1. Diffusion Tensor Imaging: Is It Ready For The Clinic ? eEdE:14
Tushar Chandra, MD1 Mohit Agarwal, MD1 Ibrahim Tuna, MD1 Laura Kohl, MD1 Andrew Klein, MD1 Leighton Mark, MD1 Mohit Maheshwari, MD1 Suyash Mohan, MD2 Sumei Wang, MD2 John Ulmer, MD1
Medical College of Wisconsin, Milwaukee1
Perelman School of Medicine, University of Pennsylvania2

2. Disclosures
Nothing to disclose

3. Educational Objectives
Succinct overview of the fundamental principles and techniques of diffusion imaging, Diffusion tensor imaging (DTI), fiber tractography and Diffusion kurtosis imaging (DKI)
Simplified interpretation of DTI metrics
Discuss clinical application of DTI in neuropathology
Overview technical limitations and pitfalls

4. Introduction Introduction
Diffusion Tensor Imaging (DTI) is a novel method which has various applications in clinical neuroimaging and research
Within the central nervous system, water diffusion is more anisotropic in white matter and isotropic in gray matter and CSF
This property can be exploited to highlight white matter changes in various pathological processes
DTI is a powerful tool for assessment of microstructural integrity of the white matter qualitatively as well as quantitatively

5. Diffusion imaging - Principle
Water molecules in biological tissues are in constant movement, governed by two major principles:
a. Fick`s Law: Random diffusion due to
concentration differences
b. Temperature and ion-ion interactions
Diffusion of water molecules can be restricted in various pathological conditions
Random
Brownian Motion
Free Diffusion
Restricted Diffusion

6. Diffusion imaging
Free Diffusion Restricted Diffusion

7. Diffusion imaging - Technique
Detects the molecular motion of water and allows for quantitative assessement of the freedom of diffusion
The addition of 2 strong, symmetric gradients to a EPI SE sequence helps in differentiation of stationary from mobile water molecules
If there is net movement of spins (i.e. if diffusion occurs) between the 2 gradients, signal attenuation occurs
MRI pulse diagram demonstrating placement of strong, symmetric diffusion encoding gradients on either side of the 1800 radiofrequency pulse that enables creation of diffusion weighted images (With permission from Hagmann P, Jonasson L, Maeder P et al. Understanding diffusion MR imaging techniques: from scalar diffusion-weighted imaging to diffusion tensor imaging and beyond. Radiographics 2006;26: S
Radiographics 2006;26: S

8. Application of gradients
Diffusion Signal
Diffusion imaging
Application of gradients
No Diffusion
Between Gradients -
More signal
Signal drop
More Diffusion
Between Gradients -
Less signal
Gradients cause a drop in signal
if diffusion is present

9. ‘b’ value Represents the strength of ‘diffusion sensitizing gradients’
Expressed in s/mm2
The larger the b value, the smaller magnitudes of water motion detected.
Since EPI is intrinsically sensitive to magnetic field inhomogeneity, paramagnetic blood breakdown products produce signal loss similar to that in T2*- weighted GRE sequence.
b 0 image:
No diffusion weighting
Poor man’s gradient or T2

10. Apparent diffusion coefficient - ADC
Measures area of water molecular diffusion in 1 second
Expressed in mm2/s
Reduced ADC - acute stroke, abscesses, cellular neoplasms, recurrent tumors
Increased ADC – benign lesions, necrosis, post radiation changes
According to Fick’s law, true diffusion is the net movement of molecules due to a concentration gradient. With MR imaging, molecular motion due to concentration gradients cannot be differentiated from molecular motion due to pressure gradients, thermal gradients, or ionic interactions. Also, with MR imaging we do not correct for the volume fraction available or the increases in distance traveled due to tortuous pathways. Therefore, when measuring molecular motion with DW imaging, only the apparent diffusion coefficient (ADC) can be calculated. The signal intensity of a DW image is best expressed as
Why Apparent?
Since MRI cannot distinguish molecular motion arising from differences
in concentration gradient from that resulting from temperature gradient
or other reasons, the coefficient is apparent and not a true value

11. Exponential Apparent diffusion coefficient -eADC
Derived from dividing DWI by T2 images to remove effects of T2 shine through
True restricted diffusion – dark on ADC, bright on eADC
ADC or eADC maps can be used depending on whether we want contrast to match, or be opposite to, the diffusion weighted images

