Coronary Artery Calcium Scoring on Low-doseProspective Electrocardiographically-triggered

64-Slice CT1

Jun Horiguchi, MD, Noriaki Matsuura, MD, Hideya Yamamoto, MD, Masao Kiguchi, RT, Chikako Fujioka, RTToshiro Kitagawa, MD, Nobuoki Kohno, MD, Katsuhide Ito, MD

Rationale and Objectives. The purpose of this prospective study was to assess image noise and variability in repeatedcoronary artery calcium (CAC) scoring on low-dose prospective electrocardiographically-triggered 64-slice multidetectorcomputed tomography.

Materials and Methods. Patients (n � 115) suspected of having coronary artery disease were scanned twice, using a tubecurrent of 10 � body mass index mA. The standard deviation (SD) of the computed tomographic value in the ascendingaorta and (mean � 2 � SD) were obtained. Repeated CAC scores (Agatston, volume, and mass) were measured by twoobservers, and the interscan and interobserver variability were determined.

Results. The mean tube current used was 246 � 36 mA. The mean tube current–time product and mean estimated effec-tive dose were 57 � 8 mA and 0.9 � 0.2 mSv, respectively. The SD and (mean � 2 � SD) computed tomographic val-ues in the ascending aorta were 16 � 3 and 75 � 10 Hounsfield units, respectively. Repeated CAC scores were correlated(r2 � 0.995–0.998). The interscan variability for observer 1 and observer 2, respectively, were 13% and 13% for Agatstonscore, 12% and 11% for volume, and 11% and 11% for mass. The interobserver variability for scan 1 and scan 2, respec-tively, were 3% and 3% for Agatston score, 5% and 3% for volume, and 3% and 3% for mass.

Conclusion. Low-dose prospective electrocardiographically-triggered 64-slice multidetector computed tomography showslow interscan and interobserver variability on CAC scoring while maintaining low image noise.

Key Words. CT; coronary artery; calcium; radiation dose; interscan variability; interobserver variability.

© AUR, 2009

The validity of serial coronary artery calcium (CAC)measurements as a method to monitor the progression ofatherosclerosis requires (1) that the progression of CAC

Acad Radiol 2009; 16:187–193

1 From the Department of Clinical Radiology, Hiroshima University Hospital,1-2-3, Kasumi-cho, Minami-ku, Hiroshima, 734-8551 Japan (J.H., M.K.,C.F.); and the Department of Radiology, Division of Medical Intelligenceand Informatics (N.M., K.I.), and Department of Molecular and InternalMedicine, Division of Clinical Medical Science (H.Y., T.K., N.K., K.I.), Pro-grams for Applied Biomedicine, Graduate School of Biomedical Sciences,Hiroshima University, Hiroshima, Japan. Received April 6, 2008; acceptedMay 13, 2008. This study was financially supported by the Tsuchiya Foun-dation, Hiroshima, Japan. Address correspondence to: J.H. e-mail:horiguch@hiroshima-u.ac.jp

©

AUR, 2009doi:10.1016/j.acra.2008.05.017

has biologic relevance to atherosclerotic activity, (2) thatthe progression of CAC can be detected relative to intert-est variability, (3) that changes in CAC severity haveprognostic relevance, and (4) that the modification of car-diovascular risk factors modulates the progression ofCAC (1). The normal progression of CAC score per yearis reported to be 14% to 27% (average, 24%) (2) and isaccelerated up to 33% to 48% with significant coronarydisease (3,4). However in previous studies, the interscanvariability of the Agatston score (5) using electron-beamcomputed tomography (CT) has been 20% to 37% (6–9).Retrospective electrocardiographically-gated overlappingscanning using 4-slice CT (10) and 16-slice CT (11) have

shown low interscan variability (12%–13%) for CAC

187

HORIGUCHI ET AL Academic Radiology, Vol 16, No 2, February 2009

scoring, but this method entails high radiation exposure.Because the progression of CAC is not clearly modifiablethrough standard risk-reducing therapies, and CAC mea-surement involves both cost and radiation exposure, theclinical monitoring of CAC progression using serial fastCT is not recommended by the American College of Car-diology and American Heart Association expert commit-tee (1). Under these circumstances, low radiation expo-sure and low interscan variability are key requirementsfor CAC scoring.

