S

HTa

DJ

A

H

i

t

e

d

d

n

c

f

c

T

o

v

v

f

v

a

f

H

a

t

FUNr

M

J A C C : C A R D I O V A S C U L A R I M A G I N G V O L . 3 , N O . 4 , 2 0 1 0

© 2 0 1 0 B Y T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N I S S N 1 9 3 6 - 8 7 8 X / 1 0 / $ 3 6 . 0 0

P U B L I S H E D B Y E L S E V I E R I N C . D O I : 1 0 . 1 0 1 6 / j . j c m g . 2 0 0 9 . 1 2 . 0 1 1

T A T E - O F - T H E - A R T P A P E R S

eart Failure With Normal Ejection Fraction:he Complementary Roles of Echocardiographynd CMR Imaging

arryl P. Leong, MBBS,*† Carmine G. De Pasquale, MBBS, PHD,*‡oseph B. Selvanayagam, MBBS, DPHIL*‡

delaide, Australia

eart failure with normal ejection fraction (HFNEF), previously referred to as diastolic heart failure, has

ncreased in prevalence as a cause of heart failure, now accounting for up to 50% of all cases. Contrary

o initial evidence, the prognostic outlook in HFNEF may be similar to that of heart failure with reduced

jection fraction. According to current consensus statements, the diagnosis of HFNEF requires the

emonstration of relatively preserved systolic left ventricular function and evidence of diastolic

ysfunction. Noninvasive imaging techniques now permit evaluation of these parameters without

eed for cardiac catheterization in the large majority of patients. Echocardiography is the modality of

hoice in the evaluation of diastolic function but suffers from limitations in its assessment of systolic

unction. Cardiac magnetic resonance (CMR) imaging is the gold standard in the volumetric quantifi-

ation of systolic function; however, it has limitations in its ability to characterize diastolic function.

his report aims to review the strengths and weaknesses of both imaging modalities in the diagnosis

f HFNEF. With regards to echocardiography, it will specifically describe limitations in measuring left

entricular ejection fraction, describe novel techniques to assess systolic function such as tissue

elocity and strain analysis, and will review the measurements used in the evaluation of diastolic

unction. With respect to CMR, this review will highlight its value in the assessment of systolic left

entricular function, will review ancillary CMR findings that may support the diagnosis of HFNEF such

s tissue characterization, and will provide a brief overview of CMR techniques to assess diastolic

unction. We propose that these 2 modalities may play a complementary role in the diagnosis of

FNEF. The importance of imaging in the diagnosis of HFNEF extends to both the individual patient

nd to clinical trials of therapies for this condition. (J Am Coll Cardiol Img 2010;3:409 –20) © 2010 by

he American College of Cardiology Foundation

rom *Flinders Medical Centre, Adelaide, SA, Australia; †University of Adelaide, Adelaide, SA, Australia; and ‡Flindersniversity, Adelaide, SA, Australia. Dr. Leong is supported by a Medical Postgraduate Scholarship funded jointly by theational Health and Medical Research Council of Australia and the National Heart Foundation of Australia, and is the

ecipient of a Dawes Scholarship from the Royal Adelaide Hospital.

anuscript received August 20, 2009; revised manuscript received December 1, 2009, accepted December 15, 2009.

Hcp2sdiocbesep

fH

P

Aftmtrw

psai

icfoa8ifmortdsiibtepp

A

A

2

3

A

C

r

H

n

H

r

L

L

e

L

L

fraction

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

410

eart failure with normal ejection fraction(HFNEF) refers to the clinical syndrome ofheart failure in the presence of normal leftventricular (LV) ejection fraction. Three

onditions must be fulfilled for its diagnosis: 1) theresence of symptoms or signs of heart failure;) the presence of normal or mildly abnormalystolic function; and 3) evidence of diastolic LVysfunction (1). With the aging population, and

ncreasing prevalence of hypertension, diabetes, andbesity, HFNEF now accounts for up to 50% of allases (2). Despite initial studies suggesting aetter prognosis than heart failure with reducedjection fraction (HFREF), more recent evidenceuggests similar prognostic outlook with bothntities (3). A recent consensus statement pro-osed replacement of the terms systolic and dia-

stolic heart failure with HFREF andHFNEF, respectively, as the former im-plies mutually exclusive mechanisms,whereas diastolic dysfunction is highlyprevalent in systolic heart failure (4), andthere is evidence to suggest reducedsystolic tissue velocities in diastolic heartfailure (5). Moreover, published guide-lines on the diagnosis of HFNEF (1)uniformly require objective evidence ofdiastolic LV dysfunction in addition toclinical features of heart failure and LVejection fraction (LVEF) �50%. Thus,imaging modalities play a pivotal role in2 of the 3 diagnostic criteria of HFNEF.Recent advances in echocardiographicand cardiac magnetic resonance (CMR)imaging techniques allow more accuratecharacterization of cardiac structure and

unction, and permit more certain diagnosis ofFNEF.

athophysiology of Diastolic Dysfunction

t a macroscopic level, HFNEF is distinguishedrom HFREF by concentric LV remodeling ratherhan eccentric. Concentric remodeling refers toaintenance or increase in LV wall thickness rela-

ive to chamber size (Fig. 1), whereas in eccentricemodeling, there is LV cavity dilation, and relativeall thickness is diminished.Diastolic dysfunction is characterized patho-

hysiologically by increased LV stiffness (6). Con-equently, with exercise, there is an inability tougment LV end-diastolic volume despite increas-

ion

ng LV end-diastolic pressure. The outcome of f

ncreased LV stiffness is reflected in the hemodynamichanges that are most widely used to assess diastolicunction. In normal diastole, LV untwisting—a vig-rous and active process—sucks blood from the lefttrium into the left ventricle. Seventy percent to0% of LV filling under these circumstances occursn early diastole, with atrial contraction accountingor the remaining 20% to 30%. With the develop-ent of early diastolic dysfunction, the proportion

f LV filling permitted to occur in early diastole iseduced. The relative importance of atrial contrac-ion is thus increased—so-called grade I diastolicysfunction. As diastolic dysfunction worsens, pres-ure mounts within the left atrium, such thatmmediately following mitral valve opening, bloods forced under positive pressure, rather than suckedy negative pressure, into the left ventricle. Al-hough the proportion of LV filling occurring inarly diastole returns to normal, the underlyinghysiology is not normal, hence the designation ofseudonormal LV filling, or grade II diastolic dys-

Figure 1. Short-Axis White Blood CMR Image DemonstratingConcentric LVH

This end-diastolic short-axis cine cardiac magnetic resonance(CMR) image demonstrates concentric thickening of the left ven-tricular walls. The high spatial resolution and signal-to-noiseratio of cardiac magnetic resonance allows tightly reproduciblemeasurement of left ventricular wall thickness and total left ven-tricular myocardial mass. See Online Videos 1 and 2. LVH � leftventricular hypertrophy.

