impact of coronary calcium score on diagnostic accuracy of multislice computed tomography coronary...

ORIGINAL ARTICLES

mogaltelow

FromveCanU

G.P.SoJ.T

Rece20

ReprLde

1071Copydoi:1

36

Impact of coronary calcium score on diagnosticaccuracy of multislice computed tomographycoronary angiography for detection of coronaryartery diseaseGabija Pundziute, MD,a,c Joanne D. Schuijf, MSc,a,d J. Wouter Jukema, MD, PhD,a,d

Hildo J. Lamb, MD, PhD,b Albert de Roos, MD, PhD,b

Ernst E. van der Wall, MD, PhD,a,d and Jeroen J. Bax, MD, PhDa

Background. The impact of the coronary calcium score on the diagnostic accuracy ofmultislice computed tomography (MSCT) to detect obstructive coronary stenoses remainscontroversial.

Methods and Results. We examined 41 patients (mean Agatston score, 340 � 530 [range,0-2546]) with coronary artery disease with 16-slice MSCT and 60 patients (mean Agatston score,446 � 877 [range, 0-6264]) with 64-slice MSCT. MSCT scans were analyzed with invasivecoronary angiography (CA) as the standard of reference. Lesions with luminal narrowing of50% or greater were considered obstructive. In total, 9% and 2% of uninterpretable segmentswere excluded from analysis in patients examined with 16- and 64-slice MSCT, respectively. Ona segment basis, the percentage of false-negative segments in the groups with Agatston scores of0 to 100, 101 to 400, and greater than 400 with 16-slice MSCT were 0%, 5.3%, and 2.9% (P �.0005), respectively; other comparisons of false-positive and false-negative segments were notsignificant. The sensitivity and specificity on a vessel and patient basis with 16- and 64-sliceMSCT were not significantly different in different calcium score groups.

Conclusions. A slight impact of coronary calcium was observed on the diagnostic accuracyof 16-slice MSCT CA on a segment basis, with no significant impact on a vessel and patient basis.No significant impact of coronary calcium was observed on the diagnostic accuracy of 64-sliceMSCT CA on a segment, vessel, or patient basis. (J Nucl Cardiol 2007;14:36-43.)

Key Words: Multislice computed tomography • coronary calcium score • diagnosticaccuracy

ies.1-4

supeinititremficitrespnersas acoroaccureselesscorocreaficainclintroprovtem

Over the past few years, multislice computed to-raphy (MSCT) has been demonstrated as a feasible

rnative to invasive coronary angiography (CA), al-ing noninvasive evaluation of the coronary arter-

the Departments of Cardiology,a and Radiology,b Leiden Uni-rsity Medical Center, Leiden, The Netherlands; Department of

ardiology, Kaunas University of Medicine, Kaunas, Lithuaniac;d The Interuniversity Cardiology Institute of the Netherlands,trecht, The Netherlands.d

is financially supported by the Training Fellowship of the Europeanciety of Cardiology and the Huygens Scholarship, The Netherlands.

D.S. is financially supported by the Netherlands Heart Foundation,he Hague, The Netherlands (grant No. 2002B105).ived for publication May 2, 2006; final revision accepted Nov 4,06.

int requests: Jeroen J. Bax, MD, PhD, Department of Cardiology,eiden University Medical Center, Albinusdreef 2, 2333 ZA Lei-n, The Netherlands; [email protected]/$32.00right © 2007 by the American Society of Nuclear Cardiology.0.1016/j.nuclcard.2006.11.002

Rapid development of the technique resulted inrior visualization of the coronary arteries, and theal published reports of 64-slice MSCT showed ex-ely promising results, with sensitivities and speci-

ies ranging from 94% to 99% and 95% to 97%,ectively.3,4 With the first generation of MSCT scan-, severe coronary calcifications have been recognizedn important factor hampering precise evaluation ofnary artery stenoses, thereby limiting diagnosticracy.5 Accordingly, coronary calcifications may rep-nt an important limitation of the technique. Nonethe-, accurate assessment in the presence of elevatednary artery calcium scores is important, as an in-sed prevalence and extent of coronary artery calci-tions have been observed in certain populations,uding the elderly and diabetic patients.6,7 With theduction of 16-slice MSCT, image quality has im-ed significantly as a result of superior spatial and

