Analysis of Heart Rate and its Variation Affecting Image Quality and Optimized Reconstruction Window in Retrospective ECG-gated Coronary Angiography Using Multi-detector Row CT Sang Ho Lee, Byoung Wook Choi, Hee-Joung Kim*, Member, IEEE, Haijo Jung, Hye-Kyung Son, Won-Suk Kang, Sun Kook Yoo, Kyu Ok Choe, Hyung Sik Yoo Abstract--It is clinically important to study the effect of heart rate and its variation on image quality and selection of optimized window in coronary angiography using multi-detector row CT (MDCT). We performed contrast-enhanced coronary angiography using MDCT in 83 patients. Sixty cases with available information of heart rate were enrolled in this study. We systemically analyzed the effect of heart rate and its variation. Two radiologists rated image quality as follows: 4, excellent; 3, good; 2, fair; 1, bad. Cardiac cycle windows at and % were routinely selected for image reconstruction. Both of % and % reconstructed images were available only to fifty-seven cases. Optimized window was rated as 1 when % reconstruction was better quality than %, as 2 when % reconstruction was the same as %, and as 3 when % reconstruction was better than %. Regression analysis was performed. The range of BPM variation and image quality were well correlated (r=-0.9, p=0.000). However the BPM variation correlation with optimized window percentage was not statistically significant (r=0.005, p=0.969). On the contrary, median BPM value and optimized window score were relatively well correlated (r=-0.239, p=0.086), indicating that % Manuscript received November, 2002. This work was supported in part by a grant from the Brain Korea (BK) Project for Medical Sciences, and in part by the Basic Research Program of the Korea Science and Engineering Foundation (KOSEF) under Grant R01-2002-000-00205-0 (2002), Yonsei University. Asterisk indicates corresponding author. Hee-Joung Kim* is with the BK Project for Medical Sciences, Research Institute of Radiological Science, Department of Radiology, Yonsei University College of Medicine, #4 Shinchon-Dong, Seodaemoon-Gu, Seoul, 120-752, Korea (phone: 822-361-5753, fax: 822-393-3035, e-mail: hjkim@yumc. yonsei.ac.kr). Sang-Ho Lee, Hye-Kyung Son and Won-Suk Kang are with the BK Project for Medical Sciences, Research Institute of Radiological Science, Yonsei University College of Medicine, #4 Shinchon-Dong, Seodaemoon-Gu, Seoul, 120-752, Korea (e-mail: shlee1@yumc.yonsei.ac.kr; hkson@yumc.yonsei.ac.kr; wskang@yumc.yonsei.ac.kr). Byoung Wook Choi, Haijo Jung, Kyu Ok Choe and Hyung Sik Yoo are with the Research Institute of Radiological Science, Department of Radiology, Yonsei University College of Medicine, #4 Shinchon-Dong, Seodaemoon-Gu, Seoul, 120-752, Korea (e-mail: bchoi@yumc.yonsei.ac.kr; hjjung1@yumc.yonsei.ac.kr; kochoe@yumc.yonsei.ac.kr; hsyoo@yumc. yonsei.ac.kr). Sun Kook Yoo is with the Department of Medical Engineering, Yonsei University College of Medicine, #4 Shinchon-Dong, Seodaemoon-Gu, Seoul, 120-752, Korea (e-mail: sunkyoo@yumc.yonsei.ac.kr). reconstruction is better in image quality than % reconstruction at higher heart rates. Median BPM value and image quality were not well correlated (r=-0.149, p=0.197). Image quality was more affected by variation of heart rate than by high heart rate. Selection of optimized reconstruction window for good image quality was mainly affected by heart rate and there was a tendency for % phase reconstruction to be better in image quality than % reconstruction at higher heart rates. Index Terms Beats per minute (BPM), coronary angiography, ECG-correlated CT, image quality, multi-detector row CT (MDCT), optimized reconstruction window, variation of heart rate. I. INTRODUCTION -ray coronary angiography is widely utilized as the Xdiagnostic gold standard for coronary artery disease. It is invasive and the catheterization procedure involves discomfort for the patient. Thus, conventional angiography should be undertaken only on strict clinical expressions. A cross-sectional computed tomography (CT) imaging technique has been tried continuously to visualize the coronary arteries in radiology. At an early stage, the use of conventional non-gated helical CT was limited in accurately displaying the coronary arteries because of the small caliber of the coronary vessels, their complicated shape, tortuous course and motion artifacts resulted from cardiac contraction [1]. For artifact-free cardiac imaging, scanning time less than ms has been suggested by Ritchie et al [2], [3]. Electron beam CT (EBCT) [4]-[], with its 100 ms scanning time, synchronized to the cardiac cycle by using prospective electrocardiographic (ECG) triggering, has been proposed as a noninvasive imaging modality for the diagnosis of coronary artery disease by Achenbach et al [4]. Reconstruction of images is possible at only one point of the cardiac cycle determined by prescribed trigger delay time at which minimum cardiac motion is expected. The results can be optimal for only one of the three major coronary arteries because each has a different motion 0-7803-7636-6/03/$17.00 2003 IEEE. 1622
