging and Neurophysiology, dropping “magic ink” in a location that can be 51
University of Rochester, Rochester, NY, USA
chosen. The application of another gradient after 52
Department of Imaging Sciences, Center for some time informs us about how much the spins 53
Advanced Brain Imaging and Neurophysiology,
University of Rochester, Rochester, NY, USA have diffused in this time. This equates to taking 54
a
400 400
t = 1000
300 t = 500
300
t = 100
200 t= 1
200
100 100
Position
0 0
–100 –100
–200 –200
–300 –300
–400 –400
0 200 400 600 800 1000 0 50 100 150 200 250 300 350
200 200 t= 1
100 100
Position
0 0
–100 –100
–200 –200
–300 –300
–400 –400
0 200 400 600 800 1000 0 50 100 150 200 250
Time (Arbitrary Units) Number of Occurences
Fig. 1.1 Evolution of random walk over discretized time when respective histogram is plotted. The distribution
steps. (a) Shows a step size of one at each time point; the widens over time, and as a function of step size, reflecting
resulting trajectories (left panel) and histogram of position processes that influence “diffusivity.” It is clear that a mol-
over time (right panel) are shown. (b) Same as (a), but ecule represented by (b) would have a higher “diffusivity”
with step size of five. Vertical bars on the left panel denote that one represented by (a)
56 comparing to the initial snapshot to quantify how be identical in different directions, based on the 71
57 much the ink has diffused. However, the analogy local cellular neighborhood. As explained later in 72
58 of introducing “magic ink” to assess diffusivity the chapter, the overall tendencies of diffusion 73
59 ends here. can be measured through metrics such as diffu- 74
sion rate, diffusion coefficient, or diffusivity 75
based on the unit used, and the asymmetry of dif- 76
60 1.3 Diffusion Imaging fusion can be measured using diffusion anisot- 77
61 of the Brain ropy, with a variety of parameters defined [4–7]. 78
86 g radients in addition to radiofrequency and gra- tells us that the intensity of the signal can be 109
87 dient pulses used in conventional MRI imaging. calculated based on the b-factor, the relation- 110
88 A simplified version of the most commonly ship is given by 111
89 used pulse- gradient spin-echo (PGSE) pulse S = S0 e − b·ADC
(1.2) 112
90 sequence for diffusion imaging is shown in
91 Fig. 1.2. Here, ADC is the apparent diffusion coeffi- 113
92 During the diffusion labeling portion, a pair cient and S0 is the signal intensity when there is 114
93 of field gradients are applied to perform no diffusion gradient used. The apparent diffu- 115
94 “diffusion-encoding” (Fig. 1.2, shaded rectan- sion coefficient is so-named because it is an aver- 116
95 gles). Each gradient in this gradient pair lasts for age measure of more complicated processes in 117
96 a time, δ, with gradient magnitude, G (usually the tissues, as discussed in the next section. 118
97 measured in mT/m). The gradient pairs are sepa- Equation (1.2) suggests that there is an expo- 119
98 rated by time Δ. Note that the diffusion gradients nential reduction in the measured signal, S, when 120
99 can be applied in different spatial orientations, diffusion-weighted gradients are applied; this can 121
100 which is relevant to measurements of anisotropic be understood with simple reasoning. The result 122
101 diffusion as we will discuss later in the chapter. of the diffusion gradients is a spatially varying 123
102 We can define the so-called diffusion-weighting magnetic field; as spins diffuse through this mag- 124
103 factor, b, as netic field, each spin is affected differently based 125
on its location and trajectory. The net outcome is 126
δ dephasing of the spins, when the spins are no lon-
b = γ 2 G 2δ 2 ∆ − (1.1) 127
104 3 ger aligned with the magnetic field. Since the sig- 128
105 where γ is the gyromagnetic ratio of a proton nal is the summation of the small signals from 129
106 (42.6 MHz/T). From Eq. (1.1), we can see that individual spins, the dephasing caused by the 130
107 b, also called the b-factor, can be increased gradient pulses leads to a drop in overall signal 131
108 through increases in G, δ, Δ. A formal analysis intensity. It is also evident that the longer the 132
90º 180º
TE/2 TE/2
RF
Acquisition:
SS
Spin Warp
Fast Spin-Echo
RO
Echo Planar Imaging (EPI)
G
PE
d
D
Echo
Fig. 1.2 Pulse sequence diagram for general diffusion which can be achieved with a variety of techniques (three
imaging. Blue portion denotes the period of time corre- examples given). TE echo time, RF radiofrequency, SS
sponding to diffusion labeling; shaded rectangles are dif- slice selection, RO readout, PE phase encoding
fusion gradients. Red portion denotes image acquisition,
6 A. Venkataraman and J. Zhong
Dephasing Signal
j0 j
j1 3
j2
Gx Time
Fig. 1.3 Effects of gradients on diffusing spins. In the there is a reduction in measured signal resulting from any
first panel, we see an example of three spins (black, blue, change in phase. Note that positive and negative phase
and red) diffusing randomly in a one-dimensional mag- changes equally attenuate signal. φ0 represents phase
netic gradient (Gx). The result of motion is dephasing, without any diffusion gradients applied
shown in the second panel. In the final panel, we see that
133 diffusion proceeds, the more dephasing takes the signal when no DW gradients are used. 162
134 place and the lower the signal (Fig. 1.3). Figure 1.2 suggests that this is actually the same 163
as would be obtained from a simple spin-echo 164
sequence. In most clinical scanners, a long TE 165
135 1.5 Apparent Diffusion time (tens of milliseconds) is needed to accom- 166
136 Coefficient modate the diffusion pulses, leading to T2 167
weighting of the S0 image. 168
137 From Eq. (1.2), we can see that when a fixed
138 b-factor is used, tissues with a higher ADC pro-
139 duce a lower signal intensity. Since brain cere- 1.6 Diffusion Represents 169
140 brospinal fluid (CSF) contains water that can a Molecular Event 170
141 move around relatively freely, it has a much
142 higher ADC value than that of other brain tissues Despite resolution of MRI images on the order 171
143 (either gray or white matter), which have diffu- of a millimeter, the information provided by dif- 172
144 sion restriction due to cellular structures. fusion imaging reflects cellular or molecular 173
145 Therefore, in a diffusion-weighted (DW) image, events in much smaller scales. This is because 174
146 one typically sees dark CSF space (due to pro- the molecular diffusion process is highly influ- 175
147 nounced dephasing) and brighter signal in the enced by these events. It can be shown that 176
148 brain tissue (less dephasing). Equation (1.2) also water spins diffuse about tens of micrometers 177
149 tells us that if we collect a series of DW images during a typical MR imaging measurement 178
150 with different b values, we can calculate ADC time, which coincides with the dimension of 179
151 value for each volumetric pixel (voxel) and obtain typical cellular structures. If spins are relatively 180
152 a parametric map, referred to as an ADC map. unrestricted by cellular structures during this 181
153 The calculated ADC map would have voxel time (as is the case in the CSF), the measured 182
154 intensities reflecting the strength of diffusion in diffusion is “free” and “isotropic” (equal in all 183
155 the pixels. Regions of CSF will, therefore, have directions), and ADC is just the intrinsic molec- 184
156 higher intensity than other brain tissues, opposite ular diffusion coefficient. On the other hand, 185
157 of what is seen in the DW images. There are rea- when diffusing spins run into cellular constitu- 186
158 sons the ADC map is advantageous to the DW ents such as cell membranes, the value of ADC 187
159 image. One is the so-called T2 shine-through will be reduced when compared with the value 188
160 effect, which will be discussed in Chap. 2. It in free space. Cellular diffusion events are rep- 189
161 should also be noted that in Eq. (1.2), S0 refers to resented schematically in Fig. 1.4. 190
1 Basics of Diffusion Measurements by MRI 7
r2
r2
Time Time
Fig. 1.4 Displacement profiles for different cell types. displacement graphs (black line, lower panel), compared
Green (left) panel depicts impermeable cell (upper) and to impermeable cell graph (gray line, lower panel).