poral resolution. However, contradictory results re-

garddiagrepootheincoseveinfluaccuthisarteand

Pati

convcoropatieevalwomMSCmeaMSCMSCchantweeratecontprovappr

MSC

64-sTokbefowasparacolliondstempelec

intramedof 4Aulnlowsor 6mA;wasrotathearcont

volua floKonmenimagapprMSCpatierecobeat

sepacardthickprespoinremopost

MSC

wasrea2areaHouan oprecwereof c

scanenceunawfollocoroagescourcorotheregation.for tlumeinsp

Con

protSeldvisublind

Stat

diffe

Journal of Nuclear Cardiology Pundziute et al 37Volume 14, Number 1;36-43 Multislice computed tomography coronary angiography

ing the effect of the coronary artery calcium score onnostic accuracy of 16-slice MSCT CA have beenrted, with some studies showing no influence8 andrs reporting high numbers of uninterpretable orrrectly diagnosed coronary lesions resulting fromre coronary calcifications.9,10 Moreover, data on theence of the coronary calcium score on the diagnosticracy of 64-slice MSCT are limited. The purpose ofstudy was to evaluate the impact of the coronary

ry calcium score on the diagnostic accuracy of 16-64-slice MSCT.

METHODS

ents and Study Protocol

A total of 101 patients who were already scheduled forentional CA, having presented with known or suspectednary artery disease (CAD), were included in the study. Allnts underwent MSCT followed by conventional CA for the

uation of the coronary arteries; 41 patients (30 men and 11en; mean age, 58 � 13 years) were examined with 16-sliceT, and 60 consecutive patients (47 men and 13 women;

n age, 60 � 11 years) underwent examination with 64-sliceT. The median interval between conventional CA andT was 4 weeks (range, 0-27 weeks). No interveningges in clinical conditions in patients have occurred be-n the 2 examinations. Only patients with a regular heartin sinus rhythm and without contraindications to iodinatedrast medium were included in this study. All patientsided written informed consent, and the study protocol wasoved by the local ethics committee.

T Data Acquisition

The MSCT examinations were performed with 16- andlice Toshiba Aquilion systems (Toshiba Medical Systems,yo, Japan). No additional �-blockers were administeredre the examinations. A prospective coronary calcium scanperformed before MSCT angiography with the same

meters for 16- and 64-slice MSCT scanners as follows:mation, 4 � 3.0 mm; gantry rotation time, 500 millisec-; tube voltage, 120 kV; and tube current, 200 mA. Theoral window was set at 75% after the R wave for

trocardiographically triggered prospective reconstruction.The 16-slice MSCT CA scan was performed after thevenous administration of 120 to 150 mL iodinated contrastium depending on the total scan time, with an injection ratemL/s, through the antecubital vein (Xenetix 300; Guerbet,ay-sous-Bois, France). The scan parameters were as fol-: collimation, 16 � 0.5 mm; tube rotation time of 400, 500,

00 milliseconds, depending on heart rate; tube current, 250and tube voltage, 120 mV. The 64-slice MSCT CA scanperformed with a collimation of 64 � 0.5 mm and a tubeion time of 400, 450, or 500 milliseconds, depending on thet rate. The tube current was 300 mA, at 120 kV. Nonionicrast material was administered in the antecubital vein at a

me of 80 to 110 mL, depending on the total scan time, andw rate of 5 mL/s (Iomeron 400; Bracco Altana Pharma,stanz, Germany). Automated detection of peak enhance-t in the aortic root was used for timing of the scan. Alles were acquired during an inspiratory breath-hold ofoximately 20 seconds and 10 seconds with 16- and 64-sliceT, respectively, with simultaneous registration of thent’s electrocardiogram. With the aid of a segmentalnstruction algorithm, data of 1, 2, or 3 consecutive heart-s were used to generate a single image.To evaluate the presence of coronary artery stenoses,rate reconstructions in diastole (typically 75% of theiac cycle) were generated with a reconstructed sectionness of either 0.3 mm or 0.4 mm. If motion artifacts were

ent, additional reconstructions were made at different timets of the R-R interval. Axial data sets were transferred to ate workstation (Vitrea2; Vital Images, Plymouth, Minn) for

processing and subsequent evaluation.

