| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Yonsei University College of Medicine, Yongdong
Severance Hospital, 146-92 Dogok-Dong, Kangnam-Ku, Seoul, South Korea.
2 Department of Cardiology, Yonsei University College of Medicine, Yongdong
Severance Hospital, Seoul, South Korea.
Received July 3, 2004; accepted after revision October 15, 2004.
Address correspondence to T. H. Kim
(thkim1{at}yumc.yonsei.ac.kr).
OBJECTIVE. The aims of our study were to investigate the clinical feasibility of a two-phase reconstruction method based on ECG to evaluate left ventricular (LV) volume and function using cardiac MDCT and to compare these results with those from echocardiography.
SUBJECTS AND METHODS. The LV end-diastolic and end-systolic volumes, stroke volume, and ejection fraction were measured using two different methods of cardiac MDCT in 19 patients who had undergone cardiac MDCT and echocardiography. The first was a two-phase reconstruction method based on retrospective ECG-triggering: The end-systolic phase was reconstructed when the reconstruction window was located halfway in the ascending T wave on ECG, and the end-diastolic phase was reconstructed when the reconstruction window was located at the starting point of the QRS complex on ECG. The second was a multiphase reconstruction method: 20 series of images were reconstructed at every 5% throughout the cardiac cycle. The LV volumes and function determined by the two reconstruction methods were compared. The results measured by cardiac MDCT were compared with those obtained by echocardiography.
RESULTS. The LV end-diastolic and end-systolic volumes, stroke volume, and ejection fraction measured by the two-phase reconstruction method correlated well with those measured by the multiphase reconstruction method (r = 0.984, 0.978, 0.969, 0.969, respectively). There were no significant differences between the results of the two different reconstruction methods (p > 0.05). The LV volumes showed moderate to good correlation between cardiac MDCT and echocardiography (0.766 < r < 0.940). Ejection fraction measured by cardiac MDCT yielded a significant overestimation of 2.9% ± 8.7% (mean ± SD) compared with that measured by echocardiography.
CONCLUSION. A two-phase reconstruction method on cardiac MDCT is relatively simple and can provide an objective standard for reconstructing the appropriate image sets for end-diastole and end-systole without the need to review serial preview images.
Left ventricular (LV) function is an important parameter for predicting the prognosis of patients with coronary artery disease. Therefore, an assessment of LV function is important for making a clinical diagnosis and for managing and following up these patients [1–3].
A variety of imaging techniques have been introduced to measure LV function. Although widely available, echocardiography is operator- and acoustic-window-dependent [4, 5]. Cineventriculography has many disadvantages such as invasiveness, the need to use iodinated contrast medium, patient exposure to ionizing radiation, and the geometric limitations that result from the projection images. However, it is still a clinically accepted standard [6, 7]. Radionuclide ventriculography is limited by its low spatial and temporal resolution [8]. The introduction of cardiac MRI allowed major progress concerning the temporal and spatial resolution and allowed image acquisition in any desired plane. Several studies have shown that the accuracy and reproducibility of cardiac MRI are excellent for making quantitative LV function measurements [6, 9, 10].
Recently, MDCT has been introduced as a new and promising tool for coronary artery imaging [11–13]. Imaging data are continuously acquired throughout the cardiac cycle using a helical scanning technique. Therefore, additional end-diastolic and end-systolic images can be reconstructed retrospectively from the same MDCT data set, thereby allowing functional analyses to be performed. However, for LV function to be evaluated, 10 or 20 series of images need to be reconstructed every 10% or 5% of the cardiac cycle, which demands a great deal of time. Therefore, we usually use end-diastolic and end-systolic phases reconstructed at the smallest and largest endocardial circumferences, respectively, after reviewing the serial reconstruction images of the cardiac cycle in 5% increments of the R-R interval on the preview console [14–16].
|
|
|
Subjects and Methods
This study was approved by the institutional review boards. Nineteen patients (11 men and eight women; mean age, 57.6 years; age range, 41–81 years) were included in this prospective study. These patients underwent cardiac MDCT as a screening examination for coronary artery disease. In all patients, an additional echocardiography examination was performed to assess the LV volumes and function. This study was undertaken after informed consent was obtained from the patients for the cardiac MDCT and echocardiography examinations.