corresponding mean displacement (<r2>) graph (lower). Diffusion barriers inside a permeable cell or in a cell
The graph shows that water in an impermeable cell is hin- much larger than mean displacement during diffusion
dered, seen by the plateau. Blue (right) panel depicts per- (blue panel) show minimal diffusion barrier effects
meable and large cells, which have identical mean
191 For patients with neurologic abnormalities, the magnitude of water movement alone, DTI 215
192 which change the water distribution in various detects both the magnitude and directionality of 216
193 cellular compartments, or change the ability for water diffusion processes in three dimensions and 217
194 water to pass through cell membranes, measured allows the investigation of the microstructure of 218
195 ADC values will be altered [4–7]. Therefore, MR the CNS by measuring the diffusion properties of 219
196 diffusion measurement offers a unique opportu- water protons in their microenvironment. It is 220
197 nity to obtain information about morphology oth- expected that water molecules move less impeded 221
198 erwise inaccessible to conventional MR imaging along white matter bundles, than perpendicular to 222
199 methods. A wide range of pathological condi- the bundles. The net effect is high directionality 223
200 tions can be explored with water diffusion mea- (high anisotropy) of movement in the white mat- 224
201 surements, as described in later chapters. The ter of the brain. In areas of high concentration of 225
202 measured ADC may also vary depending on the fluid (i.e., ventricles filled with CSF), there is little 226
203 duration of the diffusion process, the direction in preferential directionality of movement (low 227
204 which diffusion is measured, and other factors. anisotropy) and minimal hindrance to diffusion 228
205 For diffusion in an anisotropic environment (high diffusivity); these scenarios are shown in 229
206 (brain white matter consisting of axons with Fig. 1.5. Here, we define fractional anisotropy 230
207 myelin layers wrapped around them, making dif- (FA) and mean diffusivity (MD) in the context of 231
208 fusion along the bundle much easier than through DTI metrics used to measure microscopic motion 232
209 the bundle), diffusion becomes more compli- of tissue water at each image voxel. In a DTI mea- 233
210 cated, and a complete description of the process surement, six or more DWI measurements are 234
211 relies on what is called tensor analysis [8, 9]. required; tensor eigenvalues (λ1, λ2, λ3) and eigen- 235
212 Diffusion tensor imaging (DTI) is an MRI vectors are derived from these measurements [8]. 236
213 sequence that is available on clinical scanners Maps of MD and FA can then be generated from 237
214 with diffusion sequences. Instead of monitoring the eigenvalues with the following relations: 238
8 A. Venkataraman and J. Zhong
λ1 + λ2 + λ3
MD = (1.3)
239 3
2 ( λ1 − MD ) + ( λ2 − MD ) + ( λ3 − MD )
2 2 2
FA = (1.4)
240 3 λ12 + λ2 2 + λ32
241 FA represents the anisotropy of the diffusion 1.7 Requirements in Clinical 265
242 process, yielding values between 0 (perfectly iso- Diffusion Imaging 266
243 tropic diffusion) and 1 (completely anisotropic
244 diffusion). MD represents overall strength of In the clinical setting, certain requirements are 267
245 water mobility in different brain areas. A useful imposed on diffusion studies. Imaging time is 268
246 representation of DTI measurement results is often limited to a few minutes for each scan (T1- 269
247 given by a “color FA” map in which red, green, W, T2-W, diffusion, and others). Multiple slices 270
248 and blue colors are used to represent the orienta- (15–20) are required to cover most of the brain. A 271
249 tion of the largest eigenvalue along left-right good spatial resolution (~5 to 8 mm thick, 1 to 272
250 (LR), anterior-posterior (AP), and superior- 3 mm in plane) is required. A reasonably short TE 273
251 inferior (SI) directions, respectively, while image (<100 ms) to reduce T2 decay and an adequate 274
252 intensity represents FA values in each voxel. diffusion sensitivity (ADC ~ 0.2–1 × 10−3 mm2/s 275
253 By following the main direction of water for brain tissues) are also needed. The most 276
254 movement (the largest tensor eigenvalue) and its important requirement, however, is the almost 277
255 connection in adjacent voxels in the three- complete elimination of sensitivity to subject 278
256 dimensional space, fiber tractography in white motion during scanning. The best compromise so 279
257 matter can be formed. We will discuss more far in clinical diffusion imaging is the use of mul- 280
258 advanced applications, and potentials of going tidirectional (x, y, z), 2 b-factor (b = 0, b ~ 1/ 281
259 beyond the tensor model in the final chapter. The ADC = 1000) single-shot echo-planar imaging 282
260 overall impact of DTI is to increase the diagnos- (EPI) technique. In addition, multiple b = 0 283
261 tic ability of diffusion imaging in degenerative images can be acquired, usually with 6 diffusion- 284
262 neurological diseases where damage to axon weighted images being acquired for every b = 0 285
263 bundles or demyelination can be directly detected image. A clinical scan may last around 5 min, and 286
264 and visualized. allow for the acquisition of up to 60 diffusion- 287
1 Basics of Diffusion Measurements by MRI 9
288 weighted images. Sometimes, fluid attenuation of lesions with restricted diffusion, especially 330
289 with inversion recovery (FLAIR) is used to elimi- acute infarcts, but with allowing a more complete 331
290 nate signal in the highly diffusive CSF space. understanding of more subtle or complicated 332
291 Separation from relaxation effects (i.e., T2 shine- pathologies. 333
292 through) is achieved with calculation of ADC The benefits of improved diffusion contrast at 334
293 instead of just using DW. Elimination of anisotro- high b values can be seen in Eq. (1.2); an identi- 335
294 pic diffusion is achieved by averaging the diffusion cal ADC with higher b value will give a lower 336
295 measurements from three orthogonal directions. signal. The overall effect is a higher sensitivity of 337
the scan to motion. However, we can see this 338
could be problematic in calculating the 339
296 1.8 Setting the b Value in Clinical ADC. Because tissues are described by fast and 340