T Data Analysis

Coronary calcium score. The coronary calcium scoreassessed with the application of dedicated software (Vit-; Vital Images). Coronary calcium was identified as a densein the coronary artery exceeding the threshold of 130

nsfield units (HU). In addition to per-vessel measurements,verall Agatston score was recorded for each patient. Forise evaluation of the calcium score, all coronary segments

visually inspected and care was taken to avoid inclusionoronary stents.CA. MSCT angiograms obtained with 16- and 64-sliceners were retrospectively evaluated by the same 2 experi-d observers (within a limited period of time), who wereare of the results of the conventional angiogram. The

wing protocol was applied in the evaluation of the MSCTnary angiograms: the 3-dimensional volume-rendered im-were evaluated first to obtain a general impression of the

se and status of the left and right coronary arteries. Thenary arteries were divided into 17 segments according tomodified American Heart Association classification andrded as interpretable or uninterpretable by visual inspec-Subsequently, the interpretable segments were evaluated

he presence of obstructive stenoses (�50% reduction ofn diameter) by both scrolling through the axial images and

ecting curved multiplanar reconstructions.

ventional CA

Conventional CA was performed according to standardocols. For vascular access, the femoral approach with theinger technique was applied. Coronary angiograms wereally evaluated by consensus of 2 experienced observersed to the MSCT data.

istical Analysis

Continuous data are expressed as mean � SD. Therences in patient characteristics between the 2 cohorts

examthe tgrouandthe 3test.asseimpaMSCvessfalseusesegmspecdetearterinterican(ver

MSwasimatheareCADandinte(56%und

riskdiab51histprevin tpop(78%exambaseservandinderategrouexamAga(ranpatitivecoro

EvaAccAna

withsegmreasinclu

Tab

16-SNMMMM

64-SNMMMM

BMI,*P �

38 Pundziute et al Journal of Nuclear CardiologyMultislice computed tomography coronary angiography January/February 2007

ined with 16- and 64-slice MSCT were tested by use oftest. Coronary segments and patients were divided into 3

ps according to overall Agatston score (0-100, 101-400,�400). The differences in continuous variables between

groups were tested with the 1-way analysis of varianceWhen significant differences were found, they were furtherssed with the post hoc (Bonferroni) test. Analysis of thect of the calcium score on the diagnostic accuracy ofT CA was performed on 3 levels: segment by segment,

el by vessel, and patient by patient. The percentages of-positive and false-negative segments were compared byof the Fisher’s exact test, and the number of interpretableents was compared by use of the �2 test. Sensitivity,

ificity, and positive and negative predictive values for thection of luminal narrowing of 50% or greater in coronaryies are expressed as percentages with 95% confidencevals. P � .05 was considered to indicate statistical signif-ce. Statistical analyses were performed with SPSS softwaresion 12.0; SPSS, Chicago, Ill).

RESULTS

In total, 102 patients were enrolled in the study.CT was successfully performed in 101. One patient

excluded from the analysis because of insufficientge quality because of an elevated heart rate duringscan. The main characteristics of the study populationlisted in Table 1. Of all patients, 57 (57%) had known

: 53 (53%) had a history of myocardial infarction,56 (56%) had a previous percutaneous coronary

rvention. Known CAD was present in 23 patients) examined with 16-slice MSCT and 34 (57%) who

erwent 64-slice MSCT CA. The prevalence of CAD

le 1. Characteristics of study population

Variable 0–100

lice MSCT CAo. of coronary segments (men/women) 125/94ean age (y)* 51 � 15ean heart rate (beats/min) 72 � 13ean Agatston score (range) 18 � 21 (0–8ean BMI (kg/m2) 27 � 5lice MSCT CAo. of coronary segments (men/women) 188/65ean age (y)* 57 � 9ean heart rate (beats/min) 59 � 11ean Agatston score (range) 14 � 21 (0–7ean BMI (kg/m2) 27 � 3

Body mass index..05.

factors was as follows: 21 (21%) patients hadetes mellitus, 57 (57%) had hypercholesterolemia,(51%) had hypertension, 38 (38%) had a familyory of CAD, and 49 (49%) had a history of current orious smoking. There were no significant differences

he prevalence of risk factors between the 2 patientulations. �-Blocking agents were used by 32 patients

) examined with 16-slice MSCT and 43 (72%)ined with 64-slice MSCT (P � .47) as part of their

line treatment. No significant differences were ob-ed between the 2 patient cohorts examined with 16-64-slice MSCT with regard to age and body massx. However, a significant difference in mean heartduring data acquisition was noted between the 2ps of patients with low overall Agatston scoresined with different MSCT scanners. The mean

tston score in the overall population was 340 � 530ge, 0-2546) and 446 � 877 (range, 0-6264) in theents examined with 16- and 64-slice MSCT, respec-ly. Examples of heavily calcified and noncalcifiednary lesions are provided in Figures 1 and 2.

