Insights into Imaging (2018) 9:535–547

https://doi.org/10.1007/s13244-018-0624-3

REVIEW

Clinical applications of diffusion weighted imaging in neuroradiology

Marta Drake-Pérez 1,2 & Jose Boto 1 & Aikaterini Fitsiori 1 & Karl Lovblad 1 & Maria Isabel Vargas 1

Received: 18 December 2017 / Revised: 12 March 2018 / Accepted: 20 March 2018 / Published online: 30 May 2018
# The Author(s) 2018

Abstract
Diffusion-weighted imaging (DWI) has revolutionised stroke imaging since its introduction in the mid-1980s, and it has also
become a pillar of current neuroimaging. Diffusion abnormalities represent alterations in the random movement of water
molecules in tissues, revealing their microarchitecture, and occur in many neurological conditions. DWI provides useful infor-
mation, increasing the sensitivity of MRI as a diagnostic tool, narrowing the differential diagnosis, providing prognostic infor-
mation, aiding in treatment planning and evaluating response to treatment. Recently, there have been several technical improve-
ments in DWI, leading to reduced acquisition time and artefacts and enabling the development of diffusion tensor imaging (DTI)
as a tool for assessing white matter. We aim to review the main clinical uses of DWI, focusing on the physiological mechanisms
that lead to diffusion abnormalities. Common pitfalls will also be addressed.
Teaching Points
• DWI includes EPI, TSE, RESOLVE or EPI combined with reduced volume excitation.
• DWI is the most sensitive sequence in stroke diagnosis and provides information about prognosis.
• DWI helps in the detection of intramural haematomas (arterial dissection).
• In diffusion imaging, ADC is inversely proportional to tumour cellularity.
• DWI and DTI derived parameters can be used as biomarkers in different pathologies.

Keywords MRI . DWI . Stroke . Infection . Inflammation

Introduction There have been several technical improvements in DWI,

leading to reduced acquisition time and artefacts and enabling
Diffusion-weighted imaging (DWI) has revolutionised stroke the development of diffusion tensor imaging (DTI) as a tool
imaging since its introduction in the mid-1980s, but it has also for assessing white matter. DTI allows tractography for surgi-
become a pillar of current neuroimaging. Diffusion abnormal- cal planning in brain tumours and other advanced techniques
ities represent alterations in the random movement of water such as intravoxel incoherent motion (IVIM) [1] and kurtosis,
molecules in tissues, revealing their microarchitecture, and which provide information about the complexity of tissues
occur in many neurological conditions. DWI provides useful and their different compartments (intravascular, intracellular
information, increasing the sensitivity of MRI as a diagnostic and extracellular).
tool, narrowing the differential diagnosis, providing prognos- We aim to review the main clinical uses of DWI, focusing
tic information, aiding the treatment planning and evaluating on the physiological mechanisms that lead to diffusion abnor-
response to treatment. malities. Common pitfalls will also be addressed.

* Maria Isabel Vargas DWI sequences used in neuroradiology

[email protected]

1
Division of Diagnostic and Interventional Neuroradiology of Geneva
Echo-planar imaging (EPI) is the method of choice for clinical
University Hospitals, DISIM and Faculty of Medicine of Geneva, diffusion-weighted imaging with MRI because of its low sen-
Rue Gabrielle-Perret-Gentil 4, 1211 Genève 14, Switzerland sitivity to the motion-induced phase errors that occur during
2
Department of Radiology, University Hospital Marqués de Valdecilla diffusion sensitisation of the MR signal. EPI is able to capture
– IDIVAL, Santander, Spain an entire image in a very short time. However, this method is
536 Insights Imaging (2018) 9:535–547

prone to artefacts because of susceptibility changes at tissue of the movement of water molecules in 3D and whether there
interfaces and has limited spatial resolution. Major technical is a dominant direction to diffusion restriction. This allows
advances in diffusion MRI have enabled superior data fidelity, tractography studies to be performed, which are multiplanar
reduced image acquisition time and an improved signal-to- reconstructions of white matter tracts according to the domi-
noise ratio (SNR) [2]. nant direction of water movement in each voxel.
Turbo-spin echo (TSE) is an adequate sequence to mini- DTI provides several quantitative parameters, most of
mise susceptibility artefacts and distortion, and it increases the which are useful in the assessment of white matter conditions.
signal-to-noise ratio. Unfortunately, it also increases acquisi- Fractional anisotropy (FA) reflects how dominant one partic-
tion times and is therefore more sensitive to motion through- ular water movement direction in a voxel is. It varies from 0 to
out the procedure. 1 and can be considered a biomarker of axonal integrity (usu-
Readout segmentation EPI techniques (e.g. RESOLVE) al- ally decreased in white matter pathologies). Mean diffusivity
low DWI and DTI to be simultaneously generated, thus reduc- (MD) represents a more exact value than ADC because it
ing the level of susceptibility-based distortion compared with considers the three main directions of water movement.
the single-shot EPI technique [3]. These also improve the over- Axial diffusivity (AD) quantifies water movement along the
all quality and spatial resolution of the images, especially when main longitudinal direction, evaluating axonal integrity, while
air-tissue interfaces are involved. However, they have longer radial diffusivity (RD) reflects myelin integrity (myelin injury
scanning times and are more sensitive to motion. Therefore, its leads to increased RD).
main clinical use is to study small regions adjacent to air-tissue As we will discuss, all these quantitative parameters im-
interfaces such as the inner ear or spinal cord [4]. prove the diagnostic possibilities of MR in the evaluation of
EPI combined with reduced volume excitation tumours, demyelinating and infectious diseases, and vascular
(ZOOMit; Siemens Healthcare, Erlangen, Germany) is a pathologies.
focused excitation method that utilises two orthogonal
pulses for slice selection in such a way that only the spins
belonging to the intersection of the slab give rise to signal. Cerebrovascular disease
This obviates the need for oversampling or distortion re-
duction [4] and allows zoomed fields of view. Advantages Stroke
of this method are improved image quality, increased im-
age sharpness and reduced distortion, with no additional Brain ischaemia leads to variations in the intra- and extracel-
acquisition time. A major disadvantage is the reduced field lular space due to a failure of the Na+/K+ ATPase and other
of view. As a result, this technique is also useful for the ionic pumps. Water movement inside the cell becomes more
evaluation of small structures, especially if they are close limited resulting in cytotoxic oedema. In addition, the intra-
to air-tissue interfaces (orbit or spinal cord). cellular environment becomes more viscous, further
Liney et al. compared EPI, RESOLVE and EPI combined restricting diffusion and decreasing the apparent diffusion co-
with reduced volume excitation, concluding that EPI provided efficient (ADC).
significantly higher ADC values than the other two sequences Subsequently, inflammatory mediation factors are released,
and that RESOLVE had the best repeatability ratio [5]. affecting the integrity of the blood-brain barrier and leading to
water extravasation and vasogenic oedema with expansion of
the interstitial space and an increase of the total water content
DWI, DTI and their derived parameters of the tissue. Subsequently, ADC values gradually increase to
their baseline. In chronic ischaemia gliosis eventually de-
Parameters derived from DWI and DTI studies allow a quan- velops, which does not result in restricted diffusion.
titative assessment, which in turn leads to improved accuracy DWI improves detection of ischaemia because it reveals al-
and reproducibility. However, it should be noted that some of terations in water diffusion before the net increase of water con-
these parameters are still subject to clinical validation [6]. tent in the subacute phase of the infarct. DWI hyperintensity [8,
ADC is the most widely used parameter derived from the 9] and changes in ADC values [10] can be found within minutes
conventional DWI sequence, representing the level of restric- from the onset of ischaemia (Fig. 1). EPI has high sensitivity and
tion to the motion of water molecules in the extracellular com- specificity in stroke diagnosis, but a negative result does not
partment. The term Bapparent^ relates to the fact that this exclude a vascular aetiology [11]. Stroke in the posterior circu-
motion is also influenced by other physiological processes, lation is more frequently DWI negative than in the anterior
such as heartbeat, breathing or CSF pulsation. The main ad- circulation (34.9% versus 15.3%, respectively; p = 0.0019) [12].
vantage of ADC relies on its widespread availability [7]. DWI is far more sensitive than conventional imaging in the
DTI is based on the application of diffusion gradients in at detection of early ischaemia. T2WI—including fluid-
least six different directions in space, enabling the evaluation attenuated inversion recovery (FLAIR) imaging—can only
Insights Imaging (2018) 9:535–547 537

