Simultaneous Multimodal Imaging at 3T and 9.4T in Humans ...€¦ · 0LWJOLHG GHU +HOPKROW] *HPHLQVFKDIW Simultaneous Multimodal Imaging at 3T and 9.4T in Humans: Recent Advances

Simultaneous Multimodal Imaging at 3T and 9.4T in

Humans: Recent AdvancesN. Jon Shah, Director

Institute of Neuroscience and Medicine – 4Research Centre Juelich

52425 JuelichGERMANY

19 May 2014 Slide 2Institute of Neuroscience and Medicine

MRI

Anatomy

Function

Temporal Resolution

Spatial Resolution

Developmental ProspectsTechnical Maturity

Clinical Availability

Clinical Application

Sensitivity / Specificity

Molecular Imaging

Excellent

Good

Average

Low

Very Low


PETExcellent

Good

Average

Low

Very LowAnatomy

Function

Temporal Resolution

Spatial Resolution





Molecular Imaging


MR-PETExcellent

Good

Average

Low

Very LowAnatomy

Function

Temporal Resolution

Spatial Resolution





Molecular Imaging


EEGExcellent

Good

Average

Low

Very LowAnatomy

Function

Temporal Resolution

Spatial Resolution





Molecular Imaging


MR-PET-EEGExcellent

Good

Average

Low

Very LowAnatomy

Function

Temporal Resolution

Spatial Resolution





Molecular Imaging

3T MR-PET

19 May 2014 Institute of Neuroscience and Medicine

Simultaneous 3T MR-PET Hybrid Measurements

3T MR-PET hybrid scanner showing BrainPET and head

coil

Simultaneous acquisition of 18F-FDG-PET and MR

images

Si lt i iti f


CorrectedCountrate:

x103

x103

MP-RAGE

UTE

Uncorrected Countrate

Corrected Countrate

Weirich,…, Shah (2012) IEEE Trans. Med. Imaging

UTE

MP-RAGE

Correction Factor

time

DEVELOPMENT OF PROTOTYPE 3T MR PET

Cross Calibration of the PET Scanners

Randoms CorrectionScatter Correction Attenuation Correction of Head Normalisation of Crystal EfficienciesDeadtime and Pileup CorrectionAttenuation Correction of RF Coil MRI Interference Correction

Clinical Applications

K.-J. Langen et al.


T1 MPRAGE (6 min)

PET:[18F]-fluor-ethyl-tyrosine20 - 40 min p.i.

Presurgical Imaging on a 3T MR-BrainPET

BOLD imaging:Finger tappingleft hand

Fusion

Brain Tumours18F- FET PET

Clinical Studies Hybrid

MR-PET Imaging

Treatment monitoring

MRI PET


Hybrid MR-PET imaging

Dynamic FET PET 0-50

T1 / T2 PWI T1-weighted + contrastfMRI MRSIDTI

CBF15O-Water PET

Arterial Spin Labelling

K. Zhang, H. Herzog et al.

First Truly Simultaneous Comparisonof CBF Assessed by 15O-Water PET and ASL

PET

3 min

Bolus H215O

pcASL

P

3

7 min

Sequence courtesty of Tom Okell and Peter Jezzard (FMRIB, Oxford)

22nd April, 2013 Slide 16

Averaged Results

ASL-CBF

Averaged CBF images (n=10) after normalisation to MNI space

PET-CBF

ASL: 51.9 7.1 ml/100g/min

PET: 48.1 9.9 ml/100g/min

Quantitative Imaging

N. J. Shah et al.

T1 Mapping

Look and Locker, Rev. Sci. Instr. 41: 250-251

Deichmann and Haase, JMR 1992 96: 608-612

Deichmann et al., MRM 1999 42: 206-209


Implementation

TAPIR (T1 mApping of Partial Inversion Recovery)

Shah et al.,; US Patent No.: 6,803,762Shah et al., NeuroImage: 2001 14(5): 1175-85Steinhoff et al., Magn. Reson. Med.: 46(1) 131-140 2001Zaitsev, et al; Magn. Reson. Med.: 49(1) 1121-1132 2003Shah et al., Hepatology: 2003 38: 1219-26

Tapir: any perissodactyl mammal of the genus Tapirus …. of South and central America and SE Asia, having an elongated snout, three-toedhind legs, and four-toed forelegs.