MDCT Scanning Protocol and Image Reconstruction
Cardiac MDCT was performed using a 16-MDCT scanner (Somatom Sensation 16
[software version VA20], Siemens Medical Solutions). With the patient in the
supine position, cardiac MDCT was performed in the caudocranial direction
during a single breath-hold at the suspension of end-inspiration. A
ß-blocker (80 mg of Panol [propranolol hydrochloride], Dae Woong) was
administered orally 2 hr before the examination to reduce the heart rate in
patients with a heart rate of more than 70 beats per min (bpm). The duration
of the breath-hold ranged from 22 to 27 sec with an average of 24.4 sec, and
the breath-hold could be reached by all patients. The patient's ECG trace was
recorded simultaneously. The mean heart rate during scanning ranged from 45 to
116 bpm (mean, 64 bpm; median, 60 bpm). The scanning parameters were as
follows: 420-msec gantry rotation time, 120 kV, 500 mAs, 0.75-mm slice
collimation, 1-mm slice width, and 2.8-mm table feed/rotation. After 100 mL of
contrast medium (Ultravist 300 [iopromide], Schering) was mixed with 20 mL of
normal saline, a total of 120 mL of the contrast material was administered IV
at a rate of 4.5 mL for the first 20 sec followed by at a rate of 1.5 mL for
the second 20 sec using a power injector (Envision CT, Medrad). The estimated
radiation dose ranged from 4 to 6 mSv depending on the scanning range and the
patient's body weight.
Image reconstruction was performed on the scanner's workstation using commercially available software (Syngo, Somaris/5, Siemens Medical Solutions). For this study, we used the partial scan algorithm, which provided a heart-rate-dependent temporal resolution of between 70 and 210 msec from a 420-msec gantry rotation. The reconstruction parameters used were as follows: a 1-mm slice thickness, a 0.5-mm increment, a 512 x 512 pixels image matrix, a medium smooth kernel, and a 20-cm field of view.
Image sets were reconstructed using two different methods to identify the maximum systolic constriction and diastolic relaxation phases. The first was a two-phase reconstruction method based on retrospective ECG triggering: The end-systolic phase was reconstructed when the reconstruction window was located halfway in the ascending T wave on ECG, and the end-diastolic phase was reconstructed when the reconstruction window was located at the starting point of the QRS complex on ECG (Fig. 1A, 1B, 1C). The second was a multiphase reconstruction method: 20 series of images were reconstructed at every 5% throughout the cardiac cycle (Fig. 2A, 2B). The latter was used as the standard technique for the former in evaluating the LV volumes and function.
|
|
MDCT Data Analysis
Two radiologists independently evaluated the LV end-systolic and
end-diastolic volumes using commercially available software (Argus, Wizard,
Siemens Medical Solutions). The endocardial contours of all the systolic and
diastolic short-axis reformation images were manually traced on the screen of
the analysis software. Both the papillary muscles and trabeculations were
considered to be part of the LV cavity. The area of the LV outflow tract below
the aortic valve was included in the ventricular part. The ejection fraction
and stroke volume were obtained from the graph and the summarized table on the
analysis software (Fig. 3A,
3B,
3C,
3D).
|
|
|
|
 |
Statistical Analysis
Pearson's correlation coefficient and Bland-Altman analyses were performed
to determine the correlation and limits of agreement for the LV end-diastolic
and end-systolic volumes, stroke volume, and ejection fraction between the two
reconstruction methods of cardiac MDCT and between cardiac MDCT using the
two-phase reconstruction method and echocardiography. A Wilcoxon's signed rank
test was used to evaluate the statistical significance of the differences in
the LV volumes and functional data between the two reconstruction methods of
cardiac MDCT and between cardiac MDCT using the two-phase reconstruction
method and echocardiography. A p value of less than 0.05 was
considered significant.
