1 sleep apnea asthma heart disease cardiovascular disease stroke / - PDF document

TOOLBOX of METHODS for FAT MRI Water (Lean Tissue) vs. Fat Water (Lean Tissue) vs. Fat Signal Contrast Signal Contrast T1, T2 ‐ weighting RELAXOMETRY Based Inversion Recovery (e.g. T1 nulling) Frequency ‐ Selective Methods CHEMICAL ‐ SHIFT Chemical ‐ Shift ‐ Encoded Based Water ‐ Fat MRI 1H Single Voxel MRS Magnetization Transfer MICROSTRUCTURE Based Diffusion TOOLBOX of METHODS for FAT MRI TISSUE CONTRAST Water (Lean Tissue) vs. Fat Water (Lean Tissue) vs. Fat MR Signal = Water, Fat Signal Contrast Signal Contrast ? Water, Fat T1, T2 ‐ weighting (Fat is brighter) Water, Fat Inversion Recovery (STIR, FLAIR) (e.g. T1 nulling) (Fat or Water is nulled) (Fat ‐ Sat, Water, Fat Frequency ‐ Selective CHESS) Methods (Fat or Water is nulled) Water, Fat Chemical ‐ Shift ‐ Encoded Water ‐ Fat MRI (Fat & Water separated 1H Single Voxel MRS 1

sleep apnea asthma heart disease cardiovascular disease stroke / hypertension fatty liver disease gallbladder disease metabolic disorders cholestrol / glucose colon cancer insulin / type 2 diabetes back pain PCOS spine muscular dystrophy osteoarthritis muscle strength peripheral vascular diseases Courtesy of Wei Shen, MD (Columbia ‐ New York) & Steven Heymsfield, MD (Pennington ‐ Lousiana) 2

MRI of Fat by T1 ‐ WEIGHTING Fat has shortest T1, fast recovery, brightest signal. signal contrast slow T1 recovery Würslin C, et al. JMRI 2010 Peng Q, et al. JMRI 2005 Jürgen Machann, PhD (Tübingen, Germany) 3

Whole Body T1 ‐ weighted MRI (circa 2005) LE T UE I II III IV V VI VII VIII I III V VII I III V VII II IV VI VIII II IV VI VIII Courtesy of Dr. J. Machann (Univ. Hospital Tubingen), J Magn Reson Imaging 2005:21:455 ‐ 462. MRI of Fat by T2 ‐ WEIGHTING (?) T1 ‐ weighted T2 ‐ weighted Neither the shortest nor longest ... T1 ‐ weighted T2 ‐ weighted 4

Slide courtesy of Shahid Hussain, MD PhD (Nebraska) WATER versus FAT (Chemical ‐ Shift) O H symmetrical Ren et al. J Lipid Res 2008; 49:2055–2062 ‐ CH2 ‐ methylene peak 4.7ppm B only: single ‐ peak model A thru J: multi ‐ peak model Courtesy of Johan Berglund, PhD (Uppsala University, Sweden) 5

Selective Water ‐ Fat Excitation / Suppression – Develop pulse sequences to … … SUPPRESS fat / excite water � mainly water signal remain … SUPPRESS water / excite fat � mainly fat signal remain Saturation RF pulse Saturation RF pulse centered on FAT centered on WATER resonant frequency resonant frequency Water Fat (CH 2 ) Water Fat (CH 2 ) Selective Water ‐ Fat Excitation / Suppression T2 ‐ weighted T2 ‐ weighted fat sat Fat Sat Fat Sat Water Sat Water Sat residual fat signal no residual (WHY?) water signal Machann J et al., Annual Reports on NMR Spectroscopy 2003:50:1 ‐ 74. 6

☺ � � water sat Selective Water Suppression (“Water Sat”) Why Does it Fail? B0 Magnet Inhomogeneity ☺ Water Fat (CH 2 ) Peak locations shifted � But RF pulse still targets the same frequencies � water sat 7

Single vs. Multi ‐ Peak Fat Model Fat Sat Fat Sat ‐ HC=CH ‐ residual fat signal Water Sat Water Sat no residual water signal Machann J et al., Annual Reports on NMR Spectroscopy 2003:50:1 ‐ 74. W F (CH 2 ) bkgd (dark) fat (bright) musc T1 ‐ weighted water sat Peng Q et al., Machann J et al., J Magn Reson Imaging Magn Reson Med T1 ‐ weighted water sat spectroscopy 2005:21:263-271. 2006:55:913-917. 8

Slide courtesy of Russell N. Low, MD (San Diego) “Dixon” Method (2 ‐ point) Fat ‐ Water Separation MRI θ =2 π ( Δ f)(TE) {W,F} z W F θ {W,F} y in ‐ phase (IP) = W + F oppose ‐ phase (OP) = W ‐ F … … … t = 0 - t = 0 + x Fat is precessing slower, accrues phase periodically. … … … in ‐ phase (IP) = W + F 9

“Dixon” Method (2 ‐ point) Fat ‐ Water MRI Opposed ‐ Phase (OP) = W ‐ F In ‐ Phase (IP) = W+F ( ) W + F = IP − OP F OP = |W ‐ F| = |F ‐ W| η = 2 IP 50% ambiguity 50% 34% fat ‐ signal fraction ! e.g. 34% of the liver signal originates from fat 0% 100% Courtesy of Dr. S. Reeder, U. Wisconsin Madison. “Dixon” Method Evolution 1990 ‐ 1991 2003 – present 3 and 4 ‐ pt. methods 6 ‐ pt. methods towards robustness and fat ‐ signal fraction 1984 Mid ‐ 1990s, early ‐ 2000 accuracy 2 ‐ pt. methods Extensive clinical use Subtle variations MULTLI ‐ peak fat model SINGLE (CH 2 ) fat peak model COMPLEX ‐ base MAGNITUDE ‐ based Extensive use in liver • General Electric – Lava ‐ Flex (2 pt.) – commercial – IDEAL ‐ Quant, IDEAL ‐ IQ (3 ‐ 6 pt.) – commercial + research • Philips – mDixon (2 pt.) – commercial, (3 ‐ 6 pt.) – commercial + research • Siemens – Dixon (2 ‐ 3 pt.) – research 10

“N OT S O S IMPLE ” S PECTROSCOPIC I MAGING π Δ = + ⋅ 2 ( )( ) [ ] i TE f S W F e n TE n Radiology 1984; 153:189 ‐ 194 (2 echo water ‐ fat MRI) improving quantitative accuracy complex complex real unknown unknown unknown water fat ∑ * π − = + π ψ 2 2 / ( ) ( ) measured i t f i t t T s t W F a e e e 2 k MRI signal k fat chemical shift term Σ multiple fat peaks B 0 fieldmap non ‐ uniformity term ‐ CH 2 ‐ signal relaxation ‐ HC=CH ‐ ‐ CH 3 Generalized “Dixon” Method Fat Water iterative algorithm . . . TE1 TE2 TE3 TE4,5,6 Water IP = W+F B O Inhomogeneity Fat Fraction (%) T2* map (ms) Field Map 0 ‐ 100% (Hz) OP = W ‐ F Fat QUANTITATIVE !!! 11

L IVER S TEATOSIS ( A LL H ISPANIC T EENAGERS, 8 ‐ 15 Y EARS) least ‐ squares cost X Fieldmap (Hz) Image courtesy Samir D. Sharma 12