In a cardiac phantom study, prospective electrocardio-graphically-triggered 64-slice multidetector CT (MDCT),even with low radiation dose comparable with that ofelectron-beam CT, showed low variability for CAC scor-ing, comparable with that of retrospective electrocardio-graphically-gated 16-slice MDCT (12). In a clinical studywith 64-slice MDCT (13), prospective electrocardiograph-ically-triggered scanning showed low variability (Agatstonscore, 18%; volume, 12%; mass, 11%). The study used atube current of 350 mA, thereby resulting in a tube cur-rent–time product of 82 mA. In studies using 4-slice CT,however, lower tube current–time products of 40 mA (14)and 55 mA (15) have been recommended for low-doseCAC scoring. Thus, the purpose of this prospective studywas to assess image noise and variability in repeatedCAC scoring on low-dose prospective electrocardiograph-ically-triggered 64-slice MDCT.

MATERIALS AND METHODS

PatientsFor 8 months, 115 consecutive subjects (81 men, 34

women; mean age, 67 � 11 years; range, 34–87) whounderwent coronary CT for coronary risk factor or chestpain evaluation were enrolled in the study. Subjects withhistories of cardiac surgery, stents, or pacemakers wereexcluded. The study was approved by our institutionalreview committee. Written informed consent was receivedfrom all patients involved after the nature of the proce-dure had been fully explained (including radiation doseinformation).

Prospective Electrocardiographically-triggeredStep-and-Shoot Computed TomographicProtocol

Two repeated prospective electrocardiographically-triggered step-and-shoot half-scans were performed using

a 64-slice multidetector computed tomographic scanner

188

(LightSpeed VCT; GE Healthcare, Waukesha, WI) withsimultaneous electrocardiographic digitizing and record-ing. Between the 2 scans, to simulate different body posi-tions, the table was advanced by 1 mm. Scans were per-formed 4 to 5 seconds after holding the breath on mildinspiration to minimize changes in heart rate during thescan (16). Using 2.5-mm collimation width and 16 detec-tors, images 2.5 mm thick were obtained. Scans weretemporally triggered to be centered at 75% RR, becausemid-diastole reconstruction has been recommended for64-MDCT at a gantry rotation speed of 0.35 seconds (17).The gantry rotation speed was 0.35 seconds, and the tubevoltage was 120 kV. Because a significant associationbetween noise and body mass index (BMI; calculated asbody weight in kilograms divided by body mass in squaremeters) was shown in a study using electron-beam CT(18), we modified the tube current according to the fol-lowing equation: tube current � 250 � (BMI/25) mA �(10 � BMI) mA. This is based on the strategy that pa-tients with standard BMIs of 25 kg/m2 receive a tube cur-rent–time product of 58 mA, which is almost the samelevel as the recommendation for CAC scoring using low-dose 4-slice CT (19).

The tube current–time product in a typical patient witha BMI of 25 kg/m2 is equal to tube current (mA) � gan-try rotation speed (seconds) � exposure time per rotationtime � 250 mA � 0.35 seconds � (2/3) � 58 mAs.

For image reconstruction parameters, a matrix size of512 � 512 pixels, a display field of view of 26 cm, andthe “standard” kernel were used. The temporal resolutionwas 175 ms.

Image NoiseThe mean and standard (SD) of computed tomographic

values on scan 1, in regions of interest set in the aorta atthe level of the left coronary artery, were measured byobserver 1. The value (mean � 2 � SD), which is prefer-ably lower than the threshold of 130 Hounsfield units(HU) (14), was calculated.

CAC ScoringThe Agatston score (5) and calcium volume and mass

(7) were determined using a commercially available exter-nal workstation (Advantage Windows version 4.2; GEHealthcare) and CAC-scoring software (Smartscore ver-sion 3.5; GE Healthcare), according to the followingequations: (1) Agatston score � slice increment/slicethickness � �(area � cofactor); (2) volume � �(area �

slice increment); and (3) mass � �(area � slice incre-

Academic Radiology, Vol 16, No 2, February 2009 CAC SCORING ON 64-SLICE CT

ment � mean computed tomographic density) � calibra-tion factor (19).

The calibration factor was determined by measuring awater cylinder (0 HU) and a calcified cylinder (200 HU)with known calcium hydroxyapatite density. The factorwas used to calculate the absolute calcium hydroxyapatitemass from the measured mass (19).

