European Journal of Radiology

Transcription

1 European Journal of Radiology 82 (2013) e58 e63 Contents lists available at SciVerse ScienceDirect European Journal of Radiology journa l h o me pa ge: Calcium score of small coronary calcifications on multidetector computed tomography: Results from a static phantom study J.M. Groen a, K.F. Kofoed b, M. Zacho b, R. Vliegenthart a, T.P. Willems a, M.J.W. Greuter a, a Department of Radiology, University of Groningen, University Medical Center Groningen, The Netherlands b Department of Cardiology and Radiology, Rigshospitalet, University of Copenhagen, Denmark a r t i c l e i n f o Article history: Received 22 July 2012 Received in revised form 24 September 2012 Accepted 30 September 2012 Keywords: Computed tomography Coronary calcium Calcium score a b s t r a c t Introduction: Multi detector computed tomography (MDCT) underestimates the coronary calcium score as compared to electron beam tomography (EBT). Therefore clinical risk stratification based on MDCT calcium scoring may be inaccurate. The aim of this study was to assess the feasibility of a new phantom which enables establishment of a calcium scoring protocol for MDCT that yields a calcium score comparable to the EBT values and to the physical mass. Materials and methods: A phantom containing 100 small calcifications ranging from 0.5 to 2.0 mm was scanned on EBT using a standard coronary calcium protocol. In addition, the phantom was scanned on a 320-row MDCT scanner using different scanning, reconstruction and scoring parameters (tube voltage kv, slice thickness mm, reconstruction kernel FC11 FC15 and threshold HU). The Agatston and mass score of both modalities was compared and the influence of the parameters was assessed. Results: On EBT the Agatston and mass scores were between 0 and 20, and 0 and 3 mg, respectively. On MDCT the Agatston and mass scores were between 0 and 20, and 0 and 4 mg, respectively. All parameters showed an influence on the calcium score. The Agatston score on MDCT differed 52% between the 80 and 135 kv, 65% between 0.5 and 3.0 mm and 48% between FC11 and FC15. More calcifications were detected with a lower tube voltage, a smaller slice thickness, a sharper kernel and a lower threshold. Based on these observations an acquisition protocol with a tube voltage of 100 kv and two reconstructions protocols were defined with a FC12 reconstruction kernel; one with a slice thickness of 3.0 mm and a one with a slice thickness of 0.5 mm. This protocol yielded an Agatston score as close to the EBT as possible, but also a mass score as close to the physical phantom value as possible, respectively. Conclusion: With the new phantom one acquisition protocol and two reconstruction protocols can be defined which produces Agatston scores comparable to EBT values and to the physical mass Elsevier Ireland Ltd. All rights reserved. 1. Introduction Coronary calcium deposit is a powerful marker in screening studies for coronary artery disease (CAD) [1,2]. Therefore, the amount of coronary calcium is used as a risk stratification for a main cardiac event within the next 5 years [3]. However, a calcium score of zero cannot be interpreted as a reassurance of the absence of CAD [4 6]. Although the prognostic value of zero calcium is under Corresponding author at: University of Groningen, University Medical Center Groningen, Department of Radiology EB44, PO Box , 9700 RB Groningen, The Netherlands. Tel.: addresses: jaap.groen@slaz.nl (J.M. Groen), kkofoed@dadlnet.dk (K.F. Kofoed), dls332089@vip.cybercity.dk (M. Zacho), r.vliegenthart@umcg.nl (R. Vliegenthart), t.p.willems@umcg.nl (T.P. Willems), m.j.w.greuter@umcg.nl (M.J.W. Greuter). debate, it is still related to a low cardiac risk event. Because small coronary calcifications can contribute significantly to a higher risk on a major adverse cardiac event, detection and a precise score of these small calcifications is therefore important. Coronary calcifications were originally quantified by the Agatston score (AS) for patients who underwent electron beam computed tomography (EBT) [7]. Later, mass scoring (MS) was proposed as an alternative method because of a better reproducibility compared to the AS [8]. The majority of coronary artery disease studies used EBT as imaging modality because of its high temporal resolution [9 14]. However, the presence of EBT scanners has strongly diminished in recent years and patients are generally scanned on multi-detector computed tomography (MDCT) systems. Although EBT has almost become obsolete, clinical risk stratification is mostly still based upon EBT-acquired AS. Also, the recommendations from expert groups to substitute AS with MS has not yet let to the general acceptance of MS [15] X/$ see front matter 2012 Elsevier Ireland Ltd. All rights reserved.