pattern [12]-[14]. Besides, vascular sections are often displayed at an undesirable cutting angle or are affected by partial volume effects due to their rapid movements [15]. The effects of cardiac motion can be decreased by cardiac gating of CT [16], [17]. Gating of the cardiac images to the ECG test can be retrospectively executed with helical CT. Multi-detector row CT (MDCT), with 2 ms image acquisition time, enables a higher signal-to-noise ratio and spatial resolution than EBCT. Moreover, as compared with conventional single-section CT, not only is acquisition time remarkably decreased, but z-plane and in-plane resolution are also improved [1], [18]. Recently, as image quality is highly dependent on the heart rate, it has been advised to evaluate at lower heart rates of less than 65 beats per minutes (BPM) with beta-blocker in order to achieve best image quality [19]. However, the effect of heart rate and its variation on image quality has not received much systematic study. In addition, it has been reported that the image reconstruction window for CT angiography of the coronary arteries should be adapted to each coronary artery [20]. With retrospective ECG gating, multiple selection of image reconstruction window is possible for MDCT coronary angiography. The length of cardiac cycle is continuously changing depending on a patient s condition in a single breath hold time. As the heart rate increases, the difference between image acquisition time and minimum cardiac motion time increases. As this variation increases, unreliable cardiac gating of CT data is caused by cardiac cycles that appear at irregular intervals. Therefore, the purpose of this study was to evaluate the influence of heart rate and its variation on image quality and optimized window for good quality reconstruction in retrospective ECG gated coronary angiography using MDCT. II. MATERIALS AND METHODS A. Materials A total of 83 consecutive patients ( male, 23 female; mean age years ± 8; range, 35-74) underwent contrast-enhanced coronary angiography using MDCT. Sixty cases with available information of heart rate were enrolled in this study. The range of variations of heart rate (mean, 12.8±10.9; range, 2-65) and median BPM value (mean, 67.3±9.4; range, -100) were recorded for each patient. B. MDCT For MDCT, a LightSpeed Plus (GE medical systems, Milwaukee, WI) scanner was used. Generations of MDCT scanners operate at the rotation rate of 0 ms and produce up to four slices of 1.25 mm collimation simultaneously, with a constant time resolution of 2 ms per section []. Multi-sector reconstruction algorithm was not available (i.e., scanning data from only one heart cycle was used to reconstruct an image). C. Scan Protocol After a low-dose pre-contrast helical localization (collimation 2.5 mm, pitch 1.5, 142 kv, ma, rotation time 0 ms), a test bolus of 15 ml of contrast medium was injected through an 18-gauge catheter into an antecubital vein in order to determine the circulation time. For the contrast-enhanced scan (collimation 1.25 mm, 120 kv, 300 ma, rotation time 0 ms), 1 ml of contrast agent (Iopamiro 3 mg/ml, Bracco s.p.a., Milano, Italy) was injected at 4 ml/s during the first half of the total scan time plus delayed period after injection of contrast material and at 2ml/s during the second half. The start of the contrast-enhanced scan was adapted to the calculated circulation time. D. Image Reconstruction To set the best delay suited to obtaining high image quality, sections at every given z position were respectively reconstructed at % and % of the R-R interval, which represent systolic and diastolic phases. This is because the optimal delay is at % of R-R interval for the RCA and % R-R for LAD and LCx [20]. Both of % and % reconstruction images were available only to fifty-seven cases out of sixty patients. E. Evaluation of Image Quality Evaluation of the coronary arteries was performed by reading proximal and mid coronary arteries in both of % and % window reconstruction images. Image quality was rated by two radiologists