Dynamic Contrast Enhanced Magnetic Resonance Imaging

1
Dynamic Contrast Enhanced Magnetic Resonance
Imaging

• A Brief Overview
• Hematology/Oncology Grand Rounds
• October 16th 2009

2
Disclosures

• No financial interests to disclose, unfortunately

3
Overview

• DCE-MRI a potential biomarker for
  antiangiogenic therapy?
• Basic principles of MRI
• Physics of MRI
• Creating MRI images in a nutshell
• Generating contrast in MRI
• DCE-MRI
• The underlying concept
• The tricky part - implementing DCE-MRI
• Summary and conclusions

4
(No Transcript)
5
DCE-MRI A potential biomarker for
antiangiogenic therapy?

• DCE-MRI is a non-invasive imaging technique that
  can be used to derive quantitative parameters
  that (supposedly) reflect microcirculatory
  structure and function in imaged tissues
• A potential biomarker for antiangiogenic therapy?
• As pointed out by Jain et al at ASCO 2009, many
  challenges exist in terms of finding biomarkers
  of response and resistance in antiangiogenic
  therapy, including
• Angiogenesis in cancer is a heterogeneous and
  dynamic process
• Difficult to do repeated biopsies to assess
  dynamic biomarkers
• Technology for measuring various biomarkers is
  not standardized

6
DCE-MRI A potential biomarker for
antiangiogenic therapy?

• DCE-MRI is considered by some to be a promising
  biomarker of drug efficacy in clinical trials of
  angiogenesis inhibitors
• Non-invasive and may be more feasible than
  repeated biopsies to follow dynamic changes
• Spatially-resolved
• Although DCE-MRI may seem promising, its
  practical application is far from straightforward
• Concensus opinion recommends simple models
  describing parameters e.g. Ktrans and Ve, along
  with IAUC in assessing antiangiogenic and
  vascular disrupting agents in clinical trials

7
Where does the magnetic resonance signal come
from?

• The human body is mostly fat and water and have
  plenty of hydrogen nuclei
• Protons have the quantum mechanical property of
  spin
• Certain nuclei 1H (which is really a proton) have
  a non-integer spin and have a magnetic moment
• When placed in a very strong magnetic field (such
  as in an MRI scanner), protons align in two
  eigenstates - one lower energy eigenstate and
  one higher energy eigenstate
• The excess of protons in the lower energy
  eigenstate compared to the higher energy
  eigenstate gives rise to the magnetization
  signal in MRI
• Relative numbers controlled by the magnetic field
  strength and the temperature
• Higher magnetization with higher magnetic fields
• Higher magnetization with lower temperatures (but
  it may not really be such a great idea to freeze
  your patients)

8
What do MRI scanners offer?

• MRI scanners use a very strong magnetic field
  generated by a superconducting magnet to align
  hydrogen nuclei in water, i.e. magnetization
• Typically, clinical scanners operate at 1.5 Tesla
  Earths magnetic field is 0.00003 to 0.00006
  Tesla
• Radiofrequency (RF) pulses and field gradients to
  select imaging slice, and to encode spatial
  position
• Ability to image in any plane desired
• Ability to manipulate image acquisition process
  to alter contrast, depending on application, e.g.
  T1 weighted, T2 weighted, proton density weighted
  images
• Contrast agents include chelated gadolinium
  compounds
• Has potential to provide functional information
  of different types on top of anatomical imaging
  DCE MRI is an example of a functional
  application of MRI

9
What happens in the MRI scanner?
Magnetic Field Bo
Bo
Net magnetization M
Excitation with RF pulse
Precession at Larmor frequency (dependent on
magnetic field strength)
z
z
Oscillating signal detected by receiver coil
M
M
x
x
Magnetic field gradients impart differences in
magnetic field depending on spatial location, and
therefore causes precession frequency to vary
with position, which allows encoding of spatial
information
y
y
10
What are T1 and T2?