B B R E V I A T I O N S

N D A C R O N YM S

D � 2-dimensional

D � 3-dimensional

F � atrial fibrillation

MR � cardiac magnetic

esonance

FNEF � heart failure with

ormal ejection fraction

FREF � heart failure with

educed ejection fraction

A � left atrial

GE � late gadolinium

nhancement

V � left ventricular

VEF � left ventricular eject

unction. Grade III diastolic dysfunction is charac-

tiVim

A

EHd(dplatHddnttLbioaf

asTntbbmoCTe2medqap

mfi

radbestDioee(wspsrtDi

efrasttrsihotmralipCianrtgriel

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

411

erized by very abrupt flow of blood from left atriumnto left ventricle, but with reversibility with thealsalva maneuver. Grade IV diastolic dysfunction

s characterized by irreversibility with the Valsalvaaneuver.

ssessment of Systolic LV Function

chocardiography. A sine qua non in the diagnosis ofFNEF is the establishment of normal LVEF. Tra-

itional measurement of LVEF by 2-dimensional2D) echocardiography relies on clear endocardialefinition, which is lacking in up to 31% cases (7),articularly when intravenous contrast is not uti-

ized. LVEF in patients with clinical heart failure isunimodally distributed continuous variable. Thus,

he cutoff of 50% to dichotomize HFNEF fromFREF is somewhat arbitrary. Nonetheless, repro-

ucible assessment of LV volumes is crucial inelineating the predominant pathogenic mecha-isms underlying a patient’s heart failure. Even inhe hands of expert echocardiographic laboratories,here is considerable variability in the estimation ofVEF (8). Furthermore, there is poor agreementetween Simpson’s biplane technique and CMRmaging, the imaging gold standard for estimationf cardiac volumes, with Bland-Altman limits ofgreement of up to �20% among a sample of heartailure patients (7).

In 3-dimensional (3D) echocardiography, datare acquired over 4 to 6 cardiac cycles then synthe-ized into a single so-called full-volume dataset.his can then be manipulated or cropped to obtainonforeshortened planes through the left ventricle,hus permitting more accurate estimation of cham-er volumes. Three-dimensional echocardiographyoasts superior reproducibility and closer approxi-ation to CMR than 2D echo in the measurement

f LVEF (9). Despite minimal bias compared withMR, 3D echo has slightly greater variability (10).his may relate to poorer spatial resolution and

ndocardial definition, which is suboptimal in up to0% of cases, and is associated with less reliableeasurement of LV volumes compared with non–

ndocardial-dependent techniques (11). Three-imensional echocardiography and CMR both re-uire the patient to breath-hold during imagecquisition, which can prove difficult in those dys-neic at rest.Although accurate and reproducible measure-ent of LVEF is clearly desirable, it is not always

easible in clinical practice, where patient character-

stics often differ greatly from those studied in the a

esearch setting. Obstacles such as the inability todequately position bed-bound patients, ventilatorependency in the intensive care unit, and comor-idities such as chronic obstructive pulmonary dis-ase and obesity frequently render accurate mea-urement of LVEF limited using conventional echoechniques. Contrast echocardiography and tissueoppler imaging may potentially have a role to play

n this setting. Contrast-enhanced 2D echocardi-graphy has similar accuracy to noncontrast 3Dchocardiography (12). Contrast-enhanced 3Dchocardiography may yield even superior results12). Peak systolic tissue velocity (Sm) by pulseave tissue Doppler at the mitral annulus has been

hown to correlate with LVEF. Sm �7 cm/sredicts LVEF �45% with sensitivity 93% andpecificity 87% (13). Tissue velocity in this locationepresents an integral of longitudinal cardiac mo-ion from base to apex. The advantage of tissueoppler imaging is that it is not reliant on good 2D

mage quality and has good reproducibility.Speckle tracking technology provides another

chocardiographic means to evaluate systolic LVunction. It relies on tracking small myocardialegions of interest, each with a relatively uniqueppearance owing to its pattern of acoustic back-catter, through space from frame to frame. Thishen permits calculation of strain—tissue deforma-ion—on a regional basis. Current software allowsegional strain scores to be averaged to yield a globaltrain score. This score has been proven an accuratendex of systolic LV function in chronic ischemiceart disease (14). Global strain score’s advantagever LVEF as an index is its reproducibility, al-hough its prognostic role remains to be deter-ined. Furthermore, strain analysis permits inter-

ogation of myocardial deformation in any cardiacxis. Wang et al. (15) recently demonstrated thatongitudinal and radial systolic strains are reducedn diastolic heart failure although LV twist isreserved.MR imaging. CMR has rapidly become the imag-ng method of choice and the gold standard in thessessment of systolic cardiac function of bothormal and abnormal left ventricles (7). Withegard to measurement of global LV systolic func-ion, given its 3D nature and an order of magnitudereater signal-to-noise ratio, CMR is highly supe-ior to 2D echo (7). Given its greater reproducibilityn measuring LVEF over 2D echo, the potential forrroneously categorizing heart failure is markedlyessened. This imaging is typically performed in,

lthough not limited to, the conventional short-axis

vviiis

phgtaom

awtbcpo

hgcwdtp

ffsdaofrls

ddmnht(whuis

T

Tmrrcoa(he(hnadfigT

normal ejection fraction; L

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

412

iews (Online Video 1) and the 3 cardinal long-axisiews (Online Video 2). The ability of CMR tomage in any plane without the need for optimalmaging windows allows for unprecedented flexibil-ty for the interrogation of abnormal hearttructures.