2 J.M. Groen et al. / European Journal of Radiology 82 (2013) e58 e63 e59 Table 1 Scanning, reconstruction and scoring parameters used on the EBT and MDCT system. FBP = filtered back projection. EBT Scan parameters Tube voltage (kv) 130 Tube current (ma s) 600 Acquisition time (ms) 50 Reconstruction parameters Slice thickness (mm) a 3.0 Increment (mm) a 3.0 Reconstruction kernel Sharp Reconstruction method FBP MDCT Scan parameters Tube voltage (kv) b 80, 100, 120, 135 Tube current (ma s) 200 Rotation time (ms) 350 Collimation (mm) Reconstruction parameters Slice thickness (mm) b 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 Increment (mm) b 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 a Reconstruction kernel b FC11, FC12, FC13, FC14, FC15 Reconstruction method FBP Scoring parameter Threshold (HU) b 110, 120, 130, 140, 150 a The increment was equal to the slice thickness. b Parameters that were varied in a systematic way. Whereas the temporal resolution of MDCT is lower than EBT, the spatial resolution is higher [16]. Although the two techniques seem very similar, standard MDCT calcium scoring protocols do not give the same results as EBT [17 25]. In general, MDCT is associated with a 2 10% underestimation of the AS compared to EBT [17]. For large, high density coronary calcifications this underestimation may have minor clinical implications. However, for small, low density coronary calcifications this underestimation can have a significant clinical implication, e.g. an impact on the estimated cardiovascular risk of the patient [26]. Therefore especially small amounts of coronary calcium should be measured accurately when using MDCT and with outcome similar to measures performed with EBT system in the individual patient. The aim of this study was to assess the feasibility of a new phantom which enables establishment of a calcium scoring protocol for MDCT that yields a calcium score comparable to the EBT values and to the physical mass. 2. Materials and methods 2.1. Phantom A dedicated phantom was developed which contained 100 small cylindrical calcifications varying in size and density (Fig. 1a). The diameter and length of the calcifications was equal and ranged from 0.5 to 2.0 mm. The density of the calcifications ranged from 90 to 540 mg hydroxyapatite (HA) per cm 3. The phantom was inserted into a thorax phantom (QRM Thorax, QRM, Germany), which comprised of artificial lungs and a spine (Fig. 1b) EBT, MDCT scanners To establish a reference standard in terms of calcium scoring, the phantom was scanned on three EBT scanners (Imatron C300, GE, Milwaukee, USA). Data was acquired at a tube voltage of 130 kv, a tube current of approximately 600 ma s and a collimation of 3.0 mm. The data was reconstructed using a sharp kernel at a slice thickness of 3.0 mm and an increment of 3.0 mm, a standard calcium scoring protocol for EBT scanners (Table 1). Each scan was Fig. 1. (a) Schematic axial view of the phantom at one of the four planes containing 25 calcifications each (left), and schematic side view of the phantom showing the all four planes with calcifications (right). (b) Radiologic axial view of the phantom, the different calcifications are clearly visible. (c) Schematic drawing of the analysis method used. The border between detected and undetected calcification as determined by the reference method EBT is given by the visibility curve (isocurve for HU-peak = 130 HU). Two calcifications are given as example, one with a size of 1.1 mm and density of 250 mg HA (triangle) which is not detected and thus below the visibility curve, the other with a size of 1.7 mm and a density of 450 mg HA (circle) which is detected and thus above the visibility curve. repeated five times with a small translation (2 mm) and/or rotation (2 ) between each scan. Next, the phantom was scanned with a 320-row MDCT scanner (Toshiba Aquilion ONE, Toshiba Medical Systems, Japan). Data was acquired in a single rotation with an axial scan field of view of 50 cm encompassing the whole phantom. The scan was performed sequentially at different tube voltages (Table 1). Each scan was

3 e60 J.M. Groen et al. / European Journal of Radiology 82 (2013) e58 e63 repeated 5 times with a small translation (2 mm) and/or rotation (2 ) between each scan. Images were reconstructed with different slice thicknesses and reconstruction kernels were the increment was kept equal to the slice thickness (Table 1) Data analysis After reconstruction, the image data sets were analyzed using a MATLAB script with the standard scoring algorithms implemented to easily obtain the calcium scores [7,8,15,27]. This script was validated against commercially available calcium scoring software (Siemens Syngo CaScore (Siemens Medical Solutions, Erlangen, Germany)) with a 100% match. The calcium score of the individual calcifications as well as the total calcium score of the phantom obtained with the MATLAB script was compared to the calcium scores of Syngo