New Approaches to Systolic Function: Strain Imaging

Size: px

Start display at page:

Download "New Approaches to Systolic Function: Strain Imaging"

Damian Williamson
5 years ago
Views:

Cardiovascular Institute Professor of Medicine, Feinberg School of Medicine,

1 New Approaches to Systolic Function: Strain Imaging James D. Thomas, MD, FACC, FASE Director, Center for Heart Valve Disease Bluhm Cardiovascular Institute Professor of Medicine, Feinberg School of Medicine, Northwestern University Chicago, Illinois Conflicts of interest: GE, Abbott, Edwards (honoraria) Spouse employment: Bay Labs

2 The Growing Importance of Strain Imaging in Clinical Echo Where is it Useful? James D. Thomas, MD, FACC, FASE Director, Center for Heart Valve Disease Bluhm Cardiovascular Institute Professor of Medicine, Feinberg School of Medicine, Northwestern University Chicago, Illinois Conflicts of interest: GE, Abbott, Edwards (honoraria) Spouse employment: Bay Labs

3 Welcome to Manila!

4 Myocardial Strain: What is It?? Strain: dimensionless index of change in length Strain ( ) = L-L 0 / L 0 LV strain may offer a pure index of regional LV function but is difficult to measure L 0 L

5 Strain Derived from Tissue Velocities Lat Septum -20% RV -30%

6 Limitations of TDI Strain Detects only single component of strain Limited scope of radial (anteroseptum and posterior wall) and circumferential (septum and lateral wall) strain from parasternal window Subject to noise, particularly strain rate Very tedious to perform

7 Derivation of 2D Strain by Echo ZOOM ZOOM BLOCK Y dy 0 dx New location Old location X

8 Not a New Idea, Just Better Implementation

9 Longitudinal Strain from B-Mode Echo Normal Subject

10 Longitudinal Strain Dilated Cardiomyopathy

11 Normal Apical 4 Chamber Apical 2 Chamber Apical Long-Axis

12 Bull's-eye Plot from 3 Apical Views Apical 4 Apical 2 Apical 3

Normal Global Strain and Strain Rate by Age 242 Patients in Cleveland, Brisbane, and Aachen Strain: -18.6 1.

13 Normal Global Strain and Strain Rate by Age 242 Patients in Cleveland, Brisbane, and Aachen Strain: % without age variation Age groups Marwick et al. JACC Imaging 2009; 2: 80-84

14 Echo Strain Gives Only Relative Shortening, Not Absolute Supine Dias Sys = -20% 1 G Dias Sys Reduced load = -15% Greater shortening

15 % 0 5 Impact of Preload on LV Strain Tilt Study in Astronauts, n= Supine G GLS p<0.001 Smaller heart J Am Soc Echocardiogr 2017;30:

16 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

17 MRI vs Echo Strain for Infarct Size Large infarct Medium infarct 115 g 36 g r r c l c l Gjesdal et al. Circ Img 2008; 1:

18 Detection of 30 g MI Size Strain WMSI Echo EF MRI EF Area under Curve Echo EF 0.81 MRI EF 0.90 WMSI 0.90 Strain 0.95 Gjesdal et al. Clinical Science 2007; 113:

19 LAD Infarct

20 RCA Infarct

21 LCx Infarct

22 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

23 Strain Imaging Diastolic Index SI-DI = % of systolic wall thickening lost in the first 1/3 of diastole: (A-B)/A Ishii et al. JACC 2009; 53:

24 60-year-old Male: Exertional Angina Radial Strain Before and After Exercise mid-sep mid-post Baseline immediately after 5 min after Ishii et al. JACC 2009; 53:

25 60-year-old Male: Effort Angina Left coronary artery (#7: 99% #9:90% stenosis) Ishii et al. JACC 2009; 53: Right coronary artery