12. Exponential Apparent diffusion coefficient -eADC
DWI
ADC
eADC
An area of increased diffusion signal on DWI image in the left parietal lobe in a 60 y/o male with treated astrocytoma is slightly dark on ADC but not increased in signal on eADC, suggesting that there is no ‘true’ restricted diffusion.
There was no recurrent tumor on pathology

13. Diffusion Tensor imaging
ISOTROPIC- Equal diffusion in all directions
ANISOTROPIC – Diffusion preferentially increased in some directions

14. Diffusion Tensor imaging
DTI requires obtaining data from diffusion acquisitions with gradients in different directions in each acquisition to provide directional information to the diffusion data
The information is provided by 3 eigen values which represent the direction of 3 major axes of the ellipsoid and 3 eigen vectors that represent the magnitude in these directions
In the white matter, diffusion is anisotropic and is related to cell density and integrity, axonal integrity, and myelination status
Isotropic Diffusion
Anisotropic Diffusion

15. Physiological Principals of DTI
White
matter
H2O
H2O
H2O
H2O
Diffusion Gradients

16. Physiological Principals of DTI
White
matter
Voxel
H2O
H2O
Diffusion Ellipsoids
Commentaries: Mark, Ulmer. AJNR 2002, 2004

17. Diffusion Tensor Diffusion Tensor imaging λ1 λ2 λ3
Tensor is a mathematical model of directional anisotropy of diffusion
Diffusion tensor describes Gaussian diffusion distribution - a 3D ellipsoid with lengths and orientations of the 3 axes corresponding to the eigen vectors - λ1, λ2 and λ3
Acquisition in at least 6 directions is required, but clinically
up to 30 directions are used
From the tensor, we can calculate:
a. Direction of greatest diffusion
b. Degree of anisotropy
c. Diffusion constant in any direction λ1 λ2 λ3

18. Diffusion Kurtosis imagingλ1 λ2 λ3
DTI Metrics and Tensor
Diffusion Kurtosis imaging
Diffusion Kurtosis imaging λ1 λ2 λ3
ISOTROPIC
ANISOTROPIC
Mean Diffusivity (MD) = (λ1+λ2+λ3)/3 Axial Diffusivity (Da) = λ1 Radial Diffusivity (Dr) = (λ2+λ3)/2

19. Fractional anisotropy - FA
Measures the degree of anisotropic (unequal) diffusion in a voxel
Ranges from 0 to 1 (no units)
0 – isotropic (sphere-like)
1 – Purely anisotropic (straight line)
Can characterize demyelinating lesions, e.g., breakdown of myelin and axonal loss can reduce FA and remyelination can increase FA
FA value of CSF is 0.
Color coded FA map
(Red –Higher FA, Blue – Lower FA)
Note thatWM tracts showing red color have a higher FA

20. Mean Diffusivity - MD Measure of directionally averaged
magnitude of diffusion (λ1+λ2+λ3)/3
Higher MD values mean that the tissue
is more isotropic
MD is an inverse measure of membrane density and tumor cellularity
Sensitive to cellularity, edema and necrosis
Color coded MD map
(Red –lower MD
Purple – higher MD)

21. Axial Diffusivity - Da
Da is the apparent diffusion parallel to white matter tracts
Da = Prinicipal Eigen value = λ1
Da is variable in white matter pathologies
Da decreases in axonal degeneration
Color coded Da map
(Red –Higher Da Blue – Lower Da)

22. Radial Diffusivity - Dr
Apparent diffusion perpendicular to the white matter tracts
Dr = (λ2+λ3)/2
Dr generally increases in white matter demyelination and dysmyelination
Change in axonal diameter and density also affect Dr
Color coded Dr map
(Red –Higher Dr Green – Lower FA)

23. DTI - Tractography Fiber Tractography
Technique to assess direction of white matter tracts within the brain
Directional information from neighboring voxels is combined to estimate 3D structure of major white-matter pathways
Voxels are connected together taking into consideration both the direction of principle Eigen vector and FA value

24. HARDI – can assess crossing tracts in the same voxel
DTI - HARDI
DTI ellipsoid not accurate for detecting white matter tracts as it assumes one direction of axons in each voxel (in truth, there are crossing fibers in each voxel)
HARDI – can assess crossing tracts in the same voxel
For the radiologist who wants to use these techniques in clinical practice and research, it is important to understand a few key principles of diffusion imag- ing so as to select the appropriate technique for answering a specific question.

