Coronary Calcium Screening Using Low-Dose Lung Cancer Screening: Effectiveness of MDCT with Retrospective Reconstruction

Transcription

1 Cardiac Imaging Original Research Kim et al. Coronary Calcium Screening Using Lung Cancer Screening Cardiac Imaging Original Research Sung Mok Kim 1 Myung Jin Chung 1 Kyung Soo Lee 1 Yeon Hyun Choe 1 Chin A. Yi 1 Bong-Keun Choe 2 Kim SM, Chung MJ, Lee KS, Choe YH, Yi CA, Choe BK Keywords: calcium-scoring CT, coronary artery calcium, lung cancer screening, MDCT, radiation dose, screening DOI: /AJR Received August 6, 2007; accepted after revision October 19, Department of Radiology and Center for Imaging Science, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-ku, Seoul , Korea. Address correspondence to M. J. Chung. 2 Department of Preventive Medicine, School of Medicine, Kyunghee University, Seoul, Korea. AJR 2008; 190: X/08/ American Roentgen Ray Society Coronary Calcium Screening Using Low-Dose Lung Cancer Screening: Effectiveness of MDCT with Retrospective Reconstruction OBJECTIVE. The purpose of our study was to show the usefulness of nongated low-dose chest CT for coronary screening by comparing the results of coronary artery calcium measurement with that of dedicated calcium-scoring CT. MATERIALS AND METHODS. One hundred twenty-eight consecutive participants (all men; mean age, 52 ± 7 years) underwent low-dose chest CT and calcium-scoring CT with prospective ECG gating using 40-MDCT. Low-dose chest CT volume data were reconstructed as 25-cm field of view and three slice thicknesses: 1, 2.5, and 5 mm. For each examination, the lesion area, Agatston calcium score, and calcium mass were measured at 90- and 130-H thresholds. All measurements (130-H threshold) from the calcium-scoring CT were used as reference standards. Spearman s correlation test was used to compare the results. RESULTS. Among the low-dose chest CT examinations, sensitivity was best determined with a 1-mm slice thickness at 130 H and 2.5-mm slice thickness at 90 H. Specificity was best determined with a 5-mm slice thickness at 130 H. Accuracy (90%) was best determined with a 2.5-mm slice thickness at 130 H. Of all protocols, calcium area, score, and mass from a 2.5-mm slice thickness at 130 H correlated best with the reference results (r = 0.89 for all three criteria). CONCLUSION. Using a low radiation dose and nongated MDCT, we can detect coronary artery calcium and obtain results comparable to those obtained with dedicated calciumscoring CT that uses a higher dose and ECG gating. C alcifications of the coronary artery wall are regarded as a recognized marker of coronary atherosclerosis [1]. Both electron beam CT and conventional single-detector and MDCT have been used for the detection and quantification of coronary calcifications [2]. The use of MDCT has several advantages over the use of electron beam CT, including a higher signal-to-noise ratio due to a larger photon count [3, 4], thinner section thickness, and simultaneous acquisition of multiple sections, which may reduce the number of misregistration artifacts. MDCT has also been shown to be more reproducible than electron beam CT [5]. Although the imaging time (temporal resolution) of MDCT is still longer than that of electron beam CT, partial view acquisition and ECG-gated reconstruction allow fast image acquisition in the diastolic phase of the cardiac cycle [6]. However, the high image quality of MDCT is achieved at the expense of a substantial amount of ionizing radiation. Exposure to radiation is a concern, especially for individuals in the screening population because many participants undergo repeated examinations. Although the risk from CT radiation may be relatively low per patient, the added risk to the population may be substantial, with unknown long-term effects. Thus, it is essential to reduce the CT radiation dose to the minimum level that will allow an adequate quantification of coronary artery calcium. The purpose of this study was to compare prospectively the results of coronary artery calcium measurement obtained from a retrospective reconstruction of low-dose chest CT for lung cancer screening and from highdose dedicated calcium-scoring CT. Materials and Methods Our institutional review board approved this retrospective study and waived the requirement for informed consent. Participants Between March and April 2007, 128 consecutive participants (all men; mean age, 52 ± 7 years; range, years) were included. They were AJR:190, April

