PUBLISHED VERSION

Leong, Darryl P.; Madsen, Per Lav; Selvanayagam, Joseph Non-invasive evaluation of myocardial fibrosis: implications for the clinician, Heart, 2010; 96(24):2016-2024.

Originally published by BMJ – http://heart.bmj.com/content/96/24/2016

© The Authors

http://hdl.handle.net/2440/62338

PERMISSIONS

http://www.bmj.com/about-bmj/resources-authors/forms-policies-and-checklists/copyright-open-access-and-permission-reuse

Thus authors may use their own articles for the following non commercial purposes without asking our permission (and subject only to acknowledging first publication in the BMJ and giving a full reference or web link, as appropriate).

• Posting a pdf of their own article on their own personal or institutional website for which no charge for access is made.

10th January 2013

NON-INVASIVE IMAGING

Non-invasive evaluation ofmyocardial fibrosis: implications forthe clinicianDarryl P Leong,1,2 Per Lav Madsen,1 Joseph B Selvanayagam1

The presence and extent of myocardial fibrosis arekey determinants of response to treatment andprognosis in a number of cardiac conditions. Untilrecently, myocardial fibrosis could only be detectedante mortem by endomyocardial biopsy, which isassociated with procedural risk and sampling error.The development of novel cardiac imaging tech-niques and serum assays now permits the accuratedetection and quantification of myocardial fibrosis.These have yielded new insights into disease prog-nosis and response to treatment.

PATHOGENESIS OF MYOCARDIAL FIBROSISMyocardial fibrosis develops in response to a cardiacinsult, which may include ischaemia, pressure orvolume overload, viral infection, or geneticallymediated injury as in hypertrophic cardiomyo-pathy. Net collagen deposition results from animbalance of its synthesis relative to degradation. Anumber of enzymes have been identified as poten-tial mediators of myocardial extracellular matrixturnover. The matrix metalloproteinases (MMPs)are a family of at least 20 calcium dependentendopeptidases that digest interstitial constituents.The various MMPs have different substratesdMMP-1 and -13 are collagenases and MMP-2 and -9gelatinases. Left ventricular (LV) myocardial MMPactivity in idiopathic dilated cardiomyopathy andischaemic cardiomyopathy has been shown to begreater than in normal hearts.1 Abolition of MMP-9synthesis has been associated with reducedmyocardial fibrosis and improved LV function ina rodent model of pressure overload.w1 The tissueinhibitors of matrix metalloproteinase (TIMPs) area family of four proteins that bind to and inhibit theeffects of MMPs (figure 1). TIMP-1 expression isreduced in explanted hearts from patients with bothischaemic and non-ischaemic cardiomyopathy.w2

Regardless of the aetiology and/or molecularcascade resulting in collagen deposition, the pres-ence of myocardial fibrosis has both mechanicaleffects on cardiac function, mediated by increasedmyocardial stiffness, and electrophysiologicaleffects, by acting as substrate for re-entry andarrhythmia.

CARDIAC MAGNETIC RESONANCE (CMR)IMAGINGCMR imaging is established as a major techniquein clinical cardiology and cardiovascular research.

Allied to its well recognised role in accurate andreproducible measurement of LV volumes andmass, CMR offers an unprecedented ability todetect and quantify myocardial fibrosis. It providesexquisite three dimensional images allowingconcurrent assessment of myocardial structure,function, and tissue characterisation.

Late gadolinium enhancement (LGE) techniqueGadolinium-DTPA (Gd-DTPA) is a paramagneticcontrast agent that is used to delineate areas ofinjured myocardium. Gd-DTPA reduces hydrogeneproton T1 relaxation times in proportion to its localconcentration. In T1 weighted imaging, tissueswith a shorter T1 relaxation time exhibit greatersignal intensity than those with longer T1 relaxa-tion times. Gd-DTPA equilibrates rapidly betweenintravascular and interstitial spaces, but is excludedfrom the intracellular compartment by the intactcell membrane.w3 Following intravenous adminis-tration, altered wash-in/wash-out kinetics and anincreased volume of distribution in damaged tissue(owing to interstitial oedema and/or loss of cellmembrane integrity) account for its pattern ofappearance in these regions.2 Late or ‘delayed’imaging (after at least 5 min post-contrast) with T1weighted inversion recovery sequences identifiesconditions associated with expansion of theextracellular space and hence also with fibrosis.