Fig. 1 Acute stroke. a Increased

signal intensity on DWI in the
territory of the right middle
cerebral artery (MCA), which
corresponds to an area of low
ADC value in b (asterisks),
representing the extension of an
acute ischaemic stroke. c A clot as
the aetiology, presenting as a
linear hypointense abnormality in
the MCA on T2*WI (arrow)

detect ischaemia typically 1–4 h later, since signal abnormal- estimation of both the time of onset and the infarct-
ities on these sequences relate to the subsequent increase in salvageable brain proportion.
total water content in the infarcted area [13]. Despite some technical challenges (need for strong gradi-
In the setting of acute ischaemia, the mismatch between ents, the size of the spinal cord, flow artefacts), DWI is useful
DWI and FLAIR has been suggested to enable identification in assessing cord ischaemia. Although signal changes were
of patients with acute ischaemic stroke who are likely to be reported on both T2 and DWI in the majority of cases, DWI
within 4.5 h of symptom onset with high specificity and high showed a clear benefit over conventional MRI in some pa-
positive predictive value [14], which is useful for identifying tients, in whom T2 failed to identify an abnormality [19, 20].
candidates for fibrinolysis. According to Thurnher et al., the earliest changes on DWI in
ADC alone can be considered a biomarker of brain ischae- spinal cord ischaemia can be seen 3 h after symptom onset
mia as it allows differentiation between the two main causes of [21] An imaging protocol with DWI of the spinal cord with
increased DWI signal in acute stroke [15]: Btrue^ diffusion maximum b values between 500 and 700 s/mm2 and a slice
restriction and T2 signal elongation. ADC values also yield thickness of 3 mm is recommended [22] (Fig. 2).
timing information [16]: they start to decrease from the 1st
hour, reaching the minimal value around the 24th hour. Key points Ischaemia leads to DWI changes within minutes of
Subsequently, ADC values increase after the 3rd day, reaching symptom onset
higher values than the surrounding parenchyma after the 10th Pearls DWI helps in:
Diagnosis (most sensitive sequence)
day. When combined with T2WI signal changes, they allow
Prognosis and treatment planning (DWI aids in Bdating^
radiologists to estimate the approximate time of stroke onset. the stroke and in achieving better patient selection)
It should however be noted that haemorrhagic transformation Pitfalls False negatives for posterior circulation strokes possible
may be present, which will affect the signal of the lesion and
modify the expected sequence of events.
The area showing diffusion restriction is usually taken as
the Binfarct core^ (final infarct size), but some diffusion ab- Transient ischaemic attack (TIA)
normalities may be reversible, mainly if there is early (< 4.5 h)
spontaneous or therapeutic recanalisation of the occluded ves- The new definition of TIA is based on a biological concept
sels [17]. However, the size of the area of diffusion restriction (tissue injury). ATIA is characterised by a transient episode of
(and the ADC value) correlates with prognosis, and volume neurological impairment caused by focal ischaemia without
measurements and scores (ASPECTS, DRAGON) can be use- acute infarction [23].
ful tools in predicting the outcome after intravenous throm- Previously, TIA was defined based on arbitrary timing (du-
bolysis [18]. ration ≤ 24 h). However, permanent tissue damage (infarction)
Thrombolytic or endovascular therapies are the treatment is possible even when focal transient neurological symptoms
options for acute stroke. Selection of patients eligible for these last < 1 h [24]. Additionally, approximately 50% of patients
therapies is critical to achieve good outcomes and avoid fatal with classically defined TIA syndromes have corresponding
complications. Both time from stroke symptom onset to treat- ischaemic lesions on brain MRI diffusion- or perfusion-
ment and the individual therapeutic time window secondary to weighted imaging. Therefore, whether a symptomatic ischae-
variations in collateral circulation determine the long-term mic episode will result in ischaemic infarction cannot be based
functional outcome after stroke. Therefore, imaging—includ- solely on time duration.
ing DWI—has great prognostic value and is crucial to treat- The current definition takes imaging into account as DWI
ment decision making because it allows early diagnosis and an must be normal for the diagnosis of TIA to be upheld. If MRI
538 Insights Imaging (2018) 9:535–547

Fig. 2 Spinal cord ischaemia.

High T2 signal abnormality in the
conus medullaris with diffusion
restriction on DWI (arrows) com-
patible with an ischaemic lesion

shows areas of restricted diffusion in a patient whose symp- useful in the diagnosis of acute dissection, when the intramural
toms have resolved and lasted less than 24 h, the term cerebral haematoma can barely be detected on fat-saturated T1-weight-
infarction with transient symptoms is preferred [25]. Once a ed images because of obscuration of its isointense signal by the
morphological change has been demonstrated on imaging, the surrounding soft tissues with similar signal intensity.
risk of full-blown stroke is higher, and treatment should be DWI using EPI usually shows susceptibility artefacts at the
more aggressive. air-tissue or bone-soft tissue interfaces. Artefacts present as a
fuzzy linear hyperintensity or diffuse hypointensity, which are
usually easy to differentiate from the focal round or linear
Key points New definition that takes imaging into account (DWI -) well-demarcated hyperintensity typical of intramural
Pearls If DWI +, the term cerebral infarction with transient haematoma [28]. Correlation with T1- or T2-weighted images
symptoms is preferred (more aggressive treatment)
is also helpful. The use of an SE sequence, reducing the echo
Pitfalls The duration of ischaemic symptoms does not predict
whether a symptomatic ischaemic event will result in
time and increasing the acquisition matrix, may be helpful to
ischaemic infarction minimise these susceptibility artefacts.

