mapping of oxygen by imaging lipids relaxation enhancement

Mapping of Oxygen by Imaging Lipids RelaxationEnhancement: A Potential Sensitive Endogenous MRIContrast to Map Variations in Tissue Oxygenation

B�en�edicte F. Jordan,1 Julie Magat,1 Florence Colliez,1 Elif Ozel,1

Anne-Catherine Fruytier,1 Val�erie Marchand,1 Lionel Mignion,1 Caroline Bouzin,2

Patrice D. Cani,3 Caroline Vandeputte,4 Olivier Feron,2 Nathalie Delzenne,3

Uwe Himmelreich,4 Vincent Denolin,5 Thierry Duprez,6 and Bernard Gallez1*

Purpose: Because of its paramagnetic properties, oxygen mayact as an endogenous magnetic resonance imaging contrastagent by changing proton relaxation rates. Changes in tissue

oxygen concentrations have been shown to produce changesin relaxation rate R1 of water. The aim of the study was to

improve the sensitivity of oxygen enhanced R1 imaging byexploiting the higher solubility of oxygen in lipids (as comparedwith water) to sensitively monitor changes in tissue oxygen lev-

els by selectively measuring the R1 of lipids.

Methods: The method, with the acronym ‘‘MOBILE’’ (mappingof oxygen by imaging lipids relaxation enhancement), was

applied in different mouse models of hypoxic processes on a11.7 T magnetic resonance imaging system. MOBILE was

compared with R*2, R1 of water, and with pO2 measurements(using electron paramagnetic resonance oximetry). MOBILEwas also applied in the brain of healthy human volunteers

exposed to an oxygen breathing challenge on a 3 T magneticresonance imaging system.

Results: MOBILE was shown to be able to monitor changes in

oxygenation in tumor, peripheral, liver, and brain tissues. Theclinical translation was demonstrated in human volunteers.

Conclusion: MOBILE arises as a promising noninvasive andsensitive tool for diagnosis and therapeutic guidance in disor-ders involving hypoxia. Magn Reson Med 70:732–744, 2013.VC 2012 Wiley Periodicals, Inc.

Key words: oxygen mapping; tissue hypoxia; MRI; R1;endogenous contrast

There is a critical need for developing dynamic, nonin-vasive methods for oxygen mapping in clinical practice(1). Although hypoxia has been long recognized as a cru-cial factor in many disorders and in treatment, response/failure, atraumatic, ready-to-use, and reliable methods toquantify tissue oxygenation are still lacking in day-to-day clinical practice. Direct quantitative methods,including Eppendorf microelectrodes (2), electron para-magnetic resonance (EPR) oximetry (3,4), 19F relaxometry(5), or Overhauser enhanced magnetic resonance imaging(MRI) (6), are either invasive or require the injection of areporter probe, and are currently not clinically applica-ble. Nowadays, the clinically available armamentariumfor this purpose only includes radiolabeled nitroimidaz-oles detected by positron emission tomography (PET)which prominently accumulate in hypoxic areas, therebyacting as potential markers for tissue hypoxia (7,8). En-dogenous sources of contrast in MRI include variationsof T1 (longitudinal relaxation rate) and T*2 (effectivetransversal relaxation rate) values in tissue (9). T1 disclo-ses sensitivity to dissolved oxygen, which acts as a T1-shortening paramagnetic contrast agent (10). T*2 is sensi-tive to the relative deoxyhemoglobin/oxyhemoglobin(Hb/HbO2) ratio in vessels (11) (Fig. 1a). T*2 mapping,also referred to as functional MR imaging or blood oxy-gen level dependent (BOLD) imaging, is sensitive to var-iations in oxygenation in the vascular compartment, andhas been successfully applied to monitor changes in tis-sue oxygenation (12–14). However, BOLD-MRI has alsodemonstrated significant limitations in terms of quantita-tive relationships between response signal intensity andtrue changes in tissue pO2 and is sensitive to changes intotal hemoglobin content (15–17). Recently, changes intissue oxygen concentrations have been shown to pro-duce changes in relaxation rate R1 (¼1/T1) of water (18–20). Studies have demonstrated that oxygen-enhancedMRI produces measurable signal intensity changes innormal tissues using standard clinical MR systems. Thus

1Biomedical Magnetic Resonance Research Group, Louvain Drug ResearchInstitute, Universit�e Catholique de Louvain, Brussels, Belgium.2Angiogenesis and Cancer Research Laboratory, Pole of Pharmacology andTherapeutics, Institute of Experimental and Clinical Research, Universit�eCatholique de Louvain, Brussels, Belgium.3Metabolism and Nutrition Research Group, Louvain Drug Research Institute,Universit�e Catholique de Louvain, Brussels, Belgium.4Biomedical MRI/Molecular Small Animal Imaging Center, KatholiekeUniversiteit Leuven, Leuven, Belgium.5Philips Health Care, Benelux, Belgium.6Radiology and Medical Imaging, St. Luc hospital, Institute of Neuroscience(IoNS), Universit�e Catholique de Louvain, Brussels, Belgium.

Disclosure: B.F.J., J.M., and B.G. declare a possible competing financialinterest: they filed Patent applications 11163573.6-1265 and PCT/EP2012/057237.

Grant sponsor: Actions de Recherches Concert�ees-Communaut�e Francaisede Belgique; Grant number: ARC 09/14-020; Grant sponsor: Poled’Attraction Interuniversitaire; Grant number: PAI VI (P6/38); Grantsponsors: Belgian National Fund for Scientific Research (FNRS), JosephMaisin Foundation, Saint-Luc Foundation, 2011 ISMRM Seed Grant, EUFP7 Large Scale Integrating Project INMiND.

*Correspondence to: Bernard Gallez, Ph.D., Biomedical MagneticResonance Research Group, Louvain Drug Research Institute, Universit�eCatholique de Louvain, Av. Mounier, 73.08, B-1200 Brussels, Belgium.E-mail: [email protected]

Received 23 May 2012; revised 30 August 2012; accepted 31 August 2012.

DOI 10.1002/mrm.24511Published online 28 September 2012 in Wiley Online Library(wileyonlinelibrary.com).

Magnetic Resonance in Medicine 70:732–744 (2013)

VC 2012 Wiley Periodicals, Inc. 732

far, oxygen-induced increase in R1 has the potential toprovide noninvasive measurements of tissue oxygenlevel variations, in a different way than BOLD imaging.Unfortunately, the technique still suffers from insuffi-cient sensitivity and the measured DR1 may be biased byconfounds unrelated to changes in tissue oxygen level,

such as alteration in blood flow, and tissue H2O content(19).

