Transcript What calcium score is typical for a person my

De plaats van de
Calcium Score
in het Atrium Medisch Centrum
16-7-2015
©Eduard van den Berg, cardio.nl
1
CVRM ESC 2012
16-7-2015
Eduard van den Berg, cardio.nl
2
Decrease † in about 10 yrs
16-7-2015
Eduard van den Berg, cardio.nl
3
NICE report
According to the report of NICE, implementation of the population
approach may bring numerous benefits and savings:
† Narrowing the gap in health inequalities.
† Cost savings from the number of CVD events avoided.
† Preventing other conditions such as cancer, pulmonary diseases,
and type 2 diabetes.
† Cost savings associated with CVD such as medications, primary
care visits, and outpatient attendances.
† Cost savings to the wider economy as a result of reduced loss of
production due of illness in those of working age, reduced
benefit payments, and reduced pension costs from people
retiring early from ill health.
† Improving the quality and length of people’s lives.
16-7-2015
Eduard van den Berg, cardio.nl
4
A Rosetta Stone for Coronary Calcium Risk Stratification:
Agatston, Volume, and Mass Scores in 11,490 Individuals
When we matched them against a known “lesion” phantom, the Agatston and volume
scores behave nonlinearly, and the latter grossly overestimates volume. The mass method
is linear except for lesions near the edge of detectability and matches known volumes to
within a small percentage.
CONCLUSION. We provide validated risk stratification data for use with mass scoring
methods.
AJR American Journal of roentgenology
http://www.ajronline.org/content/181/3/743.full
John A. Rumberger1 and Leon Kaufman2,3
16-7-2015
© Eduard van den Berg, cardio.nl
8
Why percentile ranking
The actual calcium score, although an indicator of overall disease extent [7], may not be as
useful for predicting coronary events as its percentile ranking.
even small scores that are much higher than those anticipated for age and sex may be better
predictors of risk. For instance, a calcium score of 40 in a 40-year-old man would place him well
above the 95th percentile and engender a risk of a cardiac event during the next 3-6 years that
may well exceed that of a 70-year-old with a similar score, who would rank for that age group
below the 10th percentile.
a calcium score above the 75th percentile for age and sex may increase heart risk an order of
magnitude above that for individuals with scores below the 25th percentile [8, 9].
16-7-2015
Eduard van den Berg, cardio.nl
9
16-7-2015
© Eduard van den Berg, cardio.nl
10
16-7-2015
© Eduard van den Berg, cardio.nl
11
Agatston scale
The Agatston scale calculates a calcium burden by multiplying the area of the lesion
above a 130-H threshold (obtained from 3-mm-thick nonoverlapped slices) by a
weighting factor that is dependent on the peak signal anywhere in the lesion.
Because of the discrete nature of the weighting, the Agatston score is sensitive to
noise when the peak signal is near one of the threshold values and completely
insensitive away from them.
H = Hounsefield units
Considering that newer multi slice MDCT scanners can image the heart with much less slice
thickness, nothing in the Agatston rules allows for a consistent computational method to
translate a measurement from one scanner to another in a consistent fashion.
In other words, the Agatston method is not portable from one centre to another.
16-7-2015
Eduard van den Berg, cardio.nl
12
Volume score
the volume score was introduced by Callister et al. [12]. The volume score linearly
interpolates the data set to isotropic volumes and, as its name implies, computes the
volume of the lesion above a 130-H threshold (the nominal threshold for calcification
derived from observational studies).
the volume method, contrary to what its name implies, does not measure a value purely
dependent on volume; it is affected by calcium content and by scanner operating parameters.
Even when the latter are held fixed, the volume score reflects both lesion volume and calcium
content.
Its portability is affected by the same issues that affect the Agatston method.
16-7-2015
Eduard van den Berg, cardio.nl
13
Mass measurement methods
Basically, these consist of integration of the signal for pixels above a given threshold.
