VALVULAR HEART DISEASES

Potential role of fibrosis imaging in severe valvularheart diseasePhilippe Debonnaire,1,2 Victoria Delgado,1 Jeroen J Bax1

▸ Additional references arepublished online only. To viewplease visit the journal online(http://dx.doi.org/10.1136/heartjnl-2013-304679).1Department of Cardiology,Leiden University MedicalCenter, Leiden, TheNetherlands2Department of Cardiology,AZ Sint-Jan Bruges, Bruges,Belgium

Correspondence toProfessor Jeroen J Bax,Department of Cardiology,Leiden University MedicalCentre, Albinusdreef 2, Leiden2333 ZA, The Netherlands;j.j.bax@lumc.nl

To cite: Debonnaire P,Delgado V, Bax JJ. HeartPublished Online First:[please include Day MonthYear] doi:10.1136/heartjnl-2013-304679

Currently, the most encountered valve diseasesinclude aortic stenosis (AS) and mitral regurgitation(MR); data from population based studies haveshown that approximately 9% of individuals aged65 years or more have either MR or AS.1 w1 Whileindications for surgery are well defined in symp-tomatic patients, the optimal timing of surgery inasymptomatic severe AS or MR remains controver-sial. Currently, surgical intervention is consideredin the presence of reduced left ventricular ejectionfraction (LVEF), left ventricular (LV) dilatation, pul-monary hypertension, reduced exercise capacity,increased plasma concentrations of biomarkers(eg, N-terminal pro-brain natriuretic peptide(NT-proBNP)), and arrhythmias (eg, atrial fibrilla-tion), since all these variables are associated withworse prognosis if treated medically.w2 w3

However, most of these variables are encounteredonly when AS or MR have progressed significantly,leading to suboptimal clinical outcomes after sur-gery.w2 w3

Accordingly, markers that could identify earlystructural and functional abnormalities of the LVare needed to potentially facilitate the decision fortiming of surgery, thereby improving clinical out-comes. In both AS and MR, ultrastructural changesof the LV with expansion of the extracellularmatrix and fibrosis formation may occur due topressure and volume overload, respectively.w4 w5

Fibrosis causes increased LV stiffness, leading todiastolic dysfunction and subtle worsening of sys-tolic function, whereas overt LV systolic dysfunc-tion (reduced LVEF) will occur later in the courseof AS and MR.Focal fibrosis reflects scar tissue formation by

replacing dead myocardial cells with collagen,which is observed after myocardial infarction, forexample. In valvular heart disease, however, pres-sure or volume overload predominantly causediffuse interstitial fibrosis, a distinct type of fibro-sis.w4–w6 This type of fibrosis increases the intersti-tial collagen without notable cell loss and thereforeit may be (partially) reversible.w6

Recently, a rapid development in non-invasiveimaging technology has occurred which maypermit direct or indirect assessment of LV fibrosis.In particular, new cardiac magnetic resonance(CMR) techniques (with or without contrastagents) permit direct assessment of LV fibrosis,whereas strain imaging with advanced echocardio-graphic techniques or CMR can be used to detectsubtle systolic LV dysfunction (while LVEF is stillnormal), which provides an indirect reflection ofLV fibrosis. This article presents an overview of

these non-invasive imaging techniques and thepotential role for early detection of structural andfunctional LV abnormalities in patients with AS orMR. Table 1 provides a summary of direct andindirect imaging techniques for quantitative orqualitative fibrosis evaluation.2 The differentimaging techniques used for fibrosis detection arediscussed below.

IMAGING TECHNIQUES FOR DIRECTASSESSMENT OF LV FIBROSISDelayed contrast enhanced CMRDelayed contrast enhanced CMR has become thegold standard imaging technique to assess and quan-tify focal fibrosis in the LV. Gadolinium chelates areextracellular contrast media that do not penetrateintact cell membranes and accumulate in the myo-cardial extracellular space. Several minutes afterintravenous administration, the contrast agent istrapped into the expanded extracellular space and(after nullifying the signal of the myocardium) isvisualised as increased signal intensity (white) com-pared to normal myocardium (black) (figure 1).3

However, detection of diffuse LV fibrosis withdelayed contrast enhanced CMR standard techni-ques may be difficult since this technique relies onrelative signal differences between normal myocar-dium and interstitial collagen.

T1 weighted CMRFor assessment of diffuse LV fibrosis, where there isno clear differentiation between normal and dis-eased myocardium, T1 weighted CMR imagingtechniques are increasingly applied. T1 weightedtechniques are based on the energy released by thetissue (protons) after applying radiofrequencypulses. This relaxation process follows an exponen-tial formula that includes a time constant, theso-called T1 time. The shorter the T1 time con-stant is, the faster the relaxation process. T1 timecan be assessed before (native T1) or after (post-contrast T1) contrast administration using a varietyof possible techniques with multiple or singlebreath holds (figures 2 and 3).4

Normal native T1 values range between 900–1100 ms and increase in circumstances of increaseddensity of protons (water content or oedema).4 Themeasurement of native T1 time is also highlydependent on the magnetic field strength andacquisition techniques.5 w7

Post-contrast T1 weighted techniques have beenused more frequently to assess diffuse LV fibrosis.After continuous or bolus gadolinium contrastadministration, the volume distribution of contrast

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 1

Education in Heart Heart Online First, published on June 12, 2014 as 10.1136/heartjnl-2013-304679

Copyright Article author (or their employer) 2014. Produced by BMJ Publishing Group Ltd (& BCS) under licence.