CaScore and no differences were observed. For each scan the AS and MS (mg) for each calcification was determined. The standard scoring threshold of 130 HU was used, as well as other thresholds (Table 1). Next, the five repeated measurements were averaged and the average calcium score (AS and MS) and standard deviation were calculated for each calcification for each combination of the scan parameters. The calcium scores obtained with EBT were averaged and defined as the reference standard scores (AS ref and MS ref ). The calibration factor, required for mass scoring, was obtained from the phantom as well [15]. Five larger calcified objects with known density were used to calculate the calibration factor for each scanner and tube voltage. The calculated calcium scores and/or HU peak values of the calcifications were arranged in a formation with increasing size and density along the x- and y-axis, respectively (see Fig. 1c). Relatively large and high density calcifications were positioned in the upper right corner, whereas relative small and low density calcifications were positioned in the lower left corner of the formation. Next a visibility curve (isocurve for HU-peak = 130 HU) was drawn based upon the HU-peak values, which showed the separation between calcifications that were detected and that were not (Fig. 1c), an approach comparable to contrast-detail curves in radiography [28,29]. Calcifications were indicated as visible if their pixel value was larger than 130 HU. The position of the visibility curve depended on the visibility of calcifications and HU-peak values of the calcifications and therefore on the scanning parameters. In addition, the visibility curves could be used to compare different scanning systems Statistical analysis A second comparison between systems was made with a pair wise comparison between the MDCT-based scores (AS MDCT, MS MDCT ) and the EBT-based scores (AS ref, MS ref ) and between the MDCT-based scores and the physical masses (MS physical ). In order to minimize the difference between the MDCT-based scores and the reference standard scores for EBT, a square root of the sum of squared differences (SSD) was calculated. The SSD is defined as SSD = (x s,d y s,d ) 2 s,d in which x s,d and y s,d are the calcium scores at a given size s and density d. The larger SSD the larger the difference between the two datasets. For calcium scores the error margin was given by the standard deviation, obtained from the repeated scans. Calcium scores were assumed to be normally distributed. Table 2 Mass calibration factor as obtained from the phantom for the three EBT scanners (EBT1 to EBT3) and for the MDCT scanner per tube voltage. Scanner Calibration factor EBT EBT EBT MDCT at 80 kv MDCT at 100 kv MDCT at 120 kv MDCT at 135 kv Results On EBT the individual AS scores of the calcifications within the phantom were between 0 and 20, and the MS scores were between 0 and 3 mg. On MDCT the individual AS scores were between 0 and 20 and the MS scores were between 0 and 4 mg. The mass calibration factors for each of the three EBT systems and for the MDCT system at the different tube voltages is shown in Table 2. For MDCT, all scanning parameters showed an influence on the calcium score (Fig. 2a d). More calcifications were detected using MDCT with a lower tube voltage. The same was observed at a smaller slice thickness, a sharper kernel and a lower HU threshold. The total AS differed 52% between the highest and lowest tube voltage, 65% between the largest and smallest slice thickness and 48% between the smoothest and sharpest reconstruction kernel, respectively. The total MS differed 26%, 55% and 40% for the tube voltage, slice thickness and reconstruction kernel, respectively. The SSD analysis showed that as a function of tube voltage, the differences between AS MDCT and AS ref showed an optimum for 100 and 120 kv (Fig. 3a). The difference between MS MDCT and MS ref decreased at decreasing tube voltage. Also the difference between MS MDCT and MS physical decreased at decreasing tube voltage. As a function of slice thickness, the difference between AS MDCT and AS ref was minimal for slices larger than 2.0 mm (Fig. 3b). The differences between MS MDCT and MS ref showed an optimum for 1.5 mm. The difference between MS MDCT and MS physical decreased at decreasing slice thicknesses. As a function of reconstruction kernel, the difference between AS MDCT and AS ref was minimal for a medium kernel (Fig. 3c). The difference between MS MDCT and MS ref decreased at an increasing sharpness of the reconstruction kernel. The difference between MS MDCT and MS physical decreased at an increasing sharpness of the reconstruction kernel. Combining the observations on calcium scores, visibility curves and the SSD analysis we could define an optimal calcium scoring protocol. For Agatston scores on MDCT comparable to the Agatston scores on EBT, we found on basis of the position of the visibility curves and the SSD