• T2 decay
• Transverse relaxation time, or spin-spin
  relaxation
• Reflects time for disappearance of signal in the
  transverse plane due to interactions between the
  spinning protons

• T1 recovery
• Longitudinal relaxation time, or spin-lattice
  relaxation
• Reflects time for object to become
  re-magnetized after excitation
• Tissues with long T1 take longer to recover

M along Z axis
M in x-y plane
Short T1
Long T2
Long T1
Short T2
time
time
Recovery of magnetization Loss of signal
11
Creating MR images in a nutshell

• Select an imaging slice using radiofrequency
  pulse and slice selection gradient
• Encode position information along one axis
  (typically called the y-axis in MRI
  terminology) using phase encoding gradient
• Frequency encoding and readout of MRI signal
  (echo)
• Data for MRI images is in the form of a matrix
  (usually 256 x 256), and represents the anatomic
  image in the frequency domain or the k-space
• In conventional MRI imaging, a row of data is
  acquired with each excitation -gt thats why MRI
  takes so long
• After collection of the MRI data matrix, a
  process called inverse Fourier Transform is used
  to generate the actual images

12
What are TR and TE?

• Terminology in MRI
• TR
• Repetition time time from one excitation to the
  next
• TE
• Echo time time between excitation and data
  acquisition

TR TE
Excitation
Data acquisition
13
What are pulse sequences?

• Terminology in MRI
• Pulse sequences
• Computer programs that tell the MRI scanner how
  to acquire data -gt can be manipulated by the
  programmer to alter contrast in images, acquire
  functional information, etc

Slice selection
Example of a pulse sequence timing diagram
Phase encoding
Readout
14
Generating contrast

• Contrast in MRI generated by manipulating image
  acquisition parameters

M along Z axis
Short T1
Long T1
time
M in x-y plane
Long T2
Short T2
time
15
Gadolinium-based contrast agents in MRI

• Traditional MRI agents are typically
  gadolinium-based
• Free Gd is highly toxic chelated form used in
  contrast agents
• Variety of different chelates available, eg
• Magnevist (Gd-DTPA)
• Omniscan (Gadodiamide)
• ProHance (Gadoteridol)
• Multihance (Gadobenate dimeglumine)
• Vasovist (Gadofosveset)
• OptiMARK (Gadoversetamide)

16
Other contrast agents in MRI

• Iron oxide nanoparticles have become available as
  contrast agents
• Tend to aggregate -gt Dextran or silica coating
• Superparamagnetic -gt predominantly act by
  shortening T2 relaxation to produce negative
  enhancement
• Vary in size
• SPIO particles are usually 50-150nm in diameter
  and are mainly taken up by phagocytic cells
  within the RES and lymphatic system
• USPIO are 10-15nm in diameter taken up more
  slowly by the RES

17
Concepts underlying the development of DCE-MRI

• DCE-MRI seeks to determine the actual
  pharmacodynamics of tumor contrast enhancement,
  specifically the degree and rate of early tumor
  enhancement
• Contrast agent administered intravascularly
• Contrast travels through vascular system to
  neoplastic tissues
• AIF the time-dependent contrast agent
  concentration in the arterial blood feeding the
  tissue of interest
• Contrast agent leaks from tumor vasculature and
  accumulates in the extracellular extravascular
  space (EES) by passive diffusion
• As plasma concentration falls because of renal
  excretion, backflow of contrast agent from the
  EES to plasma will continue until contrast agent
  has been eliminated

18
Concepts underlying the development of DCE-MRI

• DCE-MRI uses T1-weighted images to detect the
  relaxivity effects of contrast agents during
  dynamic data collection
• Change in the rate of T1 relaxation is
  approximated to be proportional to the
  concentration of contrast agent -gt time
  concentration function can then be generated
  which describes concentration of gadolinium
  within tumor tissue over time
• In order to capture the pharmacodynamic
  information, imaging in DCE-MRI must occur at a
  much faster rate (on the order of 2 to 10
  seconds) than that normally performed in clinical
  MRI