2 Kim et al. smokers who came to the health care center of our institute to undergo low-dose chest CT for lung cancer screening and calcium-scoring CT for coronary risk factors. Participants with a history of cardiac surgery, coronary stent insertion, or other cardiac intervention were excluded from the study. We calculated body mass index (BMI) from the individual s height and weight. The overall mean BMI was 24.2 kg/m 2 (range, kg/m 2 ). Calcium-Scoring CT Protocol CT of the heart was performed using 40-MDCT (Brilliance 40, Philips Medical Systems). This exam ination was performed with 120 kvp, 55 mas, and prospective ECG gating. The imaging parameters included a 600-millisecond scanning time and 2.5-mm single slice thickness, with a total of 60 slices obtained during a single breathhold. The single-section mode of imaging was used, with imaging ECG-triggered to 75% of the R-R interval (for a diastolic blood pressure > 80 mm Hg, 45% of the R-R interval). Imaging was reconstructed into a matrix with a 25-cm field of view. All calcification related to a coronary artery equal to or greater than the minimum attenuation of 130 H was considered potential coronary calcium. Low-Dose Chest CT Protocol CT of the lung was also performed using 40-MDCT (Brilliance 40). Participants remained stationary on the table during both examinations without changing position. This examination was performed with 120 kvp, 30 mas, 0.5-second gantry rotation, and a table pitch of 1.3. The lowdose chest CT volume data were retrospectively reconstructed into a matrix with a 25-cm field of view (carina heart base), and three slice thicknesses: 1, 2.5, and 5 mm (Fig. 1). The thresholds for identifying calcification in the coronary arter ies were 90 H (modified Agatston score) that has been used previously with nongated helical CT [3] and 130 H (conventional Agatston score), along with the requirement for two contiguous pixels. Therefore, we acquired data with six protocols using low-dose chest CT as follows (threshold, slice thickness): 90 H, 1 mm; 90 H, 2.5 mm; 90 H, 5 mm; 130 H, 1 mm; 130 H, 2.5 mm; and 130 H, 5 mm. Data Measurement and Analysis For each examination, the lesion area, Agatston calcium score, and calcium mass were measured at two thresholds using a commercial CAD workstation (Extended Brilliance Workspace, V ; Philips Medical Systems) with dedicated cardiac analysis software. Calcifications were A identified according to the vessel (left main, left anterior descending, circumflex, right, and posterior descending coronary arteries). Using the Agatston algorithm, an attenuation factor was determined on the basis of the maximal CT attenuation of the lesion as follows: factor 1 = H (modified Agatston score, ), factor 2 = H, factor 3 = H, and factor 4 = 400 H or greater. The calcium score was calculated by multiplying the area of each calcified plaque by the corresponding attenuation factor. The calcium score for each coronary vessel was also calculated [7]. The mass score is based on calculating the mean attenuation value of a calcification and multiplying the result by the volume of the calcification and a calibration factor to obtain the calcification mass [8]. All measurements (130-H thresholds) from calcium-scoring CT were used as reference standards. Also, on the basis of the score from the best low-dose chest CT protocol, we classified participants using the risk stratification proposed by Rumberger et al. [9]: very low risk (Agatston score = 0), low risk (> 0 10), moderate risk (> ), moderately high risk (> ), and high risk (> 400). Then we compared the risk stratification of the participant with that of calcium-scoring CT. Statistical Analysis The lesion area, the calcium score, and the calcium mass for each coronary vessel obtained in the two examinations were compared using a two-tailed paired sample test. Correlation of the calcium scores between the two examinations was measured using the Spearman s correlation test. Multiple regression analysis was used to show the partial effect of BMI and the calcium score from the calcium-scoring CT on measured values in each low-dose chest CT protocol. A p value of 0.05 or less was considered to indicate a statistically significant difference. C D Fig. 1 Comparison of image quality between dedicated coronary calcium-scoring CT and retrospective reconstruction of nongated low-dose chest CT. A, Axial image of coronary calcium-scoring CT in 67-year-old man shows elongated calcified plaque (arrow) at left anterior descending coronary artery. B, Same plaque is shown on same plane of reconstructed axial image of low-dose CT with 5-mm section thickness but looks smaller than in reference image (A). C, Reconstructed axial image of low-dose CT with 2.5-mm section thickness shows coronary plaque having similar size as in reference image (A). D, Coronary plaque on axial image of low-dose CT with 1-mm section thickness is similar to that on A and C. Note higher background noise on this image, which compromises plaque conspicuity. B 918 AJR:190, April 2008