luation of Impact of Calcium on Diagnosticuracy of MSCT CA: Segment-by-Segmentlysis

Of 570 coronary segments in 41 patients examined16-slice MSCT, 30 stented segments and 47 coronaryents that were uninterpretable for calcium-unrelated

ons were excluded, resulting in 493 segments beingded in the analysis. The reasons for uninterpretability

Agatston score

101–400 >400

128/41 105/062 � 11 66 � 765 � 16 70 � 18281 � 100 (102–397) 1077 � 731 (428–2546)26 � 4 27 � 5

183/83 239/4257 � 10 66 � 957 � 14 61 � 11213 � 74 (111–336) 1088 � 1306 (410–6264)29 � 4 27 � 3

1)

0)

werinsusliceexclCA.artersegmscor

of 1scor

corobeca(n �64-ssten


e small vessel size (n � 16), motion artifacts (n � 13),fficient contrast enhancement (n � 10), and missingor trigger artifact (n � 2); in addition, 6 segments were

uded because of a lack of information on conventionalAfter exclusion of stented segments, 9% of coronaryy segments were regarded as uninterpretable. Of allents, 28 (11%) were excluded in the group with Agatston

es of 0 to 100, 16 (9%) in the group with Agatston scores

Figure 1. Example of evaluation of stenoses in hmultiplanar reconstruction of right coronary arteB, Cross-sectional reconstruction perpendicular tohighly calcified coronary plaque in proximal segmeing sites in the longitudinal view of the right coroInvasive CA confirmed multiple obstructive steno

Figure 2. Example of evaluation of stenosis recalcification in left anterior descending coronarydemonstrating a nonobstructive stenosis in thereconstruction (arrow). (Corresponding site in thecoronary artery is depicted as lines in A.) C, Thevolume-rendered reconstruction (arrow). D, Findin

01 to 400, and 3 (3%) in the group with Agatstones greater than 400 (P � .03).After exclusion of 43 stented segments and 13nary segments having insufficient image qualityuse of motion artifacts (n � 3) or small vessel size

10), 800 segments in 60 patients examined bylice MSCT were included in the analysis. Afterted segments had been excluded, 2% of coronary

alcified right coronary artery. A, Curvedonstrating multiple obstructive stenoses.ine of right coronary artery demonstratingight coronary artery (arrow). (Correspond-tery are depicted as lines in A and C.) C,the right coronary artery.

from a coronary plaque with minimaly. A, Curved multiplanar reconstruction

part of the vessel. B, Cross-sectionaldinal view of the left anterior descendingesion was observed on the 3-dimensionale confirmed by conventional CA (arrow).

ighly cry demcenterlnt of rnary arses in

sultingarter

middlelongitusame lgs wer

artecoroAgaexclandscor

segm100withdiffeami(0%

EvaAcc

Tab

withterpmen159tionobsexaove240exainedobssitivtiveracievabetwtheFigu

Tabcalc

16-SInFaFa

64-SInFaFa

*Low

Tab

16-SAAA

64-SAAA

Valu


ry segments were regarded as uninterpretable. Nonary segments were excluded in the group withtston scores of 0 to 100, whereas 8 (3%) wereuded in the group with Agatston scores of 101 to 4005 (2%) were excluded in the group with Agatstones greater than 400 (P � .03).The percentages of false-positive and false-negativeents in the groups with overall Agatston scores of 0 to

, 101 to 400, and greater than 400 in patients examined16- and 64-slice MSCT are given in Table 2. Only therence between false-negative segments in patients ex-

ned with 16-slice MSCT was found to be significant, 5.3%, and 2.9%, respectively; P � .0005).

luation of Impact of Calcium on Diagnosticuracy of MSCT CA: Vessel-by-Vessel Analysis

The results of per-vessel analysis are provided inle 3 and Figure 3. In the patient population examined

le 2. Percentages of interpretable, false-positive, andium scores

A

0–100

lice MSCT CAterpretable [No. (%)] 219/247 (89)lse positive [No. (%)] 1/219 (0.5)lse negative [No. (%)] 0/219 (0)lice MSCT CAterpretable [No. (%)] 253/253 (100)lse positive [No. (%)] 1/253 (0.4)lse negative [No. (%)] 4/253 (1.6)

est P values of comparisons between groups.