19
How do you analyze the data?

• Look at it - Visual inspection
• Subjective evaluation of time-signal intensity
  curve -gt classify with a grading system
• Semiquantitative analysis
• Parameters such as onset time (time from
  injection to first increase in tissue signal
  enhancement), initial and mean gradient of the
  upsweep of enhancement curves, maximum signal
  intensity and washout gradient
• Initial area under concentration-time curve
  (IAUC) is often used as a biomarker in drug
  trials
• Easy to calculate
• Reasonably reproducible

20
How do you analyze the data?

• Semiquantitative analysis (continued)
• Drawbacks
• Do not accurately reflect contrast medium
  concentration in tissue of interest -gt influenced
  by scanner settings
• Unclear what these parameters reflect
  physiologically and how robust they are to
  patient factors unrelated to tumor physiology
• IAUC, while easy to calculate, has a complicated
  and incompletely defined relationship with
  underlying tumor physiology

21
An example of the application of DCE-MRI in
monitoring response to NAC in breast cancer
22
Applying models and generating numbers
Quantitative DCE-MRI

• Wide range of pharmacokinetic models have been
  applied to the analysis of DCE-MRI data
• Concensus is still lacking on the exact kinetic
  model to be used
• Tofts et al proposed the terms of Ktrans, Ve, etc
  as outcome parameters from a two-compartment
  general kinetic model, which is the most widely
  accepted model
• Parameters can be depicted numerically or as
  color-encoded images

23
Schematic representation of quantitative DCE-MRI
Arterial input function
Simple model


MRI images
Ktrans
Vp
Ve
GOF
Fractional plasma volume
Fractional volume of extravascular space
Transfer constant
Goodness of fit
24
Image Acquisition in DCE-MRI

• Imaging data in DCE-MRI is generated in 3 steps
• Images are obtained which provide anatomical
  information, including localization of the tumor
• Sequences are performed that allow calculation of
  baseline T1 values
• Dynamic data are acquired, typically every few
  seconds, for a duration of 5 to 10 minutes, after
  injection of contrast agent pushing the limit
  of temporal resolution of MRI?

25
You might be a physics major if you'll assume
that a "horse" is a "sphere" in order to make the
math easier.
26
Tofts Modified Tofts models

• Tofts model
• The rate of flux of contrast agent from the
  plasma to the extra-vascular extra-cellular space
  (EES) is assumed to be proportional to the
  concentration difference between the plasma and
  the EES. Within the tissue of interest, the blood
  plasma is assumed to make a negligible
  contribution to the overall signal intensity.
• Modified Tofts model with vp term
• Includes a contribution to the signal intensity
  in the tissue of interest from the blood plasma.
  An extra term (vp - the blood plasma volume
  fraction of the whole tissue) is estimated.

27
Modeling the pharmacokinetics

• All tissues, including tumor tissues, comprise of
  three compartments
• Vascular space
• Extracellular extravascular space (EES)
• Intracellular space
• Clinically used MRI contrast agents do not pass
  into the intracellular space -gt therefore, in
  pharmacokinetic modeling, only two compartments
  are considered

28
Assumptions, assumptions

• Contrast agent concentration in a given
  compartment is uniformly distributed
• Intercompartmental flux is linearly proportional
  to the concentration in each compartment
• Parameters have not changed during data
  acquisition
• Relaxivity of gadolinium contrast agent is
  directly proportional to its concentration

29
Two Compartment Model A simple model
Ve
Vp
Ktrans
Plasma Extracellular extravascular space
Blood flow
30
What is Ktrans?

