AD

ULT

CA

RD

IAC

Dynamic 3-Dimensional EchocardiographicAssessment of Mitral Annular Geometry in PatientsWith Functional Mitral RegurgitationKamal R. Khabbaz, MD, Feroze Mahmood, MD, Omair Shakil, MD,Haider J. Warraich, MD, Joseph H. Gorman, III, MD, Robert C. Gorman, MD,Robina Matyal, MD, Peter Panzica, MD, and Philip E. Hess, MD

Division of Cardiac Surgery and Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical

Center, Harvard Medical School, Boston, Massachusetts; and Gorman Cardiovascular Research Group, University of Pennsylvania,Philadelphia, Pennsylvania

Background. Mitral valve (MV) annular dynamics havebeen well described in animal models of functional mitralregurgitation (FMR). Despite this, little if any data existregarding the dynamic MV annular geometry in humanswith FMR. In the current study we hypothesized that3-dimensional (3D) echocardiography, in conjunctionwith commercially available software, could be used toquantify the dynamic changes in MV annular geometryassociated with FMR.

Methods. Intraoperative 3D transesophageal echocar-diographic data obtained from 34 patients with FMR and15 controls undergoing cardiac operations were dynami-cally analyzed for differences in mitral annular geometrywith TomTec 4D MV Assessment 2.0 software (TomTecImaging Systems GmbH, Munich, Germany).

Results. In patients with FMR, the mean mitral annulararea (14.6 cm2 versus 9.6 cm2), circumference (14.1 cm

versus 11.4 cm), anteroposterior (4.0 cm versus 3.0 cm)

110 Francis St, Boston, MA 02215; e-mail: kkhabbaz@bidmc.harvard.edu.

© 2013 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

and anterolateral-posteromedial (4.3 cm versus 3.6 cm)diameters, tenting volume (6.2 mm3 versus 3.5 mm3) andnonplanarity angle (NPA) (154 degrees � 15 versus 136degrees � 11) were greater at all points during systolecompared with controls (p < 0.01). Vertical mitral annu-lar displacement (5.8 mm versus 8.3 mm) was reduced inFMR compared with controls (p < 0.01).

Conclusions. There are significant differences in dy-namic mitral annular geometry between patients withFMR and those without. We were able to analyze thesechanges in a clinically feasible fashion. Ready availabil-ity of this information has the potential to aid compre-hensive quantification of mitral annular function andpossibly assist in both clinical decision making andannuloplasty ring selection.

(Ann Thorac Surg 2013;95:105–10)

© 2013 by The Society of Thoracic Surgeons

Using invasive imaging techniques, 3-dimensional(3D) mitral valve (MV) annular dynamics have

been well described in large animal models of functionalmitral regurgitation (FMR) [1, 2]. Although the increasinguse of real-time 3D echocardiography has significantlyimproved our understanding of static human MV annu-lar geometry over the past decade, little if any data existon the dynamic 3D changes in annular function inhumans with FMR. The knowledge of annular changesduring the cardiac cycle is based on manual reconstruc-tion of ”dynamic” MV models from static images [3–13].Effects of chronic mitral regurgitation (MR) states such asFMR on annular behavior are also extrapolated fromannular position at a single point (end systole) in thecardiac cycle [5, 14–18].

At present, the echocardiographic assessment of MVin FMR is performed only to quantify regurgitation as

Accepted for publication Aug 27, 2012.

Address correspondence to Dr Khabbaz, Division of Cardiac Surgery,Beth Israel Deaconess Medical Center, Lowry Medical Office Bldg, 2A,

a marker of valve dysfunction and exclude stenosisafter repair. Ideally, geometric distortion of the mitralannulus incurred from annuloplasty devices shouldalso be objectively quantified and followed. However,despite recognition of the prognostic value of mitralannular changes in FMR, technologic impedimentshave precluded their clinical application [19]. Thedemonstration of differences in mitral annular geom-etry (static and dynamic) between patients with andwithout FMR in a clinically feasible fashion should bethe first step toward achieving this goal. In the currentstudy we hypothesized that real-time 3D echocardiog-raphy, in conjunction with commercially available im-aging software, could be used to compare 3D mitralannular geometry during systole in patients with FMRand those without.

