3. In 1826 a botanist named Robert Brown
was studying the seemingly random pattern
of motion that pollen grains exhibited
when suspended in water through his
microscope.
It later became clear that the motion that he
had observed was due to the buffeting of
the pollen grains by water molecules
surrounding them.
4. liquids are not static and lifeless as they
might appear at first glance.
The atoms and molecules in constant
motion, undergoing persistent collisions
and energy exchanges with other
molecules and atoms.
This led to an intuitive understanding of
phenomena of spreading of an ink drop in
a glass of seemingly motionless water.
7. ⢠In water at room temperature, the average
distance moved of a water molecule in one
second is around 100 Îźm.
⢠This length scale is of a similar magnitude to that
of many cellular structures of central nervous
system tissue.
⢠Hence if we probe the motion of water molecules
at such timescales , we can probe the geometric
structure of the tissue of the central nervous
system at the cellular level from outside the body.
8. DWI VS DTI
⢠A conventional DWI sequence evaluates diffusion in all directions.
⢠Diffusion tensor imaging (DTI) evaluates diffusion in multiple
different directions (represented by vectors with magnitude and
direction) to investigate the three-dimensional microanatomical
structure of brain parenchyma.
⢠Each point of the imaged tissue is mathematically represented as a
multidimensional diffusion vector that is known as a diffusion tensor.
9. ⢠In cerebrospinal fluid, the diffusion
of protons is unrestricted in all
directions, and therefore isotropic,
⢠In highly organized biological tissue,
diffusion often is restricted in some
directions or anisotropic. and
represented by an elongated
ellipsoid tensor
10. DTI
⢠The diffusion tensor can be fully
characterized by calculating its
âeigenvaluesâ (Îť 1 , Îť 2 , and Îť 3 ),
which describe the length of the
three axes of the ellipsoid, and the
âeigenvectorsâ (Îľ 1 , Îľ 2 , and Îľ 3 ),
which describe the orientation of
these axes in space.
⢠The eigenvectors provide
information about the direction of
maximum diffusion within a voxel
.
11. ⢠From the eigenvalues, the mean diffusivity (MD), is calculated.
⢠The most commonly used measure for diffusion anisotropy is
fractional anisotropy (FA), which is calculated from the eigenvalues
and gives a normalized value to the tensor's degree of anisotropy (0 is
completely isotropic and 1 is completely anisotropic).
⢠Diffusion tensors are commonly visualized with color encoded FA
maps, which display fiber orientation with three standard colors: red
(transverse), blue (craniocaudal), and green (anteroposterior).
DTI
14. The Apparent Diffusion
Coefficient
Apparently, the self-diffusion
coefficient of water has changed,
just because we measured it in the
grey matter.
We are not interested in the ADC for the purpose of quantifying diffusion
itself, but rather to investigate properties of the tissue that apparently caused
the diffusion process to behave in the way that we measure
15. Mean Diffusivity (MD)
the average of the eigenvalues, it
describes the overall size of the
tensor and as such represents a
invariant ADC measure.
17. DTI-based tractography
⢠DTI tractography can be classified broadly into two methods:
deterministic and probabilistic.
⢠In deterministic techniques, a starting or âseedâ point is designated in
three-dimensional space and tracking is terminated when âstop
criteria,â such as a pixel with low FA or a predetermined trajectory
angle between two contiguous vectors is attained.
⢠A distinct limitation is the inability to produce more than one
reconstructed trajectory per seed point.
18.
19. ⢠probabilistic algorithms allow for the modelling of uncertainty of
two (or more) fiber directions at each voxel.
⢠Tracking is done by launching a high number of streamlines from a
seed region, and at each voxel drawing from the previously
determined 3D probability distribution to propagate the tracts.
⢠After a sufficient number of samples, the output is a probabilistic
mapping of the uncertainty of fiber tracts at each voxel, with the
dominant streamline surfacing as most probable.
DTI-based tractography
20.
21. ⢠Probabilistic tractography techniques incorporate uncertainty of
fiber direction to create a probabilistic map of âlikelyâ tracts, which
has the benefit of allowing for branching fibers.
⢠The probabilistic method creates a three-dimensional volume of
potential connectivities that may leak into unexpected regions of the
brain.
⢠This makes visual interpretation of fibers more difficult and requires
judgment to determine relevance.
DTI-based tractography
26. ⢠Pre-surgical planning and intraoperative guidance in
regions adjacent to functional tracts.
⢠In addition to diffusion MRI, high-resolution T1- and T2-weighted
anatomic images must be acquired to coregister with the FA map and
DTI tractography for use with an intraoperative navigation system .