Cine images are obtained using steady-state freerecession sequences, which provide images ofigher signal to noise (than older sequences such asradient echo), and hence exceptional delineation ofhe blood–myocardium interface. This allows forccurate and reproducible quantitative assessmentf chamber dimensions and systolic function usinganual or semiautomated planimetry techniques.Visual inspection of LV and right ventricular

rchitecture identifies patterns of regional or diffuseall thinning and concentric or asymmetric hyper-

rophy. Pericardial thickness and calcification cane assessed. The spatial resolution with which thisan be achieved by CMR is useful in distinguishingericardial constriction from restrictive cardiomy-pathy, both of which may result in features of

c Patterns of LGE and Ancillary Findings on CMR Imaging ins That May Present as HFNEF

Pattern of LateGadoliniumEnhancement

Prevalence ofLate GadoliniumEnhancement* Ancillary Findings

-wall withinypertrophiedegments in patchyistributionst consistently atnterior andosterior junction ofnterventriculareptum and rightentricular free wall

79%–81% LV hypertrophy, usuallyasymmetric

st often basalnferolateral wallendocardiumsually spared

50% LV hypertrophy—maybe indistinguishablefrom HCM. Biopsy orgenetic testingconfirms diagnosis

useendocardial butraversing arterialerritories

69% Multiple images of thesame view usingdifferent inversiontimes may allowdemonstration ofdiffuse early nullingof myocardial signalto suggestamyloidosis

presence ofericardialnhancementuggests ongoingnflammation

— Pericardial thickness�4 mm is consistentwith the diagnosis,and �5 mm hashigh specificity

he cited references and represent the best available estimate of late gadoliniumtual prevalence may vary with the population studied.resonance; HCM � hypertrophic cardiomyopathy; HFNEF � heart failure withGE � late gadolinium enhancement; LV � left ventricular.

t

eart failure with normal ejection fraction. Tomo-raphic imaging is an important adjunct to echo-ardiographic characterization of hemodynamics,hich forms the cornerstone of diagnosing pericar-ial constriction, especially given the observationhat constriction may occur in the absence ofericardial thickening.CMR tagging is a technique by which a radio-

requency pulse is applied to LV myocardium in theorm of grid lines. The tagged grid deforms as theaturated myocardium moves throughout the car-iac cycle, allowing regional strain to be visualizednd quantified. The tag is generally applied at thenset of systole, triggered by the ECG R-wave, andades during the cardiac cycle as the magnetizationecovers towards equilibrium as a result of spin-attice, or T1, relaxation. Measurement of strain isuited to quantifying regional systolic LV function.

The tissue phase mapping technique allows theetermination of 3D velocity tensors over the car-iac cycle, i.e., for rotation, radial, and longitudinalovement, with a pixel-by-pixel spatial resolution

earing that of “conventional” cine CMR (16). Thisas been investigated in clinical studies in both pa-ients and volunteers, and newer navigator sequencesobviating the need for breath-holding, and henceith the potential for improving spatial resolution)ave been developed. Displacement-encoded imagingsing stimulated echoes (DENSE) can also providenformation on myocardial displacement, velocity, andtrain (17).

issue Characterization

he development of the late gadolinium enhance-ent (LGE) CMR technique (LGE-CMR) has

evolutionized the role of CMR in clinical andesearch practice. It has a potential role in theontext of HFNEF, in identifying specific patternsf fibrosis and scarring in many of the cardiomyop-thy states that may initially present with HFNEF18,19). This is summarized in Table 1. There is aigh prevalence (79% to 81%) of late gadoliniumnhancement in hypertrophic cardiomyopathy18,20). Its distribution is typically mid-wall withinypertrophied segments, and is particularly promi-ent at the junction of the interventricular septumnd right ventricular free wall (20) (Fig. 2). Fabryisease, caused by X-linked alpha-galactosidase de-ciency, may be difficult to distinguish echocardio-raphically from hypertrophic cardiomyopathy.his distinction is important because disease-

Table 1. CharacteristiVarious Disease State

Disease (Ref. #)

HCM (18,20) Midhsd

Moapisv

Fabry disease(21)

Moi

Subu

Amyloidosis (22) DiffSub

tt

Pericardialconstriction

Thepesi

*Quoted figures are from tenhancement, although acCMR � cardiac magnetic

argeted therapy may improve outcome in Fabry

dusbdrgsmLndoivd

A

Lm(cc(slt

au0EssdaitpcueuroMm(mbaavg

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

413

isease. LGE occurs in 50% of cases, and if present,sually affects the basal inferolateral wall, sparing theubendocardium (21). Amyloidosis has been found toe associated with qualitative global and subendocar-ial gadolinium enhancement of the myocardium. Aecent landmark study found that subendocardial lon-itudinal relaxation time (T1) in amyloid patients washorter than in controls, and was correlated witharkers of increased myocardial amyloid load, such asV mass, wall thickness, interatrial septal thick-ess, and diastolic function. Global subendocar-ial LGE was found in approximately two-thirdsf patients (22). The T2* CMR technique, whichs noncontrast-dependent, has been shown to bealuable in the quantification of myocardial ironeposition in hemochromatosis (23).

ssessment of Left Atrial Volume

eft atrial (LA) size is an important prognosticarker in a number of cardiovascular disease states

24,25). It can be considered as an end point ofumulative LV insult. LA dilation can help dis-riminate normal from pseudonormal LV filling26), and its finding in the appropriate clinicaletting supports the diagnosis of HFNEF (27). Aarge cross-sectional study has shown the associa-

Figure 2. End-Diastolic Images From a Patient With Hypertroph

(A) Short-axis white blood cardiac magnetic resonance (CMR) imagegadolinium-enhancement CMR image demonstrating increased signventricular free wall (arrows). (C) Vertical long-axis white blood CMaxis late gadolinium-enhancement CMR image demonstrating increas

ion between LA volume (indexed to body surface m

rea) and grade of diastolic dysfunction with an areander the receiver-operator characteristic curve of.81 to detect grade II diastolic dysfunction (28).chocardiography. As with assessment of LV cavityize, quantification of LA size has progressed fromingle-dimension measurement by M-mode echocar-iography to single-plane and biplane estimation ofrea and volume. Prognostic discriminatory powerncreases with each step’s increase in data. Among 2Dechniques used to assess LA volume Simpson’s bi-lane and the area-length method have the closestorrelation, whereas the prolate-ellipsoid techniquenderestimates LA volume (29). More recently, 3Dchocardiography has been used to evaluate LA vol-me (30). This technique has superior test-retesteliability and inter- and intraobserver variability tother echocardiographic techniques (31).agnetic resonance imaging. Early studies deter-ined reference ranges for LA dimensions and area

32). Subsequent studies examined the accuracy ofeasurement of LA volume by either a modified

i-plane Simpson’s technique from the horizontalnd vertical long axes (33), or by measurement fromshort axis stack (34). The inter- and intraobserver

ariability and interstudy variability tend to bereater when LA volume is measured using the

ardiomyopathy

monstrating asymmetric septal hypertrophy. (B) Short-axis latetensity at the junction of the interventricular septum and rightage demonstrating left ventricular hypertrophy. (D) Vertical long-ignal intensity in the anterior wall (arrow). See Online Video 3.