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

media is higher within the interstitium or fibroticmyocardium than in normal myocardium, resultingin a shortened T1 time.4 w7 The relative shorteningin T1 time is related to the extent of myocardialfibrosis. Renal clearance, acquisition protocol, con-trast dose, body composition, and haematocrit maysignificantly affect the absolute post-contrast T1value.5 To partially overcome these limitations,quantification of extracellular volume (ECV) hasbeen developed.4 w7 The ECV is estimated by cal-culating the volume distribution of gadolinium inthe extracellular myocardial space relative to theblood in a dynamic steady state. In myocardialtissue, the contrast exchange rate with the blood ishigher than the net clearance of the contrast from

the blood, which defines the dynamic steady state.This dynamic steady state can be achieved withcontinuous intravenous infusion of gadolinium(until the T1 in the myocardium and blood poolare constant) or following an intravenous bolus ofgadolinium (assuming equilibrium between the con-centration of contrast in the blood pool and myo-cardium at each time point after the bolus).4 5 TheT1 time is then calculated according to theformula: (1−haematocrit)×(post 1/T1myocardium−pre 1/T1myocardium)/(post 1/T1blood− pre 1/T1blood),where the factor (1−haematocrit) represents thevolume distribution of gadolinium in the bloodpool.5 Normal ECV values range between 24–28%of the myocardium.4 Both myocardial T1 and ECV

Table 1 Non-invasive imaging modalities and techniques for left ventricular fibrosis assessment

Availability Fibrosis specificity Limitations Experience in AS Experience in MR

EchocardiographyIBS ++ +++ Modest reproducibility + −TDI ++++ + Angle dependent ++ ++2D speckle tracking ++++ + Vendor variability, low frame rate ++++ ++++

Cardiac magnetic resonanceDelayed enhanced (replacement fibrosis) ++++ ++ Focal fibrosis only ++++ −T1 weighted imaging (diffuse fibrosis) ++ +++ Many confounders, expertise, standardisation ++ −Tissue tagging + + Expertise + +Collagen specific contrast ± +++ Experimental − −

Nuclear imagingPET perfusable water index ++ +++ Radiation, expertise − −PET molecular imaging ± +++ Radiation, expertise, experimental − −SPECT molecular imaging ± +++ Radiation, expertise, experimental − −

Adapted with permission from Jellis et al.2

2D, two dimensional; AS, aortic stenosis; IBS, integrated backscatter; PET, positron emission tomography; MR, mitral regurgitation; SPECT, single photon emission CT; TDI, tissueDoppler imaging.

Figure 1 Delayed contrast enhanced cardiac magnetic resonance patterns that can be observed in patients with severe aortic stenosis. (A) Absenceof delayed enhancement. (B) Anterior and septal sub-endocardial contrast enhancement, similar to that noted in infarcted myocardium. (C) Focalspots of contrast enhancement in the lateral midwall (arrows). (D) Linear septal midwall contrast enhancement. (E, F) Midwall contrast enhancementof the lateral wall (arrows). Reproduced with permission from Dweck et al.3

2 Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

can be visually represented by colour coded mapsof the LV. ECV represents a very promising bio-marker of diffuse myocardial fibrosis closely corre-lated with collagen extent. However, its assessmentis technically challenging and needs further stand-ardisation to allow accurate and reproducible mea-surements during single breath hold without theconfounding effects of heart rate and through-plane motion.4 To optimise T1 scan planning andacquisition, and to standardise analysis, a recentexpert consensus document has been produced.6

Calibrated integrated backscatter withechocardiographyCollagen and water content affect myocardial tissuereflectivity of ultrasound waves. Dedicated offlinecardiac ultrasound software permits evaluation oftissue reflectivity amplitudes, sampled at the pericar-dium and the LV myocardium at different locations.Subsequently, calibrated integrated backscatter (IBS)is calculated by subtracting the mean backscatterintensity of the pericardium from the LV myocar-dium (figure 4).2 Higher IBS values (less negative)suggest larger diffuse fibrosis burden, as validated byin vivo biopsy specimens.w8 Echocardiographic IBSoffers relatively easy and fast assessment of LV

fibrosis, but the main limitations include reducedinter- and intra-observer reproducibility due tonoise, confounding effects of the location of thesample volume, artefacts, and dependence on thesettings of the ultrasound systems.7

Molecular imagingMolecular imaging of fibrosis comprises direct visu-alisation and quantification of radionuclide tracersthat bind to specific molecular or cellular com-pounds involved in the pathogenesis of myocardialfibrosis. Integrins, matrix metalloproteinases, angio-tensin converting enzyme (ACE) inhibitors, angio-tensin II receptor antagonists, factor XIII, andcollagen constitute the main targeted agents formolecular imaging of fibrosis.w6 So far, studiesevaluating the role of molecular imaging to assessfibrosis have focused on animal models of myocar-dial infarction and ischaemic heart failure and thereare no data on fibrosis associated with valvular heartdisease.w6 w9 The use of integrins, matrix metallo-proteinases, and factor XIII as targeted agentspermits assessment of the early inflammatoryresponse and the wound healing process of theinfarcted tissue, and therefore may predict LVremodelling. In contrast, collagen targeted CMR

Figure 2 T1 relaxation time assessment by cardiac magnetic resonance. (A) Sequential images at the same cardiac phase of different heartbeatsare acquired after an inversion pulse, thereby obtaining a multitude of different inversion times. T1 recovers secondary to longitudinal magnetisationparalleled by inversion time increment. (B) T1 relaxation curves are constructed after data sorting by inversion time. Post-contrast shorter T1relaxation time is seen in areas of myocardium with increased interstitium (fibrosis, inflammation) (red curve) versus normal myocardium (greencurve). (C) Epicardial inflammation is seen on the T1 map (red region and yellow arrows). (D) By inverting the pixel values, an R1 map (1/T1) can begenerated providing visualisation of regions of fibrosis similar to conventional late gadolinium enhancement images (bright). Reproduced withpermission from Salerno et al.4