297 DW Imaging slow components, the results of a two-point mea- 341
surement will depend on the specific b values 342
298 For the sake of standardization in the clinical set- chosen. If the lower b value is set to 0 343
299 ting, it is advisable to maintain the same b value (T2-weighted image) and the upper value is 344
300 for all examinations, making it easier to interpret allowed to vary, the ADC will vary as a function 345
301 images and become aware of the findings in vari- of the upper value. Specifically, one would expect 346
302 ous disease processes. The studies and discus- the measured ADC to decrease as the upper b 347
303 sions presented in this book are limited to DW value increases. 348
304 imaging using b values of 0 and 1000 except in Advances in acquisition hardware alongside 349
305 the case of advanced implementations of DW RF coil development are now making it possible 350
306 imaging (Chap. 7). An upper b-factor of around to collect data from multiple slices of the brain at 351
307 1000 is available on most clinical scanners and once, a technique called simultaneous multislice 352
308 DW imaging at these standard values has been (SMS) imaging [11]. In SMS imaging, a special- 353
309 shown to be a sensitive tool in detecting restricted ized pulse excited multiple slices within the same 354
310 diffusion (e.g., in acute ischemic lesions of the acquisition. In combination with multiple chan- 355
311 brain). DW imaging has become clinically impor- nels in the RF coils, imaging data of those slices 356
312 tant in many other disease processes, which will are acquired. The so-called SMS factor tells us 357
313 be discussed in this book. how many slices are being excited in one acquisi- 358
tion. In diffusion imaging, factors of up to four are 359
used, resulting in an acquisition in one-fourth of 360
314 1.9 Future Trends in Clinical the original time. It should be noted that there is a 361
315 Diffusion Imaging trade-off between SMS factor and signal to noise 362
ratio (SNR); with increased SMS factors, SNR 363
316 Newer DW imaging techniques are using even decreases. In addition, SMS is not as helpful in 364
317 higher b values: 8000 s/mm2 and above. The current clinical sequences, which are already fast. 365
318 advent of “Connectome” scanners (of which SMS, instead, is extremely useful in multi-shell 366
319 there are three in the world) allows for the cre- acquisitions, which can last for up to 2 h without 367
320 ation of extremely large gradients in a small SMS. Further details about multi-shell acquisition 368
321 amount of time. These are currently being used to and its advantages are discussed in Chap. 7. 369
322 acquire extremely high-resolution (1.25 mm iso-
323 tropic) DW images [10]. The increased b values
324 may free up routine DW imaging from its most
325 pressing problem, T2 shine-through. At high b
References 370
326 values, more attention will be focused on the 1. Einstein A (1905) On the movement of small parti- 371
327 actual physiological basis of restricted and facili- cles suspended in stationary liquids required by the 372
328 tated diffusion. Clearly, much of the advantage of molecularkinetic theory of heat. Ann. d. Phys 17. 373
329 increased b values may lie not with the diagnosis 549–560:1. 374
10 A. Venkataraman and J. Zhong
375 2. R. V.P.L.S (2009) XXVII. A brief account of micro- brain and in stroke studied by NMR diffusion imag- 399
376 scopical observations made in the months of June, ing. Magnet Reson Med 19(2):327–332 400
377 July and August 1827, on the particles contained in 7. Zhong J, Petroff O, Prichard J, Gore J (1993) Changes 401
378 the pollen of plants; and on the general existence of in water diffusion and relaxation properties of rat 402
379 active molecules in organic and inorganic bodies. cerebrum during status epilepticus. Magnet Reson 403
380 Philosophical Mag Ser 2(21) (1828):161–173 Med 30(2):241–246 404
381 3. Stejskal E, Tanner J (1965) Spin diffusion mea- 8. Basser PJ, Mattiello J, Lebihan D (1994). MR dif- 405
382 surements: spin echoes in the presence of a time‐ fusion tensor spectroscopy and imaging. J Magnetic 406
383 dependent field gradient. J Chem Phys 42(1): Reson Ser B 66(1):259–267 407
384 288–292 9. Bihan D, van Zijl P (2002) From the diffusion coef- 408
385 4. Bihan LD, Breton E, Lallemand D, Grenier P, ficient to the diffusion tensor. NMR Biomed. 15. 409
386 Cabanis E, Laval-Jeantet M (1986) MR imaging of 7–8:431–434 410
387 intravoxel incoherent motions: application to diffu- 10. Essen DC, Ugurbil K, Auerbach E, Barch D, 411
388 sion and perfusion in neurologic disorders. Radiology Behrens TEJ, Bucholz R, Chang A, Chen L, Corbetta 412
389 161(2):401–407 M, Curtiss SW, Penna DS, Feinberg D, Glasser MF, 413
390 5. Moseley M, Kucharczyk J, Mintorovitch J, Cohen Harel N, Heath AC, Larson-Prior L, Marcus D, 414
391 Y, Kurhanewicz J, Derugin N, Asgari H, Norman Michalareas G, Moeller S, Oostenveld R, Petersen 415
392 D (1990) Diffusion-weighted MR imaging of acute SE, Prior F, Schlaggar BL, Smith SM, Snyder AZ, 416
393 stroke: correlation with T2-weighted and magnetic Xu J, Yacoub E, Consortium W-M (2012) The 417
394 susceptibility-enhanced MR imaging in cats. AJNR Human Connectome Project: a data acquisition per- 418
395 Am J Neuroradiol 11(3): 423–429 spective. Neuroimage 62(4):2222–2231 419
396 6. Moonen CT, Pekar J, de Vleeschouwer MH, van 11. Barth M, Breuer F, Koopmans PJ, Norris DG, Poser 420
397 Gelderen P, van Zijl PC, Despres D (1991) Restricted BA (2016) Simultaneous multislice (SMS) imaging 421
398 and anisotropic displacement of water in healthy cat techniques. Magnet Reson Med 75(1):63–81 422
1
2
Diffusion-Weighted and Tensor
Imaging of the Normal Brain
2
3 Noriko Salamon and Toshio Moritani
N. Salamon
Department of Radiological Sciences, Ronald Reagan
UCLA Medical Center, Los Angeles, CA, USA 2.2.3 Choroid Plexus 45
e-mail: nsalamon@mednet.ucla.edu
T. Moritani (*) The choroid plexus occasionally shows promi- 46
Division of Neuroradiology, University of Michigan, nent hyperintensity on DWI associated with mild 47
Ann Arbor, MI, USA elevation of ADC. In these situations the ADC is 48
e-mail: tmoritan@med.umich.edu
a b c
d e f
Fig. 2.1 Normal adult brain of a 40-year-old male with- isotropic DWI are normal and are caused by T2 contrast.