500ms<T1<3s

Phantom Results

FLASH

Acquisition parameters:TR=13ms; TE=1.9ms; TI=30ms; TD=3 6 ; 4 slices; 48 timepoints; 2562; FoV=250mm; slice-thickness = 8mm; interleaved

T1 Map


TAPIR: In vivo T1 Mapping

Large number of points affords reconstruction of accurate maps

Multi-exponential fitting is feasible

T1 mapping enables quantitative measurement of water content.

S(t) = M0 {1-2exp(t/T1)}

… life is not so simple!

Shah et al.,; US Patent No.: 6,803,762Shah et al., NeuroImage: 2001 14(5): 1175-85Steinhoff et al., Magn. Reson. Med.: 46(1) 131-140 2001Zaitsev, et al; Magn. Reson. Med.: 49(1) 1121-1132 2003Shah et al., Hepatology: 2003 38: 1219-26

1 4 7 12 35

T2* Mapping

Mansfield: 1984 Spectroscopic Imaging (EPSI)


Phantom and in vivo Results

Water Mapping

N. Jon Shah, Zaheer Abbas, Vincent Gras, Klaus Moellenhoff, Anca Oros-Peusquens

19 May 2014 Institute of Neuroscience and Medicine5/19/2014

Tubes filled with different mixtures of (doped) normal water H2O and heavy water D2O. Heavy water is not MR visible at the proton resonance frequency!

Excellent agreement (<1.5%) between known mixing ratio and MR measured water content

Neeb et al., Intern. Congr. Series, 2004 Neeb et al., 13th ISMRM, 2005 Neeb et al., NeuroImage 2006 31 1156-1168

Water Content Mapping


Neeb et al., 2006a,NeuroImage 31 1156-1168

Water Content in Grey/White MatterIn Controls

Shah et al., Hepatology: 2003 38: 1219-26Neeb H, Shah NJ. Magn Reson Med. 2006 56(1):224-9.Neeb H, Zilles K, Shah NJ. NeuroImage. 2006 31(3):1156-68.Neeb H, Zilles K, Shah NJ. NeuroImage. 2006 29(3):910-22.Shah et al., NeuroImage: NeuroImage 2008 41(3):706-17

C5

Water Content Mapping @ 1.5T

Hepatic Encephalopathy grade HE 0

Shah et al., US Patent No.: 6,803,762Shah et al., NeuroImage: 2001 14(5): 1175-85Steinhoff et al., Magn. Reson. Med.: 46(1) 131-140 2001Zaitsev et al; Magn. Reson. Med.: 49(1) 1121-1132 2003Shah et al., German Patent No.: 10028171

Hepatic Encephalopathy grade HE II

putamen putamen

thalamus

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Test-retest stability

Institute of Neuroscience and Medicine

12 measurements

Repositioning

SD of mean values: 0.3%

Voxel-based SD:1% mean value

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Test-retest stability: mean water content map

Institute of Neuroscience and Medicine

Exquisite anatomical detail, e.g. brain stem, thalamus

EPIK and Multi-Contrast EPIK

N. Jon Shah, Seong-Dae Yun

Shared-Echo EPIK

s = sparse sampling factor

k = keyhole-factor

5/19/2014

Sliding Window Reconstruction

Spatial Resolution of EPIK

EPI EPIK

Resolution Profiles

48 56 64 72 80

0.5

1.0

EPI EPIK

Imag

e in

tens

ity

pixel number in phase encoding direction

Susceptibility Artefacts in EPIK

EPI EPIK

fMRI with EPIK

fMRI with EPIK at 3T

EPI

1.25x1.25x4mmMatrix = 192x192Slices = 28TR = 3000msTE = 35msiPAT = 2Fourier = 5/6

EPIK

1x1x4mmMatrix = 240x240Slices = 32TR = 3000msTE = 35msiPAT = 2Fourier = 5/6

EPIK / M-EPI (Recent Activities)

Seong-Dae Yun, Ke Zhang M

itglied der Helm

holtz-Gem

einschaft

EPI: able to achieve a relatively high imaging speed but it is prone to field inhomogeneities => image distortions

EPIK : multi-shot EPI + Keyhole imaging

Accelerated EPIK : EPIK + parallel imaging

13/12/13 40

Imaging techniques: EPIK, Accelerated-EPIK

EPIKEPI Accelerated-EPIK

Sparse

Keyhole

Sparse

Mitglied der H

elmholtz-G

emeinschaft

AIF is of great interest for quantification in hybrid MR-PET.