Key points DWI helps in the detection of the intramural haematoma

Arterial dissection
Pearls Bright spot on the DWI sequence early on (when it might
be isointense or harder to see on other sequences)
The Bcrescent sign^, representing an intramural haematoma, Pitfalls Susceptibility artefacts at skull base may interfere with the
can be an imaging criterion for diagnosing arterial dissection. correct interpretation of this sign on DWI
It consists of an eccentric rim of hyperintensity surrounding
the hypointense arterial lumen on MRI. It has traditionally
been described on T1-weighted fat-saturation MRI sequences
[26], but may be seen on other sequences such as diffusion- Venous infarction
weighted imaging (Fig. 3).
The signal intensity of intramural haematoma in arterial Ischaemia in the setting of venous thrombosis has a different
dissection is influenced by the paramagnetic effects of blood physiology from arterial infarction [29]. Initially, there is in-
breakdown products [27]. The haematoma is isointense on creased venous pressure and vasogenic oedema (reversible)
T1WI and shows T2 prolongation a few hours after bleeding that may lead to a Na+/K+ ATPase pump failure, cytotoxic
has occurred in the hyperacute phase (oxyhaemoglobin phase). oedema and restricted diffusion. Therefore, diffusion imaging
The T2 signal becomes hypointense in the acute stage helps to define the stage of the disease because it allows dif-
(deoxyhaemoglobin phase). T1 and T2 signals become hyper- ferentiation between irreversible venous infarction (DWI hy-
intense in the subacute phase, between 7 days and 2 months perintense with restricted diffusion) and reversible venous oe-
(methaemoglobin phase). In cerebral venous thrombosis, dema (DWI isointense with increased diffusion). In contradis-
hyperintensity on DWI has been suggested to represent re- tinction to arterial stroke, DWI-positive venous stroke may be
stricted proton movement within the clot, but it is also ob- completely reversible.
served within the dissected vertebral artery wall in the hyper- In venous ischaemia, DWI may be able to show the actual
acute and early subacute phases [28]. DWI can therefore be thrombus as increased signal intensity [30] because of the
Insights Imaging (2018) 9:535–547 539

Key points Different physiology from arterial ischaemia

Pearls Depending on stage, the thrombus may be hyperintense
on DWI
Pitfalls Very variable imaging features, depending on the time
after thrombosis

Tumours

Regarding extraaxial tumours, DWI allows distinction be-

tween epidermoid and arachnoid cysts. Epidermoid cysts
show diffusion restriction, as opposed to arachnoid cysts,
which follow CSF signal intensity on all sequences. Other
MR sequences can also aid in the diagnosis, since epidermoid
cysts are bright on FLAIR, but not arachnoid cysts [34]. In
meningiomas it has been shown that ADC values correlate
with tumour cellularity [35] but this has not been found to
provide any additional value in differentiating histopatholog-
ical subtypes of meningiomas [36].
Concerning glial tumours, quantitative assessment with
DWI has mainly targeted an estimation of cellularity based
on its inverse relation with water diffusivity in the extracellular
compartment. Hence, ADC can be considered as a tumour
Fig. 3 Arterial dissection. DWI (a), ADC (b), non-contrast FAT SAT T1
(c) and coronal MIP reconstruction of the basilar artery (d) demonstrate a biomarker in gliomas: the higher the tumour grade, the lower
dissection with mural haematoma represented by the Bcrescent sign^ the mean tumour ADC values [37]. However, the range of
(arrows), which shows restricted diffusion and hyperintensity on T1WI ADC values within a given glioma varies markedly [38] and
there is an overlap between ADCs of grade II astrocytomas and
glioblastomas, probably due to the inherent tissue heterogene-
restricted movement of water molecules within the venous ity associated with gliomas across different grades, limiting the
clot (Fig. 4). However, sinus thrombosis may have different use of DWI. Despite this, in the context of low-grade tumour
appearances depending on the stage of thrombus formation, follow-up and where appropriate, a decrease in the ADC value
which has been demonstrated by ex vivo MRI measurements could be considered an early sign of tumour progression.
[31]. Another use of DWI on imaging of CNS tumours is to aid
In previous studies, the approximate frequency of the in- in the distinction between glioblastoma and primary CNS
creased DWI signal within the venous thrombus has been lymphoma, two entities with markedly different treatment
reported as 41% [32], thus not truly providing significant strategies and not always easy to differentiate with conven-
added value since all cases showed T1WI or FLAIR signal tional imaging in patients presenting with an enhancing brain
changes at the site of occlusion. mass. Due to its higher cellularity compared with GBM, pri-
Taking into account that venous infarctions have a higher mary CNS lymphoma has been shown to exhibit lower ADC
tendency to bleed, and haemorrhage can also restrict diffusion, values [39]. There are some published works aiming to find an
imaging features in venous infarction may be considerably ADC threshold for differentiating lymphoma from other tu-
variable, depending on the time after onset and the sequence mours [40] (Fig. 5). Guzman et al. reported higher ADC
of events. values in the more aggressive tumours (comparing glioblasto-
Other vascular pathologies in which diffusion may be help- ma and metastasis vs. low-grade glioma) [41].
ful are transient global amnesia, with a typical transient focal Attempts to differentiate glioblastoma from metastasis
dot-like area of diffusion restriction in the hippocampus and/ based on the FA values of the surrounding oedema have been
or parahippocampal gyrus [33], or posterior reversible en- made [42] based on the assumption that glioblastoma has a
cephalopathy syndrome (PRES), where the majority of cases tendency to grow in an infiltrative manner typically invading
do not show diffusion restriction because oedema is typically surrounding tissues. Conversely, metastatic tumours tend to
vasogenic. In the cases where DWI is positive, this may indi- grow in an expansive manner and typically displace the sur-
cate the earliest sign of non-reversibility due to progression to rounding brain tissues rather than invading them [43]. Another
cytotoxic oedema and therefore the possibility of permanent entity associated with brain tumours is intracranial hyperten-
infarction or haemorrhagic transformation. sion, which can be seen as papilloedema, bright on DWI [44].
540 Insights Imaging (2018) 9:535–547

Fig. 4 Venous thrombosis. DWI

(a), ADC map (b) and sagittal
reconstruction of T1 3D with
gadolinium (c) show thrombosis
of the left jugular vein (arrows),
presenting with restricted
diffusion and a filling defect in the
vessel after contrast
administration

DTI and tractography reconstruction can be helpful for potentially interpreted as residual tumour or tumour progres-
surgical planning in patients with brain or spinal cord masses. sion (Fig. 7).
It shows white matter tracts as well as displacement or inter- Regarding paediatric tumours, DWI can non-invasively
ruption of these tracts secondary to different pathologies. add valuable information that can be used to narrow the dif-
Slow-growing tumours displace the fibres surrounding the ferential diagnosis of the three most common tumours of the
lesion, whereas in ischaemia or trauma, representing more posterior fossa in children (medulloblastoma, ependymoma,
acute insults, the fibres are usually interrupted [45] (Fig. 6). pilocytic astrocytoma). Several studies have demonstrated that
An additional important use of DWI after surgical resection ADC values in the enhancing, non-necrotic, non-oedematous,
of a tumour is to detect areas of diffusion restriction in the solid parts of cerebellar tumours are negatively correlated with
cavity margins on MRI performed in the first 48 h, corre- tumour grade [46] (high-grade tumours such as medulloblas-
sponding to small postsurgical infarctions. The importance toma are characterised by high cellularity, low extracellular
of this imaging time window lies in the fact that the majority space, and cells with large nuclei and high nuclear-to-
of these infarcts will enhance after 1 or 2 weeks and could be cytoplasmatic ratios, causing decreased diffusion). Cutoff
values of > 1.4 × 103 mm2/s for juvenile pilocytic astrocytoma
and < 0.9 × 103 mm2/s for medulloblastoma have been sug-
gested with a high specificity [47]. An even greater accuracy
can be achieved by combining DWI and MRS to obtain more
information about the tumour proliferative potential [48].
DWI has also been reported to be useful in detecting relapse
of embryonal tumours such as medulloblastoma, being more
sensitive than contrast-enhanced MRI in subjects with the
classic variant [49].