ic C

deal inR imed s

odified Simpson’s technique (35). The appendage

in

A

TmsarcpdmirinLdEnMrdhrcrmfCmdr(ce

c(asdgp

A

EtAgtiiirdifidiipett�rooOifl

tal m

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

414

s generally included in the LA volume, but pulmo-ary veins are excluded.

ssessment of LV Mass

here is a well-recognized association between LVass and adverse cardiovascular outcomes and even

udden cardiac death (36). This relationship persistsfter adjustment for conventional cardiovascularisk factors. This association likely holds true be-ause LV mass, like LA volume, represents an endoint of cumulative LV insult. Although not aiagnostic criterion for HFNEF, increased LVass provides supportive evidence. Furthermore, it

s conceptually attractive to quantify LV mass as itelates directly to the pathological process underly-ng HFNEF, although further investigation iseeded to truly relate the anatopathologic defect ofV mass to that of the physiologic perturbation ofiastolic function.chocardiography. Early echocardiographic tech-iques for measuring LV mass that relied on-mode have been superseded by 2D and, more

ecently, 3D techniques. Each advance in echocar-iographic methodology has resulted in fewer in-erent assumptions and thus more accurate andeproducible measurement of LV mass. 3D echo-ardiographic measurement of LV mass is moreeproducible than 2D, and approximates CMRore closely (37), whereas 2D techniques suffer

rom wide limits of agreement (38).MR. CMR is the noninvasive gold standard foreasurement of LV mass (39) and has been vali-

ated against cadaveric measurements (40). Normalanges for LV mass by CMR have been published41). The technique for measurement of LV myo-ardial volume requires manual or semiautomated

Table 2. Echocardiographic Classification of Grades of Diastolic

Parameter NormalMild Dysfunction

(Grade 1)M

Transmitral PW Doppler

E/A 1–1.5 �1

DT (ms) 140–250 �250

TDI

E’ (cm/s) �7 �7

Valsalva Negative Positive

LAVI (ml/m2) 22 � 6 �28

Pulmonary venous flow S � D S �� D

Mitral in-flow propagationvelocity (cm/s)

�50 �50

D � peak D-wave velocity; DT � deceleration time; E/A � ratio of E- to A-waveS � peak S-wave velocity; TDI � pulse wave tissue Doppler imaging at the sep

ndo- and epicardial border tracing. LV mass is a

alculated by multiplying this volume by 1.05g/cm3

the specific density of myocardium). The greaterccuracy and reproducibility of CMR in the mea-urement of LV mass allow smaller changes to beetected (Online Videos 1 to 3). This providesreater sensitivity to measure the effects of diseaserogression or efficacy of disease therapy.

ssessment of Diastolic LV Function

chocardiography. Echocardiography is the imagingechnique of choice in evaluation of diastolic function.mong the parameters available for echocardio-

raphic assessment of diastolic function (Table 2),ransmitral flow velocities and tissue Doppler imag-ng are the quickest to acquire and have the lowestnterobserver variability (42). In healthy, youngndividuals, most LV filling occurs in early diastole,esulting in a prevailing E-wave. With aging oriastolic dysfunction, increasing LV stiffness andmpaired diastolic LV untwisting reduces early LVlling, thus resulting in a predominant A-wave. Asiastolic dysfunction progresses to pseudonormal-ty, LA pressure rises, and early LV filling increasesn proportion. Despite an E/A ratio �1, the patho-hysiological distinction from normality is thatarly LV filling is from blood being forced in ratherhan being sucked in. In grade III diastolic dysfunc-ion, the E/A ratio is �2 and deceleration time

140 ms (Fig. 3). Although the finding of aestrictive transmitral filling profile allows easy rec-gnition of diastolic dysfunction, this pattern isnly observed in 10% of cases of HFNEF (4).ccasionally, a triphasic pattern of transmitral fill-

ng is observed (Fig. 4) as an L-wave: antegradeow resulting in a Doppler signal between the E

function

rate Dysfunction(Grade 2)

Severe Dysfunction(Grade 3)

Severe Dysfunction(Grade 4)

1–1.5 �2 �2.5

140–250 �140 �140

�7 �5 �5

Positive Positive Negative

�28 �35 �40

S � D S �� D S �� D

�50 �50 �50

ities; LAVI � left atrial volume indexed to body surface area; PW � pulse wave;itral annulus.

Dys

ode

veloc

nd A waves, during what is normally diastasis.

Adfof

ptwmD�cvi5cwdsisfisra

ew(Eslt�at

abtsdwngCsCturrmh

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

415

lthough not necessarily pathognomonic for severeiastolic dysfunction, the L-wave has a predilectionor occurrence in those with significant LV pathol-gy, and thus may be supportive of diastolic dys-unction (43).

A challenge in diastology is the distinction ofseudonormal filling from normal. Of assistance inhis regard are (Table 2): 1) the Valsalva maneuver,hich results in reversal in E/A ratio in pseudonor-ality, but not in normality (Fig. 5B); 2) tissueoppler imaging at the mitral annulus, where E=7 cm/s suggests diastolic dysfunction, and E= �5

m/s strongly supports it (Fig. 5C); 3) pulmonaryenous Doppler (Fig. 5D); 4) propagation of mitraln-flow towards the apex using color M-mode (44);) systolic LV function—if impaired, suggests con-omitant diastolic dysfunction; 6) LV hypertrophy,hich suggests diastolic dysfunction; and 7) LAilation, which also supports the diagnosis of dia-tolic dysfunction. Caution must be exercised rely-ng solely on E/A ratio: following cardioversion orpontaneous reversion to sinus rhythm from atrialbrillation the A-wave is reduced owing to atrialtunning. This can give a misleadingly high E/Aatio. Moderate or greater mitral valve disease posessimilar caveat.Peak early diastolic mitral velocity (E) is influ-

nced by LA pressure, age, and LV relaxation,hereas peak diastolic mitral annular tissue velocity

E=) is governed by age and LV relaxation. Thus the/E= ratio is, theoretically, a measure of LA pres-

ure. E/E= has been shown to have a better corre-ation with LV end-diastolic pressure (LVEDP)han other echocardiographic indices, with a ratio

15 predicting elevated mean LVEDP (45). It isgain important to be aware of a technique’s limi-

Figure 3. Doppler Evaluation of a Patient With Restrictive LV Fi

(A) Transmitral pulsed wave Doppler at mitral valve leaflet tips demPulsed wave Doppler at the ostium of the right upper pulmonary vties; LV � left ventricular.

ations. If septal motion is abnormal, for instance as

result of right ventricular pacing or left bundleranch block, E/E= has been shown unreliable forhe prediction of pulmonary capillary wedge pres-ure (46). Nonetheless, in a study of patients 1.6ays following myocardial infarction, 42% ofhich were anterior, E/E= was a superior prog-ostic index to other clinical and echocardio-raphic parameters (47).MR. In contrast to the wealth of studies in LVystolic dysfunction, there is relative inexperience inMR assessment of diastolic LV function. None-

heless, a number of CMR techniques have beensed to evaluate diastolic function. Peak LV fillingates in early and late diastole can be measured byates of change in chamber volume—a techniqueade possible by CMR’s high spatial resolution. It

as been demonstrated that their ratio decreases

Figure 4. Diastolic L-Wave on Transmitral Doppler Echocardiogr

Transmitral pulsed wave Doppler at mitral valve leaflet tips in a patpseudonormal left ventricular filling. An L-wave is seen as antegrad

trating E/A ratio �2 and brief deceleration time (arrows). (B)showing S-wave �� D-wave. E/A � ratio of E- to A-wave veloci-

aphy

ient withe flow

lling

onsein

occurring between the E and A waves.

wo(lhdhi

Psvl9altrvb

tdTtbcw

ssaatAildtta3pa

hderglpDgdr

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

416

ith age in healthy individuals, in a manner anal-gous to E/A ratio by transmitral Doppler echo48). Volume-derived indices such as peak early andate LV filling rate and time to peak early LV fillingave been promoted as sensitive markers of diastolicysfunction (49), but their measurement remainsighly time-consuming with present software and

mpractical from a clinical perspective.Blood flow velocity can be measured by CMR.

hase-contrast imaging, in which the spin phasehift is used to estimate transmitral blood flowelocity, has demonstrated at least moderate corre-ation with echocardiography (50,51), although5% confidence limits on Bland-Altman analysisre wide. Velocities measured by CMR tend to beower than by echocardiography as a consequence ofhe lower temporal resolution of CMR. Temporalesolution may be improved by reducing number ofiews/segment; however, this requires longerreath-hold times.Studies have employed myocardial grid tagging

o demonstrate abnormalities in LV untwisting iniastole in conditions such as aortic stenosis (52).o date, the major limitation in myocardial grid

agging for the assessment of diastolic function haseen the fading of grid lines over the course of theardiac cycle. In a large population study of subjects

Figure 5. Doppler Evaluation of a Patient With Pseudonormal L

(A) Transmitral pulsed wave Doppler at mitral valve leaflet tips. (B)Valsalva maneuver illustrating E/A reversal. (C) Pulsed-wave tissue Ddiastolic E= velocity. (D) Pulsed wave Doppler at the ostium of the rventricular.

ithout known cardiovascular disease, early dia- t

tolic strain rate (the first temporal derivative oftrain) could be measured in 80% of segmentsnalyzed (53). Atrial-induced strain rate could bessessed in only 32% of patients. In the remainder,ag lines faded in intensity, precluding analysis.lthough this study found early diastolic strain rate

nversely proportional to indexed LV mass, it high-ights the current limitations of CMR in assessingiastolic function. Imaging sequences with higheremporal resolution have been shown to increasehe proportion of the cardiac cycle that can benalyzed over older fast-gradient echo sequences. A.0-T CMR has been shown to produce better tagersistence through the cardiac cycle than 1.5-Tnd might have a role in the future (54).

Velocity-encoded CMR or tissue-phase mappingas been used to evaluate regional systolic andiastolic tissue velocities in an analogous manner tochocardiography (55,56). It has superior spatialesolution to grid tagging because the number ofrid lines that can be placed on the myocardiumimits the latter. Tissue-phase mapping also encom-asses more of the cardiac cycle than grid tagging.espite its feasibility, tissue-phase mapping has not

ained widespread popularity in the evaluation ofiastolic function, mainly due to poorer temporalesolution and difficulty with the long breath-hold

lling

smitral pulsed wave Doppler at mitral valve leaflet tips duringler imaging at the septal mitral annulus demonstrating reducedupper pulmonary vein showing S-wave � D-wave. LV � left

V Fi

Tranoppight

imes required, which is of particular relevance in

tvidv

ilao

A

AelmfiAIAab

ds(dcssosomDpwLdt

F

PriswCditv

otmm

apCtaktwCawdifiimtpivneisfi

iabimt

C

TdatsTpgmwpC

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

417

he heart failure population. Bergvall et al. (57) havealidated tissue-phase mapping during free breath-ng against grid tagging. Fading of the grid lines iniastole limited the validation of diastolic tissueelocities, however.

In summary, although early studies show promisen CMR assessment of diastolic function, presentimitations and the time-consuming nature of datacquisition restrict the widespread implementationf these techniques.

trial Fibrillation

trial fibrillation (AF) poses a challenge to thevaluation of cardiac function. Variability in cycleength hinders estimation of LV function by any

odality because of beat-to-beat variation in LVlling and ejection. Furthermore, loss of the-wave precludes measurement of the E/A ratio.

t is recommended that measurements taken inF be averaged over at least 5 cardiac cycles,

lthough this may have workflow implications inusy laboratories.Modest correlation exists between transmitral

eceleration and pulmonary capillary wedge pres-ure, although the clinical utility of this is limited58). The E/E= ratio remains of value in theiagnosis of HFNEF in the presence of AF. In aross-sectional study of patients with HFNEF,eptal E/E= �13 corroborated the diagnosis withensitivity 81.8% and specificity 89.5% (59). Anbservational study of nonvalvular AF demon-trated that E/E= �15 is an independent predictorf mortality (60). Pulmonary venous flow profilesay be of use in AF. A deceleration time of the-wave �220 ms has been shown to predict mean

ulmonary capillary wedge pressure �12 mm Hgith 100% sensitivity and specificity (58). AlthoughA volume reflects prevailing LV function, itsiscriminatory role in AF is confounded by the facthat AF itself promotes LA dilation.