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 3

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

contrast agents, radiolabelled ACE inhibitors orangiotensin receptor antagonists have been devel-oped to characterise postinfarction myocardial fibro-sis and LV remodelling. In mouse models of chronicmyocardial infarction, the use of a gadoliniumbased, collagen targeted contrast agent (EP-3533)for in vivo CMR imaging of myocardial fibrosis hasbeen shown to be feasible.w10 On dynamic T1weighted CMR data, the washout time constants ofthe targeted agent (EP-3533) was significantlylonger than those for gadolinium alone in theregions of scar (194.8±116.8 min vs 25.5±4.2 minand 45.4±16.7 min vs 25.1±9.7 min, respectively,p<0.05 for both), which suggest improvement invisualisation and characterisation of scar.w10

The cardiac renin–angiotensin system isenhanced in cardiac hypertrophy and fibrosis andthe development of targeted radiolabelled agentshas permitted in vivo imaging of ACE activity inthe myocardium with positron emission

tomography (PET) and single photon emission CT(SPECT).8 By using a high affinity analogue of lisin-opril radiolabelled with 99mtechnetium, the tissueupregulation of ACE in heart failure may be visua-lised with SPECT-CT. This molecular imaging mayhelp to identify those patients at high risk of devel-oping overt heart failure in whom a more aggres-sive therapeutic strategy may be needed to improvethe prognosis. However, molecular imaging offibrosis is currently restricted to research andfurther histological validation is needed.

IMAGING TECHNIQUES FOR INDIRECTASSESSMENT OF LV FIBROSISStrain and strain rate imaging withechocardiography and CMRDeformation imaging comprises quantitative assess-ment of the magnitude of myocardial fibre contrac-tion and relaxation (myocardial mechanics).9 Strainrefers to the relative change of myocardial fibre

Figure 3 Native T1 relaxation mapping in aortic stenosis (AS). Mid-ventricular short axis colour coded T1 maps (upper row) and correspondingdelayed contrast enhanced cardiac magnetic resonance images. Examples (from left to right) of a normal individual (T1=944 ms), a patient withmoderate AS and moderate left ventricular hypertrophy (T1=951 ms), and a patient with severe AS and severe hypertrophy (T1=1020 ms).Reproduced with permission from Bull et al.16

Figure 4 Calibrated integrated backscatter (IBS) in aortic stenosis. (A) Normal subject: with values of calibrated IBS of the septum and posteriorwall of −26.9 and −29.3 dB, respectively (after subtracting the value of the pericardium, in red). (B) Patient with severe aortic stenosis showsincreased calibrated IBS (less negative) of the septum (−7.5 dB) and posterior wall (−9.3 dB), suggestive of diffuse fibrosis.

4 Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

length over time and is expressed as a percentage,with positive and negative values reflecting myocar-dial shortening or thickening and lengthening orthinning, respectively. Strain rate refers to thechange of strain per unit of time. Deformationimaging is highly sensitive to detecting subtlechanges in systolic LV function, even before LVEFbecomes reduced (figure 5) or LV dilatationoccurs.10 11

Early loss of longitudinal deformation isobserved in the majority of cardiac disorders and isinversely associated with the presence and extent offibrosis (reduced elasticity) in both ischaemic andnon-ischaemic cardiomyopathies. Intrinsic contract-ility, loading conditions, and chamber geometryinfluence myocardial deformation.12 Accordingly,deformation imaging represents an indirect andnon-specific measure of fibrosis.2

Quantitative deformation analysis can be per-formed with echocardiographic imaging techniquessuch as tissue Doppler imaging (TDI) or twodimensional (2D) speckle tracking echocardio-graphy (STE), but also with CMR using tissuetagging. TDI assesses myocardial tissue velocities tocalculate strain (rate). However, as with anyDoppler technique, TDI derived strain is dependenton the insonation angle and strain can only beassessed accurately in those LV segments which areproperly aligned along the ultrasound beam.7 STE,developed more recently, assesses myocardial dis-placement to calculate deformation, based on 2D(or even 3D) spatial tracking of ‘speckles’ (naturalacoustic markers present in bimodal echocardio-graphic images) throughout the cardiac cycle. Thistechnique operates at lower frame rates and isangle-independent. Therefore, comprehensive func-tion analysis including longitudinal, circumferential,and radial function as well as rotational mechanicscan be studied for all myocardial segments. Recentadvances also allow differentiation between epicar-dial and endocardial layers. High quality grey scaleimages are required for proper tracking and defin-ition of the region of interest confined to the LV

myocardium. Absolute values might differ betweensoftware vendors due to different speckle trackingalgorithms.7 9

CMR tissue tagging is a complex technique ofspatial modulation of cardiac tissue magnetisation.Specific magnetic field gradients and time series ofradiofrequency pulses ‘tag’ the myocardium, creat-ing a dark line grid that can be followed duringsubsequent acquisitions throughout the cardiaccycle. This permits assessment of LV displacementand strain (rate) calculation. Similar to STE, itallows comprehensive analysis of longitudinal, cir-cumferential, and rotational mechanics of the LVmyocardium.w11 The technique is not widely avail-able, requires expertise, and is currently restrictedfor research purposes.

Perfusable tissue fraction and index with PETFocal or diffuse myocardial fibrosis can also bedetected with PET using 15O-labelled water andcarbon monoxide (C15O) as radiotracers and quan-tifying the perfusable tissue fraction and index ofthe LV. The perfusable tissue fraction is defined asthe fraction of tissue capable of exchanging15O-labelled water within a region of interest,while the perfusable tissue index is the proportionof 15O-labelled water perfusable tissue within theanatomic tissue derived from the transmission scan.Increased myocardial fibrosis diminishes exchange-ability of water leading to a low perfusable tissueindex.2 A reduction in the perfusable tissue indexhas been correlated with an increasing extent of LVfibrosis after myocardial infarction.w12 In addition,compared with healthy volunteers, patients withhypertrophic cardiomyopathy show a decreasedperfusable tissue index in the non-hypertrophiedlateral wall, suggesting the presence of diffuse fibro-sis.w13 In light of these observations, it would beexpected that patients with severe AS and LVhypertrophy, for example, could also show areduced perfusable tissue index, although currentlyno such data are available.

