10.08.2013 Views

Download the full report (116 p.) - KCE

Download the full report (116 p.) - KCE

Download the full report (116 p.) - KCE

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

62 Multislice CT in Coronary Heart Disease <strong>KCE</strong> Reports 82<br />

8.2 DIAGNOSTIC ACCURACY<br />

Most published clinical trials, are dealing with <strong>the</strong> diagnostic accuracy of 64-SCT as an<br />

imaging tool (for coronary arteries), referring to CCA as <strong>the</strong> gold standard and<br />

considering a coronary stenosis ≥50% on CCA as clinically relevant. In all published 64-<br />

SCT studies, in populations at intermediate or high pre-test probability of obstructive<br />

CAD (mostly >35% and up to 100%), test sensitivity is good and ranges between 95 and<br />

100%, indicating a very good negative predictive value. At present, <strong>the</strong> main value of<br />

MSCT seems to be its use to rule out obstructiuve CAD.<br />

Test specificity on <strong>the</strong> o<strong>the</strong>r hand is less good. In Abdulla’s SR, it was 91% (87.5-94), 66<br />

and in our meta-analysis it was 82.3% (78.5-85.7) in recent studies and 85.5% (80.4-89.7)<br />

in dual-source 64-SCT studies. Positive and negative likelihood ratios in recent studies<br />

were 5.0 (3.5-7.4) and 0.03 (0.02-0.06) respectively. In <strong>the</strong> large Meijboom gender trial<br />

(n=402), test performance was compared in women vs. men: whereas sensitivity was<br />

excellent in both sexes (93-100), specificity was acceptable in men (90%; 81-95) but<br />

poor in women (75%; 95% CI: 62-85). 69<br />

All studies enrolled patients with a high probability of CAD that were selected based on<br />

several additional parameters, including regular and controlled heart rate, preserved<br />

renal function, breath-hold capacity, and hemodynamic status. Virtually all patients were<br />

already scheduled for an invasive CCA before entry in <strong>the</strong> study. Whe<strong>the</strong>r <strong>the</strong><br />

performance of MSCT can be reproduced in less selected patients at lower prevalence<br />

of CAD remains to be assessed, questioning <strong>the</strong> external validity of <strong>the</strong> results. 158<br />

Moreover, selection bias may have played a role in published trials, by exclusion of<br />

patients in whom calcified coronary arteries were expected (elderly, diabetes, noncardiac<br />

a<strong>the</strong>romatous disease, renal failure) or in whom calcified coronary arteries had<br />

been documented by prior CCA. Finally, results of studies may also be biased by <strong>the</strong><br />

fact that investigators had a better experience compared with <strong>the</strong> real-life centres which<br />

usually examine larger and more broad-spectrum populations and may be less<br />

experienced. 66<br />

8.3 DIAGNOSTIC THINKING<br />

This level of diagnostic efficacy is concerned with <strong>the</strong> assessment of <strong>the</strong> effect of a test<br />

result on diagnostic reasoning and disease categorization, or in o<strong>the</strong>r words its role in<br />

clinical decision making. Few empirical evidence is available on this subject.<br />

In a patient with a negative MSCT, i.e. in which no >50% stenosis is detected, a<br />

physician will decide that <strong>the</strong> chest pain symptoms for which <strong>the</strong> test was performed,<br />

most probably were not provoked by CAD. The question remains to what extent this<br />

test result changes <strong>the</strong> initial diagnosis put forward by <strong>the</strong> attending physician (i.e. pretest<br />

likelihood). In a patient population such as that described by Goldstein, 67 <strong>the</strong> pretest<br />

probability of ACS was very low because of a profound pre-selection of patients. In<br />

<strong>the</strong>se cases, fur<strong>the</strong>r testing becomes irrelevant because diagnosis is almost certain, and<br />

additional testing will predominantly lead to false positives. In <strong>the</strong> study by Rubinshtein<br />

et al., MSCT <strong>report</strong>edly was useful in <strong>the</strong> diagnostic work-up of patients with chest pain<br />

and an inconclusive stress test. 80 However, in a substantial proportion of patients (20<br />

out of 71, 28%) that tested negative with MSCT, treating physicians later on still<br />

proceeded to CCA, “because of clinical reasons” (sic), despite previous trials indicating<br />

a high negative predictive value of MSCT. In this trial, MSCT clearly did not alter <strong>the</strong><br />

diagnostic path followed by <strong>the</strong> physician.<br />

The “2006 appropriateness review for MSCT”, issued under <strong>the</strong> auspices of <strong>the</strong><br />

American College of Cardiology Foundation, confirms that limited data are available<br />

supporting <strong>the</strong> use of MSCT coronary angiography within patient care algorithms. 108<br />

8.4 THERAPEUTIC IMPACT<br />

MSCT of <strong>the</strong> coronary arteries up to now does nothing more than providing an<br />

anatomical image of <strong>the</strong> coronary tree. In several trials it has been shown that MSCT<br />

reliably can be used to rule out <strong>the</strong> presence of CAD in subsets of patients. In contrast<br />

to o<strong>the</strong>r imaging techniques such as MPS and DSE, MSCT does however not provide

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!