le 3. Diagnostic accuracy of MSCT CA in vessels wit

Sensitivity (%)

lice MSCT CAll vessels (n � 159) 76 (61–91)gatston score of 0–100 (n � 120) 67 (43–91)gatston score �100 (n � 39) 83 (66–100)lice MSCT CAll vessels (n � 240) 79 (68–89)gatston score of 0–100 (n � 175) 77 (61–93)gatston score �100 (n � 65) 81 (67–95)

es in parentheses represent 95% confidence intervals.

16-slice MSCT, 5 vessels were regarded as unin-retable because more than 1 uninterpretable seg-t was present in the coronary artery, resulting invessels being included in the analysis. Conven-

al CA detected 33 coronary vessels (21%) withtructive coronary lesions in the patient groupmined with 16-slice MSCT CA, resulting in anrall sensitivity of 76% and specificity of 97%. Allvessels were regarded as interpretable in patients

mined with 64-slice MSCT. Of the vessels exam-with 64-slice MSCT CA, 57 (24%) showed

tructive coronary stenoses, and the calculated sen-ity and specificity were 79% and 96%, respec-ly. No significant differences in diagnostic accu-es of 16- and 64-slice MSCT CA and combinedluation of 16- and 64-slice MSCT CA were noted

een the groups of patients with Agatston scores inrange of 0 to 100 and greater than 100 (Table 3 andre 3).

-negative segments in groups with different

on score

P value–400 >400

85 (91) 105/108 (97) .0369 (1.2) 3/105 (2.9) .1*69 (5.3) 3/105 (2.9) �.0005*

74 (97) 281/286 (98) .0366 (2.6) 5/281 (1.8) .07*66 (1.9) 7/281 (2.5) .55*

rent calcium scores

pecificity (%)

Positivepredictivevalue (%)

Negativepredictivevalue (%)

97 (94–100) 86 (73–98) 94 (90–98)99 (97–100) 91 (74–100) 95 (91–99)86 (71–100) 83 (66–100) 86 (71–100)

96 (93–99) 87 (78–96) 94 (91–97)97 (94–100) 83 (68–98) 96 (93–99)92 (85–98) 89 (77–100) 84 (72–96)

false

gatst

101

169/12/19/1

266/27/25/2

h diffe

S

EvaAccAna

tectentis dexamlesio(44%specslice

stenspecFiguof 116-of p101

sum

Tab

16-sAAAAA

64-SAAAAA

Valu


luation of Impact of Calcium on Diagnosticuracy of MSCT CA: Patient-by-Patientlysis

The diagnostic performance of MSCT CA for de-ing obstructive lesions in patient groups with differ-overall Agatston scores on a patient-by-patient basisetailed in Table 4 and Figure 4. In the patient group

ined with 16-slice MSCT CA, obstructive coronaryns were detected by conventional CA in 18 patients), resulting in an overall sensitivity of 89%, with a

ificity of 87%. Of the patients examined with 64-MSCT CA, 32 (53%) had obstructive coronary

Figure 3. Diagnostic accuracies (per-vessel analysMSCT CA. PPV, Positive predictive value; NPV,

le 4. Diagnostic accuracy of MSCT CA in patient gro

Sensitivity (%)

lice MSCT CAll patients (n � 41) 89 (75–100)gatston score of 0–100 (n � 18) 100gatston score of 101–400 (n � 14) 75 (45–100)gatston score �400 (n � 9) 100gatston score �100 (n � 23) 86 (68–100)lice MSCT CAll patients (n � 60) 91 (81–100)gatston score of 0–100 (n � 19) 83 (53–100)gatston score of 101–400 (n � 20) 83 (62–100)gatston score �400 (n � 21) 100gatston score �100 (n � 41) 92 (82–100)

es in parentheses represent 95% confidence intervals.

oses; the calculated sensitivity was 91%, and theificity was 96%. As can be derived from Table 4 andre 4, no significant differences in diagnostic accuracies6- and 64-slice MSCT CA and combined evaluation ofand 64-slice MSCT CA were noted between the groupsatients having Agatston scores in the range of 0 to 100,to 400, greater than 400, and greater than 100.