3 Coronary Calcium Screening Using Lung Cancer Screening TABLE 1: Comparison of Performance Measures for Coronary Calcium Detection Protocol (Attenuation and Slice Thickness) Sensitivity (%) Specificity (%) PPV (%) NPV (%) Accuracy (%) 130 H 1 mm 98 (53) 53 (39) mm 91 (49) 89 (66) mm 65 (35) 99 (73) H 1 mm 89 (48) 16 (12) mm 98 (53) 50 (37) mm 81 (44) 85 (63) Note Values of calcium-scoring CT were used as reference standards: sensitivity, 100% in 54 patients; specificity, 100% in 74 patients. Numbers in parentheses are numbers of patients. PPV = positive predictive value, NPV = negative predictive value. Results Fifty-four of the 128 participants had coronary artery calcium deposits seen on the calcium-scoring CT examination. Among the low-dose chest CT examinations, the 130-H, 5-mm slice thickness protocol showed the fewest detected calcifications, 35; the 130-H, 1-mm slice thickness and the 90-H, 2.5-mm slice thickness protocols showed the most calcifications, 53. Seventy-four participants had no coronary artery calcium deposits as seen on the calcium-scoring CT examination. Among the low-dose chest CT examinations, the 90-H, 1-mm slice thickness protocol showed the lowest negative number, (n = 12), and the 130-H, 5-mm slice thickness protocol showed the highest negative number, (n = 73). The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for coronary calcium detection are summarized in Table 1. Of all protocols, sensitivity was best determined with a 1-mm slice thickness at 130 H and a 2.5-mm slice thickness at 90 H. Specificity was best determined with a 5-mm slice thickness at 130 H. Accuracy was best determined with a 2.5-mm slice thickness at 130 H. The correlation coefficients of coronary artery calcium measurements between low-dose chest CT (six protocols) and calcium-scoring CT (as reference standards) are shown in Table 2. Of all protocols, calcium area, score, and mass from a 2.5-mm slice thickness at 130 H correlated best with the reference results (r = 0.89 for all three criteria). When the results of each coronary branch were evaluated at the protocol of a 2.5-mm slice thickness at 130 H, the correlation coeffi cients were poor (r = 0.215, p = 0.015) for the posterior descending artery, but the total number of samples was insufficient (the number of vessels with a positive calcium TABLE 2: Correlation Coefficients of Low-Dose CT Compared with Dedicated Calcium-Scoring CT for Measurement of Coronary Artery Calcium Protocol (Attenuation and Slice Thickness) Area Score Mass 130 H 1 mm mm mm H 1 mm mm mm Note Data are correlation coefficients. All values had p < (Spearman s correlation test). score on calcium-scoring CT was only four). The correlation coefficients were sufficiently high in all other coronary branches (r > 0.654, p < 0.001) (Table 3). On the basis of the data from the protocol of a 2.5-mm slice thickness at 130 H, we compared the risk stratification of a patient with that of calcium-scoring CT (Table 4). Among 54 participants who had coronary artery calcium deposits as seen on the calcium-scoring CT examination, five were not detected on the low-dose chest CT protocol. All five cases are low-risk category in the calcium-scoring CT examination and thus they were moved into the very-low-risk category in the low-dose chest CT protocol. In contrast, among 74 participants who had no coronary artery calcium deposits as seen on the calcium-scoring CT examination, eight were detected on the low-dose chest CT protocol. All eight participants were moved into the low-risk category in the low-dose chest CT protocol. Among all participants, no individual differed by more than the adjacent risk category. By the result of multiple regression analysis, BMI has only a very weak negative effect with no significant effect on the measured values of each low-dose chest CT protocol (p > 0.1) (Table 5). Discussion In general, high calcium scores more likely predict hemodynamically relevant stenoses than low scores [1, 10]. Calcium scoring is convenient to perform compared with conventional