Figure 5 Differences in left ventricular strain in patients with normal ejection fraction. (A) Normal subject with normal global longitudinal strain(GLS, −21.5%) and left ventricular ejection fraction (LVEF 65%). (B) Patient with severe aortic stenosis, LV hypertrophy, and subclinical LVdysfunction shown by reduced GLS (−11.3%) despite preserved LVEF (62%). (C) Impaired GLS (−18.6%) in a patient with LV dilatation, due tosevere organic mitral regurgitation, despite high normal LVEF (67%).

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 5

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

RELEVANCE OF FIBROSIS IMAGING IN AORTICSTENOSISIn AS, the increased wall stress due to pressureoverload is initially compensated by concentric LVhypertrophy to maintain normal cardiac output.This structural remodelling often coincides withthe development of predominantly reactive (diffuseinterstitial) and later replacement (focal) fibrosis asa result of a complex interplay, mainly determinedby up-regulation of the renin–angiotensin–aldosterone system, transforming growth factor ß,and tissue inhibitors of matrix metalloproteina-ses.w4 w14 w15 Interestingly, the fibrotic response(similar to the extent of LV hypertrophy) may varysubstantially between patients despite similar ASseverity, indicating multifactorial pathophysiology.3

While calcification is established as the main deter-minant of progressive valve narrowing, fibrosisinitiates transition from compensating LV hyper-trophy towards adverse LV remodelling, functionalimpairment, and poor outcome in patients withsevere AS (table 2).w4 w14 w15

The pathophysiological mechanism relating myo-cardial fibrosis to adverse prognosis in patients withAS is debated, but probably involves deteriorationof diastolic and systolic LV function as well as sub-strate formation for atrial/ventricular tachyarrhyth-mias.w14 Hence, fibrosis could be a strongbiomarker of early cardiac dysfunction in severeAS, implying increased mortality risk, and mayprove useful for selecting optimal timing for aorticvalve replacement. Currently, symptomatic status orLVEF <50% are considered a class I indication forvalve replacement in patients with severe AS.w3

However, most patients are asymptomatic and havea normal LVEF. Moreover, a few non-randomised

studies have demonstrated significant survivalbenefit with earlier intervention.w16 w17 These find-ings point to the clinical need for additionalmarkers in patients with severe AS to optimise thetiming of intervention. A non-invasive imagingtechnique that detects and quantifies fibrosis, andcorrelates to outcome, may be of use.

Experience with fibrosis imaging in aorticstenosisFocal fibrosis is detected in 30–60% of patients withsevere AS using delayed contrast enhanced CMRand comprises between 3–7% of the LV myocar-dium, depending on the population.3 13–15 w18 w19

Often, the distribution of fibrosis is patchy, multi-focal, and restricted to (often basal) subendocardialor midwall myocardial layers (figure 1).3 13 w18

Excessive activation of the cardiac renin–angiotensinsystem, direct mechanical forces, and ischaemia dueto an imbalance between the increased myocardialmass and the relatively reduced capillary flowreserve are proposed pathophysiological mechan-isms for this replacement fibrosis.3 13 w18 A positivecorrelation between the degree of fibrosis on theone hand, and both LV mass and valvuloarterialimpedance on the other, partly explains the higherfibrotic burden in the LV demonstrated in low gradi-ent severe AS patients, irrespective of LVEF.w20

More extensive fibrosis on delayed contrastenhanced CMR in AS patients is associated withheart failure symptoms and increased NT-proBNPconcentrations.15 In 58 patients with symptomaticsevere AS, Weidemann et al15 observed that patientswithout fibrosis showed significant improvements inNew York Heart Association (NYHA) functionalclass, LVEF, and LV reverse remodelling after

Table 2 Correlations between the extent and severity of left ventricular fibrosis on non-invasive imaging and clinically relevant parameters inpatients with aortic stenosis

Parameters Positive correlation with LV fibrosis Negative correlation with LV fibrosis

Valve haemodynamics Mean aortic valve pressure gradientPeak aortic valve pressureValvuloarterial impedance

LV remodelling LV mass (index)LV volumeLA volume

LV systolic function LVEFLV stroke volumeLongitudinal strain (rate)Systolic mitral annulus displacement (TDI or M mode)

LV end systolic pressure

LV diastolic function LV end diastolic pressureE/E’

Biomarkers NT-proBNP

Baseline clinical variables NYHA class

Variables after aortic valve replacement NYHA class improvementReverse LV remodellingImprovement LV systolic function

LA, left atrial, LV, left ventricular; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; NT-proBNP, N-terminal pro-brain natriuretic peptide; TDI, tissue Dopplerimaging.