out neurological deficits. (a) Isotropic DWI is obtained by (b) ADC map shows homogeneous ADC values in globi
combining b0 image and three orthogonal unidirectional pallidi, corticospinal tracts, gray and white matter. (c) b0
images (x, y, z axis). The bilateral globi pallidi have low image shows low signal in globi pallidi (arrows), high sig-
signal on DWI as a result of physiological iron deposition nal in corticospinal tracts (arrowheads), and the gray-
(arrows). Corticospinal tracts have mildly high signal on white matter contrast. (d–f) Diffusion weighting is applied
DW image (arrowheads). Gray matter shows mildly high in x axis (d), y axis (e), and z axis (f)
signal compared to white matter. These signal changes on
49 often higher than in white matter, but lower than varies markedly within different areas of the 60
50 in cerebrospinal fluid. The high DW signal is brain and is higher in white matter 61
51 believed to represent gelatinous cystic changes of (1.13 × 10−3 mm2/s) than in gray matter 62
52 the choroid plexus, which can occur with age (1.02 × 10−3 mm2/s) [13]. ADC at birth is higher 63
53 (Fig. 2.3) [9]. in subcortical white matter (1.88 × 10−3 mm2/s) 64
than in both the anterior (1.30 × 10−3 mm2/s) and 65
posterior limbs of the internal capsule 66
54 2.3 Pediatric Brain (1.09 × 10−3 mm2/s). It is also higher in cortex 67
and the caudate nucleus (1.34 × 10−3 mm2/s) than 68
55 2.3.1 Diffusion-Weighted Imaging in the thalamus and the lentiform nucleus 69
56 and ADC of the Pediatric Brain (1.20 × 10−3 mm2/s) [15]. With the exception of 70
the cerebrospinal fluid (CSF), there is a trend of 71
57 The normal brain of neonates and infants has sig- decreasing ADC with increasing maturation in 72
58 nificantly higher ADC values than the adult brain most areas of the pediatric brain. These ADC 73
59 [10–15] (Fig. 2.4). ADC in neonates and infants changes seem to reflect a combination of differ- 74
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 13
a b
Fig. 2.2 (a) Decussation of the superior cerebellar peduncle has mildly high signal on DWI (arrow). (b) ADC map
shows similar ADC values in the decussation of the superior cerebellar peduncle to those in cerebral peduncles (arrow)
a b c
d e f
Fig. 2.3 Cystic changes in the choroid plexus. (a) DWI enhanced T1-weighted image with magnetization transfer
shows hyperintensity in cystic changes of the left choroid contrast reveals no enhancement in it (arrow). (e)
plexus (arrow). (b) ADC values of the cystic changes are Macropathology of choroid plexus cyst (another case). (f)
lower than those of the CSF, which probably represent vis- Micropathology shows gelatinous content, foci of calcifi-
cous gelatinous materials, but higher than those of brain cation with a thickened fibrous wall containing blood ves-
parenchyma (arrow). (c) T2-weighted image shows the sels. (Courtesy of Kinoshita T, MD, Research Institute for
cystic changes as hyperintensity (arrow). (d) Gadolinium- Brain and Blood Vessels Akita, Japan)
14 N. Salamon and T. Moritani
a b
Fig. 2.4 Normal neonatal brain. (a) The appearance of white matter (arrows). These ADC changes seem to reflect
the pediatric brain on DWI varies with age. In neonates it a combination of factors, including a reduction of overall
is normal to have low DW signal intensities in the frontal water content, cellular maturation, and white matter
deep white matter (arrows). (b) ADC values of the corre- myelination
sponding areas are high in neonatal brain, especially in the
75 ent factors, including a reduction of overall water image voxel, yielding values between 0 (per- 92
76 content, cellular maturation, and white matter fectly isotropic diffusion) and 1 (completely 93
77 myelination. In neonates and infants, ischemia is anisotropic diffusion) (Fig. 2.5). A color FA map 94
78 usually global and can therefore resemble the represents the directionality of the anisotropy 95
79 normal image with elevated DW signal and based on the orientation of the largest eigenvalue 96
80 decreased ADC. White matter diseases can also along left-right (red), anterior-posterior (green), 97
81 be mimicked by the normal, age-related appear- and superior-inferior directions (blue) (Fig. 2.5). 98
82 ance of DWI and ADC. Out of necessity, the A fiber tractography map is made by follow- 99
83 ADC values will therefore have to be age related ing the main direction of water movement (the 100
84 for a correct interpretation of the DWI of the largest tensor eigenvalue) and its connection in 101
85 pediatric brains. adjacent voxels in the three-dimensional space 102
(Figs. 2.6, 2.7, 2.8, 2.9, 2.10, and 2.11). DTI with 103
these maps has been used to investigate the 104
86 2.4 Diffusion Tensor Imaging microstructure of the CNS, and their clinical use- 105
87 and White Matter Anatomy fulness has been reported in the literature [16– 106
18]. DTI can beautifully demonstrate white 107
88 Diffusion tensor imaging (DTI) is available in matter fiber tracts of the CNS in detail that is use- 108
89 many modern clinical scanners. A fractional ful for understanding the anatomy. 109
90 anisotropy (FA) map represents the magnitude of White matter fiber tracts consist of myelinated 110
91 the anisotropy of the diffusion process at each fiber bundles, called the fasciculus. The white 111
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 15
112 matter fibers are traditionally classified as fol- 2.4.1 Association Fibers 122
113 lows: (1) association fibers that interconnect vari-
114 ous cortical areas of the same hemisphere, (2) Short association fibers arch beneath the six lay- 123
115 projection fibers that interconnect cortical areas ers of the cortex referred to as U-fibers or arcuate 124
116 with the thalami, brain stem, cerebellum, and spi- fibers. In long association fibers, the superior lon- 125
117 nal cord, and (3) commissural fibers that inter- gitudinal fasciculus (SLF) is the largest bundle, 126
118 connect corresponding cortical areas of the two connecting the frontal lobe cortex to the temporal 127
119 hemispheres [19–22]. However, the classification and occipital lobe cortices (Fig. 2.6). The arcuate 128
120 of the fiber bundles is complex and to date has fasciculus (AF) is the part of the SLF that forms 129
121 remained unresolved [23–27]. a curved shape and connects the Broca area 130
a b c
d e f
Fig. 2.5 (a) Fractional anisotropy (FA) map, (b–f) axial AF arcuate fasciculus, ILF inferior longitudinal fascicu-
DTI color maps, (g–i) coronal DTI color maps, (j–m) sag- lus, CG cingulum, UF uncinate fasciculus, SFO superior
ittal DTI color maps, (g–i) coronal DTI color maps, (j–m) fronto-occipital fasciculus, IFO inferior fronto-occipital
sagittal DTI color maps. DTI shows the white matter lat- fasciculus, EmC extreme capsule, EC external capsule,
eral to the posterior horn as three parts: the tapetum (TP), MB subcallosal fasciculus of Muratoff, IC internal cap-
internal and external sagittal strata including optic radia- sule, CST corticospinal tract, OT optic tract, OR optic
tion (SS/OR), and SLF. SLF/AF connects the frontal lobe radiation, SS sagittal stratum, gCC genu of corpus callo-