Material and Methods

Study PopulationThe data were collected as part of a prospective institu-

tional review board–approved protocol with a waiver of

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2012.08.078

106 KHABBAZ ET AL Ann Thorac SurgDYNAMIC ANALYSIS OF MITRAL VALVE GEOMETRY 2013;95:105–10A

DU

LTC

AR

DIA

C

informed consent. We enrolled 34 consecutive patientswith FMR who were undergoing cardiac operations.FMR was defined as MR resulting from retraction andmalcoaptation of structurally normal MV leaflets in thepresence of global left ventricular dysfunction. In asubset of patients with FMR, localized ischemia-induced wall motion abnormalities can be seen to becontributing to MR; in such patients the term ischemicMR may also be used. Exclusion criteria includedpatients with structural of the mitral apparatus abnor-malities (flail leaflets, torn chordae) or technicallyinadequate studies. We also selected 15 controls. Thesewere patients scheduled for cardiac operations for anunrelated indication and who had a normal (�50%)ejection fraction, trace or no MR, and absence of anyvalvular abnormality (Table 1).

Intraoperative 3D Transesophageal EchocardiographicExaminationAfter induction of general anesthesia, a comprehensive2-dimensional (2D) transesophageal echocardiographicexamination was performed during the period beforecardiopulmonary bypass. MR quantification was per-formed by measuring the vena contracta, which is asemiquantitative method of MR severity assessment. It isbased on the measurement of the width of the narrowestpart of the MR jet in the midesophageal long-axis view toclearly identify flow convergence of the MR jet on the leftventricular side. All our cases were patients with moder-ate or greater MR (vena contracta width � 0.5 cm2 [20])and without any evidence of structural disease of theleaflets, papillary muscles, or chordae tendineae or wallmotion abnormalities. Image acquisition in 3D was per-formed with an iE-33 ultrasound system equipped withan X7-2t “matrix” transesophageal echocardiographicprobe (Philips Medical Systems, Andover, MA). Imageswere acquired with R-wave gating over 4 to 8 beatsduring brief periods of apnea and concurrent avoidanceof patient or probe movements. In patients with atrialfibrillation and other arrhythmias, the 3D live zoommode was used to acquire an en face view of the MV.(The 3D live zoom mode displays a magnified 3D image.)A technically adequate image was defined as an en faceleft atrial image of the MV devoid of artifacts. Intraoper-

Abbreviations and Acronyms

2D � 2-dimensional3D � 3-dimensionalALA � anterior leaflet angleCSV � comma separated valuesFMR � functional mitral regurgitationMR � mitral regurgitationMV � mitral valveNPA � nonplanarity anglePLA � posterior leaflet angleUSB � universal serial bus

ative image acquisition was completed in 30 seconds; the

datasets were then immediately exported by a USB flashdrive transfer to a Windows-based workstation for anal-ysis by the TomTec Image Arena Browser (TomTecImaging Systems, GmbH, Munich, Germany).

Dynamic MV Geometric AnalysisThe MV geometric analysis was performed using theTomTec Image Arena software (TomTec Imaging Sys-tems GmBH) equipped with the 4D MV Assessment 2.0program. The feasibility and methodology of intraopera-tive dynamic geometric analysis has been established pre-viously [21]. Briefly, the dynamic MV geometric analysis isperformed in a workflow arrangement of 7 sequential steps,which are initiated with identification and selection ofend-systolic and early systolic frames in the dataset.Based on the identification of the anatomic landmarksand the frames of interest (end-systolic frame � the lastframe before the MV opens; early-systolic frame � thelast frame before the MV starts to close), mitral annulus,coaptation line, leaflets, and the aortic valve position aredynamically tracked throughout the systolic phase. Thisis based on optical flow and pattern recognition of themitral annulus and leaflets [22–25]. At the conclusion ofthe workflow, both static and dynamic geometric analy-

Table 1. Characteristics of Cases and Controls

VariableCases

(n � 34)Controls(n � 15)

Age (y) 67.4 (43–88) 61.8 (32–84)Sex

Male 20 (58.8%) 9 (60.0%)Female 14 (41.2%) 6 (40.0%)

Mitral regurgitation grade0 . . .1 . . . 152 14 . . .3 15 . . .4 5 . . .