Key points ADC is inversely proportional to tumour cellularity

Pearls DWI can help differentiate GBM from lymphoma,
epidermoid from arachnoid cysts, tumour necrosis
from tumour progression and in narrowing the
differential diagnosis of posterior fossa tumours
in children, to name a few
Pitfalls Overlaps and exceptions. Always interpret in the light
of the clinical context and other radiological
information

Infection and inflammatory conditions

Infection
Fig. 5 Primary CNS lymphoma. Left frontal periventricular lesion
showing prominent diffusion restriction, presenting with hyperintensity
on DWI (a), low ADC value (b) and mild hyperintensity on T2WI (c), all
Abscesses of the brain can be bacterial, parasitic or fungal in
typical features of this type of hypercellular tumour. T1WI post origin. The appearance of bacterial abscesses on conventional
gadolinium (d) shows homogeneous and intense contrast enhancement imaging varies with the stage of abscess formation, but the
Insights Imaging (2018) 9:535–547 541

Fig. 6 DTI for surgical planning.

Patient with an ependymoma in
the lower cervical spinal cord
with areas of high T2 signal (a)
and enhancement after
gadolinium administration (b).
DTI shows displacement of the
fibres (c), which is usually
consistent with a slow-growing
lesion

typical features of the early capsule stage (ring-enhancing Aspergillus CNS infections tend to be haemorrhagic,
lesion with a T2 hypointense rim and variable degree of which is an additional cause of restricted diffusion. In general,
associated vasogenic oedema) can be similar to primary cystic it is important to include fungal abscesses in the differential
or necrotic tumours (Fig. 8) and subacute haematomas
(Fig. 9). However, pyogenic abscesses show restricted diffu-
sion with markedly decreased signal on the ADC map and
elevated fractional anisotropy (FA) within the abscess cavity
[50] because of the presence of intact inflammatory cells and
bacteria. These impede the microscopic motion of water mol-
ecules, which helps to narrow the differential diagnosis of
ring-enhancing lesions (exceptions exist: metastases with high
cellularity and haemorrhagic metastasis).
Reddi et al. showed that diffusion-weighted imaging had a
sensitivity and specificity of 96% for the differentiation of brain
abscesses from primary or metastatic cancers (positive predic-
tive value, 98%; negative predictive value, 92%) [51] (Fig. 10).
A study [52] has shown that the abscess cavity had lower ADC
and higher FA compared with the cystic cavity of glioblastomas
and metastases, possibly representing high viscosity and
organised viable inflammatory cells. In addition, FA values were
higher in the enhancing rim of the abscess than in glioblastoma
and metastasis, maybe because of the presence of concentric
layers of collagen fibres. Regarding the peritumoral zone of
oedema, this study showed that the presence of a hyperintense
FA rim (with lower FA values) was much more frequent in
glioblastomas and metastases than in abscesses. This suggests
that microstructural changes in the oedematous white matter
immediately surrounding the abscess may be different from
those surrounding glioblastoma and metastasis, possibly be-
cause of acute versus longer time to formation, respectively.
Fig. 7 Post-surgical ischaemia. Immediate follow-up MRI in a patient
Fungal abscesses can have either restricted or elevated dif- who underwent surgery for resection of a suspicious enhancing mass. In
fusion in the centre (depending on the content), in opposition the medial aspect of the resection cavity (asterisk) there is an enhancing
to pyogenic and tuberculous abscesses that typically show area on the T1 post-contrast sequence (c, arrow). This finding alone could
restricted diffusion in the core of the cavity [53]. Restricted represent residual tumour, but the presence of restricted diffusion with
high signal on DWI (a) and a low ADC value (b) meant that a small area
diffusion is usually demonstrated in the projections and wall of peri-surgical ischaemia was more likely. Three-month follow-up T1
of the fungal abscess, referred to as Bintracavitary projec- post-gadolinium MRI (d) shows absence of enhancement in the same
tions^, and represent the fungal organism itself. region (arrowhead), confirming this diagnosis
542 Insights Imaging (2018) 9:535–547

Fig. 8 GBM. Left parieto-

occipital lesion with peripheral
vasogenic oedema. DWI and
ADC (a and b, respectively) show
a clear area of increased diffusion
within the core, corresponding to
a necrotic centre. On T1 3D post
gadolinium the mass shows ring
enhancement

diagnosis in immunocompromised patients because of differ- and allows distinction of two parts in the lesions: a central core
ent targeted antifungal treatment. (whose size can correlate to the clinical status and disease
Parasitic abscesses may also show both restricted and ele- duration) and a peripheral rim (which includes a heteroge-
vated diffusion [54]. For example, cysticercosis cysts have a neous component and areas of surrounding cytotoxic oedema
similar or slightly increased DWI signal to CSF, and the sco- with low ADC and areas of vasogenic oedema and glial repair
lex may be seen as a hyperintense focus inside [55]. In toxo- with intermediate ADC) [64].
plasmosis, diffusion restriction is highly variable (possibly
overlapping with lymphoma characteristics, its main differen- Key points Inflammatory cells and bacteria within an abscess core
tial diagnosis) [56, 57]. In cerebral malaria, DWI is useful in restrict diffusion
detecting areas of brain infarction [58]. Pearls ADC values and morphology within the area showing
restricted diffusion may help in differentiating abscess
In the postsurgical setting, the accuracy of DWI in the diag-
from tumour and pyogenic from fungal abscess
nosis of CNS infectious complications is low, showing a high The location of diffusion restriction may orientate in the
false-negative rate. There are often postoperative changes with aetiology of viral encephalitis: limbic system for herpes
diffusion restriction due to haemorrhage/ischaemia, which can simplex; splenium of the corpus callosum for influenza
and Epstein-Barr viruses; thalami for West Nile virus
hinder the distinction between purulent content and
DWI most sensitive technique in the early stages of CJD
haemorrhagic/ischaemic components. The absence of restricted
Pitfalls Subacute haematoma can look like an abscess
diffusion is not sufficient to exclude the presence of pyogenic
postcraniotomy infection and should not be used as the main
determinant in patient management in this clinical setting [59].
Diffusion restriction is commonly seen in viral encephali-
tis, and the location of the affected area may guide the radiol- Creutzfeldt-Jakob disease
ogist in the identification of the infectious agent. For example,
the limbic system (medial temporal and inferior frontal cortex) Creutzfeldt-Jakob disease (CJD) is the combination of pro-
is the typical involved area in herpes simplex encephalitis. gressive dementia and pyramidal, extrapyramidal, and/or cer-
Viruses more frequently causing potentially reversible lesions ebellar signs caused by a prion (proteinaceous particle without
in the splenium of the corpus callosum are the influenza virus DNA or RNA but capable of causing infection). DWI MR has
[60], Epstein-Barr virus, HHV-6 and papovirus JC [61]. high diagnostic accuracy and is the most sensitive neuroimag-
Thalamic involvement is typically seen with the West Nile ing technique for striatal and cortical lesions from an early
virus [62], Japanese encephalitis or Eastern equine encephali- stage [65]. Typically, diffusion-weighted hyperintensity is
tis. DWI may help detect lesions earlier than conventional MR progressive and persistent over many weeks, affecting the
imaging. In patients diagnosed with West Nile virus encepha- striatum and cortex. The gyriform DWI-hyperintense areas
litis, diffusion restriction with no FLAIR or T2 signal abnor- in the cerebral cortex (Bcortical ribbon^ sign) correlate to the
malities has been reported to be a sign of good prognosis [63]. location of periodic sharp-wave complexes on EEG [66].
Progressive multifocal leukoencephalopathy (PML) is sec- DWI hyperintensity may resolve late in the disease.
ondary to the infection of oligodendrocytes by the JC polyoma The pulvinar sign (symmetrical hyperintensity in the pos-
virus, which typically remains latent until reactivation in the terior thalamic nuclei on DWI or T2WI), Bhockey stick^ sign
context of an immunocompromised state (HIV, natalizumab (signal changes affecting the dorsomedial thalamic nuclei) or
or other immunosuppressive treatments). Due to its ability to similar changes in the periaqueductal grey matter were report-
investigate white matter architecture and diseases, DWI shows ed to be specific for variant (vCJD) [67], but can also occur,
regions of active infection and cell swelling with high signal although less frequently, in sporadic (sCJD).
Insights Imaging (2018) 9:535–547 543