to temporal and occipital lobes. SFO/MB is situated sum, bCC body of corpus callosum, sCC splenium of cor-
beneath the corpus callosum (CC) and medial to the pus callosum, TP tapetum, AC anterior commissure, FX
corona radiata (CR). ILF connects temporal and occipital fornix, ST stria terminalis, SCP superior cerebellar pedun-
lobe cortices. The cingulum (CG) is situated on both sides cle, MCP middle cerebellar peduncle. (Courtesy of Kim J
of the midline on the surface of CC. The uncinate fascicu- MD, The University of Iowa Hospitals and Clinics, USA,
lus (UF) connects the ventral and lateral frontal lobe with and White ML MD and Zhang Y MD, The University of
the anterior temporal lobe. EmC/EC is hard to be sepa- Nebraska Medical Center, USA)
rately seen on DTI. SLF superior longitudinal fasciculus,
16 N. Salamon and T. Moritani
g h i
j k l
131 (motor area of speech) with the Wernicke area the parahippocampal gyrus and adjacent tempo- 144
132 (language comprehension). The inferior longitu- ral lobe cortex. The uncinate fasciculus (UF) con- 145
133 dinal fasciculus (ILF) connects the temporal and nects the ventral and lateral frontal lobe cortex 146
134 occipital lobe cortices, which are concerned with with the anterior temporal lobe cortex, which is 147
135 visual perception, object identification, and concerned with memory integration (retrograde 148
136 recognition. amnesia). 149
137 The cingulum (CG) is situated on both sides The existence of the superior or inferior fronto- 150
138 of the midline on the peripheral surface of the occipital (occipitofrontal) fasciculus is controver- 151
139 corpus callosum and courses within the cingulate sial [23–27]. The superior fronto- occipital 152
140 gyrus that is concerned with behavior and regula- fasciculus (SFO) connects the frontal and parietal 153
141 tion of emotional processing (Fig. 2.7). The cin- lobes. It is thought to be concerned with space 154
142 gulum bundle contains variable length fibers perception. Ventrally its base rests on the subcal- 155
143 which connect the frontal and parietal lobes with losal fasciculus of Muratoff (MB) [24, 26]. The 156
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 17
a b
c d
Fig. 2.6 (a, b) The superior longitudinal fasciculus forms a curved shape and connects the Wernicke area with
(SLF) interconnects frontal lobe cortex to temporal and Broca area. (d) There is asymmetry in right and left side
occipital lobe cortices. The inferior longitudinal fascicu- of arcuate fasciculi (Courtesy of White ML MD and
lus (ILF) connects temporal and occipital lobe cortices. Zhang Y MD, The University of Nebraska Medical
(c) The arcuate fasciculus (AF) is the part of the SLF that Center, USA)
157 SFO/MB is situated beneath the corpus callosum The extreme capsules (EmC) contain associa- 168
158 and medial to the corona radiata (Fig. 2.5). The tion fibers parallel to the external capsule (EC), 169
159 SFO may not exist based on analyses using fiber which are separated by a sheet of gray matter 170
160 dissection techniques [23]. The inferior fronto- known as the claustrum. 171
161 occipital fasciculus (IFO) connects the frontal and Corticostriatal fibers interconnect the cortex 172
162 temporo-occipital areas intermingled with the with the striatum including the external capsule 173
163 uncinate fasciculus antero- inferiorly, possibly and the subcallosal fasciculus of Muratoff (MB) 174
164 concerned with auditory-visual association [24] concerned with motor and cognitive–affective 175
165 (Fig. 2.7). The IFO may not exist based on analy- performance. The corticostriatal fibers are not 176
166 ses using isotope anterograde tract tracer in the exactly association fibers, based on the defini- 177
167 macaque monkey [24–27]. tion [25, 28]. 178
18 N. Salamon and T. Moritani
a b
Fig. 2.7 (a) The cingulum bundle contains variable lobe cortex with the anterior temporal lobe cortex. The
length fibers which connect frontal and parietal lobes with fornix (FX) (light blue) is also demonstrated. (Courtesy of
parahippocampal gyrus and adjacent temporal lobe cor- Aoki S, MD, The University of Tokyo, Japan). (c) The
tex. (Courtesy of White ML, MD and Zhang Y MD, inferior fronto-occipital fasciculus (IFO) connects frontal
University of Nebraska Medical Center, USA). (b) The and temporo-occipital areas intermingled with the unci-
cingulum (orange) is situated on both sides of the midline nate fasciculus antero-inferiorly. (Courtesy of White ML
on the peripheral surface of the corpus callosum and MD and Zhang Y MD, The University of Nebraska
courses within the cingulate gyrus. The uncinate fascicu- Medical Center, USA)
lus (UF) (pink) connects the ventral and lateral frontal
179 2.4.2 Projection Fibers [29, 30]. The most posterior component of the 187
posterior limb of the internal capsule contains 188
180 The corona radiata and internal capsule (IC) are fibers, connecting the lateral geniculate nucleus 189
181 composed of afferent and efferent fibers to and to the calcarine sulcus (geniculo-calcarine tract), 190
182 from the entire cerebral cortex. Efferent fibers in also known as optic radiation (OR) (Fig. 2.9). 191
183 the internal capsule arise from the cortex and The ventral bundle of the optic radiation passes 192
184 project to the brain stem and spinal cord and are anteriorly before making a sharp turn, known as 193
185 categorized as corticothalamic, corticopontine, the Meyer loop, and terminating in the inferior 194
186 corticobulbar, and corticospinal tracts (Fig. 2.8) bank of the calcarine cortex. The temporal stem 195
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 19
a b
c d
Fig. 2.8 (a, b) Anterior-posterior and oblique lateral Japan). (d) Multi-tensor DTI separates each corticospinal
views of the corticospinal tract. (Courtesy of White ML fiber projecting special areas: hand (green), tongue
MD and Zhang Y MD, University of Nebraska Medical (orange), trunk (purple), face (blue), and lower extremity
Center, USA). (c) The corticospinal tract (orange) and (yellow). (Courtesy of Yamada K MD, Kyoto Prefectural
corticobulbar tract (light blue) are demonstrated. University of Medicine, Japan) (from [29])
(Courtesy of Aoki S, MD, The University of Tokyo,
196 is a critical landmark in temporal lobe surgery, 2.4.3 Commissural Fibers 208
197 which is crossed by the uncinate fasciculus, infe-
198 rior fronto-occipital fasciculus, and Meyer’s loop The corpus callosum (CC) is a broad thick plate 209
199 [24, 31]. The optic radiation is part of the sagittal of dense myelinated fibers that reciprocally 210
200 stratum. The sagittal stratum not only conveys interconnect broad regions of the cortex 211
201 fibers from parietal, occipital, temporal, and cin- (Fig. 2.11). The genu (gCC) contains fibers 212
202 gulate regions to the thalamus and brain stem, but interconnecting rostral parts of the frontal lobes. 213
203 also conveys principally from the thalamus to the Fibers from the remaining parts of the frontal 214
204 cortex. The thalamocortical (afferent) and corti- lobe and the parietal lobe traverse the body of 215
205 cothalamic (efferent) fibers are arrayed around the corpus callosum (bCC). Fibers transversing 216
206 the thalamus, termed thalamic radiations or the splenium relate regions of the temporal and 217