All computed tomographic scans were independentlyscored by two radiologists with eight and two years ofexperience in CAC scoring (observers 1 and 2, respec-tively). For patients with positive CAC scores on the twoscans, each of the Agatston, volume, and mass scores, ona logarithmic scale to reduce skewness, were comparedbetween two repeated scans and between two observers(repeated-measures analysis of variance [ANOVA]). Therepeated CAC scores, in each CAC scoring algorithm,were correlated.

We performed monthly scanning of a calibration phan-tom (Anthropomorphic Cardio Phantom; Institute of Med-ical Physics, Erlangen, Germany and QRM GmbH,Möhrendorf, Germany) to determine the calibration factor(20).

Interscan and Interobserver VariabilityFor patients with positive CAC scores on the two

scans, interscan and interobserver variability were calcu-lated using the percentage difference in calcium scores:interscan variability � [absolute(scan 1 � scan 2)/(scan1 � scan 2) � 0.5] � 100; interobserver variability �[absolute(observer 1 � observer 2)/(observer 1 � ob-server 2) � 0.5] � 100, where observer 1 is the CACscore measured by observer 1.

Using 2-factor factorial ANOVA, the interscan vari-ability was compared between the observers and the CACscoring algorithms and interobserver variability betweenthe scans and the CAC scoring algorithms. For thesetests, CAC scores were transformed to a logarithmicscale.

Radiation DoseDose–length product (in milligray–centimeters) dis-

played on a dose report on the computed tomographicscanner was recorded. The effective dose was estimatedusing a method proposed by the European WorkingGroup for Guidelines on Quality Criteria in CT (21). Inthis method, the effective dose is derived from the dose–length product and a conversion coefficient for the chest(k � 0.017 mSv/mGy/cm averaged between male and

female models).

Statistical AnalysesAll statistical analyses were performed using a com-

mercially available software package (Statcel 2; OMSPublishing, Saitama, Japan). Data are expressed as mean �SD. ANOVA (multivariate calculation) tests were used todetermine group differences. P values � .05 were consid-ered to identify significant differences.

RESULTS

All patients were able to hold their breath for the twoscans. Scan times were 4.9 � 1.3 and 5.0 � 1.4 secondsfor scans 1 and 2, respectively.

Eighty of the overall 115 patients showed positiveCAC scores for the three algorithms on both scans. Thirty-three patients showed negative scores for the three algo-rithms on both scans, and two patients showed both posi-tive and negative scores on the algorithms (eg, negativeAgatston score and positive volume and mass scores).Baseline characteristics of patients are presented in Table1. These 35 patients were excluded, because the variabil-ity could not be obtained for negative scores on bothscans. No patient showed, for the three algorithms, posi-tive scores on one scan and negative scores on the otherscan. For patients included in further investigation (62men and 18 women), the mean heart rates and changes ofheart rate (difference between maximal and minimal heartrates), respectively, were 62 � 11 beats/min (range, 40–87) and 3 � 9 beats/min (range, 0–76) on scan 1 and62 � 10 beats/min (range, 40–85) and 2 � 6 beats/min(range, 0–56). The patients’ body weight and BMI were64 � 11 kg (range, 40–100) and 24.2 � 3.1 kg/m2

(range, 18.3–34.6).

Image NoiseThe mean and SD in regions of interest were 42 � 7

and 16 � 3 HU. The value of (mean � 2 � SD) was 75 �10 HU (range, 53–99). Representative images, illustratinghigh and low SD values, are shown in Figure 1.

CAC ScoresThe Agatston, volume, and mass scores are summa-

rized in Table 2. Repeated-measures ANOVA revealedthat there was no statistical significance of log-trans-formed CAC scores between the scans (Agatston score,P � .15; volume, P � .17; mass, P � .38) and observers

(Agatston score, P � .93; volume, P � .89; mass, P �

189

HORIGUCHI ET AL Academic Radiology, Vol 16, No 2, February 2009

.95). The CAC scores of scan 1 and scan 2 were corre-lated (r2 � 0.995–0.998).

Interscan VariabilityThe interscan variability in Agatston, volume, and

mass scores is shown in Figure 2. Among CAC scoringalgorithms, the mass algorithm showed the least variabil-ity for the two observers. However, 2-factor factorialANOVA revealed that no significance of the variabilitywas reached between observers (P � .92) or betweenscoring algorithms (P � .35).

Interobserver VariabilityThe interobserver variability in Agatston, volume, and

mass scores is shown in Figure 3. The interobserver vari-ability was small for all algorithms for both scans. Two-factor factorial ANOVA revealed that there was no statis-tical significance of the variability between scans (P �.60) or between scoring algorithms (P � .61).