Ejection fraction�50% 20 0�50% 14 15 (100%)

Functional mitral regurgitation 34 (100%) . . .Ischemic mitral regurgitation 17 (50%) . . .Coronary angiography

Yes 32 (94%) 15 (100%)No 2 (6%) 0

Coronary artery disease 28 (82%) 0Procedure

MVR � CABG 17 (50%) . . .MVR 9 (26%) . . .CABG 6 (18%) 5 (33%)AVR 1 (3%) 7 (47%)AVR � CABG 1 (3%) 2 (13%)PFO closure . . . 1 (7%)

AVR � aortic valve replacement; CABG � coronary artery bypass

graft; MVR � mitral valve replacement; PFO � patent foramenovale.

107Ann Thorac Surg KHABBAZ ET AL2013;95:105–10 DYNAMIC ANALYSIS OF MITRAL VALVE GEOMETRY

AD

ULT

CA

RD

IAC

ses are generated. The time taken for data export andanalysis was less than 5 minutes. The number of framesencompassing systole varied between 30 and 50 depend-ing on the patient’s heart rate, depth of imaging, and thenumber of heartbeats over which R-wave gatingoccurred.

Statistical AnalysisData generated from the static and dynamic analysis ofthe MV were exported to Microsoft Excel for Mac 2011(Microsoft Corp, Redmond, WA), through a CSV (commaseparated values) file format. To account for variation inframes per systolic cycle and to normalize for heart rate,time values were averaged to 5 equal points during thesystolic phase. SPSS, version 18.0 (IBM Corp, Armonk,NY) was used to analyze the data. Baseline demographicdata were compared using the t test or Fisher’s exact text,as appropriate. Comparison between MV geometricmeasurements was made using the t test for singlemeasures and linear repeated measures analysis forcomparisons over time. Pearson’s correlation was used toassess the relationship between vena contracta and dif-ferent MV geometric measurements throughout systole.Reliability of the echocardiographic evaluation was as-sessed in a random sample of 9 patients (5 cases, 4controls) by examining the interobserver and intraob-server variability for all measurements using Pearson’scorrelation. Statistical significance was determined atp less than or equal to 0.05.

Table 2. Differences in Mitral Annular Geometric Measureme

Variable

Annular dimensionsAnteroposterior diameterAnterolateral-posteromedial diameterCommissural diameterAnnular circumference2D annular area3D annular areaAnterior leaflet areaPosterior leaflet area

Annular shapeNonplanarity angleMaximum nonplanarity angleCircularity indexTenting heightTenting volumeTenting volume fractionAortomitral angle

Annular excursionMaximum vertical annular displacementMaximum vertical annular displacement velocityMaximum vertical annular acceleration

Values are described as mean � standard deviation.

FMR � functional mitral regurgitation; 2D � 2-dimensional; 3D � 3-di

Results

Baseline Patient and Imaging CharacteristicsData from 34 patients with FMR (cases) and 15 controlswere used for analysis in this study. No significantdifferences were noted in baseline characteristics be-tween the 2 groups with regard to age, sex, body massindex, and body surface area. Of the 49 patients, 5 (10%)had MV datasets acquired with live zoom imaging,whereas in the remaining 44 patients (90%) it was possi-ble to acquire R-wave gated volumetric images. In allpatients, we were able to complete MV geometric anal-ysis within 40 seconds of initiating the workflow steps.

Intraobserver and Interobserver VariabilityReliability of the assessment comparing intraobserverand interobserver correlation was 0.92 and 0.83, respec-tively (p � 0.01 for both).