Fig. 9 Subacute haematoma. Right parietal mass (arrowhead) showing differentiate haemorrhage from abscess. There is also a subacute ischae-
diffusion restriction within the core on DWI and ADC maps (a and b, mic lesion in the inferior right frontal lobe (arrows) that shows early
respectively) and a ring-enhancing pattern on T1 post gadolinium (c). pseudonormalisation of the ADC and gyriform enhancement post
This was a subacute haematoma. Clinical context is important to gadolinium

Multiple sclerosis limbic encephalitis, paraneoplastic cerebellar degeneration and

brainstem encephalitis, and the encephalitis syndromes associ-
In multiple sclerosis, the majority of active plaques show nor- ated with antibodies against neuronal cell surface/synaptic pro-
mal or increased diffusivity, but some may show restricted teins, which are commonly referred to as autoimmune
diffusion, often at the margins [68]. It is important to bear this encephalitis.
in mind since, in the appropriate context, these plaques can Diagnosis is based on electroencephalography (EEG), lum-
mimic a lacunar infarct. Possible explanations for the de- bar puncture and serological testing for appropriate biomarkers
creased ADC in active plaques are decreased extracellular (presence of autoantibodies, lymphocytic pleocytosis or
space due to myelin oedema or cytotoxic oedema and de- oligoclonal bands in CSF). Neuroimaging should be performed
creased water movement in the extracellular space secondary despite yielding a high proportion of false negatives (in auto-
to inflammation [15]. DTI demonstrates different degrees of immune encephalitis more than in paraneoplastic encephalitis
mean diffusivity increase and FA decrease in T2-hyperintense syndromes) [71]. When positive, the studies are usually non-
established lesions or even before their formation. These specific (high FLAIR signal in cortical or subcortical regions—
changes are also present in normal-appearing white or grey hippocampus, basal ganglia, white matter). Diffusion restriction
matter of patients with MS [69] (as opposed to ADEM, which may be present in the areas showing abnormal signal.
shows normal diffusivity within normal-appearing white mat- Although limbic encephalitis has been known as a
ter). Diffusivity changes in MS correlate with areas of demy- paraneoplastic syndrome, there have been recognised forms
elination and axonal loss in postmortem studies [70]. characterised by serum autoantibodies, directed against
glutamic acid decarboxylase (anti-GAD) and voltage-gated
Immune-mediated encephalitis potassium channels (VGKC), not related to neoplastic
conditions.
Immune-mediated encephalitis includes two different types of On DTI studies, a reduction of fractional anisotropy and
conditions: paraneoplastic encephalitis syndromes, comprising increased mean diffusivity have been shown in the white matter

Fig. 10 Abscess. Right occipital

mass showing marked diffusion
restriction within the core on DWI
and ADC maps (a and b,
respectively) and a peripheral
enhancing pattern on T1 post
gadolinium (c). DWI helps to
differentiate ring-enhancing le-
sions because restricted diffusion
in the centre of the mass is char-
acteristic of pyogenic abscesses.
In this case, diffusion-based se-
quences also helped to identify
ventriculitis (arrows)
544 Insights Imaging (2018) 9:535–547

of patients suffering from NMDAR encephalitis (the most com-

mon type of autoimmune encephalitis), suggesting that white
matter damage is a contributor to the pathophysiology and clin-
ical phenotype, even correlating with disease severity [72].
Another study aiming to assess white matter integrity in the
two forms of non-paraneoplastic limbic encephalitis
characterised by serum autoantibodies (GAD and VGKC)
concluded that GAD-limbic encephalitis presented with FA
reduction, probably reflecting demyelination, whereas
VGKC-limbic encephalitis caused no changes in the diffusion
properties of white matter, probably because this condition is
more limited to grey matter [73].

Neuritis

Abnormal signal intensity of the optic nerve due to diffusion

restriction may be seen in ischaemic or traumatic optic neu-
ropathy [74, 75]. DWI has been suggested as a useful tool in
differentiating these aetiologies from acute optic neuritis (pre-
sumed autoimmune aetiology) [76, 77], where there is in-
creased diffusivity in demyelinating plaques, related to disrup-
tion of myelinated axons, and decreased fractional anisotropy. Fig. 11 Optic neuritis. High DWI signal (a) and low ADC value (b)
representing restricted diffusion in the right optic nerve in a patient with
However, this should be interpreted in the clinical context, non-specific optic neuritis
since optic neuritis may also show transient decreased diffu-
sion in the acute phase (Fig. 11).
there is not sufficient evidence for nigral DTI parameters to
With the exception of the orbital apex region, DTI of the
serve as biomarkers for Parkinson’s disease [82].
optic nerve is facilitated by its anatomy and antero-posterior
ADC values of the putamen have been used to discriminate
orientation. This can be useful in assessing tumours such as
between Parkinson’s disease and atypical parkinsonism, in-
neurofibromas, schwannomas or meningiomas [78] as well as
cluding the classical phenotype of progressive supranuclear
in brain malformations such as septo-optic dysplasia or optic
palsy and corticobasal syndrome. Putaminal ADC values were
nerve hypoplasia [79].
significantly higher in both atypical parkinsonism syndromes
Other cranial nerves can also be studied with high-
compared with Parkinson’s disease [83].
resolution DTI in some clinical situations. Demonstration of
In other neurodegenerative diseases, such as Friedreich’s
atrophy or loss of anisotropy in cases of trigeminal neuralgia
ataxia, DWI may be an appropriate non-invasive tool to sys-
and fibre displacement in vestibulocochlear schwannomas are
tematically quantify the extent and distribution of brain chang-
examples of this [80].
es. DWI has also highlighted the involvement of new areas in
the pathophysiology of this condition, such as the optic radi-
ations and middle cerebellar peduncles [84].
Neurodegenerative disorders