Radiation DoseThe tube current used was 246 � 36 mA, yielding a

tube current–time product of 57 � 8 mA. Dose–lengthproducts displayed on the dose report on the computedtomographic scanner on scan 1 and the estimated effec-tive doses were 53.5 � 8.9 mGy · cm and 0.9 � 0.15mSv, respectively.

DISCUSSION

The present study is the first to demonstrate that low-dose prospective electrocardiographically-triggered 64-

Table 1Patient Demographics

Variable All Patients

No. of patients 115Women/men 81/34Age (y) (range) 64 � 12 (34–8Body mass index (kg/m2) 25 � 3 (18–35Symptom 82/115 (71%Risk factor 80/115 (70%

Quantitative variables are expressed as meanas frequency (percentage).

slice MDCT, using a BMI-adapted tube current, shows

190

low interscan and interobserver variability on CAC scor-ing while maintaining low image noise.

Image NoiseTakahashi et al (14) recommended a dose of 40 mA,

because the CAC scores at 40 and 150 mA were highlycorrelated and the interscan variability showed no differ-ence between the variability of a subgroup (80 and 150mA). Shemesh et al (15) showed that CAC scores at 55and 165 mA were highly correlated, and the SD levelswere 14.7 � 3.9 HU at 55 mA and 9.3 � 2.1 HU at 165mA. In a recent report of the International Consortium onStandardization in Cardiac CT (22), an SD level target of20 HU for small and medium-sized patients and an SDlevel target of 23 HU for large patients were recom-mended.

We determined the tube current on the basis of thestrategy that patients with BMIs of 25 kg/m2 received atube current–time product of 58 mA, and with the use ofBMI-based modulation, slender patients received less tubecurrent and obese patients received more. Although weused a mean effective dose of 0.9 � 0.15 mSv, the SDlevel was suppressed to 16 � 3 HU, which is lower thanthe previous recommendation. Moreover, the value of(mean � 2 � SD) was 75 � 10 HU, controlled withinthe CAC detecting threshold of 130 HU. This indicatesthat hyperattenuating noise is rarely misjudged as CAC(23). Ideally, the SD should be kept constant over differ-ent BMI or body weight values. In 7 of the 144 patients,however, the SD level exceeded 23 HU (24 HU, n � 5;26 HU, n � 1; 27 HU, n � 1). The BMI-based modula-

Coronary Artery Calcium (AgatstonScore)

Positive Negative

80 3562/18 19/16

67 � 11 (39–87) 57 � 13 (34–80)25 � 3 (24–35) 25 � 4 (18–34)

56/80 (70%) 26/35 (74%)56/80 (70%) 14/35 (40%)

D (range). Categorical variables are expressed

7))))

� S

tion in the present study accounted only for patients’

Academic Radiology, Vol 16, No 2, February 2009 CAC SCORING ON 64-SLICE CT

Figure 1. Representative images showing high and low SD values. Two low-noise images (a, 11 Hounsfield units [HU]; b, 11HU) and two high-noise images (c, 27 HU; d, 26 HU) are shown. (a) Image of a man with a body mass index [BMI] of 20 kg/m2

and a body height of 164 cm. (b) Image of a woman with a BMI of 22 kg/m2 and a body height of 146 cm. (c) Image of a manwith a BMI of 31 kg/m2 and a body height of 180 cm. (d) Image of a woman with a BMI of 29 kg/m2 and a body height of 160

cm, presenting with cardiomegaly.

191

98.

HORIGUCHI ET AL Academic Radiology, Vol 16, No 2, February 2009

slimness or obesity, however, not their body size or bodyhabitus.

CAC Scores: Interscan and InterobserverVariability

CAC scores on repeated scans were well correlated,and the interscan variability was comparable with the re-sults on retrospective electrocardiographically gated over-lapping scans using 4- or 16-slice computed tomographicscanners (10,11) and 64-slice prospective electrocardio-graphically-triggered CT (12). Among CAC scoring algo-rithms, although not statistically significant, mass showedthe least variability. The mass algorithm enables thequantitative comparison of CAC that has been measured

Figure 2. Interscan variability of repeated coronary artery cal-cium scoring. Graph shows the interscan variability (Agatstonscore: black; volume: white; mass: gray) between the two observ-ers. Bars and vertical lines indicate mean and SD, respectively.