as excellent when all proximal to mid coronary arteries were assessable, good when two proximal to mid coronary arteries were assessable, fair when all proximal coronary arteries were assessable, and bad when only one or two proximal coronary arteries were assessable. In addition, image quality was compared between % and % reconstructions to analyze correlations of the other factors. Optimized window was rated as 1 when % reconstruction was better quality than %, as 2 when % reconstruction was the same as %, and as 3 when % reconstruction was better than %. F. Statistics Linear regression analysis was performed. A 2-tailed P value <0.05 was considered statistically significant. We analyzed the correlation of BPM variation with optimized window score and with image quality, and of median BPM value with optimized window score and with image quality. All statistical analyses were conducted using the SPSS 9.0 statistics software package (SPSS Inc, Chicago, USA). III. RESULTS The box plots in Fig. 1a and 1b respectively represent the relation between image quality and heart rate and that between image quality and variation of heart rate. Image quality was well correlated inversely with the range of BPM variations (r=-0.9, p=0.000), indicating that image quality is degraded with wider BPM variation. However, the correlation with 0-7803-7636-6/03/$17.00 2003 IEEE. 1623
median BPM value was not good (r=-0.149, p=0.197). The box plots in Fig. 2a and 2b respectively represent the relation between better reconstruction window and median BPM value and the relation between better reconstruction window and BPM variation. Optimized window score was relatively well correlated with median BPM value (r=-0.239, p=0.086), which means that % reconstruction is better in image quality than % reconstruction at higher heart rates. On the contrary, the correlation with the range of BPM variations was not statistically significant (r=0.005, p=0.969). IV. DISCUSSION From the viewpoint of high heart rate, two elements are essential for the reduction of cardiac motion artifacts. Firstly, the reduction of acquisition time in order to provide sufficiently short temporal resolution. Based on a gantry rotation time of 0 ms and on spiral algorithms, temporal resolution of MDCT is limited to 2 ms [17], [22] which is still inferior to that of EBCT. Furthermore, in this study, the temporal resolution could not be improved to as low as 125 ms, being the time required to use image data of two consecutive heart cycles, because a 0 29 100 90 80 48 23 46 30 Median BPM Value BPM Variation 20 10 0-10 17 17 bad fair good excellent bad fair good excellent Image Quality Image Quality Fig. 1. Box plots representing the relation between, image quality and heart rate, and image quality and BPM variation. BPM variation is well correlated inversely with image quality, whereas heart rate is relatively not well correlated with image quality. Image quality is better in lower BPM variation. 0 100 90 14 30 44 24 1 80 7 5 20 34 Median BPM Value BPM Variation 10 0-10 18 26 18 26 % %=% % % %=% % Better Reconstruction Window Better Reconstruction Window Fig. 2. Box plots representing the relation between, better-optimized window and heart rate, and better-optimized window and BPM variation. Heart rate is well correlated inversely with optimized window score, whereas BPM variation is not well correlated with optimized window score. Reconstruction in % phase produces better quality than % at higher heart rates. 0-7803-7636-6/03/$17.00 2003 IEEE. 1624
Fig. 3. Comparison of % and % reconstructed images in a patient with a median heart rate of 54 BPM and a variation of 16 that was rated as bad. The right coronary artery (arrow) is well visualized only in the % image. Fig. 4. Comparison of % and % reconstructed 3-dimensional images in a patient with a median heart rate of 81 BPM. The % reconstruction image shows fewer artifacts than %. multi-sector reconstruction algorithm was not available. However, even though EBCT technique with a reduced temporal resolution of 100 ms has advantages over multi-detector row CT, its spatial resolution is significantly poorer than that of MDCT because an only single slice can be acquired in one cardiac cycle with EBCT [14]. Secondly, the reduction of heart rate. Whereas heart rate acceleration increases the systolic component of the cardiac cycle relative to diastolic, a decrease in heart rate, particularly a decreases to fewer than 75 BPM, reduces it [23]. Thus, lowering heart rate will increase the diastolic component and consequently improve