076 cardiac magnetic resonance imaging

Detecting Acute Coronary Syndrome in the Emergency Department with Cardiac Magnetic Resonance Imaging

Raymond Y. Kwong MDCo-Director, Cardiac Magnetic Resonance Imaging

Brigham and Women’s Hospital

Cardiology Grand RoundBeth Israel Deaconess Hospital

April 11th, 2003

Acknowledgments

NHLBI• Andrew E. Arai, MD• Robert S. Balaban, PhD• Anthony Aletras, PhD• Adam Schussheim, MD• Suresh Rekhraj, MD• W. Patricia Ingkanisorn, MD• Kenneth Rhoads, MD

Clinical Center• Janice Davis, RN• Grace Graninger, RN

Suburban Hospital• Cardiology

Eugene Passamani, MD

• Emergency RoomBob Rothstein, MD

• RadiologyWayne Olan, MDSusan O’Flahavan, RTPaul LeBlanc, RTGian Serafini, RTChris Mancini, RTSarah Pirie, RT

Myocardial Infarction and Chest Pain in the Emergency Room

• 1,100,000 MI per year in the USAbout 1/3 will die from their MI *

• 6,200,000 people in the United States suffer from angina

• 5,000,000 patients are evaluated in ER for chest pain each year

American Heart Association. 1999 Heart and Stroke Statistical Update.

* Atherosclerosis Risk in Communities (ARIC) study, NHLBI

The Spectrum of Chest Pain in the ER

Chest Pain

STEMI

NSTEMI

UnstableAngina

ACSTroponin

Non-cardiacChest Pain

StableAngina

Mortality of ER patients with chest pain

Pope et al. NEJM 2000; 342: 1163

Chest Pain in Emergency Department

LCE/NHLBI/NIH

ST ElevationAcute MI

PTCA orThrombolytics

IntermediateLikelihood CAD

CardiacRx

Non-STEMI

UnstableAngina

Low LikelihoodCAD

CardiacEnzymes

Regional Wall Motion: SSFP

Dobutamine Stress MRI Sensitivity 83-86% Specificity 83-86%Nagel Circ 1999; 99: 763Hundley Circ 1999; 100: 1697

Interobserver Variability for Qualitative Regional Wall Thickening by MRI

VersusConsensu

s

Observer1

Observer2

Observer3

Observer4

Correlation 0.95 0.96 0.93 0.95

Avg Dif -0.02 -0.03 0.01 -0.02

SD of Dif 1.3 1.6 2.0 1.3

Sierra-Galan et al. JCMR in press

First PassPerfusion RV

LV

Myo

LCE/NHLBI/NIH

Comparison of Endocardial Microsphere Blood Flow and Endocardial MRI Contrast Enhancement Ratio

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

y = 0.04 + 0.85x r = 0.92

Endocardial Microsphere Flow (ml/min/g)

MR

I End

ocar

dial

CER

NHLBI/NIH

Dipyridamole Stress MRI vs PET and QCA

Schwitter et al. Circ 2001; 103:2230

Dipyridamole Stress MRI vs PET and QCA

Schwitter et al. Circ 2001; 103:2230

MRI Vs PETSensitivity 91%Specificity 94%

MRI Vs QCASensitivity 87%Specificity 85%

Dipyridamole Stress First Pass Perfusion Myocardial Perfusion Reserve Index

Al-Saadi et al Circ 2000; 101:1379

MPR +

54

7

MPR -

6

35

Sensitivity 90%Specificity 83%

CAD+

CAD -

0

50

100

150

200

0 5 10 15 20

Time (s)

Sign

al In

tens

ity

Width = 0

Fermi Function

Decay rate =

Flow

0

500

1000

1500

2000

2500

3000

0 5 10 15 20

Time (s)

Sign

al In

tens

ity

Fermi Function Deconvolution

Estimates Absolute Myocardial Perfusion(ml/min/g)

InputFunction

MyocardialEnhancement

MathematicalModel

Endocardial and Epicardial AnalysisR = 0.92, P< 0.0001 R = 0.95, P < 0.0001

Infarct Size: TTC vs MRI

Kim R. et al. Circulation. 1999;100:1992-2002

Correlation between MRI Infarct Size and TTC Staining

Kim R. et al. Circulation. 1999;100:1992-2002

Kim RJ et al. NEJM 2000;343:1445

Irreversible LV Dysfunction: Pre-intervention

After revascularizationDiastole Systole

Diastole Systole

Transmural Hyper

0

20

40

60

80

100

0 1-25 26-50 51-75 76-100

0

20

40

60

80

100

0 1-25 26-50 51-75 76-100

Transmural Extent of Hyperenhancement Predicts Recovery of Function

Transmural Extent of Hyperenhancement (%)

Con

tract

ile R

ecov

ery

(%)

Con

tract

ile R

ecov

ery

(%)

Kim RJ et al. NEJM 2000;343:1445

All segments Akinetic or Dyskinetic

PET vs MRI Viability

Klein et al. Circulation 2002; 105: 162

Microinfarction after PCI associated withMinor Side Branch Occlusion

Ricciardi. Circulation 2001;103:2780-3

Patient 7: stent in the proximal LAD and minor side-branch

occlusion

Patient 2: stent in the mid-PDA and minor side-branch

occlusionMI size = 0.7 to 12 g

Comparison of MRI and SPECT for Detecting Myocardial Infarction

Wagner et al. Lancet 2003; 361: 374

Method: Acute Chest Pain Protocol

Localize Perfusion Function Viability

Time (min)0 10 20 30

Pilot phase

• To determine safety, feasibility, and logistic issues:• 37 patients presenting with chest pain to the ER were studied.

– Patients were either from the ER or within 72 hours of admission to the hospital.

» Well tolerated (1 failure to acquire any image due to clasutophobia)

» 13 cases of AMI with a wide range of troponins all have abnormal MRI

» Can perform MRI while receiving concurrent IV meds.

Case 1: 60 yo male Day 2 post acute MITroponin 2X ULN

Case 2: 80 yo female No prior CAD Chest pain with non-diagnostic EKG admitted 12 hours ago, small NQMI by troponin

Case 2: 70 yo male No prior CAD Chest pain with non-diagnostic EKG admitted 12 hours ago, small NQMI by troponin

Hypothesis

• We hypothesized that MRI could effectively triage patients presenting with possible acute coronary syndromes in the Emergency Room

Entry Criteria

• 30 minutes of chest pain compatible with myocardial ischemia

• ECG not diagnostic of ST-elevation MI

Exclusions• Pacemaker• Defibrillator• Other implanted active devices• Brain aneurysm clips

Method

Prospective observational trial.• Enrollment from 01/10/2000 to 09/27/2001

• Inclusion criteria: • Chest pain > 30 minutes without obvious non-cardiac cause

• Exclusion criteria: any of the following• ECG diagnostic of ST elevation acute myocardial infarction• Ongoing active chest pain• Hemodynamic instability• Contraindications to MRI scanning• > 12 hours since onset of last chest pain

.

Symptoms Suggestive of ACSNoncardiacdiagnosis

StableAngina

PossibleACS

DefiniteACS

ACC/AHA Practice GuidelinesUnstable Angina And NSTEMICirc 2000;102:1193

ST ElevationMI

Confirmed ACS

ECG NotDiagnostic

Troponin (-)

SerialStudies? Image

StressTest

- +

+- + -

+

UnstableAngina

ST or Tand/or

Troponin (+)

+Non-ST

ElevationMI

-

Method: Prospectively Defined Clinical Endpoints

Acute coronary syndrome (ACS): Possible or definite ACS according to ACC/AHA guidelines

Method: Prospectively Defined Clinical Endpoints

Acute coronary syndrome (ACS): Possible or definite ACS according to ACC/AHA guidelines as indicated by chest pain > 30 minutes (an entry criteria) AND either angiographically significant CAD AND/OR significantly abnormal stress testing performed during index hospitalization or the subsequent 6- 8 week follow-up period.

Non-ST elevation acute myocardial infarction (NSTEMI): Abnormal troponin-I with a temporal pattern consistent with acute MI and clinical evidence of coronary artery disease (Coronary angiography, echo, or noninvasive imaging).

Method: Definitions

Angiographically significant CAD =

1) > 50% left main lesionor 2) > 70% epicardial coronary artery stenosis

Method: Prospectively Defined Abnormal MRI

Abnormal defined by either:• Regional wall motion abnormality (RWMA), or• Myocardial hyperenhancement (Hyper).

Reading included perfusion images:• Although independent diagnoses were not made using

perfusion alone, the perfusion images helped identify abnormal regions.

TIMI risk score

Antman et al. JAMA 2000; v284;7:835

Results: Study population

• 193 consecutive patients • 11 refused to participate• 21 excluded:

– 6 hemodynamically unstable– 2 metallic implants– 3 large body size– 10 claustrophobia

• Remaining 161 formed the study cohort.• All patients underwent conventional cardiac

workup as determined by the admitting cardiologists or ER physician

161 enrolled•92 Male : 69 Female

•Age: 59 + 15 years old

•Average # of risk factors: 3.4

•History of prior MI: 25/161

•Median time since ER arrival:2.7 hours (off hours cases excluded)6.0 hours (off hours cases included)

•Average MRI scan duration: 38 + 12 minutes

Results: Study population

Patient Follow-up at 6 –8 weeks

161 patients• 158 (98%) were successfully contacted at 6-8 week follow-up

• 3 were lost to follow-up: all troponin negative, 2 clinical low risk with normal MRI and 1 high risk with abnormal MRI

Results: Demographic summary ACS

(n = 25) No ACS (n = 136)

p value

Age (yr) 68 + 13 57 + 14 p = 0.0006 Men (%) 60 57 NS CAD risk factors Advanced Age (M>45, W>55) (%) 92 68 p = 0.03 Hypertension (%) 56 43 NS Diabetes (%) 28 10 p < 0.001 Hypercholesterolemia (%) 64 47 NS CAD in family (%) 32 42 NS Hx. of smoking (%) 48 39 NS Total number of CAD risk factors 4.2 3.2 p = 0.003 Characteristics of Chest Pain Location Substernal (%) 32 31 NS Precordial (%) 24 18 NS Quality Crushing, heaviness, pressure or tightness (%) 56 54 NS Radiation to neck/arm (%) 32 31 NS Dyspnea, diaphoresis, or nausea (%) 32 40 NS Previous angina (%) 40 12 p = 0.001 Average Chest Pain Score 8.4 7.5 NS ( p= 0.06) TIMI risk score P < 0.00011.6 + 0.82.5 + 1.2

Table 1. Demographic Summary

ACS

n = 25

68 + 13

28%

40%

8.4

2.5 + 1.2

No ACS

n = 136

57 + 14

10%

12%

7.5

1.6 + 0.8

p value

0.0006

0.001

0.001

0.06

0.0001

Age

Diabetes

HTN

CP Score

TIMI Risk Score

Kwong et al. Circulation 2003; 107:531-7

11620No ACS

421

+ MRI

ACS

- MRI

Results: Qualitative MRI assessment compared with clinical evaluation ( N = 161)

    MRI Historyof MI ECG

 StrictECG

Initial Troponin-

I

Peak Troponin-

I

TIMIRisk Score

ACS

Sensitivity (%)

84 28 80 19 16 40 76

Specificity (%)

85 87 61 95 97 97 52

NonSTEMI

Sensitivity (%)

100 30 70 20 44 NA 80

Specificity (%)

79 85 56 94 97 97 50

IHD

Sensitivity (%)

91 51 79 14 10 24 81

Specificity (%)

98 97 67 96 97 97 58

Sensitivity and Specificity for ACS

0

10

20

30

40

50

60

70

80

90

100

MRI ECG ECGstrict

Trop Troppeak

TIMI

Sens

itivi

ty o

r Spe

cific

ity (%

)

NS

0.00

1

0.00

1 0.00

4

NS

0.00

1

0.00

1

0.01

1

0.00

1

0.00

1

Sensitivity

Specificity

Kwong et al. Circulation 2003; 107:531-7

Multivariate Logistic Regression Analysis: Predict Acute Coronary Syndrome

0

20

40

60

80

100

0 20 40 60 80 100False Positive Portion

(1 - specificty, %)

True

Pos

itive

Por

tion

(Sen

sitiv

ity, %

)

TIMI Risk Score

NIH with MRI

NIH without MRI

Kwong et al. Circulation 2003; 107:531-7

Detecting CAD by presence of Regional Wall Motion Abnormality (RWMA)

1231No IHD

864IHD

No RWMARWMADetect IHD

Sensitivity: 89% Specificity: 99%

Regional wall motion images were interpretable in 196/197 patients (99%)

Detecting CAD by Delayed Hyperenhancement (HYPER)

1121No IHD

2448IHD

HYPERDetect IHD

Sensitivity: 67% Specificity: 99%

No HYPER

Delayed hyperenhancement images were acquired and interpretable in 185/197 patients (94%)

Methods

Anterior Septum

Lateral wall

RWMA

Inferior Septum

D S

End-systolic thickness (S) in mm

End-diastolic thickness (D) in mm

Absolute wall thickening in mm = S - D

Systolic diastolic ratio = S/D

Percent change in wall thickening (%) = (S-D)/D x 100%

Methods

0

20

40

60

80

100

0 20 40 60 80 100False Positive Portion

(1 – specificity, %)

True

Pos

itive

Por

tion

(Sen

sitiv

ity, %

)

0

20

40

60

80

100

0

20

40

60

80

100

0 20 40 60 80 100False Positive Portion

(1 – specificity, %)

0 20 40 60 80 100False Positive Portion

(1 – specificity, %)

ACS Non-STE MI IHD

Receiver Operator Characteristic Analysis: Quantitative Wall Thickening

3220 Kwong et al. Circulation 2003; 107:531-7

84 year old female with Acute Non-ST Elevation MI

-20

0

20

40

60

80

100

120

140

160

0 10 20 30

Time (image number)

Sign

al In

tens

ity-50

050

100150

200250

300350

400450

500

0 10 20 30

Time (image number)SI

Inte

gral

33%

68 year old female, with 2 hours CP, no risk factors, nonspecific ECG, and normal troponin

68 year old female, with 2 hours CP, no risk factors, nonspecific ECG, and normal troponin

A 53 year old female with 3 days of intermittent rest chest pain

A 53 year old female with 3 days of intermittent rest chest pain

A 53 year old female with 3 days of intermittent rest chest pain

Adenosine Stress

Rest SSFP Adenosine Perfusion Gd Hyperenhancement

67 year old female with no prior CAD+ DM, FMH, Tob, HTNTroponin (-)Cath: RCA 99%, Cx 99%, LAD 75%

Cardiac MRI UnitThe General Infirmary at Leeds

Diastolic Systolic

Perfusion

LAD/LCX RCA

Delayed image

MRI conclusion: significant mid RCA lesion

X-ray angiogram

Comprehensive Approach to Chest Pain:MRI: Does It Have Any Role?

• ST Elevation MI• Post-MI risk stratification

• Detection of residual ischemia

• Viability

• Non-ST elevation MI• Same as ST elevation MI

• Also potential for more rapid diagnosis

• Unstable angina• Early rest scan to detect evidence of recent ischemia

• Alternatively, scan after excluding MI with a stress test

Conclusions

• Use of cardiac MRI in the ER in assessment of patients presenting with chest pain and a non-diagnostic ECG is feasibility and safe.

• Cardiac MRI has high sensitivity and specificity for detecting ACS, NSTEMI, and IHD in patients with chest pain, and could provide useful diagnostic information beyond clinical assessment.

• MRI parameters could be interpreted both qualitatively and quantitatively with high accuracy in prediction of clinical endpoints

• Further work is needed to differentiate acute from chronic myocardial infarction and to improve scan efficiency.

Conclusions

• Add a slide:

The additional benefit of adenosine stress testing

Cost effectiveness of early patient triage translating into savings in health care cost dollars

12 months prognostic data

Clinical Cases

50 yo male lawyer, atypical chest painFirst troponin negative

60 yo female presenting with an episode of prolonged chest pain 24 hours ago, then another episode for several minutes.

60 yo female presenting with “good story” of unstable angina. EKG: NSST changesFirst troponin 3X ULN

Cath during index hospitalization: NS stenosis Uneventful at 6 weeks on no medications