207 peduncles (Fig. 2.10). occipital lobes. The tapetum comprises the 218
20 N. Salamon and T. Moritani
a c
Fig. 2.9 (a) Optic radiations on the right (OR), (b) Center, USA). (c) The relationship between the Meyer’s
Meyer’s loop (arrows). (Courtesy of White ML MD and loop (green) and the uncinate fasciculus. (Courtesy of
Zhang Y MD, The University of Nebraska Medical Taoka T MD, Nara Medical University, Japan)
219 fibers in the splenium (sCC) which sweep infe- pocampi also known as the commissure of the 230
220 riorly along the lateral margin of the posterior fornix (FX). The fornix is the main efferent 231
221 horn of the lateral ventricle. The white matter fiber system of the hippocampal formation, 232
222 lateral to the posterior horn is divided into four including both projection and commissural 233
223 layers: the tapetum, the internal and external fibers, associated with antegrade amnesia. The 234
224 sagittal strata, and the SLF [22, 32]. The ante- stria terminalis is an efferent pathway from the 235
225 rior commissure (AC) is a small compact bundle amygdala, which arches along the medial bor- 236
226 that crosses the midline rostral to the columns of der of the caudate nucleus and terminates in the 237
227 the fornix. The posterior commissure connects hypothalamic nuclei lateral to the fornix. The 238
228 the right and left peritectal region. The hippo- stria terminalis is associated with the hypothal- 239
229 campal commissure joins the right and left hip- amus-pituitary-adrenal axis. 240
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 21
a b c
Fig. 2.10 Thalamic radiations (a anterior, b superior, c posterior) (Courtesy of White ML MD and Zhang Y MD, The
University of Nebraska Medical Center, USA)
a b c
Fig. 2.11 (a–c)The corpus callosum is a broad thick (bCC). Fibers transversing the splenium relate regions of
plate of dense myelinated fibers that reciprocally intercon- the temporal and occipital lobes. The tapetum (TP) is the
nect broad regions of the cortex. The genu (gCC) contains extension into the hemisphere of the corpus callosum
fibers interconnecting rostral parts of the frontal lobes. lying adjacent to the ventricular ependyma. (Courtesy of
Fibers from the remaining parts of the frontal lobe and the White ML MD and Zhang Y MD, The University of
parietal lobe traverse the body of the corpus callosum Nebraska Medical Center, USA)
a b
c d
Fig. 2.12 (a) The midbrain: The crus cerebri consists of pontine fibers (TPF). The middle cerebellar peduncle
corticopontine, corticobulbar, and corticospinal fibers (MCP) arises from pontine nuclei projecting to the cere-
(CPT, CBT, CST). The tegmentum and crus cerebri are bellar hemisphere. The inferior cerebellar peduncle is
separated by the substantia nigra (SN). The superior cer- noted on each side of the lower part of the forth ventricle.
ebellar peduncle (SCP) and their decussation (DSCP, (d) The medulla: The inferior olivary nuclei (ION), med-
round red spot) are noted in the caudal midbrain tegmen- ullary pyramids (CST, CBT), medial lemniscus (ML), and
tum. The dorsal tegmental portion contains the central ICP are demonstrated. (Courtesy of White ML MD and
tegmental tract (CTT) and medial lemniscus (ML). The Zhang Y MD, The University of Nebraska Medical
upper (b) and lower (c) pons: The ventral part of the pons Center, USA)
contains CPT, CBT, CST, pontine nuclei, and transverse
262 the reticular formation, medial lemniscus, and nuclei, and most of them decussate and terminate 272
263 cranial nerve nuclei. The medial lemniscus as mossy fibers projecting to the cerebellar hemi- 273
264 comprises large ascending fibers and terminates sphere. The function of the MCP includes initia- 274
265 in the ventral posterolateral nucleus of the thala- tion, planning, and voluntary movement. The 275
266 mus, which is concerned with sensation of medulla contains the inferior olivary nuclei, cra- 276
267 touch, vibration, proprioception, and two-point nial nerve nuclei, and decussation of the medul- 277
268 discrimination. lary pyramids and medial lemniscus. 278
269 The middle cerebellar peduncle (MCP) also The inferior cerebellar peduncle (ICP), seen 279
270 known as the brachium pontis consists of predom- on each side of the lower part of the fourth ven- 280
271 inantly afferent fibers arising from the pontine tricle, connects the cerebellum and medulla 281
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 23
282 (Fig. 2.13). The ICP carries mostly afferent path- 2.4.5 linical Importance of White
C 293
283 ways from the spinal cord and conveys impulses Matter Fiber Anatomy 294
284 from vestibular receptors, associated with bal-
285 ance coordination. The corpus medullare is a White matter anatomy is a developing area in 295
286 compact mass of cerebellar white matter imbed- neuroimaging, and the structure and function of 296
287 ding four deep nuclei (dentate, fastigial, globose, many white matter bundles are incompletely 297
288 and emboliform), continuous with the three cer- understood. Nevertheless, knowledge of white 298
289 ebellar peduncles [34]. The corpus medullare matter fiber tract anatomy can be clinically use- 299
290 splits in the roof of the fourth ventricle at an acute ful. The most common context in which white 300
291 angle (fastigium) and forms the superior and matter anatomy is clinically important is in pre- 301
292 inferior medullary vela. operative planning, most often to avoid unneces- 302
Fig. 2.13 a–c Axial (a), Coronal (b), and Sagittal (c) images. The corpus medullare is a compact mass of cere-
images. The ventral part of the pons consists of pyramidal bellar white matter imbedding 4 deep cerebellar nuclei
tracts (PrT), pontine nuclei, and transverse pontine fibers (dentate, fastigial, globose, and emboliform), continuous
(TPF) projecting to the cerebellum. The medial lemniscus with the three cerebellar peduncles. The superior cerebel-
(ML) is large ascending fibers and terminates in the ven- lar peduncle (SCP) is seen on each side of the upper part
tral posterolateral nucleus of the thalamus. The middle of the forth ventricle and decussates completely in the
cerebellar peduncle (MCP) consists of predominantly caudal midbrain tegmentum. (g, h) Fiber tractography.
afferent fibers from TPF and most of them decussate and SCP is predominantly efferent pathways arising from cer-
terminate as mossy fibers projecting to the cerebellar ebellar cortex and deep cerebellar nuclei to thalamus and
hemisphere. The inferior cerebellar peduncle (ICP), seen cerebral cortex (green). Pyramidal tracts (blue) and mid-
on each side of the lower part of the forth ventricle, con- dle cerebellar peduncles and transverse pontine fibers
nects the cerebellum and medulla. (d–f) Parasagittal (orange) are demonstrated
24 N. Salamon and T. Moritani
g h
a b
Fig. 2.14 A 45-year-old woman with pure alexia follow- connect the commissural fibers (red arrow) that connect
ing left PCA territory infarction. (a) DWI demonstrates the normal right occipital lobe (pink) to the visual-verbal
diffusion restriction in the left occipital lobe, left splenium association cortex in the left inferior parietal lobule.