25. Diffusion Kurtosis imaging
DKI is an extension of conventional DTI.
DTI assumes Gaussian distribution (bell shaped curve) of diffusion (not accurate), as water diffusion in biological tissues is non-Gaussian.
Due to the effects of cellular microstructure e.g., cell membranes, organelles & myelin in brain
Diffusion kurtosis – studies non-Gaussian diffusion behavior.
Kurtosis measures the "peakedness" of the probability distribution.
Qualitatively, a large diffusional kurtosis suggests a high degree of diffusional heterogeneity and microstructural complexity.
Mean kurtosis, the average apparent kurtosis along all diffusion gradient encoding directions, may offer an improved sensitivity in detecting developmental and pathological changes in neural tissues as compared to conventional DTI
Leptokurtic- K>0
Mesokurtic –K=0
Platykurtic –K<0

26. Diffusion Kurtosis imaging
From the diffusion and diffusional kurtosis tensors several rotationally invariant metrics such as the mean, axial, and radial kurtoses can be computed
The extra information provided by DKI can also resolve intra-voxel fiber crossings and thus be used to improve fiber tractography of white matter
DKI protocols require at least 3 b-values (as compared to 2 b-values for DTI) and at least 15 independent diffusion gradient directions (as compared to 6 for DTI)
Typical protocols for brain have b-values of 0, 1000, 2000 s/mm2 with 30 diffusion directions

27. Functional MRI BOLD – Blood Oxygen Level Dependent
As neural activity increases, blood flow increases
Deoxyhemoglobin (paramagnetic) concentration decreases
Magnetic field homogeneity increases
And therefore gradient echo EPI signal increases, rather than loss of signal
BOLD technique is used with DTI fiber tractography in pre-surgical mapping.
Rest: Normal flow
As neural activity increases, blood flow increases, and deoxyhemoglobin concentration decreases, deoxyhemoglobin is paramagnetic, and hence magnetic field homogeneity increases, and therefore gradient echo EPI signal increases, rather loss of signal.
Activity: High flow
- Deoxyhemoglobin
- Oxyhemoglobin

28. Clinical Applications – Normal Brain
Fiber tracking provides critical information about white matter anatomy and connections
Regions with similar tractographic features tend to be functionally co-activated - “neurons that fire together, wire together”
IQ has been positively correlated with anisotropy in white matter association areas
Reading ability has been correlated with anisotropy of left temporoparietal areas
In the visual pathway, DTI has shown the retinotopic organization of fibers

29. Clinical Applications - Tumors
Flair hyperintense mass in
Right frontotemporal region
MD decreases as tumor cellularity increases, due to decreased ECF volume
Atypical and malignant meningiomas - lower MD than typical meningiomas
Primary CNS lymphoma and Medulloblastoma also have low MD
MD increases with tumor response with treatment and can be used as a biomarker
Relationship of FA with tumor cellularity and treatment response is unclear
In the peritumoral zone, DTI metrics do not reliably differentiate edema from tumor infiltration
Increased MD
suggesting low cellularity
Gr II glioma at biopsy

30. Clinical Applications - Tumors
Edematous or tumor-infiltrated tracts lose some anisotropy
but remain identifiable
Intact WM tracts displaced by tumor retain anisotropy and remain identifiable
Destroyed WM tracts lose directional organization
and diffusion anisotropy is lost completely
Jellinson et al. AJNR 2004

31. Diffusion and Functional Imaging For Tumors
Diffusion-weighted imaging: Diffusion Image, Apparent diffusion coefficient (ADC), eADC – tumor cellularity
Diffusion tensor imaging: Fractional anisotropy, diffusivity (mean, axial and radial) – tumor biology
Diffusion and
Functional Imaging
For Tumors
Tractography: Accurate localization of white matter tracts in relationship to the tumor margins
Functional MR Imaging: Depiction of eloquent cortical areas in relationship to tumor margins