26 Predictive Accuracy of SI-DI for Detecting Coronary Stenoses Ishii et al. JACC 2009; 53:

27 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

28 Assessment of Myocardial Viability Improvement in Ejection Fraction vs Resting Radial Strain and MRI Markers Echo Radial Strain Gd-MRI Becker et al. JACC 2008; 51:

29 Prediction of Regional and Global Recovery vs Resting Radial Strain and MRI Markers Becker et al. JACC 2008; 51:

30 Post-Systolic Strain vs Recovery of Function Post NSTEMI Initial pattern LCx occlusion PSS Post-PCI r = p < Eek et al. EJE 2011; 12: 483-9

31 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

32 Myocardial Mechanical Dispersion 85 patients after myocardial infarction with ICD 2.3 ( ) years follow up Haugaa et al. JACC Img 2010; 3:

33 Myocardial Mechanical Dispersion 85 patients after myocardial infarction with ICD N=23 N=47 N=38 Haugaa et al. JACC Img 2010; 3: ( ) years follow up

34 85 post MI patients meeting ICD criteria Followed until arrhythmia requiring appropriate ICD intervention No difference in EF: 34±11% vs 35±9% Mechanical dispersion: 85±29 msec vs 56±13 msec, p<0.001 Haugaa et al. JACC Img 2010; 3:

35 Myocardial Dispersion in DCM Better Predictor than EF Haugaa et al. JASE 2012;25:667-73

36 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

37 Can Echo and Biomarkers Predict Chemo-Induced LV Dysfunction? Sawaya et al. AJC 2011; 107:

38 ROC Curves for 3 Month Findings to Predict 6 Month Cardiotoxicity GLS Troponins Sawaya et al. AJC 2011; 107:

40 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

41 Increased Strain After Percutaneous AVR Pre Post

42 Impact of Strain and BNP on Survival in AS 531 Pts w/ AVA<1.3 cm 2 and LVEF>50% vs symptoms vs STS score Goodman et al. JAHA 2016; 5: e002561

43 Impact of Strain on Post MVr LV function 233 Pts, Prediction of 12+ month LVEF<50% GLS = % 90% sensitivity, 79% specificity Predictors of LV dysfunction OR p GLS <19.9% 23.2 <0.001 LVESD>40mm GLS accounted for 3.3x more variance than EF EF<60% Symptoms AF = 69.1, p<0.001 Closer to zero = Bad? Witkowski et al. EHJ-CVI 2013; 14 :69-76

44 Impact of Strain on Post MVr LVEF 130 Pts, Prediction of fall in LVEF>10% Post-Op Farther from zero = Bad? Pandis et al. JASE 2014; 27:

45 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

46 HCM Phelan, Collier et al. Heart 2012; 98: AS

47 Amyloid: Apex > 2x Base & Mid Phelan, Collier et al. Heart 2012; 98:

48 74yo Man with Dyspnea and?lflgas Apex: -9.8% Nonapex: -1.3% Ratio: 7.4x

49 Differentiation of Amyloid, HHD, and HCM Value of Strain Maps 20 level 3 echo readers reviewed 24 cases of LVH 8 amyloid, 8 HTN, 8 HCM Full echo Doppler data presented but NOT strain or clinical info Readers chose 1 of 3 dx, and listed their confidence Exercise repeated some weeks later with reorded cases, now including GLS and strain maps Reader concordance improved with strain data ( = 0.28 to 0.57) Reader confidence improved with strain data (p<0.001) Reader accuracy improved with strain data (65% to 82%, p<0.001) Especially for amyloid! Sensitivity Specificity (%) (%) Accuracy (%) PPV (%) NPV (%) Base Read Strain Read P < <0.001 <0.001 <0.001 Phelan, et al. JASE 2014; 27:

50 25 cases each of amyloid, Fabry, Friedreich ataxia, HTN and controls CA Liu et al. Circ CVI 2013; 6:

51 25 cases each of amyloid, Fabry, Friedreich ataxia, HTN and controls Liu et al. Circ CVI 2013; 6:

52 Hypertrophic Cardiomyopathy Reduced Strain in Most Affected Areas 3.8 cm septum but nonobstructive Markedly abnormal in base-mid septum

53 Apical HCM Takotsubo Reverse Takotsubo

54 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function

55 Clinical Applications of Strain Imaging Coronary artery disease Detection and sizing of myocardial infarction Detection of myocardial ischemia Assessment of myocardial viability Prediction of post-mi arrhythmias Detection of subclinical LV dysfunction Chemotherapy Valvular heart disease Cardiomyopathy Amyloid Assessment of RV function Tune in tomorrow!

56 RV Strain by Speckle Tracking Pretending the RV is a Reversed LV Longitudinal Strain

57 Strain for PAH Prognosis Cox Proportional Hazard Analysis Event-free survival GLSRV < -15.5% P=0.022 GLSRV -15.5% Survival GLSRV < -15.5% P=0.017 GLSRV -15.5% Follow-up duration (months) Follow-up duration (months) Park JH, et al. J Cardiovasc Ultrasound 2015; 23: 91-9

58 Fine NM, et al. Circ Imaging 2013; 6:

59 Achilles Heel: Intervendor Reproducibility

60 Imagine This Situation 61 yr. old real estate agent Avid bowler, no symptoms AVA = 0.85 cm 2 Peak/mean grad. = 72/46 mm Hg No surgery now, follow-up in one year

61 One Year Later Our Bowler Returns for Follow-Up Staff: Wait we can t image him on the GE machine Sonographer: Why not. It s a good machine. Staff: That s irrelevant. We studied him on an ATL machine last year, and we know ATL and GE give different gradients. He always needs to be studied with an ATL. Sonographer: But we got rid of the ATL last week. Staff: Uh, oh

62 Intervendor Reproducibility in Normals V1: GE V2: Philips V3: Toshiba Radial Circumferential Longitudinal Vendor 1 vs 2 Vendor 1 vs 3 Vendor 2 vs 3 Circ J 2012; 76:

Jim Thomas Academics Technical Task force Biomedical / clinical Societies Japanese Society on Echocardiography Korean Society on

63 Task Force on Strain Standardization European Association on CardioVascular Imaging American Society on Echocardiography Dr. Luigi Badano Dr. Jim Thomas Academics Technical Task force Biomedical / clinical Societies Japanese Society on Echocardiography Korean Society on Echocardiography Industrial stakeholders Hardware/software vendors: Esaote, GE VingMed, Hitachi-Aloka, Mind Ray, Philips, Siemens, Toshiba, (Samsung Medison) Software only vendors: Epsilon Imaging, TomTec 1) Reach consensus on definitions 2) Reach consensus on a set of tools for QA

64 Standardization in Several Settings Reliability of the reference method Synthetic data First focus of the task force In vitro models Experimental Animals In vivo (patients) All levels have benefits and require attention Courtesy Jan D Hooge Level of realism of the data set

JASE 2015; 28: 183-93 Agreement on Nomenclature Display Timing of

65 JASE 2015; 28: Agreement on Nomenclature Display Timing of end-diastole Timing of end-systole Mathematical definition of strain Regions of interest

66 Tracking examples Philips Courtesy Jan D Hooge

67 D hooge et al. EHJ-CVI 2015 (on line)

68 One week: 22nd - 26th April 2013, Leuven 7 machines: Esaote MyLab Alpha GE Vivid E9 Hitachi-Aloka Prosound Alpha7 CV Philips ie 33 Samsung EKO 7 Siemens SC 2000 Toshiba Artida 2 independent Epsilon EchoInsight softwares: Tomtec Image Arena JASE 2015; 28:

69 Inter-Vendor Comparison Study Energetics: 7 machines (1000 W) 21 persons (250 W) 12 kw power + heat! Courtesy of Jens-Uwe Voigt