image quality. Although beta receptor-blocking agents are able to lower heart rate, their use is associated with side effects such as heart failure in patients with heart disease, headache, sleeping disorder, diarrhea, etc. Therefore more promising area of research is the development of scanning technologies to provide high temporal resolution against high heart rate. From the standpoint of variation of heart rate, the core for optimal image quality is the precise selection of cardiac cycle phase. EBCT allows only sequential prospective ECG triggering. Only a single phase can be selected for all three major coronary arteries. A prospective trigger is induced from the ECG trace to begin CT scanning for sequential imaging after user-selectable delay following the R wave. The delay for scan acquisition after an R wave is calculated by using a given phase parameter for each cardiac cycle individually on the basis of prospective assessment of the R-R intervals. The next R-R interval is estimated based on the median of the last seven R-R intervals. However, prospective ECG triggering fails with rapid changes in heart rate [24]. The only way to accomplish a reconstruction appropriated to the phase of the cardiac cycle is retrospective gating. Retrospective ECG gating can be applied to MDCT reconstruction techniques. Continuous volume coverage and better spatial resolution in longitudinal direction 0-7803-7636-6/03/$17.00 2003 IEEE. 1625
are possible because MDCT allows reconstruction of images with overlapping increments in any phase of the heart cycle. When multiple reconstructions are performed in different cardiac phases for optimal image quality of individual arteries, every helical data is used for image reconstructions. Consequently, the maintenance of ECG normality while coping effectively with wider variation of heart rate results in a significant improvement of image quality. In the present study, our data suggested that heart rate and variation of heart rate mainly affected image quality and optimized reconstruction window, respectively. In Fig. 1a, the median BPM value corresponding to bad score suggests that variation of heart rate at even relatively lower heart rates can cause further degradation of image quality. This is because variation of heart rate was better correlated with image quality, as shown in Fig. 1b. On the other hand, in Fig. 2b, the wider BPM variation than the rest, when % reconstruction is better quality than %, confirms that heart rate acceleration increases the systolic component of the cardiac cycle relative to diastolic, and that the effect of variation of heart rate in systolic phase is relatively less serious than in diastolic phase at higher heart rates. This is also because heart rate was better correlated with optimized reconstruction window, as shown in Fig. 2a. V. CONCLUSION The results of our study demonstrate the priority among the factors that affect image quality and optimized reconstruction window in retrospective ECG gated coronary angiography using MDCT. Firstly, image quality is more affected by variation of heart rate than by high heart rate. Secondly, selection of optimized reconstruction window for enhanced image quality is mainly affected by heart rate and there is a tendency for earlier phase reconstruction to be better in image quality than later phase reconstruction at higher heart rates. More advanced techniques that can essentially reduce the degree of image degradation, by dealing effectively with high heart rate and wide variation of heart rate during cardiac scanning and post-processing, are necessary to expand the amount of image information and facilitate clinical application. REFERENCES [1] C. Hong, C. R. Becker, A. Huber, U. J. Schoepf, B. Ohnesorge, A. Knez, R. Brüning, and M. F. Reiser, ECG-gated Reconstructed Multi-detector Row CT Coronary Angiography: Effect of Varying Trigger Delay on Image Quality, Radiology, vol. 220, pp. 712-717, 2001. [2] C. J. Riechie, J. D. Godwin, C. R. Crawford, W. Stanford, H. Anno, and Y. Kim, Minimum scan-speeds for suppression of motion artifacts in CT, Radiology, vol. 185, pp. 37-42, 1992. [3] R. J. Alfidi, W. J. Maclntyre, and J. R. Hagga, The effects of biological motion on CT resolution, AJR Am. J. Roentgenol, vol. 127, pp. -15, 1976. [4] S. Achenbach, W. Moshage, D. Ropers, J. Nossen, and W. G. Daniel, Value of electron-beam computed tomography for the noninvasive detection of high-grade coronary-artery stenoses and occlusions, N. Engl. J. Med., vol. 339, pp. 1964-1971, 1998. [5] D. Ropers, W. Moshage, W. G. Daniel, J. Jessl, M. Gottwik, and