or CT coronary angiography because calcium scoring does not require the use of a contrast agent and an invasive technique. However, calcium scoring is not the best method for the detection of coronary artery disease. Herzog et al. [11] investigated the accuracy of calcium scoring and MDCT coronary angiography in the assessment of coronary atherosclerosis. All findings were compared with conventional coronary angiography. In their study, calcium scoring as a single method showed the highest sensitivity for the detection of coronary atherosclerosis, but at the expense of low specificity. Electron beam CT and conventional singledetector and MDCT have been used for the detection and quantification of coronary calcification as a sign of coronary atherosclerosis [2]. The standard method for quantification of coronary artery calcium is electron beam CT. After the introduction of MDCT and rapid technical advances, the number of MDCT examinations of the coronary vasculature has AJR:190, April

4 Kim et al. TABLE 3: Correlations of Each Coronary Artery Branch Between Calcium- Scoring CT and Low-Dose CT Using the 130-H, 2.5-mm Slice Thickness Protocol Coronary Artery Branch Area Score Mass Left main a a a Left anterior descending a a a Left circumflex a a a Right a a a Posterior descending b b a p < (Spearman s correlation test). b p < 0.05 (Spearman s correlation test). TABLE 4: Risk Stratification of Patients by Calcium-Scoring CT and by Low- Dose CT Using 130-H, 2.5-mm Slice Thickness Protocol Low-Dose CT Very Low Risk (0) been rapidly increasing. Higher radiation doses are delivered with cardiac MDCT as compared with the doses delivered using electron beam CT. The effective radiation doses for calcium scoring with electron beam CT were 1.0 and 1.3 msv for men and women, respectively. The effective radiation doses for calcium scoring with MDCT were in the range of msv for men and msv for women [12]. Because the detection and quantification of coronary artery calcification is considered a screening tool in potentially healthy people, radiation exposure should be kept as low as possible. Much recent work has focused on the value of low-dose chest CT as a screening tool for detecting early asymptomatic lung cancers. The radiation dose for low-dose chest CT is msv, compared with 7.0 msv for conventional CT [13]. In low-dose chest CT screening for lung cancer, Brenner [14] reported an estimated upper limit of a 5.5% increase in the lung cancer risk attributable to annual CT-related radiation exposure. The risk of radiation-induced lung cancer associated with repeated CT screening for lung cancer or coronary artery disease may not be negligible. Therefore, one should choose a CT examination with a lower radiation dose Low Risk (> 0 10) Very low risk (0) 66 5 Calcium-Scoring CT Moderate Risk (> ) Low risk (> 0 10) Moderately High Risk (> ) Moderate risk (> ) Moderately high risk (> ) 1 5 High risk (> 400) 3 Total no. of patients (128) Note Numbers in percentages are Agatston scores for each risk category. High Risk (> 400) for a screening study and minimize the number of CT examinations. The risk group for lung cancer can overlap the risk group for coronary artery disease because smoking is a heavy risk factor for both. For these risk groups, a double CT screening examination is somewhat wasteful in terms of expense and time. A double CT screening examination also increases overall radiation exposure. A previous study has determined the frequency of calcification of the thoracic aorta and its relationship to risk factors and coronary artery disease using a CT examination for lung cancer and tuberculosis [15]. We also used low-dose chest CT examination for lung cancer screening to reduce the radiation dose. Coronary artery calcium detection and measurement from retrospective reconstruction of images produced by low-dose chest CT gave results that were well correlated with those obtained from dedicated calcium-scoring CT, which means that the risk groups for lung cancer and coronary artery disease can undergo just a single CT examination rather than two consecutive CT examinations. The main finding of our study is that application of a low milliampere-second setting of 30 mas for coronary artery calcium detection and measurement from a retrospective reconstruction from low-dose chest CT images produces results that are well correlated with those obtained using dedicated calcium-scoring CT at 55 mas. An additional finding is that the ECG-gating did not significantly affect the