EPIK has been extended to offer dual-contrast: T1 and T2*

More information with almost same temporal resolution as EPI

41

Application of EPIKEnhanced characterisation of AIF using dual-EPIK

11th scan 18th scan

T1

T2*

Mitglied der H

elmholtz-G

emeinschaft

13/12/13 43

Perfusion Maps with M-EPI

High-Field MRI


9.4T Whole-Body Scanner in Jülich

60 cm patient boreTQ-engine gradient coil50 cm FoVMagnet weight: 57 tonnes870 tonnes of iron shielding3.70 m lengthStored energy: 182.0 MJLength of wire: 750 km

Complete with Hybrid PET Capability!

Structural Imaging

A.-M. Oros-Peusquens, J. Lindemeyer et al.


Hippocampus, thalamus

fornix and plexus choroideus

corpuscallosum

thalamicnuclei

externalglobuspallidus

putamenp

internalglobuspallidus

hippocampusentorhinal cortex


Anterior Hippocampus

entorhinal cortex gyrus dentatus

cornu ammonis (CA2)

cornu ammonis(CA1)

subiculum

9komma4


Structural imaging at 9.4T with (0.5mm)3

resolution5ms 15ms


Phase Imaging at 9.4T (in vivo)Phase Unwrapped Phase (URSULA) Background-Corrected

(MUBAFIRE)

GRE 0.5mm isotropic, slab-selective excitation (central brain)

MR-PET @ 9.4T

N. J. Shah, H. Herzog, C. Weirich et al.


9.4T MR-PET


High-Energy Positron Emitter 120I in the “Iida” Brain Phantom

9.4 T 7 T 9.4 T 7 T

2 T 0 T 3 T 0 T


Positron Range at 9.4T using the “Iida” Brain Phantom

Polymer brain phantom filled with 120I

0T 3T

9.4T7T

Shah (2014) PLoS ONE (in press)

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

T2-weighted MRI @ 9.4 T

Rat Bone Scan with 18F-Fluoride

BrainPET @ 3 T

BrainPET @ 9.4 T

Inveon PET

Shah (2014) BSF (in review)


Opportunities for Hybrid MR-PET

… MRIHigher spatial resolution (structural imaging)Higher functional (BOLD) contrast => columnar resolution fMRI?Better image quality (contrast)Non-proton MRI and spectroscopy

… PETPartial volume correction with MRIAttenuation correction with MRI Motion correction with MRI (navigator echoes)

… Hybrid MR-PETPatient / volunteer compliance: 2 scans in 1 (at 3T and 9.4T)Metabolic imaging (e.g. FDG + 17O + 31P + 23Na + MP-RAGE)Accurate receptor density mapping Novel paradigms for brain function


Opportunities – Metabolic Imaging … Sodium

Na / K PumpDisturbances of the pump often leads to cell deathIntra vs extracellular sodium with TQF

… PhosphorusEnergy metabolism of the cellIn vivo pH

… OxygenIntimately involved in metabolism!…..

… GlucoseEnergy substrate of the brainFDG PET

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Sodium Imaging

S. Romanzetti, D. Fiege, N. J. Shah et al.


First In vivo 9.4 T results

Anatomy - 1HMP-RAGE 4T1 mm isotropic5 min acq. time

Sodium – 4TTPI 2 mm isotropic15 min acq. time

Sodium – 9.4TTPI 2 mm isotropic15 min acq. time


In vivo measurements (23Na)

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Sodium Imaging (2x2x2mm3 in 7 minutes)

In vivo 23Na Imaging with TPI steady state (TR/TE/flip = 50/0.4/60 [ms/ms/deg])

Fiege, Romanzetti, …, and Shah (2013) Magn. Reson. Med. Fiege, Romanzetti, …, and Shah (2013) J. Magn. Reson.Romanzetti, …, and Shah (2014) NeuroImage (in press)


23Na TQF in vivo 9.4 T results

In vivo results from three tumour patients.

