Title: Declaration of Conflict of Interest or Relationship
1Declaration of Conflict of Interest or
Relationship
Speaker Name Paul Tofts I have no conflicts of
interest to disclose with regard to the subject
matter of this presentation.
2Methods for quantitative relaxation parameter
mapping measuring T1 and T2
- Paul Tofts PhD
- Chair in Imaging Physics
- Brighton and Sussex Medical School, UK
- Download this from http//www.paul-tofts-phd.org.u
k/talks see also qmri.org
3Why measure relaxation times?
- T1 needed for Ktrans measurement in DCE MRI
- T1 needed for quantitative MT analysis
- T1 and T2 demonstrate subtle damage in
normal-appearing white matter
4(No Transcript)
5overview
- Principles of T1 measurement
- Principles of T2 measurement
- B1 effects
- Quality assurance
- Measurements on a clinical scanner
- Image analysis
6T1 measurement
- Bloch eqn
- Steady state solution
- Repeated ? pulses Mz just before a pulse
7T1 measurement 1. saturation recovery
- Spin echo repeated 90o pulses
- 2 (or more) TRs gives T1
- short and long TRs (TRlt0.5T1, TRgt2T1)
- Needs good 90o pulse (need Mz(0)0)
- Slow (compared to gradient echo GE)
8T1 measurement 2. spoilt gradient echo
- gradient echo repeated ? pulses
- 2 (or more) FAs gives T1 (nb FA?)
- Small and large ?
- (?lt?E PDw ?gt?E T1w ?E Ernst angle)
- Needs accurate ? pulses (more later)
- Fast (can be 3D)
- Inaccuracies (at short TR)
- Incomplete spoiling
- Relaxation during pulse
9T1 measurement 3. inversion recovery
- repeated 180o, TI, 90o , readout, TR-TI
- NB any TR
- arbitrary FAs
- Fix TR, vary TI
- Gives ACCURATE T1 (any TR, any FA)
- This is gold standard, used in spectroscopy
- MRI in spectroscopy mode (small uniform sample
long TR) - Slow interleaving slices
10T1 measurement 4. specialist methods
- There are a number of specialist methods (see
e.g. Gowland 2003), which differ quite
substantially from the conventional methods
described above - include Look-Locker, DESPOT, DESPOT2, and TAPIR
- often very fast, and sometimes can function in
the presence of B1 errors (although at the
expense of worse SNR) - may be vulnerable to off-resonance effects
- usually require intimate access to the MR imager
to implement the sequences
Gowland PA, Stevenson VL. T1 the Longitudinal
Relaxation Time (chapter 5). In Paul Tofts,
editor. Quantitative MRI of the brain measuring
changes caused by disease. Chichester John
Wiley, 2003 111-141.
11T2 measurement
- Bloch eqn
- Steady state solution
- Needs TR-TEgtgtT1 (or else variable recovery after
180o pulse) - If FAs inaccurate,
- T2 estimate is still ACCURATE
12T2 measurement 2. multi-echo
- Tempting to acquire gt1 echo in 1 shot
- E.g. dual echo or FSE
- Selective 180o pulse is imperfect
- Slice thinning
- Even in a long T2 substance (e.g. pure water
T2gt2s) echoes decay with each refocusing - Simple QA on T2 accuracy
- Gold standard is
- repeated single echo
- long TR
- few slices
13Hippocampal T2 in epilepsy
- Conventional multi-slice dual spin echo
TE30,120ms
Figure 2 Unilateral focal anterior hippocampal
sclerosis. The quantitative data show a left
anterior involvement which was missed by visual
inspection.
14T2 measurement 3. multiple-T2 s
- There is no such thing as T2
- In real biological tissue the transverse
magnetisation almost always decays with several
T2 values - 3 distinct compartments in brain
- Myelin water (10-50ms)
- Intra-cellular extra-cellular (70ms)
- CSF (gt1s)
- estimated (mono-exponential) T2 value depends on
TE values chosen - e.g. 4ms range (white matter T280ms)
Whittall KP, MacKay AL, et al . In vivo
measurement of T2 distributions and water
contents in normal human brain. Magn Reson Med
19973734-43.