For a well-calibrated CT scanner, in the absence of noise, this integration (scaled by pixel
volume) gives the total mineral content independently of slice thickness and spatial resolution.
In practice, the threshold necessary to avoid the inclusion of false-positive pixels changes the
measurement. Another issue of mass methods is what mass is being expressed; calcified lesions
include a complex of different calcium bone ash equivalent or calcium equivalent.
Whereas each will give a different result, all these are different scaling factors for the integral of
the Hounsfield values so that the different measures are easily related to each other.
If the scaling factor is given, translation is possible. With suitably low threshold settings, the
mass methods come closest to being portable.
16-7-2015
© Eduard van den Berg, cardio.nl
14
Chart shows age distribution population
Chart shows age distribution
for men (black bars) and
women (white bars) of patients
in this study. There were
11,490 patients, 63% male.
16-7-2015
Eduard van den Berg, cardio.nl
15
75th percentile range for men and women
Graph shows
comparison of results
for 75th percentile
range for men and
women
as reported in Hoff et
al. [10] (Kondos
database).
16-7-2015
© Eduard van den Berg, cardio.nl
16
Agatston and Volume scores men
Graph shows volume scores for men.
Graph shows Agatston scores for men.
16-7-2015
© Eduard van den Berg, cardio.nl
17
Volume and Mass scores men
Graph shows volume scores for men.
16-7-2015
Graph shows mass scores (bone ash equivalent) for men.
Eduard van den Berg, cardio.nl
18
Agatston and Volume scores Women
Graph shows Agatston scores for women.
16-7-2015
Graph shows volume scores for women
© Eduard van den Berg, cardio.nl
19
Volume and mass scores women
Graph shows volume scores for women
16-7-2015
Graph shows mass scores (bone ash equivalent) for women.
Eduard van den Berg, cardio.nl
20
reproducibility
Graph shows
reproducibility of three
scoring methods in 35
patients. Approximately
50% of all nonzero scores
fall within 25%
reproducibility. Mean is
38% for all three methods.
There is no significant
difference among them.
Black bars = Agatston score,
white bars = volume score,
gray bars = mass score.
16-7-2015
© Eduard van den Berg, cardio.nl
21
Conclusion Agatston – Mass - Volume
The single advantage of the mass method is its
better reflection of the physical properties of the
lesion and, consequently, better adaptation to
portability across and between CT scanners.
Most widely used in the Netherlands is the
Agatston score
16-7-2015
Eduard van den Berg, cardio.nl
22
Agatston Score Men and Women
Graph shows Agatston scores for men.
16-7-2015
Graph shows Agatston scores for women.
© Eduard van den Berg, cardio.nl
23
Percentile
Percentile rank is calculated by adjusting calcium scores for age and sex.
The following example illustrates how to read the next tables: a 57-year-old man with a calcium
score of 54 would be in the 50th percentile.
This means that, in his age- and sex-matched group, 50% of men have calcium scores greater
than his and 50% have scores less than his score.
A 46-year-old woman with a calcium score of 2 would be in the 75th percentile.
This means that compared with her age- and sex-matched peers, 75% have calcium scores less
than hers and
25% have scores above her score of 2.
16-7-2015
Eduard van den Berg, cardio.nl
24
What calcium score is typical for a person my age men ?
16-7-2015
Eduard van den Berg, cardio.nl
25
What calcium score is typical for a person my age women ?
16-7-2015
Eduard van den Berg, cardio.nl
26
Recommendations I
16-7-2015
© Eduard van den Berg, cardio.nl
27
Recommendations II
16-7-2015
© Eduard van den Berg, cardio.nl
28
ACC/AHA GL
First Calculate the Framingham traditional risk score
16-7-2015
© Eduard van den Berg, cardio.nl
29
AHA/ACC guidelines
16-7-2015
Eduard van den Berg, cardio.nl
30
Low Risk = < 1 % / yr, No test
• When a calculator, such as the Framingham
Risk Score, is used to evaluate risk, low risk
would be defined as a risk of CAD events less
than 1% per year. So if you are at low risk of a
CAD event, there is NO need for coronary
artery calcium scoring.
