Water Fat Separation: Cardiovascular Applications

Disclosures

I have no financial relationships to disclose.

- and -

I will discuss the following off label use in my presentation:

Use of contrast agent for late enhancement imaging

Peter Kellman, Ph.D., NHLBI/NIH

Landmarks in Water/Fat Imaging

(W.T. Dixon, Radiology 1984)

2-point Dixon

(G.H. Glover, JMRI 1991)

3-point Dixon

(Q.S. Xiang et al., JMRI 1997)

Direct Phase Encoding

(H. Yu et al., MRM 2005)

Region-growing IDEAL

(S.B. Reeder et al., MRM 2004)

IDEAL (Iterative LS)

(J. Ma, MRM 2004)

2PD with phase correction

(W. Lu et al., MRM 2008)

Golden section search

(D. Hernando et al., MRM 2008)

Variable Projection (VARPRO)

•Nonlinear least squares

•Optimal for any TEs

(D. Hernando et al., MRM 2010)

Graph Cut Optimization

basic 2-point Dixon method:

(assumes homogeneous B0-field)

water & fat separation:multi-echo Non-linear Least Squares methods

multi-echo

dataset

Signal model:

ML cost function:

Hernando D, et al., Joint Estimation of Water/Fat Images and Field Inhomogeneity Map. MagnRes Med. 2008 Mar;59(3):571-580.

Hernando D, et al. Robust water/fat separation in the presence of large field inhomogeneitiesusing a graph cut algorithm. Magn Reson Med. 2010 Jan; 63(1):79–90.

water & fat separation:improved models:

Yu H, et al. Multiecho Water-Fat Separation and Simultaneous R*2 Estimation WithMultifrequency Fat Spectrum Modeling. Magn Reson Med. 2008; 60:1122-34.

Motivation:Tissue characterization

•Intramyocardial fat

•Fibro-fatty infiltration

•Epicardial, Pericardial, & Visceral Fat

•Tumors/masses (lipomas)

Motivation:Artifact mitigation

•Bright fat signal elimination

•Chemical shift

•Out-of-phase cancellation (partial volume)

•Short T1 apparent “late enhancement”

Fat in and around the heart:

WATER

intraatrialseptum

epicardial

fat

mediastinal

fat

parietal

pericardium

FAT

interventricularseptum

coronary

lumen

AV groove

coronary lumen

(bright blood)

(dark blood)

Sequence variations:

Cine

Bright blood

Dark blood prep

PSIR-GRE

late enhancement

Single-shot

Sequence:multi-echo GRE (single phase)

Typical parameters:

acquisition: ECG triggered (1 R-R), segmented

matrix size:256x144

readout flip angle:12 degrees

number of echoes:4 (monopolar readout)

bandwidth:977 Hz/pixel

TE (TR):1.6, 4.2, 6.7, 9.2 ms (11.2 ms)

views-per-segment:19 (8 heartbeats + 1 discarded)

1 2 3 M

  

N-echoes per PE line

Optional IR or DB prep

Tissue Characterization:

Fibro-fatty infiltration1:Potential arrhythmogenic substrate

1Burke AP, et al. Arrhythmogenic RV Cardiomyopathy and Fatty Replacement of the RightVentricular Myocardium: Are They Different Diseases? Circulation. 1998 Apr 28;97:1571-80.

2Bluemke DA, et al. MR Imaging of Arrhythmogenic Right Ventricular Cardiomyopathy:Morphologic Findings and Interobserver Reliability. Cardiology. 2003;99:153-62.

Water

Fat

suppressed

Water

Fat

Current limitation:

Subjective interpretation2

Patient with myocardial lipodystrophy

water

fat

combined

water + fat

Fat-water separation

•Positive contrast

•Improved sensitivity

•Independent of shim

•Objective interpretation

Extensive fatty infiltration in themyocardium

WATER

FAT

Pre-contrast

Late enhancement

WATER

FAT

37-year-old male who presented with exertional ventricular tachycardia and a familyhistory of unexplained sudden death in his mother in her 30’s

The overall clinical presentation and findings are consistent with the diagnosis ofarrhythmogenic right ventricular dysplasia

Water

Fat

SSFP cine still frame

thin finger-like projections of fatinto the right ventricular free wall

Fat in moderator band

Fat in septum

Chronic MI Patient:Fibro-fatty infiltration

Fatty

infiltration

$D:\presentations\ISMRM\ISMRM2012 Fat Water Workshop\Cardiac Applications 2012\E04974_Anterior wall fat not seen on TSE.jpg$

TSE FAT sat

TSE

Water

Fat

conventional

chemical shift

fat-saturation

multi-echo

water-fat

separation

Patient with chonic MI

water

fat

water + fat

SSFP cine

conventional late enhancement

cancellation of fat & water in

partial volume

Intramyocardial Fat:may lead to false late enhancement

WATER

FAT

conventional PSIR

late enhancement

multi-echo PSIR

fat/water separated

late enhancement

apparent lateenhancement

inversion time(TI)

magnetization

normal

fat

Patient with nonischemiccardiomyopathy

WATER

FAT

WATER

FAT

Pre-contrast

Late enhancement

Patient with nonischemiccardiomyopathy

Patient with nonischemic cardiomyopathy

sub-epicardial late enhancement

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01505_S54_conventional_PSIR_subepicardial_enh.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01505_S97_water_FW_PSIR_subepicardial_enh.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01505_S97_fat_FW_PSIR_subepicardial_enh.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01505_S97_fused_fat_on_water_FW_PSIR_subepicardial_enh.tif$

Conventional PSIR

PSIR Water

PSIR Fat

Water + Fat

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04929_S105_108_PSIR_FW_fused.tif$