analysis, an optimal acquisition parameters at a tube voltage of 100 kv, a slice thickness/increment of 3.0/3.0 mm and a FC12 reconstruction kernel. Additionally, for the mass score on MDCT comparable to the physical mass, a second reconstruction should be made with a thinner slice thickness/increment of 0.5/0.5 mm. Although the SSD analysis showed that slightly better approximation to the physical mass was obtained at a lower tube voltage and a sharper kernel, the increased dose due to a second scan and the higher noise levels argues against the use of these settings. When this protocol with one acquisition and two reconstructions is used, we obtained a calcium score with a good similarity to EBT for risk stratification as well as a calcium score with the best possible similarity to the physical calcium burden (Table 3). Whereas the standard threshold at a tube voltage of 100 kv was 130 HU, for 120 kv the threshold should be lowered to 116 HU and

4 J.M. Groen et al. / European Journal of Radiology 82 (2013) e58 e63 e61 Fig. 2. (a d): Visibility curves on MDCT and EBT of the calcifications (a) at different tube voltages, (b) at different slice thickness, (c) at different reconstruction kernels, (d) at different scoring thresholds. The visibility curves are the 130 HU isocurves of the calcifications in the size-density plane. Calcifications above the visibility curve are detected, below the visibility curve are not. The black visibility curves represent the reference standard. for 135 kv the threshold should be lowered to 109 HU. When these thresholds were used, the same calcifications were detected at tube voltages of 120 kv and 135 kv as at 100 kv with a threshold of 130 HU (Fig. 4). This protocol therefore ensured a stable calcium score independent of scan- and reconstruction parameters. 4. Discussion This study showed that different scanning parameters may have a large influence on the outcome of calcium scoring using MDCT as compared to EBT. Calcium scores were increased at lower tube voltages, smaller slice thicknesses, sharper reconstruction kernels and lower scoring thresholds. However, with the proposed phantom we were able to construct one acquisition protocol and two reconstruction protocols which yielded Agatston and mass calcium scores which were comparable to the EBT based Agatston scores and physical mass, respectively. It is well known that MDCT shows an underestimation of calcium scores as compared to EBT [17,18]. In our study this was clearly shown by the lower number of calcifications detected by MDCT when using a standard calcium scoring protocol as suggested by the manufacturer (120 kv, 3.0/3.0 mm, medium kernel). This effect is due to the fact that the peak HU-values of the calcifications as measured on MDCT are lower than on EBT. As a consequence, the measured HU-values decrease below the standard scoring threshold of 130 HU, and thereby decreasing the overall score. The main difference between EBT and MDCT results from the fact that on EBT relatively larger calcifications with a lower density are detected, whereas on MDCT these calcifications could not be detected. A possible explanation for this discrepancy is the difference in filtering and reconstruction techniques used in both modalities. Table 3 Physical mass and reference standard EBT measures of total Agatston score (AS) and mass score (MS) of the phantom using the protocol as suggested by the manufacturer and the optimal MDCT protocols with calcium scores as close to the EBT values as possible (EBT based risk stratification) and to the physical mass (physical mass). Physical mass (mg) EBT (reference standard scores) MDCT (manufacturer) 120 kv/fc12/3.0 mm MDCT (EBT-based risk stratification) 100 kv/fc12/3.0 mm MDCT (physical mass) 100 kv/fc12/0.5 mm AS 92.1 ± ± ± ± 14.5 MS ± ± ± ± 2.41

5 e62 J.M. Groen et al. / European Journal of Radiology 82 (2013) e58 e63 Fig. 4. Adjustment of the scoring threshold for different kv settings as proposed in the scanning protocol. (a) Visibility curves with a slice thickness of 3.0 mm, the different kv/threshold combinations showed similar results. (b) Visibility curves with a slice thickness of 0.5 mm, again different kv/threshold combinations showed similar results. Fig. 3. (a c): SSD for AS MDCT, MS MDCT compared to AS ref, MS ref and MS physical as a function of tube voltage (a), slice thickness and increment (b) and reconstruction kernel (c). On the left axis the SSD for EBT is shown, on the right axis the SSD for the physical mass. The SSD for MS MDCT MS physical was multiplied by 10 for better visibility. In addition to a difference between EBT and MDCT, a difference between the physical mass and the mass score was observed. The mass score is a multiplication of three factors; the apparent volume, the apparent density (mean HU) and a calibration factor. When the partial volume effect is reduced with thinner slices and/or a sharper reconstruction kernel, the apparent volume is reduced