16-7-2015
Eduard van den Berg, cardio.nl
31
High Risk = > 2 % / yr, No test
• Also, if you already have diabetes or peripheral
vascular disease you are at high risk even if you have
absolutely no symptoms of coronary artery disease
and, again, knowing your calcium score in this setting
adds nothing to understanding your risk. In other
words, if a validated risk assessment tool, like the
Framingham Risk Score, determines you’re at high
risk of a CAD event, then you are at high risk — and
there is NO need for coronary artery calcium
scoring.
16-7-2015
© Eduard van den Berg, cardio.nl
32
Intermediate Risk = <2∧>1, Test
• So, the best use of fast CT scanning for
calcium scoring is in those individuals who
are determined to be in the intermediate-risk
category, which would apply to anyone
whose Framingham Risk Score puts them at a
10-year risk of a CHD event of 10% to 20%.
(Another way of stating this is that patients at
intermediate risk are estimated to have a 1%
to 2% risk of a CAD event every year.)
16-7-2015
© Eduard van den Berg, cardio.nl
33
Can you have EuroScore < 1%/yr ∧Ca-score up to 100
16-7-2015
© Eduard van den Berg, cardio.nl
34
Statements, accepted ?
• Any calcium means coronary atherosclerosis
• Score level = total atherosclerotic plaque burden ≠ severity particular
stenosis
• CS > 400 = P(at least one stenosis > 70 %) = 90 %
• CS alone is a more powerful predictor of future events than all other
RF combined
• CS is independent of other CV-RF and a good predictor of † elderly
• CS is related to CV events at other locations of the vascular tree
• CS = 0 means P(sign sten man) = 0,7%, P(sign sten woman)=0,0% and
P(Tl-201 = pos) = 0,0 %
http://www.euroscore.org/calc.html
http://www.newportbodyscan.com/CACrisk.htm
http://www.chestx-ray.com/coronary/corcalc.html
16-7-2015
© Eduard van den Berg, cardio.nl
35
Questions
• Is there a place for common sense next to guidelines, do you insist on having severe
arguments not to follow the protocol ?
• If a young (28) female FHC pt comes to you for screening because her sister of 32 died last
week of an AMI (F † 48, M † 52, 2 / 4 F † < 45 yr, 1/3 S † < 40 all of CAD, most with CABG or
PCI) and if she has no complaints or further RF, would you be satisfied with a negative
adequate bicycle test, would you give her primary prevention medication, would you still do
a CS and if CS = 0 would you than restrict to only chol synthese inhibitors.
To say it another way do you believe in CS = 0 fot this woman means P(sign CAD) = 0
16-7-2015
© Eduard van den Berg, cardio.nl
36
16-7-2015
Eduard van den Berg, cardio.nl
37
HeartScore Europe upgraded with HDL, BMI and Risk Age
Did you know...In addition to its standard assessment features,
HeartScore Europe now includes:
An HDL cholesterol function, which improves the accuracy of
the model
A Risk Age function, which will help patients quickly understand
their exposure to overall CVD risk
A ‘fast track’ calculator with Body Mass Index (BMI)
https://escol.escardio.org/Heartscore/pmsCenter.aspx?model=EuropeHigh
16-7-2015
Eduard van den Berg, cardio.nl
38
16-7-2015
Eduard van den Berg, cardio.nl
39
Score Card 2012, no HDL
Use the low risk charts in Andorra, Austria, Belgium*,
Cyprus, Denmark, Finland, France, Germany, Greece*, Iceland,
Ireland, Israel, Italy, Luxembourg, Malta, Monaco, The Netherlands*,
Norway, Portugal, San Marino, Slovenia, Spain*,
Sweden*, Switzerland and the United Kingdom.