• Kinetic modeling of contrast agent distribution
  is based on diffusion of solute across a
  semipermeable membrane
• Models used may either ignore or incorporate
  contribution of intravascular contrast agent to
  the MRI signal
• If the contribution of intravascular contrast
  agent to the MRI signal is ignored, the value of
  Ktrans at any location will reflect local blood
  flow, endothelial permeability, endothelial
  surface area, and the proportional blood volume
  within a given voxel
• Despite its physiologic non-specificity -gt
  relatively reproducible and widely used in
  clinical studies

31
More on pharmacokinetic modeling

• Models incorporating contribution of
  intravascular contrast agent to MRI signal
• Models that account for the effects of
  intravascular gadolinium estimate a more
  physiologically specific Ktrans
• Concentration of contrast agent in blood plasma
  (Cp) as a function of time is approximated by the
  arterial input function (AIF), also known as
  vascular input function (VIF)
• Model generates a differential equation that can
  be solved
• Ultimately, the physiologic parameters of Ktrans
  , Ve, and Vp can be estimated from the time
  courses of Cp and Ct (tissue concentration of
  contrast) obtained through dynamic measurement

32
The bottom line on Ktrans

• More on models incorporating contribution of
  intravascular contrast agent to MRI signal
• In this case, Ktrans will be affected by flow,
  capillary surface area, and endothelial
  permeability
• Less affected by changes in plasma volume
• For some applications, may want to separate the
  effects of flow on Ktrans from the effects of
  capillary surface area and capillary endothelial
  permeability

Ktrans depends on pharmacokinetic model applied
Among many things
33
Are all MR contrast agents created equal?

• Current applications of DCE-MRI are based on
  extravasation of low molecular weight contrast
  media (LMCM) such as Gd-DTPA
• Pass through normal endothelia
• Fast wash-in of contrast coupled with fast
  wash-out
• Requires high temporal resolution

34
Are all MR contrast agents created equal?

• Medium molecular weight contrast media (MMCM)
  such as albumin-(Gd-DTPA)30, Gadomer, and
  Gadomelitol leak more slowly into tissues and
  allow longer dynamic acquisition time
• Originally developed for MRA prolonged
  intravascular retention
• Do not pass through normal endothelia
• Different pharmacokinetic properties compared
  LMCM
• The increased size of MMCMs make them less
  diffusible, and Ktrans values may more accurately
  reflect permeability within tumors
• ?More suitable for imaging leaky tumor vasculature

35
DCE-MRI interesting idea, potentially useful,
but WIP

• DCE-MRI is an evolving functional imaging
  technology with the potential for allowing
  assessment of the microvasculature
• Practical application of DCE-MRI remains
  challenging at this time
• Estimation of DCE-MRI parameters is influenced by
  many technical issues, from the imaging hardware
  / software, to the use of contrast agent, and the
  data analysis tools
• Avoid temptation to compare DCE-MRI parameters
  across different trials
• Use of DCE-MRI as a prognostic / predictive
  biomarker has not been firmly established

36
References

• Paldino MJ, Barboriak DP. Fundamentals of
  quantitative dynamic contrast-enhanced MR
  imaging. Magn Reson Imaging Clin N Am. 2009
  May17(2)277-89
• Barrett T, Brechbiel M, Bernardo M, Choyke PL.
  MRI of tumor angiogenesis. J Magn Reson Imaging.
  2007 Aug26(2)235-49
• Hylton N. Dynamic contrast-enhanced magnetic
  resonance imaging as an imaging biomarker. J Clin
  Oncol. 2006 Jul 1024(20)3293-8
• Johansen R, Jensen LR, Rydland J, Goa PE, Kvistad
  KA, Bathen TF, Axelson DE, Lundgren S, Gribbestad
  IS. Predicting survival and early clinical
  response to primary chemotherapy for patients
  with locally advanced breast cancer using
  DCE-MRI. J Magn Reson Imaging. 2009
  Jun29(6)1300-7
• Jaspers K, Aerts HJ, Leiner T, Oostendorp M, van
  Riel NA, Post MJ, Backes WH. Reliability of
  pharmacokinetic parameters small vs.
  medium-sized contrast agents. Magn Reson Med.
  2009 Sep62(3)779-87