⢠Numerous arbitrary technical factors such as FA and angle of
trajectory thresholds can significantly alter the tract trajectories
generated.
⢠This may result in artifactual delineation of spurious fibers or in
nonvisualization of intact fibers within the area in question.
27. ⢠White matter shifts and deformation significantly limit accuracy of
this technique when DTI tractography is used intraoperatively, with
fiber tract shifts ranging from â8 to +15 mm during tumor resection.
⢠Brain shift may occur after dura mater opening as well as during
tumor resection and is more likely when peritumoral edema is
present.
⢠This has resulted in recommendations that continuous stimulation be
applied when the planned resection is within 1 cm of the DTI
estimated tracts
28. ⢠( A ) Coronal gadolinium-enhanced
image shows a melanoma metastasis
to the brain.
⢠( B ) Coronal functional MRI shows
the metastasis results in medial
displacement of the motor
activation area.
⢠( C ) Color-coded axial fractional
anisotropy map reveals less-robust
anisotropy in the posterior left
centrum semiovale.
⢠( D ) Tractography demonstrates
displacement of the fiber tracts
medially surrounding the area of
motor activation.
29. ⢠In summary, DTI tractography should not be used as the sole
technique for presurgical planning. Instead, directional color-coded
FA map and conventional MRI should be used in conjunction with
tractography to improve visualization of distorted, infiltrated tracts .
30. ⢠DTI metrics also have been used for diagnostic characterization of
brain masses to differentiate solitary metastatic lesions
from gliomas. Measured mean peritumoral MD of metastatic
lesions was found to be significantly greater than that of gliomas.
⢠however, FA values were similar or slightly decreased for gliomas.
31. ⢠FA values also have been used to try to assess tissue
differentiation between low- and high-grade gliomas.
Some authors have shown no significant difference in FA values
between low- and high-grade gliomas ,whereas others have found a
correlation between decreasing FA and increasing glioma grade .
33. ⢠The spinal cord is frequently involved by multiple sclerosis with
the dorsal and lateral columns most commonly affected.
⢠Decreased FA has been shown not only in the cord lesions themselves
but also in adjacent normal-appearing white matter (NAWM) on
conventional MRI in the cervical cord as compared with healthy
controls .
⢠Good correlation has also been found between the average FA and
MD and the severity of disability
34. ⢠Differentiate a spinal cord
ependymoma from an
astrocytoma .
⢠FA values have been shown to be
similar in ependymomas and
astrocytomas and do not help
characterize the tumor.
⢠However, the use of fiber tractography
has proven to be useful because it
often shows displacement of fibers
around ependymomas and infiltration
of fibers in astrocytomas
35. ⢠DTI in cervical spondylotic myelopathy, demonstrating
correlation between decrease in cord FA and clinical disease severity .
⢠DWI and DTI has been used to study multiple other diseases of the
spinal cord, including HIV-associated spinal cord abnormalities,
transverse myelitis, spinal cord injury, and spinal cord ischemia
37. ⢠Diffusion imaging has great potential value in epileptogenic
localization.
⢠In the peri- and postictal state, cortical changes in the apparent
diffusion coefficient (ADC) and MD are similar to those seen in
cerebral ischemia, with early decrease, followed by normalization
and subsequent elevation of these parameters
38. ⢠Diffusion changes typically normalize by day 14 .
⢠Chronic repetition of seizures may then result in irreversible
elevation of ADC and MD, reflecting development of gliosis and
cellular loss.
39. ⢠In cryptogenic temporal lobe epilepsy, findings of decreased FA and
increased ADC were noted in hippocampi and temporal lobe white
matter ipsilateral to the seizure onset in a significant number of
patients with normal MRI .
⢠Similarly, decreased FA and increased MD have been shown in the
normal-appearing subcortical white matter adjacent to focal cortical
dysplasia, identifying these occult abnormalities
40. ⢠Finally, because seizures can induce neuronal injury, chronic
refractory epilepsy may result in synaptic reorganization and altered
connectivity .
⢠Tractography can be used to help delineate these chronic effects such
as structural reorganization of higher cortical functions such as
language and memory .
⢠Thus, tractography may play a significant role in longitudinal studies
of the chronic effects of epilepsy on the brain.
42. ⢠DAI is the result of shear-strain deformation of the brain tissue with
disruption of the axonal membranes and cytoskeletal network.
⢠DTI may detect microstructural injury implicated in DAI that is
linked to persisting symptoms in patients after mild traumatic brain
injury (TBI).
43. ⢠( A ) Axial gradient-recalled echo image
demonstrates focal areas of hypointensity
within the bilateral centrum semiovale
consistent with microhemorrhage associated
with DAI.