uture Directions

resent-generation matrix array probes that permiteal-time 3D echocardiographic imaging are largen size, rendering them unwieldy. As their dimen-ions decrease, a greater proportion of individualsill be able to benefit from 3D echocardiography.urrent echocardiographic acquisition of myocar-ial tissue velocity or strain data utilizes sequential

mages in different 2D planes. The inherent limi-ation that results is from possible beat-to-beat

ariability in myocardial function. The development i

f 3D probes that permit real-time recording ofissue velocity or deformation information mayake accurate characterization of all aspects ofyocardial function more rapid and reliable.Nonangulated 3D CMR images of the heart

cquired even during free breathing has shownromise in overcoming some of the limitations ofMR estimation of LVEF (61). Emerging CMR

echniques that allow mapping of blood flow mayllow more accurate characterization of LV fillinginetics (62). The temporal and prognostic rela-ionship between myocardial fibrosis and HFNEFarrants further investigation. Tools such as LGE-MR and integrated echocardiographic backscatter

re available to quantitate myocardial fibrosis, buthether this adds incrementally useful data in theiagnosis of HFNEF is uncertain. The major lim-tation of LGE in the detection of myocardialbrosis is that it relies on difference in signal

ntensity between scarred regions and adjacent nor-al myocardium. It may thus have reduced sensi-

ivity for diffuse fibrosis. The calculation of aost-contrast myocardial T1 time following imag-ng a given plane with sequentially increasing in-ersion times has recently been shown to discrimi-ate heart failure patients from healthy controlsven after excluding myocardial segments display-ng LGE. This technique shows promise in moreensitive and quantitative evaluation of myocardialbrosis (63).Imaging of blood flow propagation and vorticity

s currently under development. Its advent willllow greater characterization of the relationshipetween LV filling and ejection. Lastly, fusionmaging, which integrates data from more than 1

odality, may permit the advantages of differentechniques to become fully integrated.

onclusions

wo of the 3 criteria for diagnosis of HFNEF areependent on cardiac imaging. Echocardiographynd CMR are the 2 modalities that offer the most inhis regard because they obtain data on cardiactructure and function, in contrast to scintigraphy.hey each have strengths and limitations that atresent make them complementary. Echocardio-raphy’s high temporal resolution makes it theodality of choice for assessing diastolic function,hereas CMR’s high spatial resolution permitsrecise evaluation of systolic function. AlthoughMR suffers from limited availability, and contra-

ndications such as implantable metallic devices, its

awitrwao

bduu

R

vC

R

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

418

bility to obtain accurate information in patientsith poor echocardiographic windows makes it an

mportant alternative in the diagnosis and charac-erization of heart failure. Each modality has madeecent progress to address their respective areas ofeakness. If appropriate echo windows are avail-

ble, 3D echo approaches CMR in the estimationf LVEF. Advances in CMR sequences increas-

dimensional echocardiography versus Magn Reson Imagin

oth imaging modalities are not necessary for theiagnosis of HFNEF in the majority of cases,ncertainty following 1 scan may be resolved bytilizing the strengths of the other.

eprint requests and correspondence: Dr. Joseph B. Sel-anayagam, Department of Cardiology, Flinders Medicalentre, Bedford Park, SA, Australia 5042. E-mail:

ingly allow characterization of diastole. Although Joseph.Selvanayagam@flinders.edu.au.

1

1

1

2

2

2

2

2

2

E F E R E N C E S

1. Paulus WJ, Tschope C, Sanderson JE, etal. How to diagnose diastolic heart failure:a consensus statement on the diagnosis ofheart failure with normal left ventricularejection fraction by the Heart Failure andEchocardiography Associations of theEuropean Society of Cardiology. EurHeart J 2007;28:2539–50.

2. Owan TE, Hodge DO, Herges RM,Jacobsen SJ, Roger VL, RedfieldMM. Trends in prevalence and out-come of heart failure with preservedejection fraction. N Engl J Med 2006;355:251–9.

3. Bhatia RS, Tu JV, Lee DS, et al. Out-come of heart failure with preservedejection fraction in a population-basedstudy. N Engl J Med 2006;355:260–9.

4. Bursi F, Weston SA, Redfield MM,et al. Systolic and diastolic heart fail-ure in the community. JAMA 2006;296:2209–16.

5. Yu CM, Lin H, Yang H, Kong SL,Zhang Q, Lee SW. Progression of sys-tolic abnormalities in patients with “iso-lated” diastolic heart failure and diastolicdysfunction. Circulation 2002;105:1195–201.

6. Westermann D, Kasner M, SteendijkP, et al. Role of left ventricular stiff-ness in heart failure with normal ejec-tion fraction. Circulation 2008;117:2051–60.

7. Bellenger NG, Burgess MI, Ray SG,et al. Comparison of left ventricularejection fraction and volumes in heartfailure by echocardiography, radionu-clide ventriculography and cardiovas-cular magnetic resonance; are they in-terchangeable? Eur Heart J 2000;21:1387–96.

8. Hare JL, Brown JK, Marwick TH.Performance of conventional echocar-diographic parameters and myocardialmeasurements in the sequential eval-uation of left ventricular function.Am J Cardiol 2008;101:706–11.

9. Soliman OI, Kirschbaum SW, vanDalen BM, et al. Accuracy and repro-ducibility of quantitation of left ventric-ular function by real-time three-

cardiac magnetic resonance. Am J Car-diol 2008;102:778–83.

10. Sugeng L, Mor-Avi V, Weinert L, etal. Quantitative assessment of left ven-tricular size and function: side-by-sidecomparison of real-time three-dimensional echocardiography andcomputed tomography with magneticresonance reference. Circulation 2006;114:654–61.

11. Tighe DA, Rosetti M, Vinch CS, et al.Influence of image quality on the accu-racy of real time three-dimensionalechocardiography to measure left ven-tricular volumes in unselected pa-tients: a comparison with gated-SPECT imaging. Echocardiography2007;24:1073–80.

12. Jenkins C, Moir S, Chan J, Rakhit D,Haluska B, Marwick TH. Left ventric-ular volume measurement with echocar-diography: a comparison of left ventric-ular opacification, three-dimensionalechocardiography, or both with mag-netic resonance imaging. Eur Heart J2009;30:98–106.

13. Ruan Q, Nagueh SF. Usefulness of iso-volumic and systolic ejection signals bytissue Doppler for the assessment of leftventricular systolic function in ischemic oridiopathic dilated cardiomyopathy. Am JCardiol 2006;97:872–5.

14. Gjesdal O, Hopp E, Vartdal T, et al.Global longitudinal strain measuredby two-dimensional speckle trackingechocardiography is closely related tomyocardial infarct size in chronic isch-aemic heart disease. Clin Sci (Lond)2007;113:287–96.