Use the high risk charts in other European countries. Of these,
some are at very high risk and the charts may underestimate risk
in these. These include Armenia, Azerbaijan, Belarus, Bulgaria,
Georgia, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Macedonia
FYR, Moldova, Russia, Ukraine and Uzbekistan.
*Updated, re-calibrated charts are now available for Belgium, Germany, Greece, The Netherlands, Spain, Sweden and Poland.
16-7-2015
Eduard van den Berg, cardio.nl
40
ESCardio.org 2012
Europe low risk
(Belgium, France, Italy, Luxembourg, Switzerland and Portugal)
Europe high risk
(Albania, Algeria, Armenia, Austria, Belarus, Bulgaria, Croatia, Czech
Republic, Denmark, Egypt, Estonia, Finland, Georgia, Hungary, Iceland,
Ireland, Israel, Latvia, Libanon, Libya, Lithuania, Former Yugoslav
Republic of Macedonia, Moldova, Morocco, Norway, Romania, San
Marino, Serbia and Montenegro, Slovakia, Slovenia, The Netherlands,
Tunisia, Turkey, Ukraine, United Kingdom)
16-7-2015
Eduard van den Berg, cardio.nl
41
Interpretation Score
High = > 1,5%/ yr, Intermediate = 1,0-1,5 %, Low = < 1,0 %/ yr
16-7-2015
Eduard van den Berg, cardio.nl
42
16-7-2015
Eduard van den Berg, cardio.nl
43
Reported Average Sensitivity &
Specificity of Stress Tests
Test modality
16-7-2015
Sensitivity
Specificity
• Non-Imaging ETT
65%
• Nuclear ETT
– Quantitative
– Qualitative
– Dipyridamole
– RVG
87%
87%
90%
87%
87%
77%
90%
75%
• Echo ETT
80%
87%
Eduard van den Berg, cardio.nl
85%
44
Post-Test Probability of CAD Based on
Pre-Test Symptoms - Women
Women 40 - 49
100
100
90
90
80
80
% Probabilty CAD
% Probabilty CAD
Women 30 - 39
70
60
50
40
30
20
10
0
0.0
0.5
0.0
1.0
1.5
2.0
2.5
ST Depression mm
None
Non-Angina
Atypical Angina
3.0
Typical Angina
Women 50 - 59
100
90
80
70
60
50
40
30
20
10
0
0.5
1.0
1.5
2.0
2.5
ST Depression mm
None
Non-Angina
Atypical Angina
3.0
Typical Angina
Women 60 - 69
100
% Probabilty CAD
% Probabilty CAD
90
0.0
0.5
1.0
1.5
Non-Angina
80
70
60
50
40
30
20
10
0
0.0
2.0
ST Depression mm
None
70
60
50
40
30
20
10
0
Atypical Angina
2.5
0.5
3.0
1.0
1.5
2.0
ST Depression mm
Typical Angina
None
Non-Angina
Atypical Angina
2.5
3.0
Typical Angina
Diamond and Forrester. N. Engl. J. Med. 1350-7, 1979
16-7-2015
Eduard van den Berg, cardio.nl
45
Post-Test Probability of CAD Based on
Pre-Test Symptoms - Men
Men 40 - 49
100
100
90
90
80
80
% Probabilty CAD
% Probabilty CAD
Men 30 - 39
70
60
50
40
30
20
10
0
0.0
0.5
0.0
1.0
1.5
2.0
2.5
ST Depression mm
None
Non-Angina
Atypical Angina
3.0
Typical Angina
Men 50 - 59
100
90
80
70
60
50
40
30
20
10
0
0.5
1.0
1.5
2.0
2.5
ST Depression mm
None
Non-Angina
Atypical Angina
3.0
Typical Angina
Men 60 - 69
100
% Probabilty CAD
% Probabilty CAD
90
0.0 0.5
1.0
1.5
Non-Angina
80
70
60
50
40
30
20
10
0
0.0
2.0
ST Depression mm
None
70
60
50
40
30
20
10
0
Atypical Angina
2.5
0.5
3.0
1.0
1.5
2.0
ST Depression mm
Typical Angina
None
Non-Angina
Atypical Angina
2.5
3.0
Typical Angina
Diamond and Forrester. N. Engl. J. Med. 1350-7, 1979
16-7-2015
Eduard van den Berg, cardio.nl