as well, whereas the density is increased. Due to the multiplication, the error between physical mass and mass score therefore remains present for thinner slices. Although calibration will make the mass score less dependent on the tube voltage used, it is still a method, which uses a threshold and is thus dependent on HU values. Since tube voltage influences HU values, it will influence the mass score. To fully resolve this problem a non-rigid threshold should be used based upon the scan or calcifications itself [30,31]. This optimal protocol ensures that when Agatston scores are obtained on an MDCT, the scores are comparable to the EBT based Agatston scores. In addition, with the additional reconstruction, the measured MS resembles closely the physical calcium burden. Moreover, with use of the phantom and the proposed scoring procedures outlined in this study we expect that other MDCT scanners can be calibrated as well so that validated Agatston and physical mass scores can be obtained with those scanners as well. A precise match between measured mass and physical mass remains difficult, due to the partial volume effect which has a large influence on both the apparent volume as the apparent mass, as stated in a previous paragraph. This study was limited by the fact that only one high-end CT scanner was used. Other manufacturers offer cardiac imaging as well. However, at the time of study we only had access to this scanner. But, the results and general findings, such as increased scores for thinner slices or decreased scores for a lower tube voltage, will be applicable to other MDCT scanners as well. The precise results, such as the suggested threshold for other tube voltages, will probably differ for other systems. A second limitation is the absence of an anatomic background in the phantom. The phantom consists out of well-defined, uniform calcifications in a background of 0 HU, whereas in the in vivo situation, calcifications of various

6 J.M. Groen et al. / European Journal of Radiology 82 (2013) e58 e63 e63 shapes and density composition are observed in an anatomical background. A third limitation is the absence of motion. One of the major challenges in cardiac scanning is cardiac motion. Cardiac motion is the cause of numerous artefacts and distortions in image sets and also calcium scores are influenced [18,32,33]. Our phantom, however, was stationary, so that the influence of motion was ignored. However, in order to fully understand the influence of all the acquisition, reconstruction and scoring parameters involved in the calcium score protocol on MDCT we decided to study the static situation as an initial evaluation. In future studies the influence of motion in 2D or even 3D should be elucidated. In conclusion, this study showed the influence of different acquisition, reconstruction and scoring parameters in MDCT on the calcium score. This knowledge was used to adjust the standard MDCT calcium score protocol to a new MDCT calcium score protocol. With the new phantom one acquisition protocol and two reconstruction protocols can be defined which produces Agatston scores comparable to EBT values and to the physical mass. Acknowledgements The authors would like to thank Jeroen Tijhaar and Tina Bock-Petersen for performing the numerous phantom scans. Furthermore, the authors would like to thank the three EBT clinics for scanning our phantom. References [1] O Rourke RA, Brundage BH, Froelicher VF, et al. American College of Cardiology/American Heart Association Expert Consensus document on electron-beam computed tomography for the diagnosis and prognosis of coronary artery disease. Circulation 2000;102: [2] Greenland P, Alpert JS, Beller GA, et al ACCF/AHA guideline for assessment of cardiovascular risk in asymptomatic adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Journal of the American College of Cardiology 2010;56:e [3] Rumberger JA, Brundage BH, Rader DJ, et al. Electron beam computed tomographic coronary calcium scanning: a review and guidelines for use in asymptomatic persons. Mayo Clinic Proceedings 1999;74: [4] Kim YJ, Hur J, Lee HJ, et al. Meaning of zero coronary calcium score in symptomatic patients referred for coronary computed tomographic angiography. European Heart Journal Cardiovascular Imaging 2012;13: [5] Chen CK, Kuo YS, Liu CA, et al. Frequency and risk factors associated with atherosclerotic plaques in patients with a zero coronary artery calcium score. Journal of the Chinese Medical Association 2012;75:10 5. [6] Yoon YE, Chang SA, Choi SI, et al. The absence of coronary artery calcification does not rule out the presence of significant coronary artery disease in Asian patients with acute chest pain. International Journal of Cardiovascular Imaging 2012;28: [7] Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery calcium using ultrafast computed tomography. Journal of the American College of Cardiology 1990;15: [8] Detrano R, Tang W, Kang X, et al. Accurate coronary