correlation of the results between prospective ECG-gated calciumscoring CT and nongated low-dose chest CT with retrospective reconstruction. Finally, the effect of BMI is negligible except in patients who are extremely overweight. The BMI of our participants ranged from 17.1 to 30.9 kg/m 2. The results of our study strengthen those of previous studies in which a low milliampere-second setting was used [16 22]. Shemesh et al. [22] assessed the coronary artery calcium measurement with two milliampere-second levels (55 and 165 mas) using prospective ECG-gated MDCT. Those investigators calculated the calcium mass and calcium score in 51 asymptomatic participants by performing two consecutive CT examinations, the first with a setting of 165 mas and the second with a setting of 55 mas. The total calcium score between the high- and low-dose scans was well correlated with respect to the Agatston method and calcium mass (r = 0.97, p < and r = 0.99, p < 0.001, respectively). A strong correlation was also found for each vessel. The overall correlation coefficients determined in our study were lower than the values reported by Shemesh et al. The main cause of this difference is the use of a different statistical method. We did not use a typical parametric method (Pearson s correlation test) to determine a correlation coefficient but used a nonparametric method (Spearman s correlation test) because the data of our study had too many null cases and thus correlation coefficients were not different between six different reconstruction protocols (r > 0.998, Pearson s correlation test). In our study, the optimal protocol has a sensitivity of 91%, which means the lowdose chest CT protocol did not detect calcium in five participants. But all five participants were in the low-risk category in the calcium-scoring CT examination. Fallavollita et al. [23] evaluated electron beam CT calcium scoring in 98 men and women who had no significant obstructive disease detected at coronary angiography and found that 87% with angiographically smooth coronary arteries had a calcium score of 5 or less. In another study, although a negative or extremely 920 AJR:190, April 2008

5 Coronary Calcium Screening Using Lung Cancer Screening TABLE 5: Effect of Body Mass Index (BWI) and Calcium-Scoring CT Score on the Values Measured with Low-Dose CT H Protocol Correlations b Section Thickness (mm) Model Beta a p Zero-Order Partial Note BMI = body mass index, CT = calcium-scoring CT. a Calculated using multiple regression analysis. b Calculated using Pearson s partial correlation test. low calcium score ( 10) could not totally exclude the presence of coronary atherosclerosis, it was consonant with the absence of a fixed (significant) coronary obstructive lesion, regardless of age and sex [9]. Thus, no further specific cardiac workup is recommended for this group. Therefore, the lowdose chest CT protocol can be used as the basis for the screening study. When we tried to use the parametric method forcibly, the correlation coefficients for the protocol of 2.5-mm slice thickness at 130 H were 0.99 each for area, score, and mass. Of all the protocols used for the retrospective reconstruction of low-dose chest CT images, the protocol that provided the best correlation was 130 H and 2.5-mm slice thickness (r = 0.89, p < 0.001). In this protocol, results from the posterior descending artery showed a lower correlation coefficient than the other coronary artery branches. This was statistically due to the small number of samples and physically due to the posterior descending artery being smaller than the other branches and its location being a site associated with higher motion artifacts, as seen in an nongated scan. Cardiac motion artifacts are known to artificially raise the calcium score [24]. In conclusion, our study showed that nongated helical CT with 30 mas can be an alternative setting that produces results for calcium measurement that are comparable to BMI CT BMI CT BMI CT BMI CT BMI CT BMI CT those obtained with a prospectively ECGgated axial CT examination at 55 mas. Thus, for one who undergoes screening for lung cancer and coronary heart disease simultaneously, the radiation exposure can be reduced for coronary artery calcium detection and measurement by the retrospective use of low-dose chest CT data instead of dedicated calcium-scoring CT. However, the appropriate use for CT with its accompanying lowdose radiation exposure for calcium scoring, particularly for large patients, should be further studied. References 1. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods and clinical implications a statement for health professionals from the American Heart Association. Circulation 1996; 94: Becker C, Jakobs T, Aydemir S, et al. Helical and single-slice conventional versus electron beam CT for quantification of coronary artery calcification. AJR 2000; 174: Shemesh J, Apter S, Rozenman J, et al. Calcification of coronary arteries: detection and quantification with double-helix CT. Radiology 1995; 197: Broderick LS, Shemesh J, Wilensky RL, et al. Measurement of coronary artery calcium with dual-slice helical CT compared with coronary angiography: evaluation of CT scoring methods, interobserver variations, and reproducibility. AJR 1996; 167: Ohnesorge B, Flohr T, Fischbach R, et al. Reproducibility of coronary calcium quantification in repeat examinations with retrospectively ECGgated multisection spiral CT. Eur Radiol 2002; 12: Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of electrocardiographically gated multisection spiral CT: initial experience. Radiology 2000; 217: Ulzheimer S, Kalender WA. Assessment of calcium scoring performance in cardiac computed tomography. Eur Radiol 2003; 13: Hong C, Bae KT, Pilgram TK. Coronary artery calcium: accuracy and reproducibility of measurements with multi-detector row CT: assessment of effects of different thresholds and quantification methods. Radiology 2003;227: Rumberger JA, Brundage BH, Rader DJ, Kondos G. Electron beam computed tomographic coronary calcium scanning: a review and guidelines for use in asymptomatic persons. Mayo Clin Proc 1999; 74: 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: Herzog C, Britten M, Balzer JO, et al. Multidetector-row cardiac CT: diagnostic value of calcium scoring and CT coronary angiography in patients with symptomatic, but atypical, chest pain. Eur Radiol 2004; 14: Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure during cardiac CT: effective doses at multi-detector row CT and electron-beam CT. Radiology 2003; 226: Diederich S, Wormanns D, Semik M, et al. Screening for early lung cancer with low-dose spiral CT: prevalence in 817 asymptomatic smokers. Radiology 2003; 222: Brenner DJ. Radiation risks potentially associated with low-dose CT screening of adult smokers for lung cancer. Radiology 2004; 231: Itani Y, Watanabe S, Masuda Y. Aortic calcification detected in a mass chest screening program using a mobile helical computed tomography unit: relationship to risk factors and coronary artery disease. Circ J 2004; 68: Hong C, Bae KT, Pilgram TK, Suh J, Bradley D. Coronary artery calcium measurement with multidetector row CT: in vitro assessment of effect of radiation dose. Radiology 2002; 225: Mahnken AH, Wildberger JE, Simon J, et al. Detection of coronary calcifications: feasibility of dose reduction with a body weight adapted examination AJR:190, April

6 Kim et al. protocol. AJR 2003; 181: Takahashi N, Bae KT. Quantification of coronary artery calcium with multi-detector row CT: assessing interscan variability with different tube currents pilot study. Radiology 2003; 228: Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of the heart with retrospective ECG gating: reduction of radiation exposure by ECG-controlled tube current modulation. Eur Radiol 2002; 12: FOR YOUR INFORMATION 20. Ohnesorge BM, Flohr TG, Becker CD, Kopp AF, Knez A, Reiser MF. Dose evaluation and dose reduction strategies for ECG-gated multislice spiral CT of the heart. (abstr) Radiology 2000; 217(P): Jakobs TF, Wintersperger BJ, Herzog P, et al. Ultra-low-dose coronary calcium screening using multislice CT with retrospective ECG gating. Eur Radiol 2003; 13: Shemesh J, Evron R, Koren-Morag N, et al. Coronary artery calcium measurement with multi- detector row CT and low radiation dose: comparison between 55 and 165 mas. Radiology 2005; 236: Fallavollita JA, Kumar K, Brody AS, Bunnell IL, Canty JM Jr. Detection of coronary artery calcium to differentiate patients with early coronary atherosclerosis from luminally normal arteries. Am J Cardiol 1996; 78: Brown SJ, Hayball MP, Coulden RA. Impact of motion artefact on the measurement of coronary calcium score. Br J Radiol 2000; 73: Mark your calendar for the following ARRS annual meetings: April 26 May 1, 2009 John B. Hynes Veterans Memorial Convention Center, Boston, MA May 2 7, 2010 Grand Hyatt San Diego, San Diego, CA May 1 6, 2011 Hyatt Regency Chicago, Chicago, IL April 29 May 4, 2012 Vancouver Convention Center, Vancouver, BC, Canada 922 AJR:190, April 2008