6 Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

surgical valve replacement, contrary to patients whoshowed focal fibrosis in ≥2 LV segments. Focal fibro-sis remained unchanged for all patient groups post-operatively, indicating irreversible myocardialdamage (macroscopic scar formation).15 A morerecent study (including 28 symptomatic AS patients)extended these results, showing that the degree offibrosis was independently associated with survival(irrespective of age), LVEF, and symptomaticstatus.w18 This independent relation was confirmedby Dweck et al3 in 143 patients with moderate(40%) and severe (60%) AS. Those patients withmidwall hyperenhancement had fivefold increasedall-cause mortality (HR 5.35, 95% CI 1.16 to24.56). Patients with midwall enhancement whounderwent aortic valve surgery had better survivalthan those treated conservatively, but worse survivalcompared to surgically treated patients withoutfibrosis. Interestingly, almost half of the deathsoccurred in patients with only moderate AS andmidwall fibrosis, underscoring the prognostic rele-vance of focal fibrosis.3

Focal fibrosis, however, occurs at later diseasestages, while diffuse fibrosis is the predominantpattern in AS patients. Higher septal IBS values,correlating with diffuse fibrosis histology, wereshown in 35 severe AS patients (figure 4).w21 Thepresence of significant LV fibrosis—defined as avalue of end-diastolic IBS of the septum indexed tothe pericardial IBS >56.6%—permitted identifica-tion of patients with LV dysfunction. In addition,significant LV fibrosis assessed with calibrated IBSwas associated with a lack of LVEF improvementafter valve replacement.w21 A reduction in IBSvalues after surgical replacement suggests at leastpartial reversibility of diffuse fibrosis (or reducedwater content due to cell volume reduction) in ASpatients and could predict LV reverse remodel-ling.w22 Recent development of T1 mapping tech-niques has renewed interest in diffuse fibrosisimaging in this group of patients.Various studies have correlated several parameters

derived from native and post-contrast T1 mappingCMR techniques. In 18 patients with severe AS thecorrelation between contrast enhanced T1 weightedCMR derived ECV and histological quantification ofmyocardial fibrosis was assessed.w23 ECV showed astrong linear correlation with histological collagenvolume fraction (r2=0.86), suggesting the potential

of T1 weighted CMR for fibrosis assessment.w23

However, using native T1 mapping, Bull et al16

showed a modest correlation between native T1values and histological collagen volume fraction in19 patients with severe AS undergoing aortic valvereplacement (figure 3). These findings suggest a sig-nificant variability among the different methodolo-gies to assess diffuse myocardial fibrosis (table 3).16 17

w18 w21 w23 Independently of the T1 mapping CMRtechnique used, all studies showed that patients withAS have increased diffuse myocardial fibrosis.16 w23

Bull et al16 reported higher native T1 values in symp-tomatic (1014±38 ms) versus asymptomatic (972±33 ms) AS patients. In addition, Flett et al14

showed higher ECV (18.1±8.1% vs 13.4±6.5%,p<0.05) in 63 patients with AS as compared to 30control patients, and this parameter was an importantpredictor of 6 min walking distance. After valvesurgery, ECV (as a measure of diffuse fibrosis)remained unchanged and 80% of patients who diedwithin the first 6 months were within the upper ECVtertile.14 However, the considerable overlap of theT1 mapping results between the patients with ASversus control individuals, as well as reproducibilityissues, may currently limit the use of this techniquefor individual patients.14 In addition to direct fibrosisassessment, reduction in LV function has been usedas an indirect marker of LV fibrosis. Impaired LVEFin patients with AS is inversely correlated with exten-sive histological fibrosis (r=−0.57).w15 However, thevast majority of patients with AS have preservedLVEF, since LVEF mainly reflects radial function,which is preserved in hypertrophied hearts until theend stage of the disease.w24 w25

In patients with AS, however, fibrosis mainlyaffects subendocardial to midwall layers of the myo-cardium that determine longitudinal LV function.Weidemann et al15 showed an inverse relationshipbetween the extent of fibrosis and LV longitudinalstrain (rate) in patients with severe AS. Manypatients with severe AS and preserved LVEF presentwith impaired longitudinal deformation.11 In moreadvanced disease stages, impaired LV radial and cir-cumferential strain, corresponding to more trans-mural myocardial fibrosis, have been reported.11

The main determinants of global LV longitudinalstrain in AS patients are stenotic valve severity,global afterload (valvulo-arterial impedance), LVmass, fibrosis, and contractility.11 12 15 w20 w26 w27

Since all these determinants show a significant rela-tion with outcome in patients with AS, detection ofsubtle LV dysfunction by strain imaging mayimprove risk stratification (and therapeutic decisionmaking). Indeed, reduced LV longitudinal functionrelates to exercise intolerance and was the singleindependent predictor of mortality or heart failurehospitalisation after valve replacement in symptom-atic patients with severe AS.w27 w28 Importantly, LVlongitudinal deformation was an independent pre-dictor of death or symptom driven valve replace-ment in asymptomatic patients with severe AS andpreserved LVEF. In particular, a global LV longitu-dinal strain value >−15% indicated worse event-free survival.w29

Table 3 Correlations of non-invasive imaging versus histology based fibrosisassessment in valvular heart disease patients (all p<0.05)

Study Imaging techniqueNumber of patients,type of valve disease Correlation

Di Bello et alw21 Calibrated integrated backscatter 35, AS r=0.74Cameli et al17 Longitudinal left atrial strain 46, MR r=−0.82*Azevedo et alw18 Delayed contrast enhancement 28, AS r=0.69Bull et al16 Native T1 relaxation time 19, AS r=0.66Flett et alw23 T1 based extracellular volume 18, AS r2=0.86

*Correlated to left atrial biopsy.AS, aortic stenosis; MR, mitral regurgitation.