Patient with fat infiltrating trabeculae

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04929_108_PSIR_FW_fat.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04929_105_PSIR_FW_water.tif$

Duchenne Muscular Dystrophy:Dilated Cardiomyopathy with fibro-fatty infiltration

WATER

FAT

Pre-contrast

Late

Enhancement

CINE

*Dog Model (UNC GRMD colony)

Improved visualization andmeasurement of parietal pericardium

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04828 Ch4 FW PSIR 4e 584C water PSIR.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04828 Ch4 FW PSIR 4e 584C fat on water PSIR.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04828 Ch4 FW PSIR 4e 584C fat PSIR.tif$

Water

Fat

Fat (red) fused with water

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04828 Ch4 FW PSIR 4e 584C water on fat PSIR.tif$

Water (red) fused with fat

Improved visualization andmeasurement of parietal pericardium

Water

Fat

Water (red) fused with fat

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04812 Ch3 FW PSIR 4e 584C_Fat_PSIR.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04812 Ch3 FW PSIR 4e 584C_Water Fat_PSIR.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E04812 Ch3 FW PSIR 4e 584C_Water_PSIR.tif$

Patient with constrictive pericarditis

conventional

PSIR LGE

Water

+ Fat

Water

Fat

free-breathing cine

Fat

Patient with constrictive pericarditis

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01803_PSIR_FW_4ch_fused_constrictive_pericarditis.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01803_PSIR_FW_4ch_fused2_constrictive_pericarditis.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01803_PSIR_FW_4ch_water_constrictive_pericarditis.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01803_PSIR_FW_4ch_fat_constrictive_pericarditis.tif$

$C:\Users\kellmanp.NIH\Documents\My Dropbox\WORK\FatWater\E01803_conventional_PSIR_4ch_constrictive_pericarditis.tif$

Conventional

PSIR LGE

PSIR Water

PSIR Fat

Water + Fat

A 48-year-old female underwentpreoperative stress nuclear imaginga fixed anteroseptal defect was revealed which wasreported as consistent with myocardial infarction

CMR demonstrated an intramyocardial lipomawithin the anteroseptumcorresponding to the fixed defect seen by nuclear techniques

Conventional

chemical shift

fat-saturation

method

TSE FAT sat

TSE

Water

Fat

Multi-echo Dixon

water-fat separation

Intramyocardial lipomaFat water separated cine

SSFP cine(still frame)

Water

Fat

Patient with interpericardial lipoma

Water

Fat

Echo

Patient with suspected intracardiac mass by echo

found to have prominent epicardial fat with lobulation adjacent to rightatrioventricular groove.

Patient referred by Echo to characterize mass

significant mass of fat attached to RV free wall

Patient with aortic dissection and prior omentoplasty

Significant fat within the left hemithorax

Single shot acquisition

accelerated

single shotacquisition

respiratorymotioncorrection

averaging

•Rapid, multi-slice acquisition

•Arrhythmia insensitive

•Free-breathing acquisition

conventional

segmented with

poor breath-hold

single-shot

free breathing

Sequence:single shot, multi-echo GRE

Typical parameters:

acquisition: ECG triggered (1 R-R)

matrix size:192x108 (36 PE lines acquired)

readout flip angle:20 degrees

number of echoes:2 (monopolar readout)

bandwidth:965 Hz/pixel

TE (TR):2.38, 4.76 ms (5.9 ms)

# repetitions:8

1 2 3 M

  

N-echoes per PE line

Single shot, Rate = 3 parallel imaging

   (multiple repetitions)   

Late enhancement sequence:single shot, multi-echo PSIR GRE

Typical parameters:

acquisition: ECG triggered (2 R-R), single shot

matrix size:192x108 (36 PE lines acquired)

readout flip angle:25 degrees

number of echoes:2 (monopolar readout)

bandwidth:965 Hz/pixel

TE (TR):2.38, 4.76 ms (5.9 ms)

# repetitions:8 (16 heartbeats)

1 2 3 M

  

N-echoes per PE line

IR data

(25° flip angle)

reference data

(5° flip angle)

Motion correction

apply

motioncorrection

averaging

fat-water

separated

image recon

non-rigid

image

registration

apply

motioncorrection

averaging

water

fat

motion field

Chronic MI with Fibro-fatty infiltration:Pre-contrast Fat/Water separation

WATER

FAT

Raw images

(8 repetitions)

Motion corrected

Motion corrected average

Chronic MI with Fibro-fatty infiltration:PSIR Late Enhancement with Fat/Water separation

WATER

FAT

Raw images

(8 repetitions)

Motion corrected

Motion corrected average

RV appearance had severeinvagination of right AV groove“pinching” the tricuspid annulus

“caved-in” chest wall deformity

resulting in a leftward shift of herheart and vascular structuresfrom midline

Axial localizer

SSFP cines

Patient with pectus excavatum

Required 3D volumetric acquisition in order to orient eachventricle in isolation and measure true annular diameters

Required water fat suppression to assess possibility of fattyreplacement of RV myocardium

$D:\presentations\ISMRM\ISMRM2012 Fat Water Workshop\Cardiac Applications 2012\A02037_Ser39_40_pectus_excavatum.png$

Volumetric 3D GRE Fat-Water acquisition:

Assessment of RV

Navigated 3D (6.5 min acquisition)

1.5 mm^3 true (non-interpolated) isotropic resolution (160x160x144)

Axial slab, 2D acceleration, rate 4x2 = 8

GRE 3 echoes

Benefits of water-fat separation

•fat has positive contrast

•more objective

•positive correlation of fatty infiltration &fibrosis using fat-water late enhancement

•robust in the presence of backgroundfield variation

•uniform fat suppression (water image)

•water and fat images acquired in a singlebreath-hold

•decreased fat related artifacts