Mitral Annular DimensionsMean annular dimensions were all significantly enlargedin patients with FMR compared with controls throughoutsystole (Table 2; Fig 1). Maximum 3D mitral annular areawas seen at end systole in 90% (44/49) of all patients.Measurements of mitral annular shape and left ventric-ular remodeling such as the anterior and posterior leafletareas, nonplanarity angle (NPA), circularity index, andtenting volume were also significantly increased in pa-tients with FMR (Table 2; Fig 2).

etween Controls and Patients With FMR (Cases)

trols (n � 15) Cases (n � 34) p Value

3.0 cm � 0.1 4.0 cm � 0.1 �0.013.6 cm � 0.1 4.3 cm � 0.1 �0.013.5 cm � 0.5 4.4 cm � 0.7 �0.011.4 cm � 0.5 14.1 cm � 0.4 �0.01

8.7 cm2 � 0.79 14.2 cm2 � 0.8 �0.019.6 cm2 � 3 14.6 cm2 � 5 �0.016.4 cm2 � 2 10.0 cm2 � 3 �0.016.1 cm2 � 2 9.0 cm2 � 3 �0.01

egrees � 11 155 degrees � 15 �0.01egrees � 10 158 degrees � 15 �0.01

0.85 � 0.02 0.93 � 0.01 �0.01.0 mm � 0.85 11.7 mm � 0.77 0.13

.5 mm3 � 2.0 6.2 mm3 � 3.5 0.0124% � 11 30% � 21 0.36

egrees � 14 119 degrees � 15 0.12

.3 mm � 3 5.8 mm � 2 �0.01mm/s � 13 26 mm/s � 8 0.05

mm/s2 � 15 18 mm/s2 � 12 0.04

nts B

Con

1

136 d140 d

103

112 d

832

26

mensional.

108 KHABBAZ ET AL Ann Thorac SurgDYNAMIC ANALYSIS OF MITRAL VALVE GEOMETRY 2013;95:105–10A

DU

LTC

AR

DIA

C

Dynamic Change in MV GeometrySignificant differences were noted in vertical annular dis-placement velocity and acceleration between patients withFMR and controls (Table 2). The delta changes in 3D mitralannular area, anterior leaflet angle (ALA), posterior leafletangle (PLA), and NPA were significantly reduced in pa-tients with FMR compared with controls (Figs 1, 2).

Correlation Between Vena Contracta and MVGeometric MeasurementsMean vena contracta width in patients with FMR was0.737 � 0.29 and correlated significantly with annular

Fig 1. In patients with functional mitral re-gurgitation (FMR), (A) mitral annular areawas larger and (B) experienced a smaller deltachange compared with controls. These differ-ences were seen throughout systole. (3D �3-dimensional.)

Fig 2. (A) Compared with controls, nonpla-narity angle (NPA) is much greater in patientswith functional mitral regurgitation (FMR),indicating a flatter and less saddle-shaped mi-tral annulus. (B) Also, the delta change inNPA over systole is much lower in patientswith FMR. (C) A screenshot from TomTec 4DMV Assessment 2.0 demonstrates the initialdecrease in the saddle shape of the mitral an-nulus followed by an increase toward endsystole.

dimensions such as the PLA (r � 0.624; p � 0.001),mitral annular circumference (r � 0.535; p � 0.006), 2Dannular area (r � 0.587; p � 0.002), 3D annular area (r �0.580; p � 0.001), and anteroposterior diameter (r �0.582; p � 0.02) at all points during the systolic cycle.Vena contracta width correlated slightly with tentingvolume (r � 0.404; p � 0.02) and annular displacementvelocity (r � 0.369; p � 0.045). However, measurementsof mitral annular nonplanarity such as the NPA, tent-ing height, and circularity index and ALA did not showany correlation with the vena contracta.