70 Mean GLS Values All Vendors GLS AV (%) Hitachi-A. Esaote GE Philips Samsung Siemens Toshiba Epsilon Tomtec Mean of all JASE 2015; 28:

71 Regression Analysis All Vendors vs Average Values Hitachi-A. Esaote GE -30 y = 0.94x R² = y = 1.01x 0.52 R² = y = 1.10x R² = Philips Samsung Siemens -30 y = 1.01x R² = y = 1.03x R² = y = 1.03x 0.06 R² = Toshiba Epsilon Tomtec -30 y = 0.90x - 1,00 R² = y = 0.85x 1.99 R² = y = 1.13x R² = JASE 2015; 28:

72 Hitachi-A Esaote GE Philips Samsung Siemens Toshiba Epsilon Tomtec Bland-Altman Analysis All Vendors vs Average Values JASE 2015 (on-line 7/22/15)

73 Intraobserver mean error (%) Interobserver Variability, GLS vs EF EF is worst! 0 Hitachi-A. Esaote GE Philips Samsung Siemens Toshiba Epsilon Tomtec EFbi Intervendor Variability, GLS vs Other Parameters Mean error (%) 10 5 GLS is best! GLS JASE 2015; 28: E E/A IVS LVEDD PW GLSAV

74 Yang et al. JASE 2015; 28: 642-8

75 Strain Standardization Task Force Next Steps Regional longitudinal strain Circumferential and radial strain RV strain Atrial strain Strain from 3D echoes

76 One week: Inter-Vendor Clinical Study Round 2 20th - 24th April 2015, Leuven 7 machines: Esaote MyLab Alpha GE Vivid E9 Hitachi-Aloka Prosound Alpha7 CV Philips ie 33 Samsung EKO 7 Siemens SC 2000 Toshiba Artida 2 independent Epsilon EchoInsight softwares: Tomtec Image Arena Subjects: 63 pts, many with regional dysfunction All with MRIs with Gd Focus on regional strain Courtesy of Jens-Uwe Voigt, published in JACC Imaging 2017

77 Emerging Applications Transmural strain 3D strain Seo et al. Circ Img 2009; 2: 451-9

303 healthy subjects, 18-82 years 10 centers in 8 countries (US,

Caucasian, 156M, 147F 3D strain using Toshiba Results Minor gender

strain most reproducible (5% relative standard error), vs GLS 8%, GCS

78 303 healthy subjects, years 10 centers in 8 countries (US, Spain, Germany, Netherlands, Hungary, Switzerland, UK, Austria) All Caucasian, 156M, 147F 3D strain using Toshiba Results Minor gender difference (GLS) Significant age effect (GLS falls, others rise) Area strain most reproducible (5% relative standard error), vs GLS 8%, GCS 6%, and GRS 15%» Segmental strain worse (11-28%) p = 0.04 p < p = 0.04 p = 0.005

79 The Heart is an Elegantly Complex Machine Fiber Sheet Oriented LV Strain Model J. Hassan P. Hunter

80 Principal strain Pedrizzetti et al. JASE 2014; 27: Principal strain rate

81 Subclinical LV dysfunction (GLS) Cancer chemotherapy (prior or current) Unexplained dyspnea (HFpEF vs pulmonary) Valvular heart disease (AS, AR, MR) LV hypertrophy (bullseye pattern) Highly predictive for cardiac amyloid CAD Where Do I Use Strain Imaging Localize infarct, predict arrhythmias??role in stress and dobutamine echo Assessment of RV function ANY suspicion for RV enlargement or dysfunction

82 Thanks!

The Basics of Systolic Function

The Basics of Systolic Function James D. Thomas, MD, FACC, FASE Director, Center for Heart Valve Disease Bluhm Cardiovascular Institute Professor of Medicine, Feinberg School of Medicine, Northwestern