of the corpus callosum, and left peritrigonal white matter. Therefore, the patient cannot read despite preserved abil-
(b) The left occipital lobe infarction (yellow) and the ity to see letters and write
lesion in the splenium of the corpus callosum (red X) dis-
303 sary injury to projection fibers like the pure alexia without agraphia (i.e., the inability to 323
304 corticospinal tracts and optic radiations. Injuries read with preserved ability to see letters and 324
305 to projection fibers are straightforward to under- write) that can occur with injury to the occipital 325
306 stand and can result in devastating deficits in pri- lobe and splenium of the corpus callosum in the 326
307 mary motor, sensory, and visual function. Injury language dominant hemisphere (Fig. 2.14). These 327
308 to other white matter fibers is less straightforward patients present with right hemianopsia and an 328
309 to understand but can explain deficits in higher inability to read even when written words are pre- 329
310 brain functions that result from association sented to the functioning visual field. The reading 330
311 between cortical areas. These deficits, called dis- deficit is due to a disconnection in the white mat- 331
312 connection syndromes, occur secondary to dam- ter fibers connecting the normal right occipital 332
313 age to one or more of the relevant cortical regions, lobe to the visual-verbal association cortex in the 333
314 the association white matter fibers connecting left inferior parietal lobule. 334
315 them, or both [35–37]. The prototypical discon-
316 nection syndrome is the conduction aphasia that
317 results from injury to association fibers in the 2.5 Conclusion 335
318 arcuate fasciculus that connect Broca’s and
319 Wernicke’s areas. These patients have normal Knowledge of the DWI appearance of the normal 336
320 speech fluency and language comprehension but adult and pediatric brain and variations is neces- 337
321 exhibit paraphasic speech and repetition deficits. sary to avoid misinterpretation. In children it is 338
322 Another classical disconnection syndrome is also important to match the findings with those of 339
2 Diffusion-Weighted and Tensor Imaging of the Normal Brain 27
340 normal children of the same age. DTI and fiber 14. Engelbrecht V, Scherer A, Rassek M, Witsack HJ, 395
341 tractography are useful for understanding the Modder U (2002) Diffusion-weighted MR imaging 396
in the brain in children: findings in the normal brain 397
342 normal white matter fiber tract anatomy. White and in the brain with white matter diseases. Radiology 398
343 matter anatomy is a developing area in neuroim- 222:410–418 399
344 aging, and the structure and function of many 15. Forbes KP, Pipe JG, Bird CR (2002) Changes in brain 400
345 white matter bundles are incompletely under- water diffusion during the 1st year of life. Radiology 401
222:405–409 402
346 stood. Nevertheless, knowledge of white matter 16. Nucifora PGP, Verma R, Lee SK, Melhem ER (2007) 403
347 fiber tract anatomy can be clinically useful. Diffusion-tensor MR imaging and tractography: 404
exploring brain microstructure and connectivity. 405
Radiology 245:367–384 406
17. Assaf Y, Pasternak O (2008) Diffusion tensor imaging 407
348 References (DTI)-based white matter mapping in brain research: 408
a review. J Mol Neurosci 34:51–61 409
349 1. Yoshiura T, Wu O, Sorensen AG (1999) Advanced 18. Alexander AL, Lee JE, Lazar M, Field AS 410
350 MR techniques: diffusion MR imaging, perfusion MR (2007) Diffusion tensor imaging of the brain. 411
351 imaging, and spectroscopy. Neuroimaging Clin N Am Neurotherapeutics 4:316–329 412
352 9:439–453 19. Jellison BJ, Field AS, Medow J, Lazar M, Salamat 413
353 2. Asao C, Hirai T, Yoshimatsu S et al (2007) Human MS, Alexander AL (2004) Diffusion tensor imaging 414
354 cerebral cortices: signal variation on diffusion- of cerebral white matter: a pictorial review of phys- 415
355 weighted MR imaging. Neuroradiology 50:205–211 ics, fiber tract anatomy, and tumor imaging patterns. 416
356 3. Chun T, Filippi CG, Zimmerman RD, Ulug AM AJNR Am J Neuroradiol 25:356–369 417
357 (2000) Diffusion changes in the aging human brain. 20. Wakana S, Jiang H, Nagae-Poetscher LM, van Zijl 418
358 Am J Neuroradiol 21:1078–1083 PC, Mori S (2004) Fiber tract-based atlas of human 419
359 4. Engelter ST, Provenzale JM, Petrella JR, DeLong white matter anatomy. Radiology 230:77–87 420
360 DM, MacFall JR (2000) The effect of aging on the 21. Mamata H, Mamata Y, Westin CF, Shenton ME, 421
361 apparent diffusion coefficient of normal-appearing Kikinis R, Jolesz FA, Maier SE (2002) High- 422
362 white matter. Am J Roentgenol 175:425–430 resolution line scan diffusion tensor MR imaging 423
363 5. Helenius J, Soinne L, Perkio J (2002) Diffusion- of white matter fiber tract anatomy. AJNR Am J 424
364 weighted MR imaging in normal human brains in Neuroradiol 23:67–75 425
365 various age groups. Am J Neuroradiol 23:194–199 22. Tamura H, Takahashi S, Kurihara N, Yamada S, 426
366 6. Gideon P, Thomsen C, Henriksen O (1994) Increased Hatazawa J, Okudera T (2003) Practical visualization 427
367 self-diffusion of brain water in normal aging. J Magn of internal structure of white matter for image inter- 428
368 Reson Imaging 4:185–188 pretation: staining a spin-echo T2-weighted image 429
369 7. Nusbaum AO, Tang CY, Buchsbaum MS, Wei TC, with three echo-planar diffusion-weighted images. 430
370 Atlas SW (2001) Regional and global changes in cere- AJNR Am J Neuroradiol 24:401–409 431
371 bral diffusion with normal aging. Am J Neuroradiol 23. Türe U, Yaşargil MG, Pait TG (1997 Jun) Is 432
372 22:136–142 there a superior occipitofrontal fasciculus? A 433
373 8. Abe O, Aoki S, Hayashi N et al (2002) Normal microsurgical anatomic study. Neurosurgery 434
374 aging in the central nervous system: quantitative 40(6):1226–1232 435
375 MR diffusion- tensor analysis. Neurobiol Aging 24. Kier EL, Staib LH, Davis LM, Bronen RA (2004 436
376 23:433–441 May) MR imaging of the temporal stem: anatomic 437
377 9. Kinoshita T, Moritani T, Hiwatashi A et al (2005) dissection tractography of the uncinate fascicu- 438