S. Achenbach, Visualization of coronary artery anomalies and their anatomic course by contrast-enhanced electron beam tomography and three-dimensional reconstruction, Am. J. Cardiol., vol. 87, no. 2, pp. 193-7, 2001. [6] W. Moshage, S. Achenbach, B. Seese, K. Bachmann, and M. Kirchgeorg, Coronary artery stenoses: three-dimensional imaging with electrocardiographically triggered, contrast agent-enhanced, electron-beam CT, Radiology, vol. 196, pp. 7-714, 1995. [7] G. P. Reddy, D. M. Chernoff, J. R. Adams, and C. B. Higgins, Coronary artery stenoses: assessment with contrast-enhanced electron beam CT and axial reconstructions, Radiology, vol. 209, pp. 167-172, 1998. [8] B. J. Rensing, A. Bongaerts, R. J. van Geuns, P. van Ooijen, M. Oudkerk, and P. J. de Feyter, Intravenous coronary angiography by electron-beam computed tomography: a clinical evaluation, Circulation, vol. 98, pp. 29-2512, 1989. [9] A. Schmermund, B. J. Rensing, P. F. Sheedy, M. R. Bell, and J. A. Rumberger, Intravenous electron-beam computed tomographic coronary angiography for segmental analysis of coronary artery stenoses, J. Am. Coll. Cardiol., vol. 31, pp. 1547-1554, 1998. [10] D. M. Chernoff, C. J. Ritchie, and C. B. Higgins, Evaluation of electron-beam CT coronary angiography in healthy subjects, AJR Am. J. Roentgenol, vol. 169, pp. 93-99, 1997. [] T. Nakanishi, K. Ito, M. Imazu, and M. Yamakido, Evaluation of coronary artery stenoses using electron-beam CT and multiplanar reformation, J. Comput. Assist. Tomogr., vol., pp. 1-127, 1997. [12] S. Paulin, Coronary angiography: a technical, anatomic, and clinical study, Acta. Radiol., vol. 233S(suppl.), pp. 1-5, 1964. [] M. J. Potel, J. Rubin, S. A. Mackay, A. Aisen, J. Al-Sadir, and R. E. Sayre, Methods for evaluating cardiac wall motion in three dimensions using bifurcation points of the coronary arterial tree, Invest. Radiol., vol. 18, pp. 47-57, 1983. [14] S. Achenbach, D. Ropers, J. Holle, G. Muschiol, W. Daniel, and W. Moshage. In-plane coronary arterial motion velocity: measurement with electron-beam CT, Radiology, vol. 6, pp. 457-463, 2000. [15] S. Achenbach, W. Moshage, D. Ropers, and K. Bachmann, Comparison of vessel diameters in electron beam tomography and quantitative coronary angiography, Int. J. Card. Imaging, vol. 14, pp. 1-7, 1998. [16] K. M. Baskin, W. Stanford, B. H. Thompsom, J. Tajik, S. D. Heery, and E. A. Hoffman, Helical versus electron-beam CT in assessment of coronary artery calcification (abstr.), Radiology, vol. 197(P), pp. 182, 1995. [17] B. Ohnesorge, T. Flohr, C. Becker, A. F. Kopp, U. J. Schoepf, U. Baum, A. Knez, K. Klingenbeck-Regn, M. F. Reiser, Cardiac imaging by means of electrocardiographically gated multisection spiral CT: initial experience, Radiology, vol. 7, pp. 564-71, 2000. [18] M. Kachelriess, S. Ulzheimer, and W. A. Kalender, ECG-correlated image reconstruction from subsecond multi-slice spiral CT scans of the heart, Med. Phys., vol. 27, no. 8, pp. 1881-902, 2000. [19] S. Schroeder, A. F. Kopp, A. Kuettner, C. Burgstahler, C. Herdeg, M. Heuschmid, A. Baumbach, C. D. Claussen, K. R. Karsch, and L. Seipel, Influence of heart rate on vessel visibility in noninvasive coronary angiography using new multislice computed tomography: experience in 94 patients, Clin. Imaging, vol. 26, no. 2, pp. 106-, 2002. [20] A. F. Kopp, S. Schroeder, A. Kuettner, M. Heuschmid, C. Georg, B. Ohnesorge, R. Kuzo, and C. D. Claussen, Coronary Arteries: Retrospectively ECG-gated Multi-Detector Row CT Angiography with Selective Optimization of the Image Reconstruction Window, Radiology, vol. 2, pp. 683-688, 2001. [] K. Klingenbeck-Regn, S. Schaller, T. Flohr, B. Ohnesorge, A. F. Kopp, and U. Baum, Subsecond multi-slice computed tomography: basics and applications, Eur. J. Radiol., vol. 31, pp. 0 24, 1999. [22] B. Ohnesorge, T. Flohr, C. Becker, A. Knez, A. F. Kopp, K. Fukuda, and M. F. Reiser. Cardiac imaging with rapid, retrospective ECG synchronized multilevel spiral CT (Herzbildgebung mit schneller, retrospekriv EKG-synchronisierter Mehrschichtspiral CT), Radiologe, vol., pp. 1-7, 2000. [23] H. Boudoulas, S. E. Rittgers, R. P. Lewis, C. V. Leier, A. M. Weissler, Changes in diastolic time with various pharmacologic agents: implications for myocardial perfusion, Circulation, vol., pp. 164-169, 1979. [24] S. Mao, R. Oudiz, H. Bakhsheshi, S. Wang, and B. Brundage, Variation of heart rate and electrocardiograph trigger interval during ultrafast computed tomography, Am. J. Card. Imaging, vol. 10, pp. 239-243, 1996. 0-7803-7636-6/03/$17.00 2003 IEEE. 1626