The DWI and DTI techniques allow quantitative analysis of

microstructural changes in neurodegenerative diseases, even Conclusion
when no abnormalities are seen on conventional MRI se-
quences, because they are able to identify changes in brain Diffusion-weighted sequences currently play a central role in
tissue integrity. These changes are characterised by a decrease neuroimaging. The already widespread qualitative assessment
in the number of barriers that restrict the movement of water, allowed by these sequences improves sensitivity in the depic-
thus causing ADC to increase in brain areas where neurode- tion of several central nervous system conditions, whereas
generation occurs. This has led to an increase in the use of new models of diffusion sequences provide quantitative pa-
DWI in the diagnostic investigation of neurodegenerative par- rameters, allowing potential biomarkers for diagnosis, prog-
kinsonian syndromes. Attempts have been made to use DTI of nosis and follow-up. The future of DWI will undoubtedly
the substantia nigra as a diagnostic marker for Parkinson dis- include technical improvements to enhance data fidelity, to
ease, but while some studies found highly significant FA re- achieve high isotropic resolution (e.g., submillimetre) for 3D
duction in the substantia nigra [81], others concluded that acquisitions and to reduce the image acquisition time.
Insights Imaging (2018) 9:535–547 545

Furthermore, advances in ultrahigh field technology are al- 14. Thomalla G, Cheng B, Ebinger M et al (2011) DWI-FLAIR mis-
match for the identification of patients with acute ischaemic stroke
ready being applied to DWI.
within 4·5 h of symptom onset (PRE-FLAIR): a multicentre obser-
vational study. Lancet Neurol 10(11):978–986
Acknowledgements MDP was supported by a research contract 15. Dietemann J-L (2012) Neuro-imagerie diagnostique. Elsevier
(Contrato post-MIR López Albo 2016) from the Instituto de Masson
Investigación Sanitaria IDIVAL. 16. Yang Q, Tress BM, Barber PA et al (1999) Serial study of apparent
diffusion coefficient and anisotropy in patients with acute stroke.
Stroke 30(11):2382–2390
Open Access This article is distributed under the terms of the Creative 17. Labeyrie MA, Turc G, Hess A et al (2012) Diffusion lesion reversal
Commons Attribution 4.0 International License (http:// after thrombolysis: a MR correlate of early neurological improve-
creativecommons.org/licenses/by/4.0/), which permits unrestricted use, ment. Stroke 43(11):2986–2991
distribution, and reproduction in any medium, provided you give appro- 18. Turc G, Apoil M, Naggara O et al (2013) Magnetic resonance
priate credit to the original author(s) and the source, provide a link to the imaging-DRAGON score: 3-month outcome prediction after intra-
Creative Commons license, and indicate if changes were made. venous thrombolysis for anterior circulation stroke. Stroke 44(5):
1323–1328
19. Fujikawa A, Tsuchiya K, Takeuchi S, Hachiya J (2004) Diffusion-
References weighted MR imaging in acute spinal cord ischemia. Eur Radiol
14(11):2076–2078
20. Loher TJ, Bassetti CL, Lövblad KO et al (2003) Diffusion-
1. Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E, Laval- weighted MRI in acute spinal cord ischaemia. Neuroradiology
Jeantet M (1986) MR imaging of intravoxel incoherent motions: 45(8):557–561
application to diffusion and perfusion in neurologic disorders. 21. Thurnher MM, Bammer R (2006) Diffusion-weighted MR imaging
Radiology 161(2):401–407 (DWI) in spinal cord ischemia. Neuroradiology 48(11):795–801
2. Wu W, Miller KL (2017) Image formation in diffusion MRI: a 22. Vargas MI, Gariani J, Sztajzel R et al (2015) Spinal cord ischemia:
review of recent technical developments. J Magn Reson Imaging practical imaging tips, pearls, and pitfalls. AJNR Am J Neuroradiol
46(3):646–662 36(5):825–830
3. Thian YL, Xie W, Porter DA, Weileng Ang B (2014) Readout- 23. Easton JD, Saver JL, Albers GW et al (2009) Definition and eval-
segmented echo-planar imaging for diffusion-weighted imag- uation of transient ischemic attack: a scientific statement for
ing in the pelvis at 3T-A feasibility study. Acad Radiol 21(4): healthcare professionals from the American Heart Association/
531–537 American Stroke Association Stroke Council; Council on
4. Viallon M, Cuvinciuc V, Delattre B et al (2015) Erratum to: state-of- Cardiovascular Surgery and An esthesia; Co uncil on
the-art MRI techniques in neuroradiology: principles, pitfalls, and Cardiovascular Radiology and Intervention; Council on
clinical applications. Neuroradiology 57(10):1075 Cardiovascular Nursing; and the Interdisciplinary Council on
5. Liney GP, Holloway L, Al Harthi TM et al (2015) Quantitative Peripheral Vascular Disease. The American academy of neurology
evaluation of diffusion-weighted imaging techniques for the pur- affirms the value of this statement as an educational tool for neu-
poses of radiotherapy planning in the prostate. Br J Radiol rologists. Stroke 40(6):2276–2293
88(1049):20150034 24. Sacco RL, Kasner SE, Broderick JP et al (2013) An updated defi-
6. Martín Noguerol T, Martínez Barbero JP (2017) Advanced diffu- nition of stroke for the 21st century: a statement for healthcare
sion MRI and biomarkers in the central nervous system: a new professionals from the American Heart Association/American
approach. Radiologia 59(4):273–285 Stroke Association. Stroke 44(7):2064–2089
7. Urger E, Debellis MD, Hooper SR, Woolley DP, Chen S, 25. Ay H, Koroshetz WJ, Benner T et al (2005) Transient ischemic
Provenzale JM (2013) Influence of analysis technique on measure- attack with infarction: a unique syndrome? Ann Neurol 57(5):
ment of diffusion tensor imaging parameters. AJR Am J 679–686
Roentgenol 200(5):W510–W517 26. Cuvinciuc V, Viallon M, Momjian-Mayor I et al (2013) 3D fat-
8. Moseley ME, Cohen Y, Mintorovitch J et al (1990) Early detection of saturated T1 SPACE sequence for the diagnosis of cervical artery
regional cerebral ischemia in cats: comparison of diffusion- and T2- dissection. Neuroradiology 55(5):595–602
weighted MRI and spectroscopy. Magn Reson Med 14(2):330–346 27. Rodallec MH, Marteau V, Gerber S, Desmottes L, Zins M (2008)
9. Mintorovitch J, Moseley ME, Chileuitt L, Shimizu H, Cohen Y, Craniocervical arterial dissection: spectrum of imaging findings and
Weinstein PR (1991) Comparison of diffusion- and T2-weighted differential diagnosis. Radiographics 28(6):1711–1728
MRI for the early detection of cerebral ischemia and reperfusion 28. Choi KD, Jo JW, Park KP et al (2007) Diffusion-weighted imaging
in rats. Magn Reson Med 18(1):39–50 of intramural hematoma in vertebral artery dissection. J Neurol Sci
10. Reith W, Hasegawa Y, Latour LL, Dardzinski BJ, Sotak CH, Fisher 253(1–2):81–84
M (1995) Multislice diffusion mapping for 3-D evolution of cere- 29. van den Bergh WM, van der Schaaf I, van Gijn J (2005) The
bral ischemia in a rat stroke model. Neurology 45(1):172–177 spectrum of presentations of venous infarction caused by deep ce-
11. Lövblad KO, Laubach HJ, Baird AE et al (1998) Clinical experi- rebral vein thrombosis. Neurology 65(2):192–196
ence with diffusion-weighted MR in patients with acute stroke. 30. Lövblad KO, Bassetti C, Schneider J et al (2001) Diffusion-
AJNR Am J Neuroradiol 19(6):1061–1066 weighted MR in cerebral venous thrombosis. Cerebrovasc Dis
12. Simonsen CZ, Madsen MH, Schmitz ML, Mikkelsen IK, Fisher M, 11(3):169–176
Andersen G (2015) Sensitivity of diffusion- and perfusion- 31. Toussaint JF, Southern JF, Fuster V, Kantor HL (1997) Water dif-
weighted imaging for diagnosing acute ischemic stroke is 97.5%. fusion properties of human atherosclerosis and thrombosis mea-
Stroke 46(1):98–101 sured by pulse field gradient nuclear magnetic resonance.
13. Venkatesan R, Lin W, Gurleyik K et al (2000) Absolute measure- Arterioscler Thromb Vasc Biol 17(3):542–546
ments of water content using magnetic resonance imaging: prelim- 32. Favrole P, Guichard JP, Crassard I, Bousser MG, Chabriat H (2004)
inary findings in an in vivo focal ischemic rat model. Magn Reson Diffusion-weighted imaging of intravascular clots in cerebral ve-
Med 43(1):146–150 nous thrombosis. Stroke 35(1):99–103
546 Insights Imaging (2018) 9:535–547