Table 2Agatston, Volume, and Mass Scores on Two

Agatston Score

Observer 1*Scan 1 688 (223), 1–5,676Scan 2 683 (211), 1–6,130

Observer 2†

Scan 1 709 (229), 1–5,849Scan 2 701 (211), 1–6,262

Data are expressed as mean (median), range.*Agaston score on scan 2 � 1.03 � (Agaston

score on scan 2 � 1.02 � (volume score on sc1.03 � (mass score on scan 1) � 2.74; r2 � 0.9

†Agaston score on scan 2 � 1.02 � (Agastonscore on scan 2 � 1.02 � (volume score on sc1.02 � (mass score on scan 1) � 3.34; r2 � 0.9

at different times and using different computed tomo-

192

graphic scanner models. This facilitates a meaningfulcomparison of patient data both over time within a patientand across patients and institutions (22). The interob-server variability in the present study was small. Artificiallesions are known to affect interobserver variability, be-cause interpreters must decide whether to take them intoaccount or not. Such lesions were very much reduced inthe study, with low noise levels.

Radiation DoseThe radiation dose in this study was comparable with

that on electron-beam CT (1.0 mSv for men and 1.3 mSvfor women) (20). Lowering of the tube voltage also con-tributes to the reduction of dose. Thomas et al (24), in a

Figure 3. Interobserver variability of repeated coronary arterycalcium scoring. Graph shows the interobserver variability (Agat-ston score: black; volume: white; mass: gray) between the twoscans. Bars and vertical lines indicate mean and SD, respectively.

s Measured by Two Observers

Volume Mass

38 (186), 2–4,270 143 (40), 0.2–1,15833 (180), 3–4,658 143 (40), 0.3–1,185

52 (189), 2–4,392 144 (40), 0.2–1,16746 (183), 3–4,748 144 (40), 0.3–1,187

e on scan 1) � 23.9; r2 � 0.997. Volume� 17.0; r2 � 0.995. Mass score on scan 2 �

e on scan 1) � 22.4; r2 � 0.998. Volume� 18.1; r2 � 0.998. Mass score on scan 2 �

Scan

55

55

scoran 1)97.scor

an 1)

study using a static anthropomorphic heart phantom with

Academic Radiology, Vol 16, No 2, February 2009 CAC SCORING ON 64-SLICE CT

calcium cylinders, showed a dose reduction of 57% on an80-kVp protocol in CAC scoring, without compromisingreproducibility and accuracy. Jakobs et al (25) showed ina clinical study that an 80-kVp protocol significantly re-duced dose, without affecting the detection and quantifi-cation of CAC. This method needs calibration for the cal-culation of a scanner-specific adapted threshold for theidentification of calcium and can be used only for calciummass scoring.

A limitation of this study was that we set the tube cur-rent using BMI-based modulation. We believe that thismethod is better than using fixed tube current, but thetube current needs further optimization so that the noiselevel is kept at almost the same level across patients. Inconclusion, low-dose prospective electrocardiographicallytriggered 64-slice MDCT shows low interscan and inter-observer variability for CAC scoring while maintaininglow image noise.

REFERENCES

1. Greenland P, Bonow RO, Brundage BH, et al. ACCF/AHA 2007 clinicalexpert consensus document on coronary artery calcium scoring bycomputed tomography in global cardiovascular risk assessment and inevaluation of patients with chest pain. J Am Coll Cardiol 2007; 49:378–402.

2. Maher JE, Bielak LF, Raz JA, Sheedy PF II, Schwartz RS, Peyser PA.Progression of coronary artery calcification: a pilot study. Mayo ClinProc 1999; 74:347–355.

3. Janowitz WR, Agatston AS, Viamonte M Jr. Comparison of serial quan-titative evaluation of calcified coronary artery plaque by ultrafast com-puted tomography in persons with and without obstructive coronaryartery disease. Am J Cardiol 1991; 68:1–6.

4. Fischbach R, Heindel W. Detection and quantification of coronarycalcification: an update. Rofo 2000; 172:407–414.

5. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, De-trano R. Quantification of coronary calcium using ultrafast computedtomography. J Am Coll Cardiol 1990; 15:827–832.

6. Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease: im-proved reproducibility of calcium scoring with an electron-beam CTvolumetric method. Radiology 1998; 208:807–814.

7. Yoon HC, Greaser LE III, Mather R, Sinha S, McNitt-Gray MF, GoldinJG. Coronary artery calcium: alternate methods for accurate and repro-ducible quantitation. Acad Radiol 1997; 4:666–673.