DISCUSSION

The most important findings of this study can bemarized as follows: A negative impact of an elevated

confidence intervals for 16- and 64-sliceve predictive value.

ith different calcium scores

Specificity (%)

Positivepredictivevalue (%)

Negativepredictivevalue (%)

87 (73–100) 84 (68–100) 91 (79–100)93 (80–100) 80 (45–100) 10083 (53–100) 86 (60–100) 71 (37–100)67 (29–100) 86 (60–100) 10078 (51–100) 86 (68–100) 78 (51–100)

96 (89–100) 97 (91–100) 90 (79–100)100 100 93 (80–100)

88 (65–100) 91 (74–100) 78 (51–100)100 100 100

93 (80–100) 96 (88–100) 88 (72–100)

is) andnegati

ups w

coroof MysisobseonaimpMSpati

64-slinesensandcorowithtivitsensbeinlineet arepoOfaccuof Cpatiund

MSmayMSpretspat

decrquarespwith16-ssegmcomanalstudmen

MSfecthavtissuthefindinflufalssignagewithprevnificfromcalccaus

calcaccuanalAlth


nary artery calcium score on the diagnostic accuracySCT CA was observed in the segment-based anal-by use of 16-slice MSCT; this effect was not

rved with 64-slice MSCT. Moreover, elevated cor-ry artery calcium scores did not have a negativeact on the diagnostic accuracy of 16- and 64-sliceCT CA when data were analyzed on a vessel andent basis.In this analysis the accuracies of both 16- andlice MSCT to detect obstructive CAD were high, inwith previous publications.3-5,10,11 In particular, theitivity and specificity of 16-slice MSCT were 89%87%, respectively, for the detection of obstructivenary artery stenoses, and pooled data from 11 studies681 patients using 16-slice MSCT revealed a sensi-

y and specificity of 88% and 96%, respectively.2 Theitivity and specificity of 64-slice MSCT were higher,g 91% and 96%, respectively. These values are inwith recently published data. For example, Molletl4 evaluated 52 patients with 64-slice MSCT andrted a sensitivity of 99%, with a specificity of 95%.interest, the results of our study confirm the highracies in populations with a much higher prevalenceAD as compared with the previous studies (44% in

ents undergoing 16-slice MSCT and 53% in patientsergoing 64-slice MSCT).In general, the sensitivity and specificity of 64-slice

CT appear higher than those of 16-slice MSCT. Thisbe related to improved image quality with 64-slice

CT, which is reflected by the reduction in uninter-able segments. Increased temporal resolution andial resolution with 64-slice MSCT have led to a

Figure 4. Diagnostic accuracies (per-patient analyMSCT CA. PPV, Positive predictive value; NPV,

ease in the number of segments with impaired imagelity, being 9% and 2% for 16- and 64-slice MSCT,ectively, in this study. These observations are in line

the literature; pooled analysis of 11 studies withlice MSCT revealed that 4% (range, 0%-17%) ofents were of uninterpretable image quality,2 as

pared with all coronary segments included in theysis in most 64-slice MSCT studies,3,4 although 1y reported exclusion of as high as 12% of seg-ts.11

In addition, the reduced voxel size with 64-sliceCT also positively influences the partial-volume ef-s of coronary calcifications. Coronary calcificationse a high density and dominate the density of adjacentes in the same voxel. With a reduction in voxel size,blooming of calcium thus decreases. Indeed, theings in our study demonstrate the absence of anence of the calcium score on the percentage of

ely diagnosed segments with 64-slice MSCT. Aificant influence of the calcium score on the percent-of false-negative segments was observed, however,

16-slice MSCT. These results are in line withious observations that, with 16-slice MSCT, a sig-ant number of coronary segments had to be excludedanalysis because of severe calcifications.9 Coronary

ifications were also demonstrated to be the majore of false-positive findings with 16-slice MSCT.10

Another important finding in our study is that theium score did not significantly impair the diagnosticracy of a vessel-based analysis and a patient-basedysis, for both 16- and 64-slice MSCT examinations.ough more false-negative segments were observed

d confidence intervals for 16- and 64-sliceve predictive value.