109Ann Thorac Surg KHABBAZ ET AL2013;95:105–10 DYNAMIC ANALYSIS OF MITRAL VALVE GEOMETRY

AD

ULT

CA

RD

IAC

Comment

In this study, we were able to compare dynamic changesin mitral annular geometry between patients with FMRand controls in a clinically feasible fashion. Our analysesdemonstrated significant differences between the 2groups in static and dynamic annular geometry (Table 2).Our results have shown that the annulus in patients withFMR is flatter, more circular, and larger in area andundergoes less vertical displacement than that in controlsthroughout systole. We were also able to dynamicallyanalyze changes in planarity and shape of mitral annulusand leaflet areas, factors that have not been previouslyanalyzed either at end-systole or dynamically [22, 26, 27].Importantly, the geometric differences in dimensionsbetween the 2 groups were maintained throughout sys-tole (Figs 1, 2) [3, 4, 27, 28]. Mitral annular area alsoprogressively increased, peaking at end systole (Fig 1) [4,13, 26–28]. Even though the baseline mitral annular areain patients with FMR was larger compared with controls,it underwent a much smaller change (Fig 1) [3, 4, 28, 29].Compared with late systole, the NPA decreased duringearly systole (ie, the annulus assumed a more saddle-shaped configuration) (Fig 2) [13, 30] and assumed amore circular shape in patients with FMR. The PLA inpatients with FMR demonstrated a greater change oversystole than did the ALA, with a significant correlationwith vena contracta (r � 0.624; p � 0.001). Interestingly,although tenting volume was much higher in patientswith FMR, no significant difference was noted in tent-ing height (Table 2). This finding emphasizes that3D tenting volume is a better predictor of mitral valvulartenting than 2D measures such as tenting height ortenting area [31, 32].

The results of our study have important clinical impli-cations for a comprehensive echocardiographic assess-ment of FMR. Although the static annular dimensions atend systole represent structure, their dynamic naturerepresents the functional aspect of mitral annulus. Ourresults show that there are significant changes in annularstructure and function in patients with FMR comparedwith those without FMR. Hence, a case can be made thatthe current model of echocardiographic interrogation offlow-dependent variables, without taking into accountthe dynamics of function, is far from comprehensive. Acomprehensive assessment of mitral annular geometryshould take into account the entire spectrum of changesover time [21]. Our results also raise the possibility offollowing the annular function as a marker of ventricularreverse remodeling after revascularization therapy (sur-gical or percutaneous). Therefore the ability to track theannulus through the cardiac cycle in a clinically feasiblefashion is a significant advance from the current para-digm of MV assessment.

Additionally, the demonstration of altered mitral an-nular dynamics throughout systole calls into question theuse of flexible ring annuloplasty in the treatment of FMR.Proponents of flexible ring annuloplasty believe thatthese devices preserve mitral annular function and pro-

vide a more anatomically correct repair [33]. The results

of this study demonstrate the flaw in this belief and thepresumed benefit [34] and support the concept that thefunction of ring annuloplasty is one of restoration andnot preservation of annular geometry. This is the goal ofthe latest generation of saddle-shaped annuloplastyrings, which have been designed to reestablish a morenormal systolic human annular and leaflet geometry [35,36]. Data from animal studies have been used to improveannuloplasty ring design, but technologic limitationshave precluded the performance of such analyses forroutine clinical use [37]. Our performance of these anal-yses in a timely fashion brings us a step closer toincorporating this information into clinical decision mak-ing and objectively assessing the concept of offering an”annular solution to a ventricular problem” [19].

We acknowledge certain limitations in our study. First,our control group represented patients undergoingtransesophageal echocardiography for clinical indica-tions and therefore might not represent a normal popu-lation. However, given that there was no clinical orechocardiographic evidence of MV disease, we believe itrepresented an adequate control group. Second, our 3Dechocardiographic data were collected in real time, andthe geometric analyses were performed off-line. How-ever, the lag time between data acquisition and exportand analysis was less than 5 minutes and the results werereadily available.

In conclusion, mitral annular geometry in patients withFMR is significantly altered throughout systole as com-pared with patients without FMR. It is now clinicallyfeasible to perform dynamic analysis of mitral annulargeometry with the ready availability of information. Ap-preciation of this knowledge has the potential to objec-tively quantify mitral annular function and possibly fol-low results of therapy for FMR. In the future this mayhelp in the design and selection of annuloplasty rings forFMR.

This study was supported by the following grant: Echocardiog-raphy to Predict Recurrent Ischemic Mitral Regurgitation afterSurgical Mitral Valve Repair (RC Gorman, PI), National Instituteof Health, R01-HL 103723.