33. Weon YC, Kim JH, Lee JS, Kim SY (2008) Optimal diffusion- 52. Toh CH, Wei KC, Ng SH, Wan YL, Lin CP, Castillo M (2011)
weighted imaging protocol for lesion detection in transient global Differentiation of brain abscesses from necrotic glioblastomas and
amnesia. AJNR Am J Neuroradiol 29(7):1324–1328 cystic metastatic brain tumors with diffusion tensor imaging. AJNR
34. Osborn AG, Preece MT (2006) Intracranial cysts: radiologic- Am J Neuroradiol 32(9):1646–1651
pathologic correlation and imaging approach. Radiology 239(3): 53. Luthra G, Parihar A, Nath K et al (2007) Comparative evaluation of
650–664 fungal, tubercular, and pyogenic brain abscesses with conventional
35. Cabada T, Caballero MC, Insausti I et al (2009) The role of and diffusion MR imaging and proton MR spectroscopy. AJNR Am
diffusion-weighted imaging in the evaluation of meningiomas: J Neuroradiol 28(7):1332–1338
radio-pathologic correlation. Australas Radiol 51(4):411–419 54. Abdel Razek AA, Watcharakorn A, Castillo M (2011) Parasitic
36. Sanverdi SE, Ozgen B, Oguz KK et al (2012) Is diffusion-weighted diseases of the central nervous system. Neuroimaging Clin N Am
imaging useful in grading and differentiating histopathological sub- 21(4):815–841 viii
types of meningiomas? Eur J Radiol 81(9):2389–2395 55. Gupta RK, Prakash M, Mishra AM, Husain M, Prasad KN, Husain
37. Surov A, Meyer HJ, Wienke A (2017) Correlation between appar- N (2005) Role of diffusion weighted imaging in differentiation of
ent diffusion coefficient (ADC) and cellularity is different in several intracranial tuberculoma and tuberculous abscess from cysticercus
tumors: a meta-analysis. Oncotarget. https://doi.org/10.18632/ granulomas—a report of more than 100 lesions. Eur J Radiol 55(3):
oncotarget.17752 384–392
38. Cha S (2006) Update on brain tumor imaging: from anatomy to 56. Camacho DL, Smith JK, Castillo M (2003) Differentiation of toxo-
physiology. AJNR Am J Neuroradiol 27(3):475–487 plasmosis and lymphoma in AIDS patients by using apparent dif-
39. Lin X, Lee M, Buck O et al (2017) Diagnostic accuracy of T1- fusion coefficients. AJNR Am J Neuroradiol 24(4):633–637
weighted dynamic contrast-enhanced-MRI and DWI-ADC for differ- 57. Schroeder PC, Post MJ, Oschatz E, Stadler A, Bruce-Gregorios J,
entiation of glioblastoma and primary CNS lymphoma. AJNR Am J Thurnher MM (2006) Analysis of the utility of diffusion-weighted
Neuroradiol 38(3):485–491 MRI and apparent diffusion coefficient values in distinguishing
40. Doskaliyev A, Yamasaki F, Ohtaki M et al (2012) Lymphomas and central nervous system toxoplasmosis from lymphoma.
glioblastomas: differences in the apparent diffusion coefficient eval- Neuroradiology 48(10):715–720
uated with high b-value diffusion-weighted magnetic resonance 58. Sakai O, Barest GD (2005) Diffusion-weighted imaging of cerebral
imaging at 3T. Eur J Radiol 81(2):339–344 malaria. J Neuroimaging 15(3):278–280
41. Guzman R, Altrichter S, El-Koussy M et al (2008) Contribution of 59. Farrell CJ, Hoh BL, Pisculli ML, Henson JW, Barker FG, Curry
the apparent diffusion coefficient in perilesional edema for the as- WT (2008) Limitations of diffusion-weighted imaging in the diag-
sessment of brain tumors. J Neuroradiol 35(4):224–229 nosis of postoperative infections. Neurosurgery 62(3):577–583 dis-
42. Lu S, Ahn D, Johnson G, Cha S (2003) Peritumoral diffusion tensor cussion 577-583
imaging of high-grade gliomas and metastatic brain tumors. AJNR 60. Fluss J, Ferey S, Menache-Starobinski C, Delavelle J, Van Bogaert P,
Am J Neuroradiol 24(5):937–941 Vargas MI (2010) Mild influenza-associated encephalopathy/
43. Lee EJ, Ahn KJ, Lee EK, Lee YS, Kim DB (2013) Potential role of encephalitis with a reversible splenial lesion in a Caucasian child with
advanced MRI techniques for the peritumoural region in differen- additional cerebellar features. Eur J Paediatr Neurol 14(1):97–100
tiating glioblastoma multiforme and solitary metastatic lesions. Clin 61. Fitsiori A, Nguyen D, Karentzos A, Delavelle J, Vargas MI (2011)
Radiol 68(12):e689–e697 The corpus callosum: white matter or terra incognita. Br J Radiol
44. Viets R, Parsons M, Van Stavern G, Hildebolt C, Sharma A (2013) 84(997):5–18
Hyperintense optic nerve heads on diffusion-weighted imaging: a 62. Guth JC, Futterer SA, Hijaz TA et al (2014) Pearls & oy-sters:
potential imaging sign of papilledema. AJNR Am J Neuroradiol bilateral thalamic involvement in West Nile virus encephalitis.
34(7):1438–1442 Neurology 83(2):e16–e17
45. Vargas MI, Delavelle J, Jlassi H et al (2008) Clinical applications of 63. Ali M, Safriel Y, Sohi J, Llave A, Weathers S (2005) West Nile virus
diffusion tensor tractography of the spinal cord. Neuroradiology infection: MR imaging findings in the nervous system. AJNR Am J
50(1):25–29 Neuroradiol 26(2):289–297
46. Poretti A, Meoded A, Cohen KJ, Grotzer MA, Boltshauser E, 64. Cosottini M, Tavarelli C, Del Bono L et al(2008) Diffusion-
Huisman TA (2013) Apparent diffusion coefficient of pediatric cer- weighted imaging in patients with progressive multifocal
ebellar tumors: a biomarker of tumor grade? Pediatr Blood Cancer leukoencephalopathy. Eur Radiol 18(5):1024–1030
60(12):2036–2041 65. Koeller KK, Shih RY (2017) Viral and prion infections of the cen-
47. Rumboldt Z, Camacho DL, Lake D, Welsh CT, Castillo M (2006) tral nervous system: radiologic-pathologic correlation: from the ra-
Apparent diffusion coefficients for differentiation of cerebellar tu- diologic pathology archives. Radiographics 37(1):199–233
mors in children. AJNR Am J Neuroradiol 27(6):1362–1369 66. Mao-Draayer Y, Braff SP, Nagle KJ, Pendlebury W, Penar PL,
48. Schneider JF, Confort-Gouny S, Viola A et al (2007) Shapiro RE (2002) Emerging patterns of diffusion-weighted MR
Multiparametric differentiation of posterior fossa tumors in chil- imaging in Creutzfeldt-Jakob disease: case report and review of the
dren using diffusion-weighted imaging and short echo-time 1H- literature. AJNR Am J Neuroradiol 23(4):550–556
MR spectroscopy. J Magn Reson Imaging 26(6):1390–1398 67. Summers DM, Collie DA, Zeidler M, Will RG (2004) The pulvinar
49. Morana G, Alves CA, Tortora D et al (2017) Added value of diffu- sign in variant Creutzfeldt-Jakob disease. Arch Neurol 61(3):446–
sion weighted imaging in pediatric central nervous system embry- 447
onal tumors surveillance. Oncotarget 8(36):60401–60413 68. Miller TR, Mohan S, Choudhri AF, Gandhi D, Jindal G (2014)
50. Guzman R, Barth A, Lövblad KO et al (2002) Use of diffusion- Advances in multiple sclerosis and its variants: conventional and
weighted magnetic resonance imaging in differentiating purulent newer imaging techniques. Radiol Clin N Am 52(2):321–336
brain processes from cystic brain tumors. J Neurosurg 97(5): 69. Filippi M, Riccitelli G, Mattioli F et al (2012) Multiple sclerosis:
1101–1107 effects of cognitive rehabilitation on structural and functional MR
51. Reddy JS, Mishra AM, Behari S et al (2006) The role of diffusion- imaging measures—an explorative study. Radiology 262(3):932–
weighted imaging in the differential diagnosis of intracranial cystic 940
mass lesions: a report of 147 lesions. Surg Neurol 66(3):246–250 70. Mottershead JP, Schmierer K, Clemence M et al (2003) High field
discussion 250-241 MRI correlates of myelin content and axonal density in multiple
Insights Imaging (2018) 9:535–547 547