sis) annegati

in pcreaincra rehighimptherthan

edgwerknoencematpatilimieffea smwasmayaccutionof tfeasshoumaiof i

thatdiagbasivesscoroaccupati

Ack

of in

Ref

1. ASes

2. SEmp4

3. LMw1

4. MEct

5. Kecc

6. WPsC

7. AJu

8. Cedt

9. KcmI

10. HLtev

11. Ras


atients with higher calcium scores, indicating de-sed sensitivity on a segmental basis, sensitivityeased when shifting to a patient level, most likely assult of the higher prevalence of CAD in patients wither calcium scores. These findings are particularlyortant in clinical practice, as further diagnostic andapeutic decisions are based on a patient-based (rathera segment- or vessel-based) analysis in daily practice.Several limitations of this study need to be acknowl-

ed. The protocols used for 16- and 64-slice MSCTe not precisely identical; therefore it remains un-wn to what extent small differences may have influ-d our results, in addition to the difference in colli-

ion of the 2 scanner generations. Also, inclusion ofents examined with the 16-slice MSCT scanner wasted and may have influenced results. Finally, thect of the presence of dense calcifications restricted toall area (resulting in a low overall calcium score)not explored in this study, and these calcificationspotentially also negatively influence the diagnosticracy despite a low overall calcium score. A limita-of MSCT in general is that only qualitative analysishe severity of coronary stenoses with MSCT isible, and more quantitative analysis techniquesld be developed. In addition, coronary MSCT re-

ns limited by the high radiation exposure and the useodinated contrast.In conclusion, the findings of this study demonstratethe coronary calcium score only slightly affected thenostic accuracy of 16-slice MSCT on a segments and showed no statistically significant impact on ael and patient basis. No significant impact of thenary calcium score was observed on the diagnosticracy of 64-slice MSCT CA on a segment, vessel, or

ent basis.

nowledgment

The authors have indicated they have no financial conflictsterest.

erences

chenbach S, Giesler T, Ropers D, Ulzheimer S, Derlien H,chulte C, et al. Detection of coronary artery stenoses by contrast-nhanced, retrospectively electrocardiographically-gated, multi-lice spiral computed tomography. Circulation 2001;103:2535-8.chuijf JD, Bax JJ, Shaw LJ, de Roos A, Lamb HJ, van der WallE, et al. Meta-analysis of comparative diagnostic performance ofagnetic resonance imaging and multi-slice computed tomogra-

hy for non-invasive coronary angiography. Am Heart J 2006;151:04-11.eschka S, Alkadhi H, Plass A, Desbiolles L, Grunenfelder J,arincek B, et al. Accuracy of MSCT coronary angiographyith 64-slice technology: first experience. Eur Heart J 2005;26:482-7.ollet NR, Cademartiri F, van Mieghem CA, Runza G, McFadden

P, Baks T, et al. High-resolution spiral computed tomographyoronary angiography in patients referred for diagnostic conven-ional coronary angiography. Circulation 2005;112:2318-23.uettner A, Kopp AF, Schroeder S, Rieger T, Brunn J, Meisner C,

t al. Diagnostic accuracy of multidetector computed tomographyoronary angiography in patients with angiographically provenoronary artery disease. J Am Coll Cardiol 2004;43:831-9.ong ND, Sciammarella MG, Polk D, Gallagher A, Miranda-

eats L, Whitcomb B, et al. The metabolic syndrome, diabetes, andubclinical atherosclerosis assessed by coronary calcium. J Amoll Cardiol 2003;41:1547-53.gatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M

r, Detrano R. Quantification of coronary artery calcium usingltrafast computed tomography. J Am Coll Cardiol 1990;15:827-32.ademartiri F, Mollet NR, Lemos PA, Saia F, Runza G, Midiri M,t al. Impact of coronary calcium score on diagnostic accuracy for theetection of significant coronary stenosis with multislice computedomography angiography. Am J Cardiol 2005;95:1225-7.uettner A, Burgstahler C, Beck T, Drosch T, Kopp AF, Heus-

hmid M, et al. Coronary vessel visualization using true 16-rowulti-slice computed tomography technology. Int J Cardiovasc

maging 2005;21:331-7.offmann U, Moselewski F, Cury RC, Ferencik M, Jang IK, DiazJ, et al. Predictive value of 16-slice multidetector spiral computed

omography to detect significant obstructive coronary artery dis-ase in patients at high risk for coronary artery disease: patient-ersus segment-based analysis. Circulation 2004;110:2638-43.aff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnosticccuracy of noninvasive coronary angiography using 64-slicepiral computed tomography. J Am Coll Cardiol 2005;46:552-7.

impact of coronary calcium score on diagnostic accuracy of multislice computed tomography coronary...

Documents