MP-RAGE

PET

SISTINA SQ mmol/l

SISTINA TQF

FLAIR

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Oxygen Imaging

K. Moellenhoff and N. J. Shah


First in vivo 17O 9.4 T results

TPI sequence with parameters:

TE=0.4msTR=11.4ms, readout time=5.12msFA = 30 degree Voxel size = 5mm isotropic;Averages = 10Acq. time = 11.4 min

Natural Abundance 17O Imaging (0.037%)

K. Möllenhoff and N.J. Shah


Method must beAccurate & reproducibleSuitable for patientsNon- or minimally-invasive

Method will ideally Have the potential for widespread useNot require radiotracers

15O PET does not meet these requirementsRequires arterial sampling for true quantitative valuesNot widely available despite a technique existing for decadesRequires onsite cyclotron to produce the short half-life 15O

Goal: Non-invasive Quantitative Metabolic Imaging of Oxygen

17O

MR

I

Slide: Atkinson et al. U. of Chicago


Where We Are: Data Collection

70% 17O to produce “large” signal change9.4 T static field to maximize sensitivityOptimized coil, preamp, TR switch

flexTPI acquisition 20 cm FOV, 8 mm isotropic resolution30-40 s volume repeat time40-60 min total scan time (5-15 min 17O inhalation)

Co-registered water maps to correct for tissue mass

Reservoir

CO2Scrubber

valveRoom

Air

Subject Air In/Out

17O-enrichedgas

DemandValve

Affinity.imaging


Where We Are: Human Results

White Matter Grey Matter

CMRO2umoles/g brain/min

Slide: Atkinson et al. U. of Chicago

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

MR Spectroscopy

D. Tse and N. J. Shah

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

19. Mai 2014 INM – 4 78

1H MRS at INM – 4


2-Hydroxygluterate in Brain Tumours @ 3T18F-FET PET + 1H MRS Detection of 2HG

2HG, normally in vanishingly small amounts, can be elevated by orders of magnitude in gliomas in IHD1 and IHD2 mutations (Choi et al)Detectable in oligodendroglioma (WHO: 2&3), astrocytoma (WHO: 2&3), oligoastrocytoma (WHO: 3) and secondary glioblastoma (WHO: 4), but not in glioblastomaCase study: non-significant FET uptake, contrast agent in T1 and FLAIR, positive 2HG => astrocytoma (WHO: 2)

Choi et al., Nat. Med. 18, 624-629 (2012)

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

19. Mai 2014 INM – 4 80

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

In Vivo Proton Spectroscopy at 9.4T

Shah (2014) BSF (in review)

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

19. Mai 2014 INM – 4 82

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

19. Mai 2014 INM – 4 83

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

19. Mai 2014 INM – 4 84

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Temporal Aspects

I. Neuner, J. Arrubla and N. J. Shah

Mitg

lied

der H

elm

holtz

-Gem

eins

chaf

t

Temporal Aspects: EEG @ 9.4T

ECG / EEG @ 0T

ECG / EEG @ 9.4T

Neuner, …, and Shah (2013a) NeuroImage Arrubla, …, and Shah (2013) PLoS ONENeuner, ....., Shah (2013b) NeuroImage

Acknowledgements

Dr. J. FelderA. CelikK. VahedipourC. Mirkes

MR Group

T. Okell / P. Jezzard (Oxford)

SIEMENS / BMBF

Dr. J. ScheinsDr. E. Rota-KopsL. Tellmann

PET Group

Dr. N. GaldiksDr. G. StoffelsDr. C. Filß

Brain Tumour Group

Thank You!!

Collaboration Opportunities

Visiting Scholars / Guest ScientistsPostdocsGraduate Students

Documents

Simultaneous Multimodal Imaging at 3T and 9.4T in Humans ...€¦ · 0LWJOLHG GHU +HOPKROW] *HPHLQVFKDIW Simultaneous Multimodal Imaging at 3T and 9.4T in Humans: Recent Advances