15B1 effects and FA errors
- Many RT measurement methods rely on accurate FA
- sources of FA error
- Scanner prescan procedure
- Slice selection (in 2D sequence)
- Dielectric resonance
- Worse at 3T
- Transmit coil nonuniformity
- Body coil better than head coil
- Solutions
- Use FA-insensitive methods
- Map FA
Dowell NG, Tofts PS. Fast, accurate, and precise
mapping of the RF field in vivo using the 180
degrees signal null. Magn Reson Med
200758622-630.
16FA errorsand slice selection
Selective 90o pulse gives reasonable profile at
correct B1 value Poor profile at incorrect
B1 Effective FA varies across slice Out-of-phase
component
17Quality assurance for RT mapping
- Needs specific tests
- Distinct from e.g. resolution, SNR tests
Phantoms (test objects) Healthy controls
Convenient to scan -
Well characterised (measure accuracy) -
need temperature control? Yes (2-3 / oC)
realism -
Manufacture/purchase -
18Accuracy and reproducibility
- Accuracy
- Closeness to truth
- Important for multi-centre studies
- Small systematic errors are acceptable in typical
(well-designed) cross-sectional and serial
studies - Reproducibility
- Test-retest
- Bland-Altmann analysis from pairs of measurements
- Instrumental variance usually
- sets limit to smallest biological change that can
be measured - determines power of a study
- Improved reproducibility gives smaller studies
- faster, cheaper
Tofts PS, Collins DJ. Multicentre imaging
measurements for oncology and in the brain. (in
press) 2009.httpwww.paul-tofts-phd.org.uk/CV/repr
ints/multicentre-2009.pdf
19Cross-sectional study
- Normal tissue value 100
- sd_normal 3
- disease effect 5
- perfect instrument
- sd ltlt sd_normal
- sd ltlt 3
- Serial study
- perfect instrument
- sd ltlt sd_within subject
- sd ltlt 1
- An instrument which is so precise that it does
not introduce any significant variation to the
existing biological variation
20Measurement using a clinical MR imager
- Implementation not straightforward
- Commercial sequences for RT are undeveloped
- Speed accuracy B1 errors
- Home-made sequence needs research agreement
- Literature
- Not much!
- cf MT, DTI, MRS, Ktrans, volume
- Scope for research
21Measurement using a clinical MR imager T1
- Spoilt Gradient Echo
- Brain 3D 1.5 x 1.5 x 1.5 in 10 minutes
- Body multi-slice breath-hold?
- B1 and FA nonuniformity
- Body coil nonuniformity fixed for a given
subject - Fast Inversion Recovery
- IR-prepared SPGR IR-EPI (MP-RAGE)
- Accurate (fixed readout short readout, long TR)
- Typical 3 accuracy in 10 min
- Single slice, long time is gold standard
- Slow limited coverage
22Measurement using a clinical MR imager T2
- Spin echo
- Multi-slice (2D)
- Beware slice cross-talk
- Dual echo
- Single echo more accurate but slower
- Faster readout with RARE (multiple phase encodes
per shot) - FSE, turbo SE
- Multiple-T2 analysis
- No commercial software
23Image analysis for relaxation time measurement
- Region of interest
- Test a specific location (prior information and
hypothesis) - biased
- Histogram
- Whole brain unbiased
- good for diffuse disease in Normal Appearing
Tissue - Voxel-Based Morphometry VBM
- Unbiased testing of many locations
- Each location can be correlated with external
score (clinical, genetic, proteomic, cognitive) - Texture
- dirty white matter
- tissue often becomes more heterogeneous in
disease
24More
E-book
Haacke et al 2009
- http//www.paul-tofts-phd.org.uk/talks
- This talk (presentation and document)
- Multicentre measurements 2009 (in press)
- qmri.org
- Out of print (2nd hand market 130 288 512)
- E-book 110 through university library
- Book updates planned on website
- ISMRM list serve Standards in Quantitative MR
http//ismrm.org/sqmr/sqmrrules.htm