sclerosis—a post-mortem study of the spinal cord. J Neurol 79. Salmela MB, Cauley KA, Nickerson JP, Koski CJ, Filippi CG
250(11):1293–1301 (2010) Magnetic resonance diffusion tensor imaging (MRDTI)
71. Pillai SC, Hacohen Y, Tantsis E et al (2015) Infectious and and tractography in children with septo-optic dysplasia. Pediatr
autoantibody-associated encephalitis: clinical features and long- Radiol 40(5):708–713
term outcome. Pediatrics 135(4):e974–e984 80. Becker M, Kohler R, Vargas MI, Viallon M, Delavelle J (2008)
72. Finke C, Kopp UA, Scheel M et al (2013) Functional and structural Pathology of the trigeminal nerve. Neuroimaging Clin N Am
brain changes in anti-N-methyl-D-aspartate receptor encephalitis. 18(2):283–307 x
Ann Neurol 74(2):284–296 81. Cochrane CJ, Ebmeier KP (2013) Diffusion tensor imaging in par-
73. Wagner J, Schoene-Bake JC, Witt JA et al (2016) Distinct white kinsonian syndromes: a systematic review and meta-analysis.
matter integrity in glutamic acid decarboxylase and voltage-gated Neurology 80(9):857–864
potassium channel-complex antibody-associated limbic encephali- 82. Schwarz ST, Abaei M, Gontu V, Morgan PS, Bajaj N, Auer DP (2013)
tis. Epilepsia 57(3):475–483 Diffusion tensor imaging of nigral degeneration in Parkinson’s dis-
74. He M, Cestari D, Cunnane MB, Rizzo JF (2010) The use of diffu- ease: a region-of-interest and voxel-based study at 3 T and systematic
sion MRI in ischemic optic neuropathy and optic neuritis. Semin review with meta-analysis. Neuroimage Clin 3:481–488
Ophthalmol 25(5–6):225–232 83. Rizzo G, Martinelli P, Manners D et al (2008) Diffusion-weighted
75. Bodanapally UK, Shanmuganathan K, Shin RK et al (2015) brain imaging study of patients with clinical diagnosis of
Hyperintense optic nerve due to diffusion restriction: diffusion- corticobasal degeneration, progressive supranuclear palsy and
weighted imaging in traumatic optic neuropathy. AJNR Am J Parkinson’s disease. Brain 131(Pt 10):2690–2700
Neuroradiol 36(8):1536–1541
84. Rizzo G, Tonon C, Valentino ML et al (2011) Brain diffusion-
76. Bender B, Heine C, Danz S et al (2014) Diffusion restriction of the
weighted imaging in Friedreich’s ataxia. Mov Disord 26(4):705–
optic nerve in patients with acute visual deficit. J Magn Reson
712
Imaging 40(2):334–340
77. Fatima Z, Motosugi U, Muhi A, Hori M, Ishigame K, Araki T
(2013) Diffusion-weighted imaging in optic neuritis. Can Assoc
Radiol J 64(1):51–55 Publisher’s Note
78. Purohit BS, Vargas MI, Ailianou A et al (2016) Orbital tumours and
tumour-like lesions: exploring the armamentarium of Springer Nature remains neutral with regard to jurisdictional claims in
multiparametric imaging. Insights Imaging 7(1):43–68 published maps and institutional affiliations.