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 7

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

RELEVANCE OF FIBROSIS IMAGING IN MITRALREGURGITATIONAssessment of myocardial fibrosis in patients withorganic MR is of interest since early detection ofassociated LV and left atrial (LA) structural changesmay help to identify patients who may benefit fromsurgical valve repair while still asymptomatic. Incontrast, in functional (ischaemic) MR, there isoften extensive focal fibrosis with severe structuralLV changes due to scar formation after previousinfarction, and therefore detection of diffuse fibro-sis may not help in the timing of surgery.Organic MR comprises primary abnormalities of

the mitral valve morphology and causes volumeoverload of both the LV and the LA, leading tochamber dilatation and eccentric LV hypertrophy.In contrast to concentric pressure overload LVhypertrophy (as noted with AS), the eccentric LVhypertrophy caused by chronic volume overloadassociated with organic MR is less pro-fibrotic.w30

A recent study involving rat surgical models ofpressure and volume overload attempted to explainthis phenomenon.w31 Compared with pressureoverload models, volume overload models showedless myocardial ischaemia and replacement fibro-sis.w31 In addition, proteolytic activity in largeanimal models of MR has been related to reducedsupport and content of extracellular matrix ineccentric hypertrophy, thereby facilitating LV dilata-tion.w31 w32 Although MR is less pro-fibrotic thanAS, it is suggested that increased interstitial tissuein the LV (found in biopsies obtained in patientsundergoing mitral valve repair/replacement) contri-butes to the development of cardiac failure.w33

Moreover, histology revealed a larger content ofsubendocardial diffuse LV interstitial fibrosis inpatients with MR as compared to patients withoutMR (18% vs 4%, p<0.05).w34 Similar to AS, com-pelling evidence indicates that LV fibrosis in MRrelates to LV dilatation, functional impairment, andadverse outcome. These findings may have animpact on timing of surgery in patients with severeorganic MR. Development of symptoms or LV dys-function (LVEF ≤60% or LV end-systolic diameter≥40–45 mm) are class I indications for mitral valvesurgery in patients with severe organic MR.w3

However, these indications should probably not beawaited, as surgical outcome seems better in theabsence of these characteristics.18 w35 w36

Additional risk stratification is therefore warrantedand (indirect) fibrosis imaging may be ofimportance.

Experience with fibrosis imaging in organic MRIn organic MR, indirect myocardial fibrosis assess-ment has been achieved using strain imaging. Incontrast, CMR, IBS or molecular imaging studies inpatients with MR are scarce. Similar to patientswith AS, most MR patients have preserved LVEF.Of note, in significant MR supra-normal LVEFvalues define normal systolic function due to thecoincidence of increased preload and reduced after-load during the initial compensated phase.

Subclinical LV dysfunction despite preserved LVEF,assessed by impaired LV longitudinal strain,however, is often encountered in asymptomaticpatients with severe MR (figure 5).10 Moreadvanced disease stages are associated withimpaired LV circumferential and radial strain.w37

The main determinants of LV longitudinal strain inMR are regurgitant volume, LV geometry (dilata-tion) and intrinsic contractility, related to myocar-dial fibrosis.10 w38 w39 Reduced strain in MRpatients may therefore parallel LV geometricchanges (dilatation) rather than contractility impair-ment; thus, some authors have advocated correct-ing the LV strain values for LV dimensions.10 12

In a large series of 233 patients with moderate-severe MR and overall preserved LVEF, Witkowskiet al19 showed that impaired longitudinal LV straintogether with LV end-systolic dimension predictpostoperative LV dysfunction (LVEF <50%), inde-pendently of baseline LVEF ≤60%, symptoms, andatrial fibrillation. In particular, a global LV longitu-dinal strain value ≥ −19.9% predicted long termLV dysfunction after mitral valve repair with a sen-sitivity and specificity of 90% and 79%, respect-ively.19 Additionally, impaired recruitment oflongitudinal contractility by strain imaging duringexercise was also associated with postoperative LVdysfunction in patients with preserved LVEF under-going surgery for severe MR.w40 Preoperative high-normal LVEF may be misleading and mask latentLV dysfunction that is only observed postopera-tively when acute preload reduction leads to anLVEF decrease below normal values.19 w41

Whether impaired LV longitudinal strain also pre-dicts worse survival after valve interventionremains to be demonstrated.The LA has also been the subject of research.

Recently, a close inverse correlation between LAglobal (reservoir) strain and histological interstitialfibrosis in patients with severe organic MR(r=−0.82) was shown (figure 6).17 The potentialclinical relevance of such findings was explored inanother study involving 121 patients with severeMR.20 Impaired global LA longitudinal (reservoir)strain was related to long term mortality aftermitral valve surgery, incremental to guideline basedindications for mitral surgery and LA size.20 w3 LAstrain could potentially predict postoperative sur-vival in patients with severe MR, without guide-lines based risk factors (figure 7).20 Apart frommyocardial deformation, both delayed contrastenhancement CMR and, more recently, T1weighted imaging of the LA have shown a correl-ation with LA fibrosis.w42 w43 However, no dataexist on the assessment of LA fibrosis with CMRtechniques in patients with severe MR.

UNRESOLVED ISSUESBased on these experimental and clinical studies, thequestion remains whether fibrosis may help in reach-ing a decision about when to intervene in asymp-tomatic AS patients with preserved LVEF—specifically since most of the currently availablestudies on direct or indirect fibrosis imaging have

8 Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

been performed in patients with indications forsurgery according to current guideline criteria.w2 w3

Various other issues need further study to define theprecise role of LV fibrosis for risk stratification anddecision-making strategies in asymptomatic patientscurrently not considered for surgery. For example, isfibrosis independent of and superior to other cur-rently used methods of risk assessment, such as

NT-proBNP? In addition, for risk estimation,should LV fibrosis be considered as a continuum orshould a threshold be applied (if more than a certainpercentage fibrosis is present in the LV, then surgicalintervention should be considered)? Which fibrosisneeds to be detected—focal or diffuse fibrosis?Which non-invasive imaging technique is optimalfor the assessment and quantification of LV fibrosis?