378 Clinically silent choroid plexus cyst: evaluation by lus, inferior occipitofrontal fasciculus, and Meyer’s 439
379 diffusion-weighted MRI. Neuroradiology 47:251–255 loop of the optic radiation. AJNR Am J Neuroradiol 440
380 10. Sakuma H, Nomura Y, Takeda K et al (1991) Adult 25(5):677–691 441
381 and neonatal human brain: diffusional anisotropy and 25. Schmahmann JD, Pandya DN (2006) Fiber pathways 442
382 myelination with diffusion-weighted MR imaging. of the brain. Oxford University Press, Oxford, NY 443
383 Radiology 180:229–233 26. Schmahmann JD, Pandya DN, Wang R, Dai G, 444
384 11. Morriss MC, Zimmerman RA, Bilaniuk LT, Hunter D’Arceuil HE, de Crespigny AJ, Wedeen VJ (2007) 445
385 JV, Haselgrove JC (1999) Changes in brain water dif- Association fibre pathways of the brain: parallel 446
386 fusion during childhood. Neuroradiology 41:929–934 observations from diffusion spectrum imaging and 447
387 12. Tanner SF, Ramenghi LA, Ridgway JP et al (2000) autoradiography. Brain 130:630–653 448
388 Quantitative comparison of intrabrain diffusion in 27. Schmahmann JD, Pandya DN (2007) The com- 449
389 adults and preterm and term neonates and infants. Am plex history of the fronto-occipital fasciculus. J Hist 450
390 J Roentgenol 174:1643–1649 Neurosci 16:362–377 451
391 13. Neil JJ, Shiran SI, McKinstry RC et al (1998) Normal 28. Utsunomiya H, Nakamura Y (2007) MR features of 452
392 brain in human newborns: apparent diffusion coeffi- the developing perianterior horn structure includ- 453
393 cient and diffusion anisotropy measured by using dif- ing subcallosal fasciculus in infants and children. 454
394 fusion tensor MR imaging. Radiology 209:57–66 Neuroradiology 49:947–954 455
28 N. Salamon and T. Moritani
456 29. Yamada K, Sakai K, Hoogenraad FG, Holthuizen 32. Kitajima M, Korogi Y, Takahashi M, Eto K (1996) 476
457 R, Akazawa K, Ito H, Oouchi H, Matsushima S, MR signal intensity of the optic radiation. AJNR Am 477
458 Kubota T, Sasajima H, Mineura K, Nishimura T J Neuroradiol 17:1379–1383 478
459 (2007) Multitensor tractography enables better depic- 33. Salamon N, Sicotte N, Alger J, Shattuck D, Perlman 479
460 tion of motor pathways: initial clinical experience S, Sinha U, Schultze-Haakh H, Salamon G (2005) 480
461 using diffusion-weighted MR imaging with standard Analysis of the brain-stem white-matter tracts 481
462 b-value. AJNR Am J Neuroradiol 28:1668–1673 with diffusion tensor imaging. Neuroradiology 482
463 30. Aoki S, Iwata NK, Masutani Y et al (2005) 47:895–902 483
464 Quantitative evaluation of the pyramidal tract seg- 34. Salamon N, Sicotte N, Drain A, Frew A, Alger JR, 484
465 mented by diffusion tensor tractography: feasibility Jen J, Perlman S, Salamon G (2007) White matter 485
466 study in patients with amyotrophic lateral sclerosis. fiber tractography and color mapping of the normal 486
467 Radiat Med 23:195–199 human cerebellum with diffusion tensor imaging. J 487
468 31. Taoka T, Sakamoto M, Nakagawa H, Nakase H, Neuroradiol 34:115–128 488
469 Iwasaki S, Takayama K, Taoka K, Hoshida T, Sakaki 35. Catani M, Mesalum M (2008) What is a disconnec- 489
470 T, Kichikawa K (2008) Diffusion tensor tractography tion syndrome? Cortex 44:911–913 490
471 of the Meyer loop in cases of temporal lobe resec- 36. Catani M, ffytche D (2005) The rises and falls of dis- 491
472 tion for temporal lobe epilepsy: correlation between connection syndromes. Brain 128:2224–2239 492
473 postsurgical visual field defect and anterior limit of 37. Catani M, Mesalum M (2008) The arcuate fasciculus 493
474 Meyer loop on tractography. AJNR Am J Neuroradiol and the disconnection theme in language and aphasia: 494
475 29(7):1329–1334. [Epub ahead of print] history and current state. Cortex 44:953–961 495
1
2
Pitfalls and Artifacts
of Diffusion-Weighted Imaging
3
3 Toshio Moritani, Akio Hiwatashi,
4 and Jianhui Zhong
19 called T2 washout. Also the effect of T2 shorten- SIb = 0 is the signal intensity on the spin-echo echo- 34
20 ing, or T2 blackout, and magnetic susceptibility planar image (b0 image) [1, 2, 5, 7, 8, 10, 12, 16]. 35
21 effects will influence the DW appearance in many To evaluate the tissue T2 and ADC, we should 36
pay attention to the images discussed below as 37
well as isotropic DWIs and b0 images [3–5, 7, 8, 38
T. Moritani (*)
Division of Neuroradiology, Department of Radiology, 10, 11, 13–16]. 39
University of Michigan, Ann Arbor, MI, USA
e-mail: tmoritan@med.umich.edu
A. Hiwatashi 3.2.2 Apparent Diffusion 40
Department of Clinical Radiology, Kyushu Coefficient Maps 41
University, Fukuoka, Japan
e-mail: hiwatasi@radiol.med.kyushu-u.ac.jp
To evaluate the diffusibility, ADC is calculated 42
J. Zhong
as: 43
University of Rochester Medical Center,
Rochester, NY, USA ADC = − ln ( SI / SI b = 0 ) / b 44
e-mail: jianhui.zhong@rochester.edu
a b c
d e
Fig. 3.1 T2 shine-through in a 35-year-old female patient On DWI the lesion is hyperintense (arrow). (d) ADC map
with multiple sclerosis and weakness of the lower extrem- also shows hyperintensity in the lesion (1.2 × 10−3 mm2/s;
ities. (a) T2-weighted image shows several hyperintense arrow). (e) Exponential image eliminates the T2 effect
lesions, with the largest in the right frontal lobe (arrow). and shows the lesion to be hypointense (arrow). This con-
(b) On T1-weighted image the lesion was hypointense firms that the hyperintensity on DWI is due to a T2 shine-
(arrow) and did not enhance with contrast (not shown). (c) through effect
3 Pitfalls and Artifacts of Diffusion-Weighted Imaging 31
a b c
d e
Fig. 3.2 T2 shine-through in a 45-year-old female patient shown), (c) DWI shows hyperintensity (arrow). (d) ADC
with seizures caused by an anaplastic astrocytoma. (a) map also shows hyperintensity in the lesion (0.98–
T2-weighted image shows a hyperintense lesion in the left 1.35 × 10−3 mm2/s; arrow). (e) Exponential image elimi-
frontal lobe (arrow). (b) On T1-weighted image the lesion nates the T2 effect and shows the lesion to be hypointense
is hypointense with a peripheral hyperintense area (arrow). This confirms that the hyperintensity on the DWI
(arrow). The lesion did not enhance with contrast (not is due to a T2 shine-through effect