References

1. Gorman JH, 3rd, Gupta KB, Streicher JT, et al. Dynamicthree-dimensional imaging of the mitral valve and leftventricle by rapid sonomicrometry array localization. J Tho-rac Cardiovasc Surg 1996;112:712–26.

2. Nguyen TC, Itoh A, Carlhall CJ, et al. The effect of puremitral regurgitation on mitral annular geometry and three-dimensional saddle shape. J Thorac Cardiovasc Surg 2008;136:557–65.

3. Daimon M, Saracino G, Fukuda S, et al. Dynamic change ofmitral annular geometry and motion in ischemic mitralregurgitation assessed by a computerized 3D echo method.Echocardiography 2010;27:1069–77.

4. Little SH, Ben Zekry S, Lawrie GM, et al. Dynamic annulargeometry and function in patients with mitral regurgitation:insight from three-dimensional annular tracking. J Am SocEchocardiogr 2010;23:872–9.

5. Veronesi F, Corsi C, Sugeng L, et al. Quantification of mitral

apparatus dynamics in functional and ischemic mitral regur-

110 KHABBAZ ET AL Ann Thorac SurgDYNAMIC ANALYSIS OF MITRAL VALVE GEOMETRY 2013;95:105–10A

DU

LTC

AR

DIA

C

gitation using real-time 3-dimensional echocardiography.J Am Soc Echocardiogr 2008;21:347–54.

6. Watanabe N, Ogasawara Y, Yamaura Y, et al. Quantitation ofmitral valve tenting in ischemic mitral regurgitation bytransthoracic real-time three-dimensional echocardiogra-phy. J Am Coll Cardiol 2005;45:763–9.

7. Rausch MK, Bothe W, Kvitting JP, et al. In vivo dynamicstrains of the ovine anterior mitral valve leaflet. J Biomech2011;44:1149–57.

8. Jassar AS, Brinster CJ, Vergnat M, et al. Quantitative mitralvalve modeling using real-time three-dimensional echocar-diography: technique and repeatability. Ann Thorac Surg91:165–71.

9. Vergnat M, Jassar AS, Jackson BM, et al. Ischemic mitralregurgitation: a quantitative three-dimensional echocardio-graphic analysis. Ann Thorac Surg 2011;91:157–64.

10. Mahmood F, Gorman JH 3rd, Subramaniam B, et al.Changes in mitral valve annular geometry after repair:saddle-shaped versus flat annuloplasty rings. Ann ThoracSurg 2010;90:1212–20.

11. Mahmood F, Karthik S, Subramaniam B, et al. Intraoperativeapplication of geometric three-dimensional mitral valve as-sessment package: a feasibility study. J Cardiothorac VascAnesth 2008;22:292–8.

12. Mahmood F, Subramaniam B, Gorman JH 3rd, et al. Three-dimensional echocardiographic assessment of changes inmitral valve geometry after valve repair. Ann Thorac Surg2009;88:1838–44.

13. Grewal J, Suri R, Mankad S, et al. Mitral annular dynamicsin myxomatous valve disease: new insights with real-time3-dimensional echocardiography. Circulation 2010;121:1423–31.

14. Parish LM, Jackson BM, Enomoto Y, et al. The dynamicanterior mitral annulus. Ann Thorac Surg 2004;78:1248–55.

15. Ahmad RM, Gillinov AM, McCarthy PM, et al. Annulargeometry and motion in human ischemic mitral regurgita-tion: novel assessment with three-dimensional echocardiog-raphy and computer reconstruction. Ann Thorac Surg 2004;78:2063–8; discussion 2068.

16. Gorman JH 3rd, Jackson BM, Enomoto Y, et al. The effect ofregional ischemia on mitral valve annular saddle shape. AnnThorac Surg 2004;77:544–8.

17. Ray S. The echocardiographic assessment of functional mi-tral regurgitation. Eur J Echocardiogr 2010;11):i11–7.

18. Silbiger JJ, Bazaz R. Contemporary insights into the func-tional anatomy of the mitral valve. Am Heart J 2009;158:887–95.