46
The Likelihood Ratio
• Ratio of two proportions:
The proportion who have a particular test
result (e.g. positive, negative, high
probability) among those with a disease
divided by
The proportion who have the same test result
among those without the disease
LR = TEST RESULT/DISEASE +
TEST RESULT/DISEASE –
16-7-2015
Keyvan
Concepts
Taught
Eduard
den Berg,
cardio.nl
47
Incremental value of a test ROC curve
16-7-2015
Eduard van den Berg, cardio.nl
48
A global χ2 model showing the incremental value of ischaemic burden assessed by stress
echocardiography over clinical assessment and stress ECG in all patients (A) and in the
exercise cohort (B).
Chelliah R et al. Eur J Echocardiogr
2010;11:875-882
Published on behalf of the European Society of
Cardiology. All rights reserved. © The Author 2010.
For permissions please email:
[email protected]
16-7-2015
Eduard van den Berg, cardio.nl
49
Sensitivity and Specificity
a
Sens 
ac
d
Spec 
bd
16-7-2015
Disease D
Test
Result
+
Eduard van den Berg, cardio.nl
“Gold standard”
+
a
c
b
d
50
Positive & Negative
Predictive Value
• PV (+): positive
predictive value
• PV (-): negative
predictive value
a
PV () 
ab
d
PV () 
cd
16-7-2015
Test
Result
+
-
Disease D
+
a
b
c
d
a /(a  c)
LikelihoodRatio( LR) 
b /(b  d )
Eduard van den Berg, cardio.nl
51
Posterior odds
When combined with information on the prior
probability of a disease*, LRs can be used
to determine the predictive value of a
particular test result:
Posterior odds = Prior odds x Likelihood ratio
*expressing the prior probability [p] of a disease as the
prior odds [p/(1-p)] of that disease. Conversely, if the odds
of a disease are x/y, the probability of the disease is x / (x
+ y)
16-7-2015
Eduard van den Berg, cardio.nl
52
Choice of a cut-off point for continuous results
Consider the implications of the two possible
errors:
• If false-positive results must be avoided (such
as the test result being used to determine
whether a patient undergoes dangerous
surgery), then the cutoff point might be set to
maximize the test's specificity
• If false-negative results must be avoided (as
with screening for neonatal phenylketonuria),
then the cutoff should be set to ensure a high
test sensitivity
16-7-2015
Eduard van den Berg, cardio.nl
53
Choice of a cut-off point
for continuous results
• Using receiver operator characteristic
(ROC) curves:
– Selects several cut-off points, and determines
the sensitivity and specificity at each point
– Then, graphs sensitivity (true-positive rate) as
a function of 1-specificity (false-positive rate)
• Usually, the best cut-off point is where the
ROC curve "turns the corner”
16-7-2015
Eduard van den Berg, cardio.nl
54
RECEIVER OPERATING
CHARACTERISTIC (ROC) curve
• ROC curves (Receiver
Operator Characteristic)
• Ex. SGPT and Hepatitis
SGPT
D+
D-
Sum
< 50
10
190
200
50-99
15
135
150
100-149
25
65
90
150-199
30
30
60
200-249
35
15
50
250-299
120
10
130
>300
65
5
70
Sum
300
450
750
Sensitivity
1
1
1-Specificity
16-7-2015
Eduard van den Berg, cardio.nl
55