32. Clinical Applications - Presurgical Brain Mapping
Progression free survival is directly related to the extent of resection
However, benefits of cytoreduction must be weighed against risk of damage to eloquent structures and white matter tracts
Pre-surgical mapping with DTI and fMRI results in more informed presurgical planning and decreases the risk of post operative neurological deficits
Motor
Area
Tumor
White Matter
Tracts
Fused image with functional motor areas and white
matter tracts superimposed on FLAIR depict relationship
of tumor to eloquent cortex and white matter tracts

33. Fiber Tractography - Presurgical Brain Mapping
Track-ball filtering of whole brain fiber tracking DTI data reveals better detail of spatial relationships between tumor and SLF HB, SLF IV, IFOF, ILF, and OR
IFOF = Inferior fronto-occipital fasciculus, ILF = Inferior longitudinal fasciculus,
SLF HB = Superior longitudinal fasciculus horizontal bundle,
SLF IV = Superior longitudinal fasciculus IV, OR = Optic radiation, UF = Uncinate fasciculus
Ulmer et al Neuroradiology Clinics of North America 2014
34-year-old, right-handed woman with a posterior
parasylvian low-grade glioma. SPGR gadolinium-enhanced
underlays with 50% faded Colorcoded fractional anisotropy
(CC-FA) diffusion tensor imaging (DTI) map

34. Tumors - Which functional systems are at risk ?
Motor
(Corticospinal
Tract)
Vision
(ILF, IFOF
Optic Radiations)
Language
(SLF,ILF,IFOF)
Vision
(Optic Radiation)

35. Clinical Applications - Demyelination
MD image: High MD
FLAIR : MS plaque 
MS lesions have higher ADC and lower FA values than Normal Appearing White Matter (NAWM)
Significantly increased ADC and lower FA values are seen in acute (enhancing) MS lesions than chronic (non enhancing) lesions
Non enhancing TI hypointense lesions have higher ADC and lower FA values than T1 isointense lesions
Color FA Map : low FA
Tractography:
Decreased WM fibers

36. Clinical Applications - Epilepsy
Gray Matter
Heterotopia
Increased MD and lower FA values are seen in hippocampi of patients with mesial temporal sclerosis
In patients with malformations of cortical development, increased MD and lower FA values are seen in abnormal areas within MCD and also in the normal appearing areas on MR
Increased MD and low FA can be used to localize lesions in MR negative cases of epilepsy
Displaced WM Tracts
Decreased
Radial
Diffusivity

37. Clinical Applications – Congenital Anomalies
Schizencephaly
White matter abnormalities in congenital brain malformations can be assessed with DTI
Pertinent applications include callosal agenesis, cortical dysplasia, holoprosencephaly, schizencephaly, Chiari II malformation etc
Improved understanding of white matter abnormalities in developmental lesions
DTI
Disrupted
WM Tracts
fMRI –
Motor
Cortex
Along
the cleft

38. Clinical Applications – Traumatic Brain Injury
DTI is a useful technique to evaluate microstructural injury to the white matter fiber tracts in patients with TBI
Decreased FA and increased MD are seen in areas afflicted by TBI, that are occult on conventional MRI
Studies suggest some correlation between findings on DTI with EEG and neuropsychological testing
In the future, DTI may serve as a surrogate marker for closed head injury
Cingulum
Temporal White Matter

39. Interpretative Challenges of Clinical DTI
Tumor, edema and radiation-induced decrease in anisotropy.
Tumor-induced geometric distortions of fiber tracts.
Anatomic constraints
Distinguishing functionally different pathways in the same white matter bundle.
Acute angulations and blending of white matter pathways.
DTI data are imperfect!

40. Conclusion
 DTI is a powerful tool to investigate microstructural white matter changes and brain connectivity
DTI is currently being clinically used in conjunction with functional MRI for presurgical brain mapping and is gradually becoming the standard of care
For indications such as demyelination, trauma, epilepsy and congenital anomalies, DTI provides useful information that is clinically helpful and often helps in diagnostic interpretation and clinical decision making
As the technique becomes more robust, it will be increasingly applied in clinical practice for other indications

41. Thank You Author: Tushar Chandra Clinical Instructor, Radiology
Medical College of Wisconsin
9200 W Wisconsin Avenue,
Milwaukee WI 53226