⢠( B ) Axial fluid-attenuated inversion recovery
image
⢠( C ) Axial diffusion-weighted image shows
several areas of restricted diffusion near the
gray-white junction bilaterally compatible
with DAI.
⢠( D ) Axial fractional anisotropy (FA) map
demonstrates decreased FA (0.494) in the
posterior body of the right corpus collosum
(black arrow) compared with the same region
on the left (0.646). however, the other
conventional MR sequences (white arrows)
show no abnormality in this region.
45. ⢠MS lesions typically demonstrate increased MD
and decreased FA compared with contralateral
normal appearing white matter .
⢠These abnormalities correlate with axonal and
myelin disruption.
⢠The greatest MD values and lowest FA values
are seen in lesions that are hypointense on
T1WI, which represent chronic destructive
changes where diffusion is least restricted
46. ⢠DTI also has been used to differentiate acute lesions and chronic
lesions.
⢠FA values were found to be relatively lower in enhancing lesions
(especially in lesions with ring enhancement) compared with non-
enhancing lesions .
⢠It is important to recognize that DTI metrics, including FA and MD,
do not differentiate axonal disruption from demyelination or from
focal tissue edema, which are frequently seen in MS.
47. ⢠Normal-appearing gray matter (NAGM), including the basal ganglia
in patients with MS, has also been shown to demonstrate DTI
abnormalities .
⢠A recent study demonstrated correlation between increased FA in
NAGM, cortical lesion volume, and clinical disability in patients with
MS .
48. ⢠DTI can distinguish MS from other secondary demyelinating white
matter diseases such as acute disseminated encephalomyelitis and
neurosarcoidosis.
⢠Increased ADC values in the corpus callosum of patients with MS,
with no significant corresponding ADC elevation in the corpus
callosum of patients with secondary demyelinating disorders.
50. ⢠In addition to evaluation of patients with clinical AD, DTI has been
investigated for detection of incipient AD, which is commonly
clinically classified as amnestic mild cognitive impairment (MCI) .
⢠DTI has demonstrated decreased FA values in white matter of frontal
and occipital lobes in patients with a diagnosis of MCI and AD .
⢠This finding likely corresponds to microarchitectural derangements
in NAWM of patients with AD and also has been shown recently to
correlate with executive function tests in AD and MCI patients
51. ⢠DTI evaluation of hippocampal formations correlates more
consistently with MCI and AD .
⢠These abnormalities include increased MD and decreased FA
in hippocampal formations (especially left hippocampus) of
AD and MCI patients.
⢠Furthermore, these DTI abnormalities in hippocampi were
detected in MCI before significant atrophy was detected .
52. Comparison of conventional T1- (A) and
T2- (B) weighted images, and DTI-derived
mean diffusivity (MD) (C), fractional
anisotropy (FA) (D), and color-coded
orientation (E) maps of cognitively normal
72-year-old woman (upper row) and 70-
year-old woman with Alzheimerâs disease.
The areas surrounded by yellow rectangles
in (E) are magnified and shown in (F) [left
(F-1): from the cognitively normal woman;
right (F-2): the Alzheimerâs disease patient].
The yellow arrows indicate the cingulum
hippocampal part.
54. ⢠ADC values decrease within minutes after onset of cerebral ischemia,
allowing the region of infarction to be delineated within the first 3â6
hours after the onset of symptoms when therapeutic interventions
are most effective .
⢠A potential application for DTI in early stroke is using directionally
encoded color anisotropy images and fiber tractography to delineate
the location of functionally important white matter pathways in
relation to the acute infarct.
⢠For example, studies have shown that the severity of damage to the
arcuate fasciculus as estimated by DTI may predict language
function in the chronic stage after an acute infarct.
55. ⢠DTI has also been used to characterize Wallerian degeneration of
white matter tracts that appear normal on conventional MRI.
⢠A decrease in FA can be detected in the pyramidal tract as early as 2
weeks after an infarct, and such applications may also provide useful
prognostic information
56. Directionally encoded color (DEC) map of DTI
images and the PLIC white-matter tract.
(A) A DTI image of a representative patient (CI001)
with damage to PLIC. Color codes to give diffusion
tensor directions: red represents tracts running left
to right; green is anterior to posterior; blue is
superior to inferior.
(B) Axial view of the PLIC white-matter tract (in
red) of patient CI001.
(C) A DEC map of DTI image of a representative
patient (CI003) with no damage to PLIC. Color also
represents diffusion tensor directions.
(D) Axial view of the PLIC white-matter tract (in
red) of patient CI003.