calcium phosphate mass measurements from electron beam computed tomograms. American Journal of Cardiac Imaging 1995;9: [9] Arad Y, Spadaro LA, Goodman K, et al. Prediction of coronary events with electron beam computed tomography. Journal of the American College of Cardiology 2000;36: [10] Budoff MJ, Georgiou D, Brody A, et al. Ultrafast computed tomography as a diagnostic modality in the detection of coronary artery disease: a multicenter study. Circulation 1996;93: [11] Rumberger JA, Sheedy PF, Breen JF, et al. Electron beam computed tomography and coronary artery disease: scanning for coronary artery calcification. Mayo Clinic Proceedings 1996;71: [12] Rumberger JA, Simons DB, Fitzpatrick LA, et al. Coronary artery calcium area by electron-beam computed tomography and coronary atherosclerotic plaque area. A histopathologic correlative study. Circulation 1995;92: [13] Schmermund A, Baumgart D, Görge G, et al. Coronary artery calcium in acute coronary syndromes: a comparative study of electron-beam computed tomography, coronary angiography, and intracoronary ultrasound in survivors of acute myocardial infarction and unstable angina. Circulation 1997;96: [14] Wong ND, Hsu JC, Detrano RC, et al. Coronary artery calcium evaluation by electron beam computed tomography and its relation to new cardiovascular events. American Journal of Cardiology 2000;86: [15] McCollough CH, Ulzheimer S, Halliburton SS, et al. Coronary artery calcium: a multi-institutional, multimanufacturer international standard for quantification at cardiac CT. Radiology 2007;243: [16] Budoff MJ, McClelland RL, Ching H, et al. Reproducibility of coronary artery calcified plaque with cardiac 64-MDCT: the multi-ethnic study of atherosclerosis. American Journal of Roentgenology 2009;192: [17] Greuter MJW, Dijkstra H, Groen JM, et al. 64 Slice MDCT generally underestimates coronary calcium scores as compared to EBT: a phantom study. Medical Physics 2007;34: [18] Groen JM, Greuter MJW, Vliegenthart R, et al. Calcium scoring using 64-slice MDCT, dual source CT and EBT: a comparative phantom study. International Journal of Cardiovascular Imaging 2008;24: [19] Oudkerk M, Stillman AE, Halliburton SS, et al. Coronary artery calcium screening: current status and recommendations from the European Society of Cardiac Radiology and North American Society for Cardiovascular Imaging. International Journal of Cardiovascular Imaging 2008;24: [20] Rumberger JA. Clinical use of coronary calcium scanning with computed tomography. Cardiology Clinics 2003;21: [21] Horiguchi J, Shen Y, Akiyama Y, et al. Electron beam CT versus 16-slice spiral CT: how accurately can we measure coronary artery calcium volume? European Radiology 2006;16: [22] Detrano RC, Anderson M, Nelson J, et al. Coronary calcium measurements: effect of CT scanner type and calcium measure on rescan reproducibility MESA study. Radiology 2005;236: [23] Stanford W, Thompson BH, Burns TL, et al. Coronary artery calcium quantification at multi-detector row helical CT versus electron-beam CT. Radiology 2004;230: [24] Daniell AL, Wong ND, Friedman JD, et al. Concordance of coronary artery calcium estimates between MDCT and electron beam tomography. American Journal of Roentgenology 2005;185: [25] Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium measurement agreement of multirow detector and electron beam CT. American Journal of Roentgenology 2001;176: [26] Greuter MJW, Tukker WGJ, Willems TP. Effect of calcification on diagnostic image quality of coronary angiography on dual source computed tomography. In: European Congress of Radiology [27] Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease: improved reproducibility of calcium scoring with an electron-beam CT volumetric method. Radiology 1998;208: [28] Thijssen MAO, Thijssen HOM, Merx JL, et al. A definition of image quality: the image quality figure. BIR Report 1989;20: [29] Geijer H, Beckman KW, Andersson T, et al. Image quality vs. radiation dose for a flat-panel amorphous silicon detector: a phantom study. European Radiology 2001;11: [30] Arnold BA, Xiang P, Mao SS, et al. Peak SNR in automated coronary calcium scoring: selecting CT scan parameters and statistically defined scoring thresholds. Medical Physics 2010;37: [31] Groen JM, Dijkstra H, Greuter MJW, et al. Threshold adjusted calcium scoring using CT is less susceptible to cardiac motion and more accurate. Medical Physics 2009;36: [32] Greuter MJW, Groen JM, Nicolai LJ, et al. A model for quantitative correction of coronary calcium scores on multidetector, dual source, and electron beam computed tomography for influences of linear motion, calcification density, and temporal resolution: a cardiac phantom study. Medical Physics 2009;36: [33] Groen JM, Greuter MJ, Schmidt B, et al. The influence of heart rate, slice thickness, and calcification density on calcium scores using 64-slice multidetector computed tomography: a systematic phantom study. Investigative Radiology 2007;42: