Optimizing Assessment of Cardiovascular Risk: Selecting a

Screening Test for Atherosclerosis

Nathan D. Wong, Ph.D., F.A.C.C.

Professor and Director

Heart Disease Prevention Program

University of California, Irvine, CA, USA

Why assess subclinical atherosclerosis?

• CHD events often occur with no clinical history and often normal risk factors (e.g., total cholesterol is a poor predictor)

• Most heart attacks result from coronary lesions of less than 50% stenosis, often undetectable by traditional methods (exercise test, angiogram)

• Need to identify disease early enough to intervene to prevent clinical events

Total Cholesterol Distribution: CHD vs Non-CHD Population

Adapted from Castelli WP. Atherosclerosis. 1996;124(suppl):S1-S9.

35% of CHD 35% of CHD Occurs in Occurs in People with People with TC<200 mg/dLTC<200 mg/dL

150 200

Total Cholesterol (mg/dL)

250 300

No CHD

CHD

Framingham Heart Study—26-Year Follow-up

Most Myocardial Infarctions Are Causedby Low-Grade Stenoses

Pooled data from 4 studies: Ambrose et al, 1988; Little et al, 1988; Nobuyoshi et al, 1991; and Giroud et al, 1992.(Adapted from Falk et al.)

Falk E et al, Circulation, 1995.

PDAY: Percentage of Right Coronary Artery Intimal Surface Affected With Early Atherosclerosis

PDAY=PDAY= Pathobiological Determinants of Atherosclerosis in Youth.Strong JP, et al. JAMA. 1999;281:727-735.

Fatty streaks

Raised lesions

White15-19 20-24 25-29 30-34

0

10

20

30 Women

0

10

20

30

15-1920-2425-2930-34Black

Age (y)

0

10

20

30

White15-19 20-24 25-29 30-34

Men

Black15-19 20-24 25-29 30-34

0

10

20

30

Intimalsurface

(%)

Features of Using Surrogate Measures of Atherosclerosis

• Use as intermediate endpoints rather than waiting until hard endpoints occur

• May have implications for identifying and tracking earlier, subclinical disease

• Compared to clinical event studies, studies of surrogate endpoints are lower cost, require fewer subjects, less-follow-up

• Use to follow progression of disease and effects of medical intervention

Atherosclerosis Screening: Required Characteristics of a Good Test

• Identify high and low risk groups more accurately--low false (+) and and false (-)

• Reproducible measures• Better identification of high

risk individuals (improved discrimination)

• Predict outcomes• Safe for subjects• Cost-effective• Educate the public

Primary Modalities for Measuring Atherosclerotic Burden

• Carotid B-mode ultrasonography: intimal medial thicknesses

• Ankle-Brachial Index (ABI) for assessment of peripheral vascular disease

• CT (EBT or multislice detectors): coronary calcium score or volume

• Magnetic resonance imaging of carotid plaques: vessel wall area

• Intravascular ultrasound (invasive)

Other Noninvasive Measures Associated with Cardiovascular Risk• Echocardiographic left ventricular mass /

hypertrophy, systolic (incl. Ejection fraction) and diastolic function

• Pulse wave velocity / arterial compliance for assessing large artery stiffness

• Brachial artery reactivity testing (BART)

Carotid B-Mode Ultrasonography

• Measurement of intimal medial thickness

• Non-invasive, inexpensive, no radiation

• Well-established as an indicator of cardiovascular risk from epidemiologic studies

• Published clinical trials on utility of carotid IMT as measure of progression of atherosclerosis and effects of therapy

Cardiovascular Health Study: Combined intimal-medial thickness

predicts total MI and stroke

Cardiovascular Health Study (CHS) (aged 65+): MI or stroke rate 25% over 7 years in those at highest quintile of combined IMT (O’Leary et al. 1999)

Carotid IMT: Clinical Trials

• ACAPS showed lovastatin therapy in those aged 40-79 with elevated LDL-C to be associated with regression in maximal IMT (Furberg et al., Circ 1994; 90: 1679-87)

• KAPS showed in men 45-65 a 45% lower rate of progression in those treated with pravastatin (Salonen et al., Circ 1995; 92: 1758-64).

Coronary Calcium and Atherosclerosis: Pathology Evidence

• Coronary calcium invariably indicates the presence of atherosclerosis, but atherosclerotic lesions do not always contain calcium (1-3).

• Calcium deposition may occur early in life, as early as the second decade, and in lesions that are not advanced (4-5).1) Wexler et al., Circ 1996; 94: 1175-92, 2) Blankenhorn and Stern, Am J Roentgenol 1959; 81: 772-7, 3) Blankenhorn and Stern, Am J Med Sci 1961; 42: 1-49, 4) Stary, Eur Heart J 1990; 11(suppl E): 3-19, 5) Stary, Arteriosclerosis 1989; 9 (suppl I): 19-32.

Coronary Calcium Evaluation: EBT and Multidetector Scanners

• Non-invasive, moderate cost, convenient• Detects location and quantity (score, mass, volume)

of coronary calcium, estimating “burden of atherosclerosis”

• Highly sensitive for angiographic disease, specificity variable depending on standard

• Reproducibility good (best at higher scores) but similar in EBT vs. multidetector scanners

• Artifact “noise” greater in EBT scanners

Significant Coronary Artery Calcium (Score >400)

Case #1 – 77 yo female w/atypical angina, hypercholesterolemia (LDL 179) and family history

of early CAD: CAC=633, Aortic Ca=2637.

Siemens VZ scanner, Cedars-Sinai Medical Center, Los Angeles, CA

52.6

40.9

29.8

20.420.6

37.6

53.5

25.424.5

50.8

58.5

75.3

0

10

20

30

40

50

60

70

80

Men Women Men Women

Pre

vale

nce

of

CA

C (

%)

Neither MetS Diabetes

ANY CALCIUM CALCIUM > 75th %tile

P=.0002

P=.02

P=.01

P=.19

Prevalence of Coronary Artery Calcium (CAC) by Disease Category and Gender (Wong et al., J Am Coll Cardiol 2003; 41: 1547-3)

Multiple Logistic Regression Odds of Calcium by Metabolic Syndrome Risk Factor Status

1.3 1.4 1.4 1.5

1.8

1.4

1.61.5 1.4

1.9

2.8

2.3 2.3

2.5

1.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

Od

ds

Ra

tio

(O

R)

1 MetS RF 2 MetS RF 3 Mets RF 4-5 MetS RF Diabetes

Any Coronary Calcium Coronary Calcium >=75th%tileAortic Calcium

Odds of calcium in comparison to those with no MetS RF, adjusted for age, sex, cigarette smoking, and total cholesterol; all comparisons 0.00001<p<0.05

24.721.2

40.9

0

10

20

30

40

50

Pro

po

rtio

n o

f S

ub

ject

s

CAC>=75th%tile10-Year CHD Risk >20%Either or Both

Figure 4. Proportion of Persons Identified with the Metabolic Syndrome at >20% 10-Year Estimated Risk of CHD and/or Significant Coronary Artery Calcium (CAC) (>75th percentile)

Wong et al., J Am Coll Cardiol 2003; 41: 1547-53

Risk of Total Cardiovascular Events by Calcium Quartile (n=881)

(compared to those with no calcium; age and risk-factor adjusted)Wong ND et al., Am J Cardiol; 86: 295-8

0123456789

Relative Risk (RR)

1 to 15 16 to 80 81 to 270 271 +

Total Calcium Score

Raggi Am Heart J 2001; 141: 375

Electron Beam Tomograpy Improves Prediction of Hard CHD

Risk of Total Mortality by Calcium Category in 10,377 Asymptomatic IndividualsShaw LJ et al., Radiology 2003; 228: 826-33

Risk of Total Mortality by Calcium Category in 10,377 Asymptomatic Individuals

(compared to those with no calcium; age and risk-factor adjusted)Shaw LJ et al., Radiology 2003; 228: 826-33

00.5

11.5

22.5

33.5

44.5

Relative Risk (RR)

11-100 101-400 401-1000 >1000

Total Calcium Score

Calcium was an independent predictor of mortality (p<.001); ROC concordance index increased from 0.72 to 0.78 (p<.001) for calcium added to model containing risk factors

Coronary Calcium Progression

Calcium Score: 56 Calcium Score: 90 Calcium Score: 128 Volume Score: 45 Volume Score: 78 Volume Score: 113

1993 1995 1997

Progression of Right coronary artery calcium score over 5 years

Coronary Calcium Progression

• Studies of serial EBT scanning show annual progression of 22-52% per year

• Observational study of statin therapy in 149 patients showed calcium volume over 1 year:

– 52% progression in those not treated

– regression (-7%) calcium volume among those treated aggressively to LDL-C <120 mg/dl, and

– moderate progression (25%) in those treated less aggressively (LDL-C >120 mg/dl)

– (Calister et al., NEJM 1998; 339: 1972-8).

Flow Diagram Showing Interaction Between EBCT Results and Clinical Management

(Taylor et al., Western J Med 1999; 171: 339-41)

L ow risk fo r C V D even tsR isk fac to r m od ifica tion p er exis t in g g u id e lin es

N eg a tive (n o ca lc iu m )

S u rvey fo r a ll C V D ris k fac to rsR isk fac to r m od ifica tion p er

exis t in g g u id e lin esA sp irin

B e low 7 5 th % tile

S core > 4 0 0S c reen fo r S ilen t Isch em ia

P os it ive exerc ise p erfu s ion s tu d y:con s id er b e ta b lockers

O p tim ize ris k fac to r trea tm en tsA g g ress ive lip id low erin g (L D L < 1 0 0 m g /d l)

A sp irinC on s id er n on -trad it ion a l ris k fac to rs

A b ove 7 5 th % tile

A g e-A d ju s ted C a lc iu m S core

E n h an ce p a tien t ris k fac to r m od ifica tion(m otiva tion an d ed u ca tion )

w ith ab n orm a l resu lts

P os it ive (co ron ary ca lc iu m )

S creening E BC T in A sym p tom atic PatientsM en ag ed 4 0 + o r W om en ag ed 5 0 +

1 o r M ore C V D R isk F ac to rs

Aortic and Carotid Magnetic Resonance Image (MRI) Imaging• Can identify plaque components such as fibrous cap, lipid core,

calcium, hemorrhage, and thrombosis (vunerable plaques have thin fibrous cap and large lipid core)

• Non-invasive and no radiation• Computerized morphometric analysis involves following edge of

significant contrast, providing measures of total vascular and lumen area, the difference being the vessel wall area (Image Pro-Plus, Media Cybernetics).

• Image-specific error of 2.6% for aortic and 3.5% for carotid plaques allows accurate measurement of changes in plaque size of >5.2% for aortic lesions and >7% for carotid lesions (Corti et al., 2001)

MRI Assessment of Thoracic Aorta Plaque

• Challenges include obtaining sufficient sensitivity for sub-mm imaging and exclusion of artifacts from respiratory motion and blood flow.

• Multicontrast approaches include performing T1-,PD-, and T2-weighted images with high resolution “black blood” spin used to visualize adjacent vessel wall.

• Matched MRI and TEE cross-sectional aortic images show strong correlation for plaque composition and maximum plaque thickness.

In Vivo MRI imaging of Coronary Artery Plaque

• Difficulties include cardiac and respiratory motion, nonlinear course of coronary arteries, and small size and location of coronary arteries.

• Inter- and intraobserver variability assessed by intraclass correlation ranged from 0.96-0.99.

• Wall thickness in human coronaries can be differentiated between normal and >40% stenosis; breathholding can minimize respiratory motion.

• Fayad and Fuster, Am J Cardiol 2001; 88 (suppl): 42E-45E.

Lipid-Lowering by Simvastatin and Reduction in MRI Vessel Wall Area

• 18 asymptomatic hypercholesterolemic patients studied, with a total of 35 aortic and 25 carotid plaques measured

• Serial black-blood MRI of aorta and carotid artery performed at baseline, 6, and 12 months

• At 12 months (but not 6 months), significant reductions in vessel wall thickness and area (8% reduction in aorta and 15% reduction in carotid artery vessel wall area), without lumen area changes, were observed.

• Corti et al., Circulation 2001: 104: 249-52

MRI Serial T2-Weighted Images During Simvastatin Treatment: Coronary vessels

(top) and descending aorta (bottom)(Corti et al., Circulation 2001; 104: 249-52)

High-frequency Brachial Ultrasonography

• The endothelium regulates vascular tone through release of vasodilators and vasoconstrictors.

• Brachial artery flow-mediated vasodilation (FMD) is assessed by high-frequency ultrasound assessment of changes in brachial artery diameter after 5-minute blood pressure cuff arterial occlusion.

• Endothelial dysfunction demonstrated as reduced FMD, and associated with coronary risk factors.

• Brachial artery FMD correlates with coronary artery FMD.

Brachial Ultrasonography (cont.)

• Brachial or coronary artery flow mediated vasodilation (FMD) predict long-term cardiovascular events.

• Clinical applicability not well-established, but measures frequently used to measure endothelial function.

• FMD decreases after age 40 in men and 50 in women, reduced at SBP>100 mmHg, LDL > 75 mg/dl, and in diabetics

• Cholesterol reduction rapidly improves FMD.

Brachial Artery Images Pre-Post Pressure Cuff Occlusion

Initial Follow-up

-50

-40

-30

-20

-10

0

10

20

30

Initial Follow-upPlacebo group Lovastatin group

Dilatation

Constriction

Effects of Lipid-Lowering Therapy on Endothelial Function in CHD Patients

Change in diameter

(%)

Treasure CB et al. N Engl J Med. 1995;332:481-487.

Arterial Stiffness•There are three separate processes increasing stiffness in the arterial system:

– Structural Breakdown of Elastin Fibers • occurs primarily in the aorta

• age driven (cum. millions of cardiac cycles)

– Damage to Endothelial Function • occurs primarily in the smooth muscle conduit arteries

• disease driven (e.g., Type II diabetes, hypercholesterolemia, atherosclerosis)

– Increase in Mean Arterial Pressure• occurs systemically throughout the arterial system.

•If the patient arteries get stiffer…. then pulse wave velocity increases, and reflected wave arrives back at the heart sooner.

•Now there is a very different aortic root pressure waveform (the green wave).

• there are three important clinical implications.

Figure 5

Pressure Wave Reflection at the Heart

Pressure Wave Reflection at the Heart

Arterial Stiffness•How should you Measure Arterial Stiffness? • Pulse Wave Velocity

• Carotid-Femoral PWV: - mainly aorta, but some smooth muscle.

• Carotid-Radial PWV: - limited aortic, mainly smooth muscle• Tr from SphygmoCor: pure aortic stiffness measure ?

• Augmentation Index• Aortic AIx from SphygmoCor is:

- a systemic measure, integrated appropriately- it is the clinically-relevant arterial stiffness measure

• Mean Pressure Changes• Do not forget them !

•The role of arterial stiffness as the major cause of cardio-vascular risk can be seen in recent outcome data.•Pulse Wave Velocity• aortic pulse wave velocity on entry was used to stratify arterial stiffness in a cohort of ESRD patients into tertiles• 1st tertile has almost normal results; 3rd tertile has 6x risk of “all cause’ mortality

Outcome StudiesOutcome Studies

Probability of survival (all-cause mortality) in ESRD - London et al

Central Pulse Augmentation (AIx)

•central pressure AIx on entry was used to stratify arterial stiffness in the cohort of ESRD patients into quartiles

• 1st quartile has normal results; 4th quartile has 8x risk of “all cause’ mortality

Outcome StudiesOutcome Studies

Recommendations for Noninvasive Screening

• AHA Prevention V (Greenland et al., Circ. 2000) indicated persons at intermediate risk may be suitable for screening by noninvasive tests, including ABI and carotid US for those over age 50 years, and coronary calcium screening.

• ATP III has suggested CAC scores above 75th percentile indications for more aggressive treatment (e.g., as CHD risk equivalent).

ACC 34th Bethesda Conference Task Force 4: How do We Select Patients for Atherosclerosis Imaging?

• The ability to select higher risk asymptomatic subsets from the population that would benefit from an earlier or more aggressive risk factor intervention is a key advantage of subclinical disease screening

• Persons with diabetes are considered CHD risk equivalents already warranting aggressive treatment as such; screening for atherosclerosis is not needed

Wilson, Smith, Blumenthal, Wong, 34th Bethesda Conference Task Force 4, J Am Coll Cardiol 2003 (in press)

• Patients at intermediate risk for total CHD comprise about 40% of the adult population.

• They have at least 1 major risk CHD factor and have a 6-20% 10-year risk of a hard CHD event, possibly warranting further risk stratification by noninvasive tests to assess atherosclerotic burden.

Wilson, Smith, Blumenthal, Wong, 34th Bethesda Conference Task Force 4, J Am Coll Cardiol June 2003

Risk Assessment

Count major risk factors

• For patients with multiple (2+) risk factors– Perform 10-year risk

assessment

• For patients with 0–1 risk factor– 10 year risk assessment not

required– Most patients have 10-year

risk <10%

Major Risk Factors (Exclusive of LDL Cholesterol) That Modify LDL Goals

• Cigarette smoking• Hypertension (BP 140/90 mmHg or on

antihypertensive medication)• Low HDL cholesterol (<40 mg/dL)† • Family history of premature CHD

– CHD in male first degree relative <55 years– CHD in female first degree relative <65 years

• Age (men 45 years; women 55 years)

† HDL cholesterol 60 mg/dL counts as a “negative” risk factor; its presence removes one risk factor from the total count.

Note: Risk estimates were derived from the experience of the Framingham Heart Study, a predominantly Caucasian population in Massachusetts, USA.

Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. JAMA. 2001;285:2486-2497.

Assessing CHD Risk in MenStep 1: Age

YearsPoints

20-34 -935-39 -440-44 045-49 350-54 655-59 860-64 1065-69 1170-74 1275-79 13

Step 2: Total CholesterolTC Points at Points at Points at Points at

Points at(mg/dL) Age 20-39 Age 40-49 Age 50-59 Age 60-69

Age 70-79 <160 0 0 0 0

0160-199 4 3 2 1

0200-239 7 5 3 1

0240-279 9 6 4 2

1280 11 8 5 3

1

HDL-C(mg/dL) Points

60 -1

50-59 0

40-49 1

<40 2

Step 3: HDL-Cholesterol

Systolic BP PointsPoints

(mm Hg) if Untreated if Treated

<120 0 0120-129 0 1130-139 1 2140-159 1 2160 2 3

Step 4: Systolic Blood Pressure

Step 5: Smoking Status

Points at Points at Points at Points atPoints at

Age 20-39 Age 40-49 Age 50-59 Age 60-69Age 70-79

Nonsmoker 0 0 0 00

Smoker 8 5 3 11

Age

Total cholesterol

HDL-cholesterol

Systolic blood pressure

Smoking status

Point total

Step 6: Adding Up the Points

Point Total 10-Year Risk Point Total 10-Year Risk

<0 <1% 118%

0 1% 1210%

1 1% 1312%

2 1% 1416%

3 1% 1520%

4 1% 1625%

5 2% 1730%

6 2%7 3%8 4%9 5%

10 6%

Step 7: CHD Risk

ATP III Framingham Risk Scoring

© 2001, Professional Postgraduate Services®

www.lipidhealth.org

Point Total 10-Year Risk Point Total 10-Year Risk

<9 <1% 2011%

9 1% 2114%

10 1% 2217%

11 1% 2322%

12 1% 2427%

13 2% 25 30%

14 2%15 3%16 4%17 5%18 6%19 8%

Assessing CHD Risk in Women

Note: Risk estimates were derived from the experience of the Framingham Heart Study, a predominantly Caucasian population in Massachusetts, USA.

Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. JAMA. 2001;285:2486-2497.

Step 1: Age

YearsPoints

20-34 -735-39 -340-44 045-49 350-54 655-59 860-64 1065-69 1270-74 1475-79 16

TC Points at Points at Points at Points atPoints at(mg/dL) Age 20-39 Age 40-49 Age 50-59 Age 60-69

Age 70-79 <160 0 0 0 0

0160-199 4 3 2 1

1200-239 8 6 4 2

1240-279 11 8 5 3

2280 13 10 7 4

2

HDL-C(mg/dL) Points

60 -1

50-59 0

40-49 1

<40 2

Step 3: HDL-Cholesterol

Systolic BP PointsPoints

(mm Hg) if Untreated if Treated

<120 0 0120-129 1 3130-139 2 4140-159 3 5160 4 6

Step 4: Systolic Blood Pressure

Step 5: Smoking Status

Points at Points at Points at Points atPoints at

Age 20-39 Age 40-49 Age 50-59 Age 60-69Age 70-79

Nonsmoker 0 0 0 00

Smoker 9 7 4 21

Age

Total cholesterol

HDL-cholesterol

Systolic blood pressure

Smoking status

Point total

Step 6: Adding Up the Points

Step 7: CHD Risk

Step 2: Total Cholesterol

ATP III Framingham Risk Scoring

© 2001, Professional Postgraduate Services®

www.lipidhealth.org

Estimated 10 Hard CHD RiskFramingham Offspring and Cohort

Men

0%

20%

40%

60%

80%

100%

30-39 40-49 50-59 60-69 70-79

Age (years)

Pe

rce

nt >20%

10-20%

6-10%

<6%

10 Year Risk

Estimated 10 Hard CHD RiskFramingham Offspring and Cohort

Women

0%

20%

40%

60%

80%

100%

30-39 40-49 50-59 60-69 70-79

Age (years)

Pe

rce

nt >20%

10-20%

6-10%

<6%

10 Year Risk

Shifting Patients to a Secondary Prevention Management Strategy

• ATP III--ABI <0.90 is indicative of PADRecommended secondary prevention approach

• ATP III– Coronary Ca++ >=75th%tile May be suitable for aggressive risk factor Rx

• Carotid IMT cutpoints?(studies show significant risk >=1 mm)?

• MRI cutpoints?Lack of large studies with follow up results

CHD Risk Equivalents

• Other clinical forms of atherosclerotic disease (peripheral arterial disease [incl ABI <0.90] abdominal aortic aneurysm, and symptomatic carotid artery disease)

• Diabetes

• Multiple risk factors that confer a 10-year risk for CHD >20%

Implications for Cardiovascular Risk Stratification and Treatment: NCEP III

and JNC-VII• Should we add to this high-risk group

requiring more intensive LDL-C treatment (to <100 mg/dl) or BP treatment (to <130/80 mmHg) , the following?– Persons with significant carotid disease (e.g., >1

mm or above 5th quintile of combined IMT)– Persons with “significant” coronary calcium

(e.g., those above 75th %tile or score >400)

Conclusions• Mounting data show surrogate measures of

atherosclerosis predict CHD risk and are sensitive to monitoring effects of therapeutic interventions.

• Noninvasive methods to measure subclinical atherosclerosis and its progression provide an opportunity to enhance primary prevention efforts

• Noninvasive identification of the vulnerable plaque (e.g., using MRI) may help identify those at highest risk.

• Patient compliance to risk-reduction may be enhanced by knowledge of disease (e.g., CAC)

Conclusions (cont.)• Identification of those with the greatest amount of

subclinical atherosclerosis may provide a better rationale for aggressive treatment (lipids, HTN) of those with borderline levels, allowing us to better target limited resources.

• Surrogate measures of atherosclerosis can also allow:– 1) testing of epidemiologic hypotheses related to CHD– 2) designing clinical trials testing efficacy of therapies– 3) monitoring preventive therapies to reduce risk of

clinical events

Considerations for Special Populations

• Women and Younger Persons Few persons reach “intermediate” (10% risk) thresholdlower threshold of risk (<6%) or with substantial elevations of

a single risk factor may be appropriate for noninvasive testing.

• Premature Family Hx of CHD or Metabolic Syndrome May be suitable for noninvasive testing as Framingham

equations do not fully account for these factors

• Diabetes or ESRDConsidered CHD risk equivalentsNot recommended for noninvasive testing

Potential Disadvantages of Atherosclerosis Screening

• Most data from specialized research centersTrained personnel and detailed protocols needed Testing not widely available

• False positives and False Negatives Misclassification of individualsPotential worry--false (+) False assurance--false (-) Many with a (+) test will not suffer eventsSome with a (-) test will suffer events

• Incidental Findings Incidental pathology--more testing, greater cost

• Effects on InsurabilityUnclear how information will be used

How should tests be accessed? • Self-referral and commercialization

Common for the CT calcium “heart scan” screening

Increasing for carotid IMT “stroke screening”

(+) test--may worry more, but may make healthy changes or comply with prescribed risk-reducing medicines

(-) test--may be falsely reassured, possibly less likely to undertake healthy lifestyle changes

• Physician referral may help to reduce screening of those unlikely to benefit

How should tests be accessed? (continued)

• ABI screeningAppropriate >50 years of age (AHA prevention V)Mostly a research toolLow costPersons may benefit from increased, selected screening

• MRI assessmentAtherosclerotic plaque studiesLimited to research centersLittle outcome dataNot suitable as a screening tool.

Receiver Operating Characteristic Curvesand Disease Prediction

0 0.2 0.4 0.6 0.8 1

1-Specificity(False Positives)

0

0.2

0.4

0.6

0.8

1

Sen

siti

vity

(Tru

e P

osi

tive

s)

Better test

Good test

Chance Line