We hereby propose a method based on higher solubilityof oxygen in lipids than in water (21). By monitoringchanges in R1 relaxation rate of the lipid peak rather thanthose of water, estimates of variations in tissue oxygen-ation are obtained. We have developed an MR method formapping variations in oxygenation based on changes inrelaxation properties of the tissue lipids, which we calledmapping of oxygen by imaging lipids relaxation enhance-ment (MOBILE). Since equilibration with 760 mmHg ofO2 was demonstrated to have a significantly higher effecton lipid protons than on water protons (21), we hypothe-sized that the R1 of lipids should be more sensitive thanthe R1 of water to variations in tissue oxygenation (Fig.1b). To assess the R1 value of lipids as a sensitive univer-sal endogenous marker of tissue oxygenation, we recordedand analyzed changes in this parameter in hypoxic mousemodels mimicking physiopathological conditions such astumor hypoxia, peripheral ischemia, cerebral ischemicstroke, and liver steatosis, before and during a hyperoxicbreathing challenge.

The choice of tumor models to assess the sensitivity ofMOBILE to variations in oxygenation was done becausetumor hypoxia is acknowledged as a major factor ofresistance of solid tumors to treatment, in particular toexternal irradiation (22,23). Quantitative follow up ofchanges in tumor oxygenation can find relevant applica-tions in radiation therapy planning (24–26) as well aswith regard to antiangiogenic and antivascular treat-ments optimization (27,28).

Peripheral vascular disease is responsible for theincreased risk of lower limb amputation observed in dia-betic patients and/or in patients presenting systemic ath-erosclerosis (29–31). A significantly better outcome isexpected for these patients, based on improved technicaloptions in peripheral revascularization (31,32). For thatpurpose, techniques able to assess quantitative peripheralpO2 are mandatory to verify the efficacy of therapeuticapproaches aimed at salvaging limbs in diabetic patientswith critical limb ischemia (CLI). In this context, wehypothesized that the MOBILE sequence could be asensitive tool for assessing peripheral tissue oxygenation.

Mapping brain tissue oxygenation shortly after acuteischemic stroke may be crucial in selecting patients whomay still benefit from thrombolytic treatment beyondconventional time-based guidelines (33,34). In this con-text, it seems mandatory to explore potential applica-tions and limitations of oxygenation/perfusion mappingtechniques in patient triage and patient eligibility forthrombolytic therapy in acute ischemic stroke (35). It istherefore useful to establish the potential relevance ofMOBILE imaging in a stroke model as a complementarytool to the classic imaging protocol currently used toimage acute or chronic strokes.

The assessment of liver oxygenation could be particu-larly important in the early monitoring of the liver aftertransplantation. Measurements of the global T1 beforeand after oxygen inhalation have been also reportedrecently as a quantitative imaging biomarker for the stag-ing of liver cirrhosis (36). MOBILE was used to assesschanges in liver tissue oxygenation, presenting abnormal

FIG. 1. MR oxygen-dependent endogen contrasts: principles of

T*2 and T1 contrasts. a: T*2 (functional imaging): the contrast isbased on the relative change in deoxyhemoglobin (Hb), which is

paramagnetic; vs. oxyhemoglobin (HbO2), which is diamagnetic.Hb is able to decrease the effective transversal relaxation time T*2

of protons that will vary under different blood oxygenation levels.

Usually, R*2 (¼1/T*2) is presented and is reported to be decreasedwith an increase in oxygenation. T1 (Oxygen enhanced longitudinal

relaxation): the contrast is based on the paramagnetic propertiesof oxygen that induces a decrease in the longitudinal relaxationtime T1 of tissue protons, which therefore varies under different

tissue oxygenation levels. R1 (¼1/T1) is reported to be elevatedwith an increase in tissue oxygenation. b: MOBILE: the contrast is

based on the change in the T1 of lipid protons instead of waterprotons. Indeed, as oxygen is more soluble in lipids than in water,the concentration of oxygen will be higher in a lipid phase than in

a water phase for the same atmospheric pressure. In a tissue(which is composed of water and lipids), by sampling the T1 of

lipid protons only and removing the signal coming from the waterprotons, a higher sensitivity to variations in tissue oxygenation canbe achieved.

MRI Contrast to Map Variations in Tissue Oxygenation 733

accumulation of lipids (steatosis being frequent in partic-ular in obese and overweight patients) in response to acarbogen breathing challenge.

Finally, to assess the potential for translating MOBILEinto clinical settings, we also report initial human dataacquired from healthy volunteers breathing alternativelyair and oxygen in a clinical 3 T MRI system.

METHODS

In Vitro Samples

Calibrations (R1 as a function of percent oxygen) of thewater and lipid peaks were performed by measuring R1

at 37�C in sealed tubes containing water, sunflower oil,or tumor homogenates (excised MDA tumors homoge-nized with NaCl 0.9% solution, 1 mL/g) bubbled withnitrogen (0% O2), air (21% O2), or carbogen (95% O2) for20 min in a 37�C water bath. The temperature was con-trolled with OxyLiteTM (Oxford Optronix, Oxford, UK)probe and kept constant using the warm water circulat-ing system. The pH was monitored using a pH micro-electrode (Microelectrodes, Inc.).

Animal Models

Animals were anesthetized by inhalation of isoflurane(Forene, Abbot, England) mixed with 21% oxygen (air)in a continuous flow (1.5 L/h). This anesthesia has beenshown not to interfere with tissue hemodynamic (37).Respiration rate and body temperature (37.0�C 6 1.0�C)were monitored and maintained with a circulating waterblanket. Studies were undertaken in accordance with thenational and local regulations of the ethical committee(agreement number LA 1230467). In vivo, four hypoxicmouse models mimicking physiopathological conditionssuch as tumor hypoxia, peripheral ischemia, and cere-bral ischemic stroke, were used in this study.

Tumor Models

A total of 5 � 106 NT2 cells (provided by E. Jaffee) or 3 �106 MDA-MB-231 (LGC Promochem), amplified in vitro,were collected by trypsinization, washed three times withHanks balanced salt solution and resuspended in 100-mLHanks balanced salt solution. These mammary tumorcells were injected subcutaneously into the right uppermammary fat pad of 6-week-old FVB/N or nude NMRIfemale mice (Janvier), respectively (n ¼ 5 per model).Tumors were analyzed when reaching 6 mm in diameter.

Peripheral Ischemia Model

Femoral vein ligation was performed on C3H femalemice 2 h before imaging. The femoral artery and veinwere exposed and ligated with three sutures, one aboveand two after the origin of the natural collateral on theleft hind leg of the mouse. Sutures were secured withsurgical glue (Histoacryl; 3M). Three mice were studied(three ligated legs and three contralateral normal legs).

Stroke Model

Cortical photothrombosis was induced in six C57BL6/Jmice (Jackson Laboratories; Bar Harbor, ME) 24 h before

imaging according to (38). A vertical incision was madebetween the right orbit and the external auditory canal.The upper part of the temporalis muscle was cauterizedso that the muscle could be displaced. Photoillumina-tion with green light (wave length, 540 nm; bandwidth,80 nm) was achieved using a Xenon lamp (modelL-4887; Hamamatsu Photonics, Hamamatsu City, Japan)with heat-absorbing and green filters. The irradiation atintensity of 0.68 W/cm2 was directed with a 3-mm opticfiber, which was placed on the exposed skin above thesensory motor cortex. Photoillumination was performedfor 5 min after intravenous injection of the photosensi-tizer Rose Bengal (20 mg/kg, Sigma-Aldrich) in a tailvein.

Liver Model

Five ob/ob (B6.V-Lepob/J) leptin deficient C57Bl6 mice(Charles River, France) were used for the liver study.

Lipids Content

Total lipids were measured by gravimetry using an adap-tation of Folch method (39). Lipids were extracted frombrain and liver of healthy mice (n ¼ 3), and from subcu-taneous NT2 tumors (n ¼ 3) in CHCl3-MeOH (2:1) as pre-viously described (40).

Preclinical Experiments

Three MR measurements of each type (R1 H2O, R1 lipids,R*2) were acquired sequentially and repeated three timesduring air breathing. Then, breathing gas was switchedto carbogen, and MR measurements were repeated at 10,25, and 40 min after the gas switch. Respiratorytriggering was employed to acquire images during theexpiration cycle to avoid motion artifacts. In tumorexperiments, we used the OxyLiteTM fiber-optic microp-robes for continuously monitoring tumor oxygenationsimultaneously with MRI (41). Independent experimentswere also performed to compare MOBILE with EPROximetry using the similar breathing challenge.

In vitro and animal experiments were performed withan 11.7 T (Bruker, Biospec), and with a quadrature vol-ume coil (inner diameter of 40 mm).

In Vitro MR Data

The relaxation properties of two common lipid peaks(1.3 and 4.0 ppm) were exploited. The lipid peak with achemical shift of �1.3 ppm was ascribed to methyleneprotons from fatty acid chains and the peak with a chem-ical shift of �4.0 ppm was ascribed to protons from theglycerol backbone of triglycerides (42–44).

T1 Measurements

A segmented inversion-recovery fast imaging with steadystate precession (IR FISP; 45) sequence (SSFP FID orSteady State Free Precession Free Induction Decaymode) was used to acquire parametric images of T1

relaxation time (Fig. 2). The acquisition parameters wererepetition time (TR)/TE/FA/BW/matrix ¼ 4 ms/1.2 ms/5�/100 kHz/64 � 64, four segments, and a total

734 Jordan et al.

acquisition time of 1 min 20 s. For the total protonexperiment (essentially reflecting the water peak), tosample the recovery of the signal, a series of 100 imageswere acquired (spaced by a scan TR ¼ 120 ms to obtainimages at different inversion times) with a slice thick-ness of 1 mm. For the lipid experiments, we first eval-uated the difference in Hertz between water and lipidpeaks in the 1H spectrum with a single pulse sequence.These offsets were then used as an imaging frequencyoffset in the same IR FISP protocol. Typically, the lipidpeak of interest was �4.0 ppm for tumor and brain tis-sues, and �1.3 ppm for peripheral and liver tissue, asidentified experimentally. We added a p/2 hermite satu-ration pulse with a bandwidth factor of 5400 Hz ms tospoil the water signal. A series of 40 images (spaced byscan TR ¼ 100 ms) with a slice thickness of 3 mm wereacquired. Then, images were fitted using a home-madeprogram written in Matlab (The MathWorks, Inc., Natick,MA) to determine the T1 relaxation in regions of interest(ROIs). For that purpose, a nonlinear fit was used todetermine the T1 relaxation in each pixel of the ROI.Raw data were filtered so as to disregard nonrelevant fitswith a T1 error/T1 > 30%. T1 values were then calcu-lated from the effective T1* values measured experimen-tally (46–48): Mz(ti) ¼ A � Bexp(�TI/C), with A ¼M0(T1*/T1), B ¼ M0(1 þ T1*/T1), and C ¼ T1*, and thenT1 ¼ (B/A � 1). T1*. Finally, R1 (1/T1) data were filteredaccording to the calibration curves obtained from tissueextracts (R1 0% O2 < measured R1 < R1 100% O2).

T*2 Measurements

For T*2 measurements, a multigradient echo sequencewas performed with eight echoes (between 3.5 and 31.5ms) with a total acquisition time of 4 min 48 s. A 256 �

256 matrix was used with TR/flip angle/slice thickness¼ 1500 ms/30�/1 mm.

EPR Oximetry

In vivo tumor pO2 was monitored using EPR oximetryusing charcoal as the oxygen-sensitive probe (49,50).EPR spectra were recorded using a 1.1 GHz EPR spec-trometer (Magnettech, Berlin, Germany). Calibrationscurves were made by measuring the EPR line width as afunction of the pO2. Mice were injected in the center ofthe tumor using the suspension of charcoal (100 mg/mL,60 mL injected). The tumor under study was placed inthe center of the extended loop resonator whose sensi-tive volume extends 1 cm into the tumor mass. The pO2

measurements correspond to an average of pO2 values inthis volume of the NT2 tumor models. Charcoal wasinjected 24 h before experiments. MOBILE and EPRmeasures were performed the same day, and it was veri-fied that charcoal did not perturb MOBILE or T1 MRImeasurements.

Clinical Experiments

MR imaging was performed at 3.0 T (Achieva; PhilipsMedical System) using a transmit/receive head coil.Three different imaging sequences were used duringbaseline acquisition (air breathing) and 5 min after initia-tion of oxygen (15 L/min). The oxygen was delivered bya mask placed by an elastic band around the subject’shead.

T1 Measurement

A look locker sequence (T1 TFE, T1 turbo field echosequence) was applied for 2 min with TR/TE/NA/flip

FIG. 2. MOBILE sequence diagram.Sequence used to measure selectively

the T1 of the lipid component: IR FISPsequence (SSFP FID mode). Frequencywas centered on the lipid frequency

(shift evaluated on spectra using a sin-gle pulse spectrum). Saturation pulse

was added to spoil the water signal. a:Global measurements. b: Lipids meas-urements. c: Evolution of signal inten-

sity (a.u.) in function of time (ms), fromwhich T1 is calculated using the men-tioned equation.


angle/TFE ¼ 3.467 ms/1.45 ms/1/5�/10. A total of 140images with 20 mm of thickness and 80 � 80 pixelswere obtained. The lipid peak of interest in the brain ofhuman volunteers was �1.3 ppm, as identifiedexperimentally.

MOBILE Measurements

The same sequence was used and a 90� SPIR pulse wasadded (spectral saturation by Inversion recovery) to spoilwater with a BW of 300 Hz centered on the water peak.The acquisition lasted 4 min and 117 images with 20mm of thickness and 80 � 80 pixels were obtained.

T*2 Measurements

A multiecho fast field echo sequence was performed with15 echoes with a total acquisition time of 41.8 s. A 320 �320 pixels matrix was obtained with TR/flip angle/slicethickness¼ 250 ms/18�/4 mm.

Statistical Analysis

Paired t-tests were used to compare mean changesbetween groups (carbogen or oxygen vs. air breathing; is-chemic vs. control, R1 H2O vs. R1 lipids during carbogenchallenge) for each parameter. Histogram distributions,linear fits, and Pearson correlation (with P values < 0.05(*), <0.01 (**), or <0.001 (***) were considered signifi-cant) were performed using the Graphpad software.

RESULTS

In Vitro Validation of MOBILE

To establish the sensitivity of R1 of lipids to variationsin oxygenation, we assessed in vitro the relaxation prop-erties of water and lipid components in pure aqueousand oil phases, in mixed systems, and in tissue homoge-nates, equilibrated in different oxygen environments(0%, 21%, and 95% O2). The higher solubility of oxygenled to a higher sensitivity when considering theevolution of R1 of lipids as a function of oxygenation,compared with the R1 of water, resulting in a higher sen-sitivity of R1 to oxygen measured in a lipophilic phasethan in water (Fig. 3a–c). Typically, the gain in sensitiv-ity (compared with water) corresponds to a factor of 7.5in pure oil, 2.1 and 11.4 in tissue homogenates corre-sponding to 2 distinct common lipid peaks with a chem-ical shift of �1.3 and �4.0 ppm, respectively. The peakswere identified by comparing the liver and tumor tissuehomogenate’s spectra with published chemical shifts oflipids (42–44), as well as with MR spectra of in vitrophantoms consisting of water/triglyceride (tripalmytoyl-glycerol) mixture. Inversion Recovery curves used to esti-mate the R1 parameter show the ability of the MOBILEsequence to suppress the water signal and selectivelymeasure the lipid component of the signal on in vitrophantoms composed of pure water and oil (Fig. 3d–g).Of note, the calibration curves were not significantlymodified by changes in temperature or pH within thephysiological range, namely, between 33 and 37�C andbetween 6.1 and 7.4 as pH values.

Sensitivity of MOBILE to Variations in Tumor Oxygenation

We used two mammary tumor models, syngeneic NT2tumors and MDA human xenografts implanted orthotopi-cally in mice. These models were selected for their largerange of hypoxic fractions. We performed MR measure-ments before and during an hyperoxic breathingchallenge protocol with carbogen. This gas induces sig-nificant transient increases in oxygenation in experimen-tal and in some human tumors (12,13,51). The increasein oxygenation was demonstrated using a direct invasivetechnique simultaneously in the same tumors. Indeed,the measurements were simultaneously correlated with :(i) a quantitative local measurement of tumor pO2 usinginvasive MR compatible fiber optic fluorescence quench-ing probes (OxyLiteTM) (Fig. 4a), and (ii) two currentlyused MR parameters reflecting changes in tissue oxygen-ation, namely, R*2 and R1 of water. Fluorescencequenching measurements are direct and quantitative, en-abling longitudinal follow-up studies performed simulta-neously to MR measurements (52). The basal pO2 inMDA tumors ranged from 2 to 25 mmHg (mean of pO2 ¼14.15 6 5.46 mmHg) whereas basal pO2 in NT2 tumorswas 37 6 0.8 mmHg. Tumor oxygenation was increasedin the majority of the tumors (7 of 8) under carbogenbreathing conditions. A typical OxyLite graph shows thedynamic and the range of response for a typical tumor(Fig. 4b). After carbogen breathing, MDA and NT2tumors reached a pO2 around 41.2 6 2.3 and 87.7 6 14.1mmHg, respectively. Mean relative changes in relaxationrates with respect to the hyperoxic challenge were com-pared (pooled results for both tumor models) between R1

H2O, R1 lipids, and R*2, with relative changes of 2.1 6

1.0% (P < 0.05; DR1 H2O), 10.1 6 4.8 % (P < 0.05; DR1

lipids), and 9.8 6 4.6 % (P < 0.05; DR*2) (Fig. 4c).Although values are rather dispersed, an increase by afactor of 4.8 was achieved for R1 lipids compared withR1 H2O. Typical maps of the relative difference in relaxa-tion rates before and during the carbogen challenge weregenerated to compare the sensitivity of the R1 of lipids(MOBILE), R1 of H2O, and R*2 methods (Fig. 4d–f), indi-cating that DR1 lipids are more sensitive than DR1 H2Oon the same tumor. The corresponding histograms weregenerated to visualize and quantify the shift in the distri-butions for each parameter (Fig. 4g–i), showing a largershift in the median value for DR1 lipids compared withDR1 H2O. In a parallel study, we compared EPR oximetryand R1 lipids and R1 H2O in the NT2 tumor model. Thebasal pO2 was 8.4 6 1.1 mmHg and reached 26.1 6 6.1mmHg after carbogen breathing. The correlation betweenR1 H2O and pO2 was not significant (P ¼ 0.8611, positivelinear fit 0.000409 6 0.002241; r2 ¼ 0.005529) (Fig. 5a),while a positive linear significant correlation was foundbetween R1 lipids and pO2 (0.01744 6 0.00656; r2 ¼0.5407, P ¼ 0.0376) (Fig. 5b).

Application of MOBILE in a Peripheral Ischemia Model

In the limb ischemia model, we tested the sensitivity ofMOBILE to (i) basal peripheral tissue oxygenation bycomparing R1 of lipids in tissues downstream the arteryligation and in the control leg, and (ii) relative changes

736 Jordan et al.

during a carbogen breathing challenge. We hypothesizedthat the ischemic tissues may have less capacity torespond to the hyperoxic challenge than the control leg.In the management of CLI, it was suggested that a 10min 100% oxygen challenge could be performed todetermine the outcome of wound cicatrization, based onthe absence or presence of response to the hyperoxicchallenge (53). Some groups also focused in the past onhyperbaric oxygen therapy to improve the outcome ofnonhealing wounds (54). Here, we show that R1 lipids ismore sensitive than R1 H2O and R*2 to evidence (i) dif-ferences in basal tissue oxygenation (control vs. ischemicleg) (Fig. 6a) and (ii) relative changes in response to carb-ogen breathing (Fig. 6b,c). Significant changes in meanvalues were observed in the control leg in response tocarbogen breathing for DR1 lipids (7.4 6 2.0%, P < 0.05),whereas no significant changes were observed for DR1

H2O (1.3 6 0.8%) or DR*2 (9.6 6 3.8%, P > 0.05) (Fig.6b). A sensitivity increase by a factor of 5.7 was achievedfor DR1 lipids compared with DR1 H2O. As expected, alack of response to carbogen breathing was observed inthe ischemic leg (Fig. 6c). Typical maps of the relativedifference in relaxation rates before and during the carb-ogen challenge were generated to compare the sensitivityof the R1 lipids (MOBILE), R1 H2O, and R*2 methods(Fig. 6d). This confirms highest sensitivity for changes inR1 lipids compared with R1 H2O and inhomogeneouschanges in R*2.

Application of MOBILE in a Stroke Model

We used the photothrombotic stroke model with unilat-eral lesion in mice to test the sensitivity of MOBILE to:(i) basal brain tissue oxygenation by comparing the intact

FIG. 3. In vitro validation of MOBILE. a: Relation between changes in R1 (¼1/T1) vs. changes in pO2 in vitro in water and in pure oil.

Note the higher sensitivity achieved in the oil system. b: Relation between changes in R1 (¼1/T1) (6SEM) vs. changes in pO2 in tumorextracts in the aqueous phase and in the lipid phase (lipids with a chemical shift of � 1.2 ppm). Note the gain in sensitivity of a factorof 2.1 in the lipid phase vs. the water phase. c: Relation between changes in R1 (¼1/T1) (6SEM) vs. changes in pO2 in tumor extracts in

the aqueous phase and in the lipid phase (lipids with a chemical shift of � 4.0 ppm). Note the gain in sensitivity of a factor of 11.4 inthe lipid phase vs. the water phase. d, f: Series of IR-FISP images (transversal slice) of an oil (left) and a water (right) tube used to sam-ple the global T1 (essentially reflecting the T1 of water). The signal intensity in each tube at each inversion time is reported on the T1

curve (f). Note the faster inversion (signal annihilation) for oil than for water. e, g: Series of ‘‘MOBILE’’ images (IR FISP with water sup-pression and with adapted imaging frequency on the lipid peaks) of an oil (left) and a water (right) tube used to sample the lipid T1.

Note that the water tube is completely suppressed on the images so that the estimated T1 is mainly influenced by lipid protons (g).


and the insulted hemispheres, and (ii) response to carb-ogen breathing in the intact (control area) and insulted(stroke area) hemispheres. We hypothesized that thestroke infarct would not be able to respond to a carbo-gen challenge as well as the control contralateral area.We demonstrated that MOBILE was able to identify dif-ferences in basal brain tissue oxygenation (Fig. 7a) butwas also highly sensitive to normal brain tissue hyper-oxygenation during the carbogen challenge (Fig. 7b), incontrast to the insulted area (Fig. 7c). The relativechange in R1 lipids was 4.5 times higher than the rela-tive change in R1 H2O in response to carbogen breath-ing in the control area (Fig. 7b). The stroke area did not

respond to carbogen breathing using MOBILE or BOLDMRI but showed a significant increase using R1 H2O(Fig. 7c).

Application of MOBILE in a Liver Steatosis Model

Liver of mice were imaged before and during a carbogenbreathing challenge (n ¼ 10). The sensitivities to changesin liver oxygenation were compared in terms of R1 H20and R1 lipids. We observed a significative higher sensi-tivity of R1 lipids compared with the R1 H2O during thecarbogen breathing challenge (P ¼ 0.023) that resulted inincreased tissue oxygenation (Fig. 8).

FIG. 4. Sensitivity of MOBILE to variations in tumor oxygenation. a: Anatomical transversal MR image of a mouse with an MDA tumor

(shown by the arrow) with the OxyLite tip at the center of the tumor (black spot). b: Typical monitoring of tumor pO2 under air and carb-ogen breathing conditions assessed via the OxyLite MR compatible fiber optic probes. Note the dramatic increase in tumor pO2 imme-diately at the time of switch between air and carbogen. c: Pooled results of MDA (empty symbols) and NT2 (filled symbols) tumors (n ¼12): Relative changes over the whole tumors in relaxation times under air and carbogen breathing conditions for global R1 (water), R1 oflipids, and R*2 [R ¼ 1/T]. Note the higher change in R1 of lipids vs. global R1 in response to carbogen breathing. d–f: Typical maps of

changes in relaxation times in response to carbogen breathing (DR ¼ Rcarbogen � Rair). Color scales highlight the higher sensitivity of theMOBILE technique with a higher proportion of ‘‘hot’’ colors (h). g–i: Corresponding histograms (to g–i) of global DR1 (j), DR1 lipids (k),and DR*2 (l). Note the shift of the median of DR1 lipids to the right whereas histograms (j) and (k) do not show any significant shift in the

medians.

738 Jordan et al.

FIG. 5. Comparison of MOBILE with EPR to assess tumor oxygenation. a, b: Correlation between global pO2 measurements and global

R1 values (a) or lipids R1 values (b).

FIG. 6. Application of MOBILE in a peripheral ischemia model. a: Basal relaxation times (global R1 and R1 lipids 6 SEM) in control legsand in ischemic legs. Note the higher sensitivity of R1 lipids (MOBILE) compared with global R1 while comparing control and ischemic sta-

tus. b: Pooled results of control legs (n ¼ 3): relative changes in relaxation times under air and carbogen breathing conditions for global R1

(water), R1 of lipids, and R*2. Note the higher change in R1 of lipids vs. global R1 in response to carbogen breathing. c: Pooled results of is-chemic legs (n ¼ 3): relative changes in relaxation times under air and carbogen breathing conditions for global R1 (water), R1 of lipids, and

R*2. Note the lack of significant change in all parameters after femoral artery ligation. d: Anatomical images of a typical control leg (up) andischemic leg (bottom) and their corresponding maps of differences in relaxation times (global R1, R1 of lipids, and R*2).


Lipids Gravimetry Results

The total lipid content ranges from 60.6 6 0.2 mg/mgliver to 93.6 6 11.8 mg/mg brain. The mean value estab-lished in the NT2 tumors is 106.1 6 20.4 mg/mg tumor.

Application of MOBILE in Clinical Settings

The potential for clinical translation of the MOBILEsequence was assessed by: (i) implementing the sequence

on a clinical 3 T MRI system; (ii) verifying appropriatewater suppression and T1 sampling of lipids on phan-toms, and (iii): assessing R1 H2O, R1 lipids, and R*2 inbrains of healthy volunteers under air and oxygenbreathing conditions. A similar axial-transverse slicelocation in the brain of three volunteers was imaged dur-ing air breathing and after 5 min of 100% oxygen breath-ing. Typical results are shown in Figure 9 wherematched transversal brain maps of R1 H20, R1 lipids, and

FIG. 7. Application of MOBILE in a stroke model. a: Basal relaxation times 6 SEM (global R1 and R1 lipids) in control (n ¼ 10) and insulted(stroke) (n ¼ 6) hemispheres. Only regions with more than 10 pixels were included in the data (nonrelevant fit with a T1 error/T1 > 30% were

discarded). b: Pooled results of control hemispheres: relative changes in relaxation times under air and carbogen breathing conditions forglobal R1 (water), R1 of lipids, and R*2. Note the higher change in R1 of lipids vs. global R1 in response to carbogen breathing. c: Pooled results

of ‘‘stroke’’ hemispheres: relative changes in relaxation times under air and carbogen breathing conditions for global R1 (water), R1 of lipids,and R*2. d: Typical anatomical image (transversal slice) of a mouse brain with the arrow pointing to the stroke area and the line delimitatingthe ROI surrounding the stroke area. e: Corresponding typical global R1 map of the brain under air breathing conditions. f: Corresponding typ-

ical ‘‘R1 lipids’’ map of the brain under air breathing conditions. g: Corresponding typical R*2 map of the brain under air breathing conditions.

FIG. 8. Application of MOBILE in a liver steatosis model. a: Mean relative changes over the entire liver in relaxation times under air and

carbogen breathing conditions for global R1 (water), R1 of lipids, and R*2 [R ¼ 1/T] (n ¼ 10 per group). Note the higher change in R1 oflipids vs. global R1 in response to carbogen breathing. b: Typical R1 lipids map in response to carbogen breathing.

740 Jordan et al.

R*2 along with the corresponding anatomical image areshown (Fig. 9a–d). The mean relative changes are shownin three arbitrary distinct ROIs in the left and right hemi-spheres, and in the center (Fig. 9e–g). A high heterogene-ity of response is observed for all three parameters, as

indicated by the maps and histograms of ROI 2 and ofthe right hemisphere (Fig. 9i–k,m–o). Nevertheless, wecan observe more ‘‘positive responding regions’’ in theDR1 lipids map than in the DR1 H2O map. This is alsoillustrated in the histograms, where the median of the

FIG. 9. Application of MOBILE in the clinical setting on healthy volunteers. a,h,l: Anatomical image of a typical volunteer brain. Threearbitrary ROIs have been defined. b–d: Typical map of changes in relaxation times in response to 5 min 100% oxygen breathing (DR ¼Rcarbogen – Rair) for global DR1 (b), DR1 lipids (c), and DR*2 (d). e–g: Mean relative changes in relaxation rates (global R1 [water], R1 of lip-

ids, and R*2) in ROI 1 (e), ROI 2 (f), and ROI 3 (g). Note the systematic higher sensitivity of R1 of lipids vs. global R1. i–k: Correspondinghistograms (of relative changes in relaxation rates) for ROI 2. Note the clear shift to the right of the median of DR1 lipids whereas the

change in DR1 of water is slightly negative (not responding to oxygen breathing). The change in DR*2 is slightly negative, as expected(responding to oxygen breathing). m–o: Histograms of relative changes in relaxation rates for the right hemisphere. Note the clear shiftto the right of the median of DR1 lipids whereas the change in DR1 of water is slightly negative (not responding to oxygen breathing).

The change in DR*2 is slightly positive (not responding to oxygen breathing).


DR1 lipids is shifted to the right (vs. the ‘‘zero value"bin). We showed that both MOBILE (supposed torespond positively to an increase in tissue oxygenation)and BOLD (R*2) imaging (supposed to respond nega-tively to an increase in blood oxygenation) are moresensitive to the changes in oxygenation than R1 H2O.Similar data were obtained on two additional volunteers.Our first clinical results suggest that MOBILE usingchanges of R1 lipids and BOLD are complementary innature with a higher sensitivity than R1 H2O.

DISCUSSION

Our goal was to implement and validate a new MR tech-nique using an endogenous contrast to sensitively detectand map variations in tissue oxygenation. Starting fromthe previously described phenomenon of oxygen-enhanced relaxation of water protons (T1), we developeda new measurement method based on the relaxation oflipid protons with the knowledge that oxygen is sixtimes more soluble in lipophilic media than in water.Using the MOBILE sequence, we achieved (i) the devel-opment of an MR pulse sequence allowing longitudinalsampling of the relaxation time of lipid protons, andwithout any contamination from water signal; (ii) in vivovalidation of the method on tissues undergoing varia-tions in oxygenation, by comparing results from the newtechnique to those obtained by the standard ‘‘T1 H2O’’and ‘‘BOLD’’ methods as reference; and (iii) comparisonwith results from direct quantitative reference methods(OxyLiteTM and EPR) in the mammary tumor models.

We demonstrate that MOBILE is sensitive to changesin tissue oxygenation. The mammary tumor, peripheralischemia, and cerebral stroke models were subjected to ahyperoxic breathing challenge using carbogen, whoseresponse may be an index of the vascular functionalityof those tissues. MOBILE demonstrates sensitivity to dis-criminating between vascularly dysfunctional tissues (tu-mor areas, ischemic limbs, or stroke) and normofunc-tional ones (i.e., contralateral leg or brain hemisphere).MOBILE is also sensitive to the baseline tissue oxygen-ation when comparing diseased with healthy tissues (inthe peripheral ischemic and cerebral stroke models). Cor-relations between data sets from MOBILE and direct O2

measurements demonstrated that R1 lipids was signifi-cantly correlated with pO2 measured by EPR. Neverthe-less, we should bear in mind that R*2 contrast arises pri-marily from hemodynamic responses whereas R1 islikely to be mainly related to tissue oxygenation levels(Fig. 1). Therefore, information provided by the tworelaxation parameters should be complementary.

This proof-of-concept study encompasses some poten-tial applications of the MOBILE technique, by showingconsistency of semiquantitative data (relative changes inR1 lipids in response to an induced change in oxygen-ation in different tissues), although further crossvalida-tions with other quantitative in vivo or ex vivo techni-ques should be warranted to further assess therobustness of quantitative data yielded by the method.Nevertheless, MOBILE presents the advantage of allow-ing longitudinal studies, contrary to positron emissiontomography imaging. Yet, the hereby presented models

of peripheral ischemia and stroke which enable compari-son between damaged areas and normal contralateralones unequivocally demonstrate significant and constantdifferences of the basal R1 lipids within the two areas.Also, oxygen enriched gases inhalation was used as atechnique to modulate tissue oxygenation. Yet, actualeffects of such gases might be variable from one tissue tothe other e.g. according to the presence or absence ofCO2 (i.e. for carbogen). This may explain the dispersionof the results, particularly in tumors where tissue hetero-geneity is high. Local pO2 was assessed quantitativelyusing EPR and OxyLite in tumors to address this limitat-ing concern. Interestingly, the short acquisition time ofthe MOBILE sequence should allow the follow-up ofacute hypoxia in tumors in a quantitative and dynamicway, similarly to previous work using 19F-relaxometry orEPR oximetry (55,56).

From a technical point of view, a current limitation ofthe technique is that the minimal amount of lipidsrequired within a tissue to be enable accurate measure-ments using the MOBILE technique yet remains to bedetermined, although consistent data were obtained fromsuch different tissues as malignant tumors, liver, muscle,and brain. This would also depend on extrinsic parame-ters such as the magnetic field strength and coils andcould be a limitation to the application of the techniqueto some body’s areas. Also, at 11.7 T, two different lipidpeaks were considered depending on the tissue of inter-est (at �1.2 and �4.0 ppm). This was established experi-mentally with respect to the peaks that could be reliablyfitted in each tissue and was defined as an ‘‘operationalprotocol.’’ Nevertheless, at lower magnetic fields (i.e. inthe clinical setting), only the 1.2-ppm peak is amenablefor analysis. Finally, it would be relevant to compare theSSFP method used in this study to assess T1 of waterand lipids with T1 fits issued from fat and water imagesextracted using the Dixon method or the alternative‘‘IDEAL’’ method (57). Indeed, the use of a saturationpulse of the water peak might affect the lipid peak at 4.0ppm and the resulting T1 fit, although in vitro validationon phantoms shows that the SSFP method allows recov-ering reliable T1 of lipids. Yet, the in vivo comparison ofthe methods would be required to select the bestsequence to assess proper T1 lipids values in vivo, espe-cially since the maps of R1 lipids sometimes show poorresolutions in some tissues because of the exclusion ofsome pixels in the postprocessing of the data (i.e., pixelswith high T1 fit errors).

Endogenous contrast arising from variations in lipidsR1 relaxation under different oxygenation conditionscould find straightforward applications in any lipid con-taining tissue and could be readily translated intopatient management on clinical MR scanners, withoutspecial requirements for software/hardware (i.e., dedi-cated coil systems, etc.). The technique is noninvasive,more sensitive than that using T1 of the water signal,and should not be influenced by other hemodynamic fac-tors, unlike T1 H2O or T*2. The spatial and temporal res-olutions (�0.6 mm, acquisition time 1 min 20 s) of thetechnique are relevant for dynamic mapping of oxygen-ation in both research (small animals) and clinicalframes. Potential applications of the technique could be

742 Jordan et al.

found in any pathophysiological condition where oxygenplays a key role, such as tumor oxygenation, cerebralstroke, peripheral ischemic diseases consecutive to, e.g.,diabetes, or liver transplantation, as illustrated by thepresent work. Another possible field of application is thechange in tissue oxygenation after physiological stimula-tion. An initial translational study on a clinical MRscanner was successfully performed on three healthyvolunteers with spatial and temporal resolutions of �3.0mm and 4 min, respectively, thereby assessing a widerange of potential applications of the technique in clini-cal practice.

In conclusion, these initial studies suggest that MO-BILE has the potential to provide a noninvasive andquantitative method for obtaining fast and robust meas-urements of the variations in tissue oxygenation in dif-ferent pathological conditions, with a unique transla-tional aspect ‘‘from laboratory to patient’s bed.’’

ACKNOWLEDGMENTS

B.F.J. and P.D.C. are Research Associates of the FNRSand O.F. is a Senior Research Associate of the FNRS. B.F. Jordan and J. Magat contributed equally to this work.

REFERENCES

1. Tatum JL, Kelloff GJ, Gillies RJ, et al. Hypoxia: importance in tumor

biology, noninvasive measurement by imaging, and value of its mea-

surement in the management of cancer therapy. Int J Radiat Biol

2006;82:699–757.

2. Milosevic M, Fyles A, Hedley D, Hill R. The human tumor microen-

vironment: invasive (needle) measurement of oxygen and interstitial

fluid pressure. Semin Radiat Oncol 2004;14:249–258.

3. Gallez B, Baudelet C, Jordan BF. Assessment of tumor oxygenation

by electron paramagnetic resonance: principles and applications.

NMR Biomed 2004;17:240–262.

4. Ahmad R, Kuppusamy P. Theory, instrumentation, and applications

of electron paramagnetic resonance oximetry. Chem Rev 2010;110:

3212–3236.

5. Yu JX, Cui W, Zhao D, Mason RP. Non-invasive physiology and

pharmacology using 19F Magnetic Resonance. In: Tressaud A, Haufe

G, editors. Fluorine and health. The Netherlands, Elsevier; 2008.

pp. 197–276.

6. Matsumoto S, Yasui H, Batra S, et al. Simultaneous imaging of tumor

oxygenation and microvascular permeability using Overhauser

enhanced MRI. Proc Natl Acad Sci USA 2009;106:17898–17903.

7. Padhani AR, Krohn KA, Lewis JS, Alber M. Imaging oxygenation of

human tumours. Eur Radiol 2007;17:861–872.

8. Sharma R. Nitroimidazole radiopharmaceuticals in bioimaging. I. Syn-

thesis and imaging applications. Curr Radiopharm 2011;4:361–378.

9. Pacheco-Torres J, L�opez-Larrubia P, Ballesteros P, Cerd�an S. Imaging

tumor hypoxia by magnetic resonance methods. NMR Biomed 2011;

24:1–16.

10. Matsumoto K, Bernardo M, Subramanian S, Choyke P, Mitchell JB,

Krishna MC, Lizak MJ. MR assessment of changes of tumor in

response to hyperbaric oxygen treatment. Magn Reson Med 2006;56:

240–246.

11. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance

imaging with contrast dependent on blood oxygenation. Proc Natl

Acad Sci USA 1990;87:9868–9872.

12. Robinson SP, Howe FA, Griffiths JR. Noninvasive monitoring of carb-

ogen-induced changes in tumor blood flow and oxygenation by func-

tional magnetic resonance imaging. Int J Radiat Oncol Biol Phys

1995;33:855–859.

13. Karczmar GS, Kuperman VY, River JN, Lewis MZ, Lipton MJ. Mag-

netic resonance measurement of response to hyperoxia differentiates

tumors from normal tissue and may be sensitive to oxygen consump-

tion. Invest Radiol 1994;29 S2:S161–S163.

14. Baudelet C, Ansiaux R, Jordan BF, Havaux X, Macq B, Gallez B.

Physiological noise in murine solid tumours using T*2-weighted gra-

dient-echo imaging: a marker of tumour acute hypoxia? Phys Med

Biol 2004;49:3389–3411.

15. Baudelet C, Gallez B. How does blood oxygen level-dependent

(BOLD) contrast correlate with oxygen partial pressure (pO2) inside

tumors? Magn Reson Med 2002;48:980–986.

16. McPhail LD, Robinson SP. Intrinsic susceptibility MR imaging of

chemically induced rat mammary tumors: relationship to histologic

assessment of hypoxia and fibrosis. Radiology 2010;254:110–118.

17. Howe FA, Robinson SP, McIntyre DJ, Stubbs M, Griffiths JR. Issues

in flow and oxygenation dependent contrast (FLOOD) imaging of

tumours. NMR Biomed 2001;14:497–506.

18. O’Connor JP, Jackson A, Buonaccorsi GA, et al. Organ-specific effects

of oxygen and carbogen gas inhalation on tissue longitudinal relaxa-

tion times. Magn Reson Med 2007;58:490–496.

19. O’Connor JP, Naish JH, Parker GJ, et al. Preliminary study of oxygen-

enhanced longitudinal relaxation in MRI: a potential novel biomarker

of oxygenation changes in solid tumors. Int J Radiat Oncol Biol Phys

2009;75:1209–1215.

20. Winter JD, Estrada M, Cheng HL. Normal tissue quantitative T1 and

T*2 MRI relaxation time responses to hypercapnic and hyperoxic

gases. Acad Radiol 2011;18:1159–1167.

21. Bennett HF, Swartz HM, Brown RD III, Koenig SH. Modification of

relaxation of lipid protons by molecular oxygen and nitroxides.

Invest Radiol 1987;22:502–507.

22. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev

Cancer 2011;11:393–410.

23. Begg AC, Stewart FA, Vens C. Strategies to improve radiotherapy

with targeted drugs. Nat Rev Cancer 2011;11:239–253.

24. Jordan BF, Gallez B. Surrogate MR markers of response to chemo- or

radiotherapy in association with co-treatments: a retrospective analysis

of multi-modal studies. Contrast Media Mol Imaging 2010;5:323–332.

25. Rodrigues LM, Howe FA, Griffiths JR, Robinson SP. Tumor R*2 is a

prognostic indicator of acute radiotherapeutic response in rodent

tumors. J Magn Reson Imaging 2004;19:482–488.

26. Elas M, Bell R, Hleihel D, et al. Electron paramagnetic resonance ox-

ygen image hypoxic fraction plus radiation dose strongly correlates

with tumor cure in FSa fibrosarcomas. Int J Radiat Oncol Biol Phys

2008;71:542–549.

27. Robinson SP, Kalber TL, Howe FA, McIntyre DJ, Griffiths JR, Blakey

DC, Whittaker L, Ryan AJ, Waterton JC. Acute tumor response to

ZD6126 assessed by intrinsic susceptibility magnetic resonance imag-

ing. Neoplasia 2005;7:466–474.

28. Li SP, Padhani AR, Makris A. Dynamic contrast-enhanced magnetic

resonance imaging and blood oxygenation level-dependent magnetic

resonance imaging for the assessment of changes in tumor biology

with treatment. J Natl Cancer Inst Monogr 2011;2011:103–107.

29. Rathur HM, Boulton AJ. The neuropathic diabetic foot. Nat Clin

Pract Endocrinol Metab 2007;1:14–25.

30. Mohler ER III. Therapy insight: peripheral arterial disease and diabe-

tes-from pathogenesis to treatment guidelines. Nat Clin Pract Cardio-

vasc Med 2007;4:151–162.

31. Uccioli L, Gandini R, Giurato L, Fabiano S, Pampana E, Spallone V,

Vainieri E, Simonetti G. Long-term outcomes of diabetic patients

with critical limb ischemia followed in a tertiary referral diabetic

foot clinic. Diabetes Care 2010;33:977–982.

32. Weinberg MD, Lau JF, Rosenfield K, Olin JW. Peripheral artery dis-

ease. II. Medical and endovascular treatment. Nat Rev Cardiol 2011;

8:429–441.

33. Merino JG, Warach S. Imaging of acute stroke. Nat Rev Neurol 2007;

6:560–571.

34. Asdaghi N, Coutts SB. Stroke: neuroimaging in acute stroke-where

does MRI fit in?Nat Rev Neurol 2011;7:6–7.

35. Duffis EJ, Al-Qudah Z, Prestigiacomo CJ, Gandhi C. Advanced neuro-

imaging in acute ischemic stroke: extending the time window for

treatment. Neurosurg Focus 2011;30:E5.

36. Kim KA, Park MS, Kim IS, Kiefer B, Chung WS, Kim MJ, Kim KW.

Quantitative evaluation of liver cirrhosis using T1 relaxation time

with 3 tesla MRI before and after oxygen inhalation. J Magn Reson

Imaging 2012;36:405–410.

37. Baudelet C, Gallez B. Effect of anesthesia on the signal intensity in

tumors using BOLD-MRI: comparison with flow measurements by

Laser Doppler flowmetry and oxygen measurements by lumines-

cence-based probes. Magn Reson Imaging 2004;22:905–912.

38. Vandeputte C, Thomas D, Dresselaers T, Crabbe A, Verfaillie C, Bae-

kelandt V, van Laere K, Himmelreich U. Characterization of a


photothrombotic stroke model by MRI: implication for stem cell

transplantation. Mol Imag Biol 2011,13:663–671.

39. Folch J, Lees M, Sloane Stanley GH. A simple method for the isola-

tion and purification of total lipides from animal tissues. J Biol

Chem 1957,226:497–509.

40. Cani PD, Amar J, Iglesias MA, Knauf C, Neyrinck AM, Chabo C,

Delm�ee E, Delzenne NM, Burcelin R. Metabolic endotoxemia initiates

obesity and insulin resistance. Diabetes 2007;56:1761–1772.

41. Jordan BF, Cron GO, Gallez B. Rapid monitoring of oxygenation by

19F magnetic resonance imaging: simultaneous comparison with flu-

orescence quenching. Magn Reson Med 2009;61:634–638.

42. Delikatny EJ, Chawla S, Leung DJ, Poptani H. MR-visible lipids and

the tumor microenvironment. NMR Biomed 2011;24:592–611.

43. Srivastava NK, Pradhan S, Mittal B, Gowda GA. High resolution

NMR based analysis of serum lipids in Duchenne muscular dystro-

phy patients and its possible diagnostic significance. NMR Biomed

2010;23:13–22.

44. Ren J, Dimitrov I, Sherry AD, Malloy CR. Composition of adipose tis-

sue and marrow fat in humans by 1H NMR at 7 Tesla. J Lipid Res

2008;49:2055–2062.

45. Scheffler K, Henning J. T1 quantification with inversion recovery

True FISP. Magn Reson Med 2001;45:720–723.

46. Nekolla S, Gneiting T, Syha J, Deichmann R, Haase A. T1 maps by

K-space reduced snapshot-FLASH MRI. J Comput Assist Tomogr

1992;16:327–332.

47. McGrath DM, Naish JH, O’Connor JP, Hutchinson CE, Waterton JC,

Taylor CJ, Parker GJ. Oxygen-induced changes in longitudinal relaxa-

tion times in skeletal muscle. Magn Reson Imaging 2008;26:221–227.

48. Deichmann R, Haase A. Quantification of T1 values by SNAPSHOT-

FLASH NMR Imaging,J Magn Res 1992;96:608–612.

49. Jordan B, Baudelet C, Gallez B. Carbon-centered radicals as oxygen

sensors for in vivo electron paramagnetic resonance: screening for an

optimal probe among commercially available charcoals. Magn Reson

Mat Phys Biol Med 1998;7:121–129.

50. Gallez B, Jordan BF, Baudelet C, Misson PD. Pharmacological modifi-

cations of the partial pressure of oxygen in murine tumors: evalua-

tion using in vivo EPR. Magn Reson Med 1999;42:627–630.

51. Padhani A. Science to practice: what does MR oxygenation imaging

tell us about human breast cancer hypoxia? Radiology 2010;254:1–3.

52. Griffiths JR, Robinson SP. The OxyLite: a fibre-optic oxygen sensor.

Br J Radiol 1999;72:627–630.

53. J€orneskog G, Djavani K, Brismar K. Day-to-day variability of transcu-

taneous oxygen tension in patients with diabetes mellitus and pe-

ripheral arterial occlusive disease. J Vasc Surg 2001;34:277–282.

54. Oubre CM, Roy A, Toner C, Kalns J. Retrospective study of factors

affecting non-healing of wounds during hyperbaric oxygen therapy. J

Wound Care 2007;16:245–250.

55. Magat J, Jordan BF, Cron G, Gallez B. Non invasive mapping of spon-

taneous fluctuations in tumor oxygenation using 19F MRI. Med Phys

2010;37:5434–5441.

56. Matsumoto S, Yasui H, Mitchell JB, Krishna MC. Imaging cycling tu-

mor hypoxia. Cancer Res 2010;70:10019–10023.

57. Kim H, Taksali SE, Dufour S, Befroy D, Goodman TR, Petersen KF,

Shulman GI, Caprio S, Constable RT. Comparative MR study of hepatic

fat quantification using single-voxel proton spectroscopy, two-point

dixon and three-point IDEAL. Magn Reson Med 2008;59:521–527.

744 Jordan et al.

mapping of oxygen by imaging lipids relaxation enhancement

Documents