8. Wang SJ, Detrano BC, Secci A, et al. Detection of coronary calcifica-tion with electron-beam computed tomography: evaluation of interex-

amination reproducibility and comparison of three image-acquisitionprotocols. Am Heart J 1996; 132:550–558.

9. Achenbach S, Ropers D, Mohlenkamp S, et al. Variability of repeatedcoronary artery calcium measurements by electron beam tomography.Am J Cardiol 2001; 87:210–213.

10. Ohnesorge B, Flohr T, Fischbach R, et al. Reproducibility of coronarycalcium quantification in repeat examinations with retrospectively ECG-gated multisection spiral CT. Eur Radiol 2002; 12:1532–1540.

11. Horiguchi J, Yamamoto H, Akiyama Y, et al. Variability of repeated cor-onary artery calcium measurements by 16-MDCT with retrospectivereconstruction. AJR Am J Roentgenol 2005; 184:1917–1923.

12. Horiguchi J, Kiguchi M, Fujioka C, et al. Variability of repeated coro-nary artery calcium scoring and radiation dose on 64-slice and 16-sliceCT by prospective electrocardiograph-triggered axial and retrospectiveelectrocardiograph-gated spiral CT—a phantom study. Acad Radiol2008; 15:958–965.

13. Horiguchi J, Matsuura N, Yamamoto H, et al. Variability of repeatedcoronary artery calcium measurements by 1.25-mm- and 2.5-mm-thickness images on prospective electrocardiograph-triggered 64-sliceCT. Eur Radiol 2008; 18:209–216.

14. Takahashi N, Bae KT. Quantification of coronary artery calcium withmulti-detector row CT: assessing interscan variability with differenttube currents-pilot study. Radiology 2003; 228:101–106.

15. Shemesh J, Evron R, Koren-Morag N, et al. Coronary artery calciummeasurement with multi-detector row CT and low radiation dose: com-parison between 55 and 165 mAs. Radiology 2005; 236:810–814.

16. Horiguchi J, Shen Y, Hirai N, et al. Timing on 16-slice scanner and im-plications for 64-slice cardiac CT: do you start scanning immediatelyafter breath-hold? Acad Radiol 2006; 13:173–176.

17. Matsuura N, Horiguchi J, Yamamoto H, et al. Optimal cardiac phasefor coronary artery calcium scoring on single-source 64-MDCT scan-ner—least interscan variability and least motion artifacts. AJR Am JRoentgenol 2008; 190:1561–1568.

18. Sevrukov A, Pratap A, Doss C, Jelnin V, Hoff JA, Kondos GT. Electronbeam tomography imaging of coronary calcium: the effect of bodymass index on radiologic noise. J Comput Assist Tomogr 2002;26:592–597.

19. Hong C, Bae KT, Pilgram TK, Suh J, Bradley D. Coronary artery cal-cium measurement with multi-detector row CT: in vitro assessment ofeffect of radiation dose. Radiology 2002; 225:901–906.

20. Halliburton SS, Stillman AE, Lieber M, Kasper JM, Kuzmiak SA, WhiteRD, Potential clinical impact of variability in the measurement of coro-nary artery calcification with sequential MDCT. AJR Am J Roentgenol2005; 184:643–648.

21. Menzel H, Schibilla H, Teunen D. European Guidelines for Quality Cri-teria for Computed Tomography. Brussels, Belgium: EuropeanCommission; 2000.

22. McCollough CH, Ulzheimer S, Halliburton SS, et al. Coronary arterycalcium: a multiinstitutional, multimanufacturer international standardfor quantification at cardiac CT. Radiology 2007; 243:527–538.

23. Bielak LF, Kaufmann RB, Moll PP, MacCollough CH, Schwartz RS,Sheedy PF II. Small lesions in the heart identified at electron beam CT:calcification or noise? Radiology 1994; 192:631–636.

24. Thomas CK, Mühlenbruch G, Wildberger JE, et al. Coronary artery cal-cium scoring with multislice computed tomography. In vitro assess-ment of a low tube voltage protocol. Invest Radiol 2006; 41:668–673.

25. Jakobs TF, Wintersperger BJ, Herzog P, et al. Ultra-low-dose coronary

artery calcium screening using multislice CT with retrospective ECGgating. Eur Radiol 2003; 13:1923–1930.

193