Ischaemic heart diseaseIn the clinical realm, contractile abnormalities inpatients with ischaemic heart disease may occur asa consequence of stunning, hibernation and scar,with the relative importance of these factorsvarying both between and within myocardialsegments and dynamically over time. Detectingdysfunctional and scarred myocardium (figure 2) asopposed to dysfunctional but viable myocardiumis of scientific and clinical significance and there isnow a reasonable body of non-randomised evidencesupporting revascularisation of hibernatingmyocardium.w4

In the setting of chronic ischaemic cardiomyo-pathy, to date there have been two single centreclinical studies examining the utility of the trans-mural extent of LGE in predicting recoveryof contractile function. The first was performedby Kim et al in a cohort of 41 patients under-going revascularisation by either percutaneous

< Additional references arepublished online only. To viewthese files, please visit thejournal online (http://heart.bmj.com).1Department of CardiovascularMedicine, Flinders University ofSouth Australia, FlindersMedical Centre, Adelaide,Australia2Discipline of Medicine,University of Adelaide, Adelaide,Australia

Correspondence toDr Joseph B Selvanayagam,Department of CardiovascularMedicine, Flinders MedicalCentre, Bedford Park, SA,Adelaide 5042, Australia;joseph.selva@health.sa.gov.au

2016 Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

transluminal coronary angioplasty or coronaryartery bypass grafting. They found that the likeli-hood of improvement in regional function afterrevascularisation decreased progressively as thetransmural extent of LGE before revascularisationincreased.3 This was subsequently confirmed ina study by Selvanayagam et al, which exclusivelyexamined patients after surgical revascularisation.4

The ability of LGE-CMR to evaluate thosesegments that have severe dysfunction (and oftenthe most difficult to evaluate with other imagingtechniques) with high diagnostic accuracy is one of

the strengths of the LGE-CMR technique. In addi-tion, with excellent spatial resolution and contrastnoise ratio, LGE-CMR has high sensitivity to detectviable myocardium and thus may provide moresensitive (albeit less specific) prediction of recoveryof segmental function than inotropic contractilereserve.The analysis of what constitutes LGE positivity

in human CMR studies of ischaemic cardiomyo-pathy has been controversial. In earlier animal modelstudies performed by Kim et al, a signal intensitycut-off 2SD above normal myocardium identifiedaccurately the extent of myocardial infarction.w5 Arecent elegant study by Amado et al in a caninemodel of myocardial infarction demonstrated thatthe full width at half-maximum technique forquantification of volume of myocardial enhance-ment was the most accurate.w6 This methodrequires selection of a seed point within hyper-enhanced myocardium. Software then identifies allpixels with signal intensity >50% of this point.The maximal signal intensity within this region isdetermined, and the final scar extent is defined astissue exhibiting signal intensity >50% that of themaximal signal intensity. Quantifying scar size bytissue with signal intensity 2SD above normalmyocardium, as originally described by Kim et al,was found to overestimate infarct size, whereas fiveand six SD thresholds were closer to histopathology.Myocardial infarct size as quantified by the late

gadolinium technique has been demonstrated to bea powerful predictor of mortality and adverse LVremodellingdmore powerful than left ventricularejection fraction (LVEF).w7 w8 Scar burden asquantified by CMR, or indeed single photon emis-sions computed tomography (SPECT) imaging, hasbeen shown to be a powerful predictor of responseto cardiac resynchronisation therapy in ischaemiccardiomyopathy patients.5 w9

Kwong et al reported a high prevalence ofmyocardial fibrosis among diabetic patients withclinical features suspicious of coronary heartdisease, but no history of, or ECG findings consis-tent with, prior myocardial infarction.6 The pres-ence of LGE on CMR was associated with a highrisk of future adverse cardiac events, whereas itsabsence portended a favourable 2 year outlook.

Non-ischaemic dilated cardiomyopathyAutopsy studies suggest that interstitial orreplacement fibrosis is found in at least 57% ofcases of non-ischaemic dilated cardiomyopathy(NICM) and that up to 20% of the LV myocardialmass may be scar in these cases.w10 McCrohon et alreported the CMR findings of 90 patients withsystolic LV dysfunction, all of whom underwentcoronary angiography.7 They observed 100% prev-alence of LGE in either a subendocardial or trans-mural distribution among those with significantcoronary disease. In contrast, LGE was absent in59% of those with NICM, had a patchy, mid-walldistribution in 28% of cases (figure 3), anddisplayed a subendocardial pattern indistinguish-able from ischaemic cardiomyopathy in the

Figure 1 Enzymatic cascade resulting in collagen synthesis and degradation.MMP, matrix metalloproteinase; PICP, carboxy-terminal propeptide of procollagen type I;PINP, amino-terminal propeptide of procollagen type I; TIMP, tissue inhibitor of matrixmetalloproteinase. The dashed line represents inhibition.

Figure 2 Horizontal long axis late gadolinium cardiacmagnetic resonance image demonstrating hyperenhance-ment (arrow) typical of scar following myocardialinfarction. The pattern of enhancement ranges fromsubendocardial to transmural, and is limited to an arterialterritory.

Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335 2017

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

remaining 13%. The authors suggested that LGE-CMR might be an acceptable alternative to coro-nary angiography in determining the aetiology ofsevere cardiomyopathy because of its non-invasivenature, and because the absence of luminal coro-nary abnormalities on angiography does notexclude an ischaemic cause of LV dysfunction. Theypostulate that some of those cases without severecoronary disease may represent re-canalisedmyocardial infarction.Assomull et al found mid-wall LGE in 35% of

patients with established NICM recruited froma large tertiary hospital setting.8 They reported thatfibrosis extent (using a 2SD signal intensitythreshold) was a significant predictor of the devel-opment of death or hospitalisation, and was supe-rior to LV volumes and ejection fraction. Morerecently, Wu et al reported that among 65 patientswith NICM diagnosed a median of 4 years before,42% exhibited LGE, involving on average 10% ofmyocardial mass. In contrast to the earlier study,‘fibrosis’was defined as signal intensity greater thanthe peak signal intensity of a remote normal regionof myocardium.w11 The presence of LGE on CMRwas independently associated with a higher risk ofthe composite primary outcome of cardiacmortality, appropriate implantable cardioverterdefibrillator discharge, and hospitalisation for heartfailure.w11 Both these studies were performed inlarge tertiary referral centres with patients evalu-ated often after years of treatment. It is unknown ifthe prevalence and extent of LGE detectedmyocardial fibrosis is similar in an ambulatorypopulation of patients with NICM and, further-more, if this is detected at first clinical presentation.Lastly, it is unclear if the presence of myocardialfibrosis on LGE-CMR and/or its quantity influencesthe clinical response to medical and cardiacresynchronisation therapy in NICM.

Hypertrophic cardiomyopathyHypertrophic cardiomyopathy (HCM) is a geneticdisorder characterised by the development of

cardiac muscle fibre disarray, dysplasia of smallintramural coronary arterioles, and myocardialfibrosis. It must be distinguished from hypertensiveLV hypertrophy and athlete’s heart, particularly inits early stages. Moon et al reported the ability ofLGE-CMR to accurately identify and quantifyreplacement fibrosis in HCM.w12 A cross-sectionalstudy of 21 HCM patients demonstrated fibrosis bythe LGE technique in 81% of cases.9 Scarring waspatchy but occurred predominantly within hyper-trophied segments (typically observed at the junc-tion of the right ventricle and interventricularseptum) (figure 4). Findings from another seriescorroborate this pattern of LGE in HCM.10 Thepresence of LGE has been shown to be associatedwith the incidence of ventricular tachycardia inHCM patients.11

Valvular heart diseaseChronic valvular heart disease may impose eithera volume load (in the case of mitral or aorticincompetence) or pressure load (in the case ofaortic stenosis) on the left ventricle. These two

Figure 3 Short axis late gadolinium cardiac magneticresonance image demonstrating hyperenhancementtypical of fibrosis in a patient with non-ischaemic dilatedcardiomyopathy (arrow). Fibrosis may be seen mid-wall.

Figure 4 (A) Short axis late gadolinium cardiacmagnetic resonance (CMR) image demonstrating hyper-enhancement typical of fibrosis in a patient withhypertrophic cardiomyopathy (arrow). Fibrosis is typicallyseen within hypertrophied segments, particularly at thejunction of the right ventricle with the interventricularseptum. (B) Corresponding short-axis cine CMR image ofthe left ventricle.

2018 Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

pathophysiological processes result in differentpatterns of chamber remodelling: volume overloadresults in eccentric remodelling, in which wallthickness is reduced relative to chamber volume;pressure overload causes concentric remodelling,where wall thickness is maintained or increasedrelative to chamber size. Myocardial fibrosis isa feature of longstanding valvular heart diseaseirrespective of the mechanism. CMR has demon-strated sensitivity of 74% and specificity of 81% fordetection of myocardial fibrosis in patients withsevere aortic valve disease.w13 Fibrosis quantity wasassociated negatively with LVEF. Rudolph et alreported a 62% prevalence of LGE in patients withaortic stenosis and LV hypertrophy.10 Weidemannet al confirmed that myocardial fibrosis is commonin patients with severe, symptomatic aorticstenosis, and could be detected accurately byCMR.12 Not surprisingly, they found thatmyocardial fibrosis was not reversible followingaortic valve replacement.

Other conditionsMyocardial fibrosis may also be seen in a number ofother cardiac conditions, such as sarcoidosis andarrhythmogenic right ventricular cardiomyopathy.The ability of CMR to identify fibrofatty replace-ment and/or infiltration, as well as to provideimages of high spatial resolution in any plane,permit it a role in their diagnosis.

T1 mappingThe major limitation of late gadolinium enhance-ment CMR in the detection of myocardial fibrosisis its reliance on differences in signal intensitybetween scarred regions and adjacent normalmyocardium. It thus has reduced sensitivity for thedetection of diffuse myocardial fibrosis, which isthe pathological hallmark of NICM and volumeoverloaded conditions. T1 mapping, the calculationof a post-contrast myocardial T1 time by imaginga given plane with sequentially increasing inversiontimes, has been validated in animal studies asshowing a good correlation with ex vivo fibrosiscontent. Reproducibility has been defined, andmore recently this technique has been able todiscriminate heart failure patients from healthycontrols even after excluding myocardial segmentsdisplaying late gadolinium enhancement.13 In thisstudy, Iles et al performed CMR on 25 patientswith heterogeneous causes of heart failure, and 20healthy controls. The mid-chamber short axis slicewas imaged at a range of inversion times 15 minfollowing the administration of 0.2 mmol/kggadolinium-DTPA. The average T1 in regions ofinterest was measured using specialised software(figure 5). They found a significant difference in T1time between LGE positive and negative myocar-dium among heart failure patients (330630 ms vs429622 ms, respectively, p¼0.02) and a significantdifference in T1 time between the LGE negativemyocardium of heart failure patients and healthycontrols (429622 ms vs 564623 ms, respectively,p<0.001). Although more work needs to be done on

the robustness of the technique, especially withrespect to multicentre, multi-vendor application,this technique shows promise in the quantitativeevaluation of diffuse myocardial fibrosis, and hencemay have a potentially wide array of applicationsin heart failure, cardiomyopathy, and valvular heartdisease.

ECHOCARDIOGRAPHYMyocardial composition influences its acousticproperties. Collagen is an important cause of ultra-sound scattering and attenuation. The measure-ment of peak integrated backscatter and cyclicvariation in integrated backscatter may thus reflecta degree of myocardial fibrosis (figure 6). Hoyt et aldemonstrated a linear relationship betweenmyocardial hydroxyproline content as a marker offibrosis and magnitude of echocardiographic inte-grated backscatter in autopsy specimens fromvictims of myocardial infarction.w14 Amonga cohort of hypertensive patients, integrated back-scatter decreased and cyclic variation in integratedbackscatter increased following treatment withblood pressure lowering agents.14 Myocardial scar,as defined by two dimensional echocardiographicwall thickness <6 mm, was associated withincreased acoustic reflectance, and predictedresponse to cardiac resynchronisation therapyamong patients with ischaemic cardiomyopathy.w15

This technique has a limited role in the assessmentof fibrosis in non-ischaemic cardiomyopathies.Myocardial strain refers to its degree of defor-

mation through the cardiac cycle. Strain may bemeasured by either tissue Doppler imaging orspeckle tracking. The latter technique relies uponechocardiographic software recognition of myocar-dial points by their acoustic characteristics, andmeasurement of their displacement over the cardiaccycle (figure 7). Given the effect of collagen depo-sition on myocardial deformation, strain measure-ment may be well suited to indirect evaluation ofmyocardial fibrosis. Gjesdal et al and Roes et al haveboth demonstrated a high sensitivity and specificityfor longitudinal strain scores to predict extent ofmyocardial scar among patients with ischaemic LVdysfunction as compared with contrast enhancedCMR.15 16 Weidemann et al quantified myocardialfibrosis histologically in biopsies from the LVoutflow tract in 58 aortic valve prosthesis recipientsfor severe, symptomatic aortic stenosis.12 Patientsunderwent comprehensive echocardiographic eval-uation (including radial and longitudinal strain andstrain rate imaging using tissue Doppler) preoper-atively and 9 months postoperatively. The authorsdemonstrated uniform improvement in theseparameters only among subjects with no or mildfibrosis. Baseline radial LV function and LVEF weresimilar irrespective of myocardial fibrosis burden,whereas longitudinal indices of LV function weresignificantly reduced in those with more extensivefibrosis. This study suggests that quantitativeevaluation of longitudinal LV function may bea sensitive tool for detection of the mechanicaleffects of myocardial fibrosis.

Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335 2019

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

NUCLEAR IMAGINGCardiac nuclear imaging encompasses the tech-niques of positron emission tomography (PET) andSPECT, and with these techniques there is nowa wealth of evidence in the assessment of myocar-dial viability and therefore indirectly of scarred andfibrotic myocardium. The basic principles of PETare similar to those of SPECT (both form images byway of intravenous radioisotopes concentrating inviable myocardium, giving off g-radiation which isdetected by an external g-camera); however, forscar detection, PET systems are generally moresensitive than SPECT systems, having betterspatial resolution, and providing more accurateattenuation correction.In patients with ischaemic cardiomyopathy, scar

tissue as detected by PET is a good predictor of lackof recovery of LV systolic function after coronaryartery bypass surgery, though precise individualprognostication is difficult.16 17 Also, as expected,PET provides information on outcome. In a meta-analysis of 10 such studies (1046 patients),annualised mortality rates were 4% for those withviable myocardium who underwent revascularisa-tion versus 17% for those with viability who didnot.w17 If no viability was demonstrated with PET,revascularisation was not associated with changes

in mortality rates (annual mortality rates of6e8%). Such evidence has encouraged studies inwhich treatment is based on PET findings, but sofar the studies have only been partly successful. Inthe PARR-2 trial, treatment based on 18F-FDG PETfindings was equivalent to standard care after1 year of follow-up.18 In a post-hoc analysis of thePARR-2 trial data, however, the patients withischaemic cardiomyopathy with larger amounts ofperfusionemetabolism mismatch (>7%) did haveimproved outcome with revascularisation,w18 andPET may find a place to assess outcome in patientswith critical ischaemic cardiomyopathy in whomsurgery is of risk.The water perfusable tissue index has emerged as

a candidate for in vivo detection of myocardialfibrosis by PET.w19 Briefly, the ratio of the waterperfusable tissue fraction to the anatomic tissuefraction is determined using the tracers H2

15O andC15O. In normal individuals, the perfusable tissueindex should be 1.0, but fibrotic tissue is unable toexchange water rapidly and hence the tissue indexdecreases in fibrotic myocardium. In dogs a reduc-tion of the perfusable tissue index correlates wellwith the extent of fibrosis after infarction. Inhumans, comparisons have been made to LGE-CMR in different patient populations. In patients

Figure 5 T1 mapping. (A) Regions of interest are drawn around the left ventricular myocardium at different inversion times. (B) Signal intensities foreach region of interest are fitted to an exponential recovery curve to obtain the myocardial T1 time. Reproduced with permission from Iles et al.13

2020 Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

with scarring due to myocardial infarction, lategadolinium enhancement was indeed negativelycorrelated (r¼�0.65) with the perfusable tissueindex,w20 but the water perfusable tissue indexsystematically underestimates the amount of scartissue if the latter becomes extensive.w21 Withrespect to replacement fibrosis in patients withHCM where Gd late enhancement was seen, theperfusable tissue index was essentially unaffected.The perfusable tissue index was slightly lowered inthe LV free wall of patients with HCM, but not atall in the septum where myocardial disarray,oedema, and later fibrosis predominantly takes

place (figure 8). The authors suggest that hyper-enhancement by LGE-CMR may not be solelygoverned by fibrosis but also (especially so early inthe natural history of HCM) by oedema; however,in light of missing histology, the finding mayequally well be attributed to a lack of sensitivity bythe perfusable tissue index technique.In clinical practice, SPECT is still the more robust

and more often applied technique and a wealth ofevidence testifies to its ability to demonstrate(larger) viability defects and provide for importantprognostic information. In comparison with PET,SPECT has more limited spatial resolution. In

Figure 6 Segmental curves representing regional integrated backscatter in the parasternal short axis view by speckletracking echocardiography.

Figure 7 Segmental curves representing regional longitudinal left ventricular strain in the apical four chamber view asdetermined by speckle tracking echocardiography.

Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335 2021

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

a study by Wagner et al19 that compared LGE-CMRwith SPECT in a canine model of myocardial injury,CMR identified 92% of all subendocardial infarctswhereas SPECT only identified 28%. The relativelack of spatial resolution with SPECT can poten-tially lead to erroneous conclusions with respect toviability. Thus, in dysfunctional segments, trans-mural LGE of 9%, 33%, and 80%, respectively,corresponds to segments that by combined 99TcSPECT/18F-FDG PET (and accepted cut-off values)are classified as normal, mismatched (hibernating),and matched (fibrotic).w22 The classification of‘hibernating’ versus ‘fibrotic’ myocardium bycombined 99Tc SPECT/18F-FDG PET may havemore to do with the extent of transmurality of scarand appropriateness of chosen cut-off values forflow (by SPECT) and glucose metabolism (by PET).

BIOMARKERS OF MYOCARDIAL FIBROSISIn comparison with cardiac imaging, serumbiomarkers of myocardial fibrosis have yet toachieve general acceptance in clinical practice. Thismay relate to the semi-quantitative nature ofmeasurement techniques, and to inferior specificityin a general population, in whom collagen turnovermay be increased due to comorbid conditions.Nonetheless, serum biomarkers of myocardialfibrosis are conceptually appealing as they allowmeasurement of diffuse fibrosis, the quantificationof which remains elusive using current imagingtechniques.Martos et al found higher concentrations of

serum carboxy-terminal telopeptide of procollagentype I, amino-terminal propeptide of procollagentype III, MMP-1, -2, and -9, and TIMP-1 in patientswith diastolic heart failure compared with thosewithout.20 Serum TIMP-1 concentration has beenassociated with echocardiographic indices ofdiastolic dysfunction in hypertensive patients.w23

This study demonstrated that serum MMP-2concentration was the most sensitive and specificbiomarker of heart failure with preserved ejectionfraction, superior to B-type natriuretic peptide(BNP).w24 Based on this finding, the authorspropose an adjunctive role for MMP-2 in the diag-nosis of this condition, although this is yet to gainwidespread clinical acceptance.In a cross-sectional study of 1069 subjects from

the Framingham Heart Study, serum TIMP-1concentration was positively related to LV mass,end-systolic diameter, and left atrial diameter afteradjustment for age, sex and height.w25 Althoughadjustment for further clinical covariates attenu-ated the strength of association between TIMP-1concentration and echocardiographic measures, theauthors concluded that this evidence supports thehypothesis that cardiovascular risk factors promoteLV remodelling by influencing turnover of theextracellular matrix.In a study by Yan et al of systolic heart failure

patients, serum MMP-9 concentration wasdemonstrated to have a positive linear relationshipwith LV end-systolic volume and a negative asso-ciation with LVEF.w26 Over 43 weeks follow-up,there was a negative relationship between changein serumMMP-9 concentration and change in LVEF.In contrast, Vorovich et al have recently shown thatserum MMP-9 concentration is a poor marker of LVremodelling and clinical outcome compared withBNP.21

At the present time, although associationsbetween serum fibrosis biomarker concentration andindices of myocardial disease have been demon-strated, the clinical role of serological testing remainsto be defined. Moreover, the robustness of theseassays outside of research laboratories is unknown.

FUTURE DIRECTIONSNovel imaging techniquesThere is preliminary research to suggest a potentialrole for molecular imaging in fibrosis detection andquantification. Collagen targeting molecules have

Figure 8 (A) Mid ventricular short axis view of a gadolinium delayed enhanced cardiacmagnetic resonance image of a patient with hypertrophic cardiomyopathy. Note patchyhyperenhancement located at hypertrophied interventricular septum at junctions ofseptum and right ventricular free walls (arrows). (B) Anatomic tissue fraction image ofsame patient derived by positron emission tomography. Reproduced with permissionfrom Knaapen et al.w20

Non-invasive evaluation of myocardial fibrosis: key points

< Myocardial fibrosis develops in response to a cardiac insult, and may havedeleterious long term effects on cardiac function and remodelling.

< Collagen deposition in the myocardium is the result of the interplay of variousenzymes including matrix metalloproteinases (MMPs) and tissue inhibitors ofmatrix metalloproteinase (TIMPs).

< Cardiac magnetic resonance (CMR) imaging using the late gadoliniumtechnique can demonstrate scar tissue in ischaemic cardiomyopathy, and canyield prognostic information.

< Approximately one third of cases of established non-ischaemic dilatedcardiomyopathy will exhibit myocardial fibrosis by the late gadolinium CMRtechnique; when present, this is associated with adverse prognosis.

< Late gadolinium enhancement CMR can also demonstrate myocardial fibrosisin other cardiac conditions, such as hypertrophic cardiomyopathy.

< Strain and backscatter echocardiographic techniques have an emerging rolein the evaluation of focal and diffuse myocardial fibrosis.

< Nuclear imaging techniques are well established for the detection of largeregions of scar tissue; however, their limited spatial resolution prohibitsdetection of small regions or diffuse fibrosis.

< Serum biomarkers are yet to gain widespread acceptance in the clinicalevaluation of myocardial fibrosis.

2022 Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

been appended to Gd-DTPA to image animalmodels of myocardial infarction.w27 Such tech-niques may increase the specificity for the identifi-cation of fibrosis and help its distinction frommyocardial oedema.Diffusion tensor magnetic resonance images

microstructural organisation by use of the propertythat the main orientation of microstructuresparallels the main diffusivity of water molecules,causing signal attenuation in the presence ofa magnetic field. Diffusion tensor CMR has beenused post-myocardial infarction to demonstratealtered fibre architecture that correlates withinfarct size using the LGE technique.w28

Evaluation of response to treatmentA number of therapeutic agents have been associ-ated with a reduction in myocardial fibrosis. Themeasurement of fibrosis burden may therefore beused as a surrogate end point to guide the ongoingdevelopment of treatments for chronic cardiacdisease. Given the prognostic implications of thepresence and extent of myocardial fibrosis, moni-toring response to treatment may allow individu-alisation of therapy, although this remains to be ofproven value in clinical trials.