19. Cheng A, Nguyen TC, Malinowski M, et al. Effects ofundersized mitral annuloplasty on regional transmural leftventricular wall strains and wall thickening mechanisms.Circulation 2006;114(1 Suppl):I600–9.

20. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommen-dations for evaluation of the severity of native valvularregurgitation with two-dimensional and Doppler echocardi-ography. J Am Soc Echocardiogr 2003;16:77–802.

21. Warraich HJ, Shahul S, Matyal R, Mahmood F. Bench tobedside: dynamic mitral valve assessment. J CardiothoracVasc Anesth 2011;25:863–6.

22. Veronesi F, Corsi C, Caiani EG, et al. Tracking of leftventricular long axis from real-time three-dimensional echo-

cardiography using optical flow techniques. IEEE Trans InfTechnol Biomed 2006;10:174–81.

23. Malpica N, Santos A, Zuluaga MA, et al. Tracking of regions-of-interest in myocardial contrast echocardiography. Ultra-sound Med Biol 2004;30:303–9.

24. Caban J, Joshi A, Rheingans P. Texture-based feature track-ing for effective time-varying data visualization. IEEE TransVis Comput Graph 2007;13:1472–9.

25. Foroughi P, Abolmaesumi P, Hashtrudi-Zaad K. Towardsreal-time registration of 4D ultrasound images. Conf ProcIEEE Eng Med Biol Soc 2006;1:404–7.

26. Kaplan SR, Bashein G, Sheehan FH, et al. Three-dimensional echocardiographic assessment of annularshape changes in the normal and regurgitant mitral valve.Am Heart J 2000;139:378–87.

27. Mihalatos DG, Joseph S, Gopal A, et al. Mitral annularremodeling with varying degrees and mechanisms ofchronic mitral regurgitation. J Am Soc Echocardiogr 2007;20:397–404.

28. Daimon M, Gillinov AM, Liddicoat JR, et al. Dynamic changein mitral annular area and motion during percutaneousmitral annuloplasty for ischemic mitral regurgitation: pre-liminary animal study with real-time 3-dimensional echo-cardiography. J Am Soc Echocardiogr 2007;20:381–8.

29. Sughimoto K, Takahara Y, Mogi K, et al. Annular excursioncontributes to efficient cardiac output: a three-dimensionalechocardiographic approach. J Heart Valve Dis 2010;19:244–8.

30. Mihaljevic T, Lam BK, Rajeswaran J, et al. Impact of mitralvalve annuloplasty combined with revascularization in pa-tients with functional ischemic mitral regurgitation. J AmColl Cardiol 2007;49:2191–201.

31. Song JM, Fukuda S, Kihara T, et al. Value of mitral valvetenting volume determined by real-time three-dimensionalechocardiography in patients with functional mitral regur-gitation. Am J Cardiol 2006;98:1088–93.

32. Tibayan FA, Wilson A, Lai DT, et al. Tenting volume:three-dimensional assessment of geometric perturbations infunctional mitral regurgitation and implications for surgicalrepair. J Heart Valve Dis 2007;16:1–7.

33. Cosgrove DM 3rd, Arcidi JM, Rodriguez L, et al. Initialexperience with the Cosgrove-Edwards annuloplasty sys-tem. Ann Thorac Surg 1995;60:499–503; discussion 503–4.

34. Rausch MK, Bothe W, Kvitting JP, et al. Mitral valve annu-loplasty: a quantitative clinical and mechanical comparisonof different annuloplasty devices. Ann Biomed Eng 2012;40:750–61.

35. Jimenez JH, Liou SW, Padala M, et al. A saddle-shapedannulus reduces systolic strain on the central region of themitral valve anterior leaflet. J Thorac Cardiovasc Surg 2007;134:1562–8.

36. Vergnat M, Levack MM, Jassar AS, et al. The influence ofsaddle-shaped annuloplasty on leaflet curvature in patientswith ischaemic mitral regurgitation. Eur J Cardiothorac Surg2012;42:493–9.

37. Rausch MK, Bothe W, Kvitting JP, et al. Mitral valve annu-loplasty: a quantitative clinical and mechanical comparison

of different annuloplasty devices. Ann Biomed Eng 2012;40:750–61.