15. Wang J, Khoury DS, Yue Y, Torre-Amione G, Nagueh SF. Preserved leftventricular twist and circumferentialdeformation, but depressed longitudi-nal and radial deformation in patientswith diastolic heart failure. EurHeart J 2008;29:1283–9.

16. Hennig J, Schneider B, Peschl S, MarklM, Krause T, Laubenberger J. Analysisof myocardial motion based on velocitymeasurements with a black blood pre-pared segmented gradient-echo se-quence: methodology and applicationsto normal volunteers and patients. J

g 1998;8:868–77.

7. Gilson WD, Yang Z, French BA,Epstein FH. Measurement of myocar-dial mechanics in mice before andafter infarction using multislicedisplacement-encoded MRI with 3Dmotion encoding. Am J Physiol HeartCirc Physiol 2005;288:H1491-7.

8. Moon JC, McKenna WJ, McCrohon JA,Elliott PM, Smith GC, Pennell DJ. To-ward clinical risk assessment in hypertro-phic cardiomyopathy with gadoliniumcardiovascular magnetic resonance. J AmColl Cardiol 2003;41:1561–7.

9. McCrohon JA, Moon JC, Prasad SK, etal. Differentiation of heart failure relatedto dilated cardiomyopathy and coronaryartery disease using gadolinium-enhancedcardiovascular magnetic resonance. Circu-lation 2003;108:54–9.

0. Choudhury L, Mahrholdt H, Wagner A,et al. Myocardial scarring in asymptomaticor mildly symptomatic patients with hy-pertrophic cardiomyopathy. J Am CollCardiol 2002;40:2156–64.

1. Moon JC, Sachdev B, Elkington AG, etal. Gadolinium enhanced cardiovascularmagnetic resonance in Anderson-Fabrydisease. Evidence for a disease specificabnormality of the myocardial intersti-tium. Eur Heart J 2003;24:2151–5.

2. Maceira AM, Joshi J, Prasad SK, et al.Cardiovascular magnetic resonance incardiac amyloidosis. Circulation 2005;111:186–93.

3. Anderson LJ, Holden S, Davis B, etal. Cardiovascular T2-star (T2*) mag-netic resonance for the early diagnosisof myocardial iron overload. EurHeart J 2001;22:2171–9.

4. Tsang TS, Abhayaratna WP, BarnesME, et al. Prediction of cardiovascularoutcomes with left atrial size: is vol-ume superior to area or diameter?J Am Coll Cardiol 2006;47:1018–23.

5. Ristow B, Ali S, Whooley MA,Schiller NB. Usefulness of left atrialvolume index to predict heart failurehospitalization and mortality in am-bulatory patients with coronary heartdisease and comparison to left ventric-ular ejection fraction (from the Heartand Soul Study). Am J Cardiol 2008;

102:70–6.

2

2

2

2

3

3

3

3

3

3

3

4

4

5

5

5

5

5

5

5

5

5

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

419

6. De Castro S, Caselli S, Di Angelan-tonio E, et al. Relation of left atrialmaximal volume measured by real-time 3D echocardiography to demo-graphic, clinical, and Doppler vari-ables. Am J Cardiol 2008;101:1347–52.

7. Yoshida C, Nakao S, Goda A, et al.Value of assessment of left atrial vol-ume and diameter in patients withheart failure but with normal left ven-tricular ejection fraction and mitralflow velocity pattern. Eur J Echocar-diogr 2009;10:278-81.

8. Pritchett AM, Mahoney DW, JacobsenSJ, Rodeheffer RJ, Karon BL, RedfieldMM. Diastolic dysfunction and left atrialvolume: a population-based study. J AmColl Cardiol 2005;45:87–92.

9. Ujino K, Barnes ME, Cha SS, et al.Two-dimensional echocardiographicmethods for assessment of left atrial vol-ume. Am J Cardiol 2006;98:1185–8.

0. Maddukuri PV, Vieira ML, DeCastro S,et al. What is the best approach for theassessment of left atrial size? Comparisonof various unidimensional and two-dimensional parameters with three-dimensional echocardiographically deter-mined left atrial volume. J Am SocEchocardiogr 2006;19:1026–32.

1. Jenkins C, Bricknell K, Marwick TH.Use of real-time three-dimensionalechocardiography to measure leftatrial volume: comparison with otherechocardiographic techniques. J AmSoc Echocardiogr 2005;18:991–7.

2. Anderson JL, Horne BD, Pennell DJ.Atrial dimensions in health and leftventricular disease using cardiovascu-lar magnetic resonance. J CardiovascMagn Reson 2005;7:671–5.

3. Sievers B, Kirchberg S, Addo M, BakanA, Brandts B, Trappe HJ. Assessmentof left atrial volumes in sinus rhythmand atrial fibrillation using the biplanearea-length method and cardiovascu-lar magnetic resonance imaging withTrueFISP. J Cardiovasc Magn Reson2004;6:855–63.

4. Hudsmith LE, Cheng AS, Tyler DJ, etal. Assessment of left atrial volumes at 1.5Tesla and 3 Tesla using FLASH andSSFP cine imaging. J Cardiovasc MagnReson 2007;9:673–9.

5. Hudsmith LE, Petersen SE, FrancisJM, Robson MD, Neubauer S. Nor-mal human left and right ventricularand left atrial dimensions using steadystate free precession magnetic reso-nance imaging. J Cardiovasc MagnReson 2005;7:775–82.

6. Haider AW, Larson MG, BenjaminEJ, Levy D. Increased left ventricularmass and hypertrophy are associatedwith increased risk for sudden death.

J Am Coll Cardiol 1998;32:1454–9.

37. Mor-Avi V, Sugeng L, Weinert L, etal. Fast measurement of left ventricularmass with real-time three-dimensionalechocardiography: comparison withmagnetic resonance imaging. Circula-tion 2004;110:1814–8.

38. Myerson SG, Montgomery HE,World MJ, Pennell DJ. Left ventric-ular mass: reliability of M-mode and2-dimensional echocardiographic for-mulas. Hypertension 2002;40:673–8.

39. Grothues F, Smith GC, Moon JC, etal. Comparison of interstudy repro-ducibility of cardiovascular magneticresonance with two-dimensionalechocardiography in normal subjectsand in patients with heart failure orleft ventricular hypertrophy. Am JCardiol 2002;90:29–34.

40. Katz J, Milliken MC, Stray-GundersenJ, et al. Estimation of human myocar-dial mass with MR imaging. Radiol-ogy 1988;169:495–8.

41. Lorenz CH, Walker ES, Morgan VL,Klein SS, Graham TP Jr. Normalhuman right and left ventricular mass,systolic function, and gender differ-ences by cine magnetic resonance im-aging. J Cardiovasc Magn Reson1999;1:7–21.

42. Bess RL, Khan S, Rosman HS, Co-hen GI, Allebban Z, Gardin JM.Technical aspects of diastology: whymitral inflow and tissue Doppler im-aging are the preferred parameters?Echocardiography 2006;23:332–9.

43. Ha JW, Oh JK, Redfield MM, UjinoK, Seward JB, Tajik AJ. Triphasic mi-tral inflow velocity with middiastolicfilling: clinical implications and associ-ated echocardiographic findings. J AmSoc Echocardiogr 2004;17:428–31.

44. Garcia MJ, Ares MA, Asher C, Ro-driguez L, Vandervoort P, ThomasJD. An index of early left ventricularfilling that combined with pulsedDoppler peak E velocity may estimatecapillary wedge pressure. J Am CollCardiol 1997;29:448–54.

45. Ommen SR, Nishimura RA, Apple-ton CP, et al. Clinical utility ofDoppler echocardiography and tis-sue Doppler imaging in the estima-tion of left ventricular filling pres-sures: a comparative simultaneousDoppler-catheterization study. Cir-culation 2000;102:1788 –94.

46. D’Souza KA, Mooney DJ, RussellAE, MacIsaac AI, Aylward PE, PriorDL. Abnormal septal motion affectsearly diastolic velocities at the septaland lateral mitral annulus, and im-pacts on estimation of the pulmonarycapillary wedge pressure. J Am SocEchocardiogr 2005;18:445–53.

47. Hillis GS, Moller JE, Pellikka PA, et al.Noninvasive estimation of left ventricu-

lar filling pressure by E/e’ is a powerfulpredictor of survival after acute myocar-

dial infarction. J Am Coll Cardiol 2004;43:360–7.

8. Maceira AM, Prasad SK, Khan M,Pennell DJ. Normalized left ventricu-lar systolic and diastolic function bysteady state free precession cardiovas-cular magnetic resonance. J Cardio-vasc Magn Reson 2006;8:417–26.

9. Kudelka AM, Turner DA, LiebsonPR, Macioch JE, Wang JZ, BarronJT. Comparison of cine magnetic res-onance imaging and Doppler echocar-diography for evaluation of left ven-tricular diastolic function. Am JCardiol 1997;80:384–6.

0. Rubinshtein R, Glockner JF, Feng D,et al. Comparison of magnetic reso-nance imaging versus Doppler echo-cardiography for the evaluation of leftventricular diastolic function in pa-tients with cardiac amyloidosis. Am JCardiol 2009;103:718–23.

1. Rathi VK, Doyle M, Yamrozik J, et al.Routine evaluation of left ventricular dia-stolic function by cardiovascular magneticresonance: a practical approach. J Cardio-vasc Magn Reson 2008;10:36.

2. Nagel E, Stuber M, Burkhard B, et al.Cardiac rotation and relaxation in pa-tients with aortic valve stenosis. EurHeart J 2000;21:582–9.

3. Edvardsen T, Rosen BD, Pan L, et al.Regional diastolic dysfunction in indi-viduals with left ventricular hypertro-phy measured by tagged magnetic res-onance imaging—the Multi-EthnicStudy of Atherosclerosis (MESA).Am Heart J 2006;151:109–14.

4. Valeti VU, Chun W, Potter DD, et al.Myocardial tagging and strain analysisat 3 Tesla: comparison with 1.5 Teslaimaging. J Magn Reson Imaging2006;23:477–80.

5. Petersen SE, Jung BA, Wiesmann F,et al. Myocardial tissue phase map-ping with cine phase-contrast MRimaging: regional wall motion analysisin healthy volunteers. Radiology 2006;238:816–26.

6. Jung B, Foll D, Bottler P, Petersen S,Hennig J, Markl M. Detailed analysis ofmyocardial motion in volunteers andpatients using high-temporal-resolutionMR tissue phase mapping. J Magn Re-son Imaging 2006;24:1033–9.

7. Bergvall E, Cain P, Arheden H, SparrG. A fast and highly automated ap-proach to myocardial motion analysisusing phase contrast magnetic reso-nance imaging. J Magn Reson Imag-ing 2006;23:652–61.

8. Chirillo F, Brunazzi MC, BarbieroM, et al. Estimating mean pulmonarywedge pressure in patients withchronic atrial fibrillation from trans-thoracic Doppler indexes of mitral and

pulmonary venous flow velocity. J AmColl Cardiol 1997;30:19–26.

5

6

6

Kche

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 3 , N O . 4 , 2 0 1 0

A P R I L 2 0 1 0 : 4 0 9 – 2 0

Leong et al.

Echocardiography and Cardiac MR in HFNEF

420

9. Arques S, Roux E, Sbragia P, et al.Usefulness of bedside tissue Dopplerechocardiography and B-type natri-uretic peptide (BNP) in differentiat-ing congestive heart failure fromnoncardiac cause of acute dyspnea inelderly patients with a normal leftventricular ejection fraction and per-manent, nonvalvular atrial fibrilla-tion: insights from a prospective,monocenter study. Echocardiogra-phy 2007;24:499 –507.

0. Okura H, Takada Y, Kubo T, et al.

Tissue Doppler-derived index of left

ventricular filling pressure, E/E’, pre-dicts survival of patients with non-valvular atrial fibrillation. Heart 2006;92:1248–52.

61. Uribe S, Tangchaoren T, Parish V, et al.Volumetric cardiac quantification by us-ing 3D dual-phase whole-heart MRimaging. Radiology 2008;248:606-14.

62. Bolger AF, Heiberg E, Karlsson M, etal. Transit of blood flow through thehuman left ventricle mapped by car-diovascular magnetic resonance.J Cardiovasc Magn Reson 2007;9:

741–7. h

3. Iles L, Pfluger H, Phrommintikul A,et al. Evaluation of diffuse myocar-dial fibrosis in heart failure withcardiac magnetic resonance contrast-enhanced T1 mapping. J Am CollCardiol 2008;52:1574 – 80.

ey Words: echocardiography yardiac magnetic resonance yeart failure with normaljection fraction y diastolic

eart failure y imaging.