Figure 6 Relation between extent of left atrial fibrosis and left atrial reservoir strain in patients with severe mitral regurgitation. Four patientsoperated on for severe mitral regurgitation with (from left to right) progressive left atrial dilatation corresponding to more impaired left atrialfunction (reservoir strain, decreasing from 45% to 9%) and larger extent of fibrosis on histology of the left atrial free wall (haematoxylin and eosin,and Masson’s trichrome staining). Reproduced with permission from Cameli et al.17

Figure 7 Prognostic value of left atrial (LA) strain in patients undergoing mitral valve surgery for severe organic mitral regurgitation.Dichotomising the population based on a cut-off value of LA strain of 24%, patients with an LA strain >24% had better cumulative survivalcompared with patients with an LA strain ≤24%, independently of symptoms or the presence of guidelines based surgical criteria. Reproduced withpermission from Debonnaire et al.20

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 9

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

And what will be the precise role of indirect fibrosisassessment using LV strain?Moreover, it is not clear whether a potential rela-

tion between fibrosis and outcome can be extrapo-lated to patients with different AS severity. Finally,does infarct related LV fibrosis alter the predictivevalue in AS patients?Similarly, for asymptomatic patients with severe

organic MR, the key issue is whether strainimaging is specific enough for risk stratification ofindividual patients in order to justify early interven-tion in asymptomatic patients with preserved LVEF.In addition, will LV or LA strain provide theoptimal risk stratification and therapy guidance in

these patients? Will strain imaging be independentand superior to other risk factors, includingNT-proBNP? And if considered for risk stratifica-tion and therapy guidance, will a threshold of LVand/or LA strain be used or a continuum?Moreover, the value of imaging techniques such ascontrast enhanced CMR or T1 mapping in MRshould be explored both for the LV and the LA.

CONCLUSIONIf successfully performed before complicationsarise, surgical correction of AS or MR may result inpatients achieving a normal life expectancy. Timingof surgery in asymptomatic patients is currently con-sidered in the presence of reduced LVEF, LV dilata-tion, pulmonary hypertension, reduced exercisecapacity, increased plasma values of biomarkers oratrial fibrillation. These markers, however, occurrelatively late in the disease progression, andoutcome after surgery is suboptimal. LV fibrosis mayoccur at an earlier stage in the valve disease process,and could be a future marker for improving thera-peutic decision making. Novel non-invasive imagingtechniques have been developed to detect fibrosisdirectly or to assess subtle changes in LV systolic dys-function (while LVEF is still preserved), secondaryto LV fibrosis. Future research is needed to deter-mine whether LV fibrosis assessment with theseimaging techniques may further refine the timing ofsurgery in severe AS or MR, and whether this willimprove outcome after surgery.

Contributors All authors contributed to this paper.

Competing interests In compliance with EBAC/EACCMEguidelines, all authors participating in Education in Heart havedisclosed potential conflicts of interest that might cause a bias inthe article. PD was partially supported by a Sadra Medical ResearchGrant (Boston Scientific) and with a European Association ofCardiovascular Imaging (EACVI) Research Grant, and receivesspeaker fees from Abbott Vascular (MitraClip Crossroads Faculty)and Philips Healthcare. VD received consulting fees from St JudeMedical and Medtronic. The Department of Cardiology of LeidenUniversity Medical Centre received research grants from Biotronik,Medtronic, Boston Scientific, Lantheus Medical Imaging, EdwardsLifesciences, St Jude Medical, GE Healthcare. No specific financialsupport for this work is involved.

Provenance and peer review Commissioned; externally peerreviewed.

REFERENCES1 Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke

statistics—2013 update: a report from the American HeartAssociation. Circulation 2013;127:e6–245.

▸ Landmark report on prevalence and incidence of cardiovasculardiseases.

2 Jellis C, Martin J, Narula J, et al. Assessment of nonischemicmyocardial fibrosis. J Am Coll Cardiol 2010;56:89–97.

3 Dweck MR, Joshi S, Murigu T, et al. Midwall fibrosis is anindependent predictor of mortality in patients with aortic stenosis.J Am Coll Cardiol 2011;58:1271–9.

▸ Large study correlating focal left ventricular fibrosis and prognosisof patients with aortic stenosis.

4 Salerno M, Kramer CM. Advances in parametric mapping withCMR imaging. JACC Cardiovasc Imaging 2013;6:806–22.

5 Kellman P, Wilson JR, Xue H, et al. Extracellular volume fractionmapping in the myocardium, part 1: evaluation of an automatedmethod. J Cardiovasc Magn Reson 2012;14:63.

6 Moon JC, Messroghli DR, Kellman P, et al. Myocardial T1mapping and extracellular volume quantification: a Society forCardiovascular Magnetic Resonance (SCMR) and CMR Working

Fibrosis imaging in severe valvular heart disease: key points

▸ In left sided valvular heart disease, pressure overload (aortic stenosis) ismore pro-fibrotic than volume overload (mitral regurgitation).

▸ Fibrosis causes adverse LV remodelling and functional impairment, and isassociated with poor clinical outcome.

▸ Non-invasive imaging techniques to assess direct LV fibrosis include:– Contrast enhanced CMR:

W focal (replacement) fibrosis– T1 weighted cardiac magnetic resonance:

W diffuse fibrosis: T1 mapping, extracellular volume– Echocardiographic calibrated integrated backscatter– Molecular imaging:

W collagen targeted agents with CMRW radiolabelled ACE inhibitors and angiotensin receptor antagonists

with SPECT or PET▸ Non-invasive imaging techniques to assess indirect LV fibrosis include:

– Strain and strain rate imaging:W tissue Doppler imagingW speckle tracking echocardiographyW tagged CMR

– Perfusable tissue fraction and index with PET

You can get CPD/CME credits for Education in Heart

Education in Heart articles are accredited by both the UK Royal College ofPhysicians (London) and the European Board for Accreditation in Cardiology—youneed to answer the accompanying multiple choice questions (MCQs). To accessthe questions, click on BMJ Learning: Take this module on BMJ Learning fromthe content box at the top right and bottom left of the online article. For moreinformation please go to: http://heart.bmj.com/misc/education.dtl▸ RCP credits: Log your activity in your CPD diary online (http://www.

rcplondon.ac.uk/members/CPDdiary/index.asp)—pass mark is 80%.▸ EBAC credits: Print out and retain the BMJ Learning certificate once you

have completed the MCQs—pass mark is 60%. EBAC/ EACCME Credits cannow be converted to AMA PRA Category 1 CME Credits and are recognisedby all National Accreditation Authorities in Europe (http://www.ebac-cme.org/newsite/?hit=men02).

Please note: The MCQs are hosted on BMJ Learning—the best available learningwebsite for medical professionals from the BMJ Group. If prompted, subscribersmust sign into Heart with their journal’s username and password. All users mustalso complete a one-time registration on BMJ Learning and subsequently log in(with a BMJ Learning username and password) on every visit.

10 Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

Group of the European Society of Cardiology consensusstatement. J Cardiovasc Magn Reson 2013;15:92.

▸ Consensus document on acquisition and postprocessing of CMRdata to assess myocardial T1 mapping and extracellular volumequantification.

7 Mor-Avi V, Lang RM, Badano LP, et al. Current and evolvingechocardiographic techniques for the quantitative evaluation ofcardiac mechanics: ASE/EAE consensus statement on methodologyand indications endorsed by the Japanese Society ofEchocardiography. J Am Soc Echocardiogr 2011;24:277–313.

▸ Consensus document on current status of echocardiographictechniques for the assessment of cardiac mechanics.

8 Dilsizian V, Zynda TK, Petrov A, et al. Molecular imaging ofhuman ACE-1 expression in transgenic rats. JACC CardiovascImaging 2012;5:409–18.

9 Blessberger H, Binder T. Two dimensional speckle trackingechocardiography: basic principles. Heart 2010;96:716–22.

10 Marciniak A, Claus P, Sutherland GR, et al. Changes in systolicleft ventricular function in isolated mitral regurgitation. A strainrate imaging study. Eur Heart J 2007;28:2627–36.

11 Ng AC, Delgado V, Bertini M, et al. Alterations in multidirectionalmyocardial functions in patients with aortic stenosis and preservedejection fraction: a two-dimensional speckle tracking analysis. EurHeart J 2011;32:1542–50.

12 Bijnens BH, Cikes M, Claus P, et al. Velocity and deformationimaging for the assessment of myocardial dysfunction. Eur JEchocardiogr 2009;10:216–26.

13 Debl K, Djavidani B, Buchner S, et al. Delayed hyperenhancementin magnetic resonance imaging of left ventricular hypertrophycaused by aortic stenosis and hypertrophic cardiomyopathy:visualisation of focal fibrosis. Heart 2006;92:1447–51.

14 Flett AS, Sado DM, Quarta G, et al. Diffuse myocardial fibrosis insevere aortic stenosis: an equilibrium contrast cardiovascularmagnetic resonance study. Eur Heart J Cardiovasc Imaging2012;13:819–26.

15 Weidemann F, Herrmann S, Stork S, et al. Impact of myocardialfibrosis in patients with symptomatic severe aortic stenosis.Circulation 2009;120:577–84.

16 Bull S, White SK, Piechnik SK, et al. Human non-contrast T1values and correlation with histology in diffuse fibrosis. Heart2013;99:932–7.

17 Cameli M, Lisi M, Righini FM, et al. Usefulness of atrialdeformation analysis to predict left atrial fibrosis and endocardialthickness in patients undergoing mitral valve operations for severemitral regurgitation secondary to mitral valve prolapse. Am JCardiol 2013;111:595–601.

18 Suri RM, Enriquez-Sarano M. Association between early surgicalintervention vs watchful waiting and outcomes for mitral regurgitationdue to flail mitral valve leaflets. JAMA 2013;310:609–16.

▸ Large registry comparing the long term prognosis of patients withsevere mitral regurgitation due to flail mitral leaflets whounderwent early surgical mitral valve repair and patients whowere medically treated.

19 Witkowski TG, Thomas JD, Debonnaire PJ, et al. Globallongitudinal strain predicts left ventricular dysfunction aftermitral valve repair. Eur Heart J Cardiovasc Imaging 2013;14:69–76.

20 Debonnaire P, Leong DP, Witkowski TG, et al. Left atrial functionby two-dimensional speckle-tracking echocardiography in patientswith severe organic mitral regurgitation: association withguidelines-based surgical indication and postoperative (long-term)survival. J Am Soc Echocardiogr 2013;26:1053–62.

Debonnaire P, et al. Heart 2014;0:1–11. doi:10.1136/heartjnl-2013-304679 11

Education in Heart

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

doi: 10.1136/heartjnl-2013-304679 published online June 12, 2014Heart

 Philippe Debonnaire, Victoria Delgado and Jeroen J Bax valvular heart diseasePotential role of fibrosis imaging in severe

http://heart.bmj.com/content/early/2014/06/12/heartjnl-2013-304679.full.htmlUpdated information and services can be found at:

These include:

Data Supplement http://heart.bmj.com/content/suppl/2014/06/12/heartjnl-2013-304679.DC1.html

"Supplementary Data"

References http://heart.bmj.com/content/early/2014/06/12/heartjnl-2013-304679.full.html#ref-list-1

This article cites 20 articles, 9 of which can be accessed free at:

P<P Published online June 12, 2014 in advance of the print journal.

serviceEmail alerting

the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from

CollectionsTopic

(1905 articles)Echocardiography   � (4347 articles)Clinical diagnostic tests   �

(2480 articles)Acute coronary syndromes   � (203 articles)Mitral valve disease   � (330 articles)Aortic valve disease   �

(2651 articles)Hypertension   � (7843 articles)Drugs: cardiovascular system   �

(23 articles)Valvular heart disease   � (476 articles)Education in Heart   �

 Articles on similar topics can be found in the following collections

Notes

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on July 14, 2014 - Published by heart.bmj.comDownloaded from