CONCLUSIONThe formation and degradation of myocardialcollagen is a dynamic process. Although netcollagen deposition may serve an as yet poorlydefined physiological purpose, its excess in thelonger term appears to be detrimental to myocar-dial mechanical function and provides substrate forcardiac arrhythmia. Hence the detection andquantification of myocardial fibrosis has becomea key focus of recent research. Focal scar, as may beseen following myocardial infarction, can bevisualised with high spatial resolution using the

LGE-CMR technique, and can be inferred throughits functional consequences in terms of echocar-diographic measures of LV strain or PET metabolicimaging. In contrast, diffuse myocardial fibrosisremains difficult to identify using currently avail-able imaging modalities, although T1 mapping byCMR holds promise. There is preliminary evidenceto support the ability of serum markers of collagenturnover to identify myocardial fibrosis. The utilityand clinical benefit of these assays in ‘real world’practice remains to be demonstrated.

Funding Dr Leong is supported by a Medical PostgraduateScholarship funded jointly by the National Health and MedicalResearch Council of Australia and the National Heart Foundation ofAustralia.

Competing interests In compliance with EBAC/EACCME guidelines,all authors participating in Education in Heart have disclosed potentialconflicts of interest that might cause a bias in the article. Theauthors have no competing interests.

Provenance and peer review Commissioned; not externally peerreviewed.

REFERENCES1. Spinale FG, Coker ML, Heung LJ, et al. A matrix

metalloproteinase induction/activation system exists in the humanleft ventricular myocardium and is upregulated in heart failure.Circulation 2000;102:1944e9.

< Early work suggesting a role for MMP activation in dilatedcardiomyopathy.

2. Kim RJ, Chen EL, Lima JA, et al. Myocardial Gd-DTPA kineticsdetermine MRI contrast enhancement and reflect the extent andseverity of myocardial injury after acute reperfused infarction.Circulation 1996;94:3318e26.

< Rabbit model study of myocardial infarction that revealedabnormal gadolinium kinetics in infarcted myocardium.

3. Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhancedmagnetic resonance imaging to identify reversible myocardialdysfunction. N Engl J Med 2000;343:1445e53.

< Seminal work demonstrating the ability of LGE-CMR todistinguish reversible from irreversible myocardialdysfunction in ischaemic cardiomyopathy.

4. Selvanayagam JB, Kardos A, Francis JM, et al. Value of delayed-enhancement cardiovascular magnetic resonance imaging inpredicting myocardial viability after surgical revascularization.Circulation 2004;110:1535e41.

5. Ypenburg C, Schalij MJ, Bleeker GB, et al. Impact of viability andscar tissue on response to cardiac resynchronization therapy inischaemic heart failure patients. Eur Heart J 2007;28:33e41.

< Important SPECT study of CRT recipients that revealed thatscar extent, particularly in the site subtended by the leftventricular pacing lead, predicted response to cardiacresynchronisation therapy.

6. Kwong RY, Sattar H, Wu H, et al. Incidence and prognosticimplication of unrecognized myocardial scar characterized bycardiac magnetic resonance in diabetic patients without clinicalevidence of myocardial infarction. Circulation 2008;118:1011e20.

7. McCrohon JA, Moon JCC, Prasad SK, et al. Differentiation ofheart failure related to dilated cardiomyopathy and coronary arterydisease using gadolinium-enhanced cardiovascular magneticresonance. Circulation 2003;108:54e9.

< Seminal paper suggesting that CMR may be a suitablealternative to coronary angiography in diagnostic work-upof dilated cardiomyopathy.

8. Assomull RG, Prasad SK, Lyne J, et al. Cardiovascular magneticresonance, fibrosis, and prognosis in dilated cardiomyopathy. J AmColl Cardiol 2006;48:1977e85.

< Key study measuring the prevalence of myocardial fibrosisby LGE-CMR in non-ischaemic dilated cardiomyopathy andillustrating its prognostic importance.

9. Choudhury L, Mahrholdt H, Wagner A, et al. Myocardial scarringin asymptomatic or mildly symptomatic patients with hypertrophiccardiomyopathy. J Am Coll Cardiol 2002;40:2156e64.

10. Rudolph A, Abdel-Aty H, Bohl S, et al. Noninvasive detection offibrosis applying contrast-enhanced cardiac magnetic resonance in

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 Cardiologydyouneed to answer the accompanying multiple choice questions (MCQs). Toaccess the questions, click on BMJ Learning: Take this module on BMJLearning from the content box at the top right and bottom left of the onlinearticle. For more information 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)dpass mark is 80%.< EBAC credits: Print out and retain the BMJ Learning certificate once you

have completed the MCQsdpass 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/?hit¼men02).

Please note: The MCQs are hosted on BMJ Learningdthe best availablelearning website for medical professionals from the BMJ Group. If prompted,subscribers must sign into Heart with their journal’s username and password. Allusers must also complete a one-time registration on BMJ Learning and subse-quently log in (with a BMJ Learning username and password) on every visit.

Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335 2023

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

different forms of left ventricular hypertrophy relation toremodeling. J Am Coll Cardiol 2009;53:284e91.

11. Kwon DH, Smedira NG, Rodriguez ER, et al. Cardiacmagnetic resonance detection of myocardial scarring inhypertrophic cardiomyopathy: correlation with histopathologyand prevalence of ventricular tachycardia. J Am Coll Cardiol2009;54:242e9.

12. Weidemann F, Herrmann S, Stork S, et al. Impact of myocardialfibrosis in patients with symptomatic severe aortic stenosis.Circulation 2009;120:577e84.

13. Iles L, Pfluger H, Phrommintikul A, et al. Evaluation of diffusemyocardial fibrosis in heart failure with cardiac magneticresonance contrast-enhanced T1 mapping. J Am Coll Cardiol2008;52:1574e80.

< Recent key paper on the use of T1 mapping CMR techniqueto quantify diffuse myocardial fibrosis in dilatedcardiomyopathy.

14. Di Bello V, Giorgi D, Talini E, et al. Incremental value of ultrasonictissue characterization (backscatter) in the evaluation of leftventricular myocardial structure and mechanics in essential arterialhypertension. Circulation 2003;107:74e80.

< Paper illustrating the potential for ultrasonic tissuecharacterisation by backscatter analysis in hypertensivepatients.

15. Gjesdal O, Hopp E, Vartdal T, et al. Global longitudinal strainmeasured by two-dimensional speckle tracking echocardiographyis closely related to myocardial infarct size in chronic ischaemicheart disease. Clin Sci (Lond) 2007;113:287e96.

< Study demonstrating the ability of strain as measured byspeckle tracking echocardiography to identify regions ofmyocardial infarction.

16. Tillisch J, Brunken R, Marshall R, et al. Reversibility of cardiacwall-motion abnormalities predicted by positron tomography.N Engl J Med 1986;314:884e8.

17. Beanlands RS, Ruddy TD, deKemp RA, et al. Positron emissiontomography and recovery following revascularization (PARR-1): theimportance of scar and the development of a prediction rule for thedegree of recovery of left ventricular function. J Am Coll Cardiol2002;40:1735e43.

18. Beanlands RS, Nichol G, Huszti E, et al. F-18-fluorodeoxyglucosepositron emission tomography imaging-assisted management ofpatients with severe left ventricular dysfunction and suspectedcoronary disease: a randomized, controlled trial (PARR-2). J AmColl Cardiol 2007;50:2002e12.

19. Wagner A, Mahrholdt H, Holly TA, et al. Contrast-enhanced MRIand routine single photon emission computed tomography (SPECT)perfusion imaging for detection of subendocardial myocardialinfarcts: an imaging study. Lancet 2003;361:374e9.

20. Martos R, Baugh J, Ledwidge M, et al. Diastolic heart failure:evidence of increased myocardial collagen turnover linked todiastolic dysfunction. Circulation 2007;115:888e95.

< Comprehensive serological study of hypertensive patientswith varying degrees of diastolic dysfunction.

21. Vorovich EE, Chuai S, Li M, et al. Comparison of matrixmetalloproteinase 9 and brain natriuretic peptide as clinicalbiomarkers in chronic heart failure. Am Heart J 2008;155:992e7.

2024 Heart 2010;96:2016e2024. doi:10.1136/hrt.2009.183335

Education in Heart

group.bmj.com on January 9, 2013 - Published by heart.bmj.comDownloaded from

doi: 10.1136/hrt.2009.183335 2010 96: 2016-2024Heart

 Darryl P Leong, Per Lav Madsen and Joseph B Selvanayagam fibrosis: implications for the clinicianNon-invasive evaluation of myocardial

http://heart.bmj.com/content/96/24/2016.full.htmlUpdated information and services can be found at:

These include:

Data Supplement http://heart.bmj.com/content/suppl/2012/02/01/96.24.2016.DC1.html

"Supplementary Data"

References

http://heart.bmj.com/content/96/24/2016.full.html#related-urlsArticle cited in:  

http://heart.bmj.com/content/96/24/2016.full.html#ref-list-1This article cites 21 articles, 9 of which can be accessed free at:

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

CollectionsTopic

(450 articles)Heart failure   � (232 articles)Dilated cardiomyopathy   � (3598 articles)Clinical diagnostic tests   �

(205 articles)Hypertrophic cardiomyopathy   � (6394 articles)Drugs: cardiovascular system   �

(19 articles)Non-invasive imaging   � (422 articles)Education in Heart   �

 Articles on similar topics can be found in the following collections

Notes

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 January 9, 2013 - Published by heart.bmj.comDownloaded from