The interobserver variability (Var) for the LV end-diastolic and
end-systolic volumes measured by cardiac MDCT using the two-phase
reconstruction method was assessed using the following equation:
 |
Results
Cardiac MDCT short-axis reformations allowed clear delineation of the endocardial contours of the left ventricle in all patients. Minor stair-step artifacts were visualized, which were more prominent in the systolic reconstruction images than in the diastolic reconstruction images. Although these artifacts were more severe in the systolic 2D reformations from patients with a heart rate of more than 70 bpm, there were no obstacles in evaluating the LV volumes and function.
The measurement results are summarized in Tables 1 and 2. The mean (± SD) LV end-diastolic volume, LV end-systolic volume, LV stroke volume, and LV ejection fraction measured by cardiac MDCT using the two-phase reconstruction method correlated well with the respective measurements of cardiac MDCT using a multiphase reconstruction method (Table 1). The mean differences in the LV volumes and the functional measurements between the two-phase and multiphase reconstruction methods of cardiac MDCT were –0.4 ± 6.2 mL/m2 for LV end-diastolic volume, –0.3 ± 5.0 mL/m2 for LV end-systolic volume, –0.4 ± 4.6 mL/m2 for LV stroke volume, and 0.4% ± 4.1% for LV ejection fraction. These results show good agreement between the two reconstruction methods of cardiac MDCT. There were no significant differences in the mean LV volumes and the functional measurements between the two reconstruction methods of cardiac MDCT (LV end-diastolic volume, p = 0.52; LV end-systolic volume, p = 0.64; LV stroke volume, p = 0.47; LV ejection fraction, p = 0.45) (Table 1 and Fig. 4A, 4B, 4C, 4D).
|
|
|
|
|
|
The LV end-diastolic volume and LV end-systolic volume measured by cardiac MDCT using the two-phase reconstruction method correlated well with those measured by echocardiography (r = 0.931 and r = 0.940, respectively), and there were no significant differences in their mean volumes between cardiac MDCT and echocardiography (LV end-diastolic volume, p = 0.39; LV end-systolic volume, p = 0.10). The LV stroke volume measured by cardiac MDCT using the two-phase reconstruction method showed a moderate correlation with that measured by echocardiography (r = 0.766), and the mean difference in the LV stroke volume between cardiac MDCT and echocardiography was not significant (p = 0.08). However, the LV ejection fraction measured by cardiac MDCT using the two-phase reconstruction method showed a moderate correlation with that measured by echocardiography (r = 0.846), and the mean difference in the LV ejection fraction between cardiac MDCT and echocardiography was statistically significant (p = 0.02). This represents a significant overestimation of 2.9% ± 8.7% (mean ± SD) for the LV ejection fraction using cardiac MDCT with the two-phase reconstruction method (Table 2).
The interobserver variability for the LV volumes measured by cardiac MDCT using the two-phase reconstruction method was 5.8% of the mean for the LV end-diastolic volume and 10.5% of the mean for the LV end-systolic volume.
Discussion
This study showed that the LV end-diastolic volume, end-systolic volume, stroke volume, and ejection fraction measured by cardiac MDCT with the two-phase reconstruction method based on ECG correlated well with those measured by cardiac MDCT using the multiphase reconstruction method. There were no significant differences in the LV volumes and function between the two reconstruction methods of cardiac MDCT. The LV volumes and function showed a moderate correlation between cardiac MDCT using the two-phase reconstruction method and echocardiography. The LV ejection fraction measured by cardiac MDCT using the two-phase reconstruction method based on ECG gave a significant overestimation of 2.9% ± 8.7% (mean ± SD) from that measured by echocardiography.
The recently introduced MDCT scanners have proven their ability to perform high-resolution CT coronary angiography through the subsecond rotation times and dedicated cardiac reconstruction algorithms [11–13]. Data acquisition covers the cardiac cycle during a single breath-hold at the end-inspiratory suspension using a helical CT technique. The images are retrospectively reconstructed using ECG gating. Because a freely selectable distance from the preceding or following R peak defines the data segment from the cardiac cycle, cardiac MDCT can then be used to calculate the ventricular volumes and function.
To acquire the appropriate image sets for end-systole and end-diastole, investigators in previous studies selected the smallest and largest cross-sectional LV cavity areas, respectively, on the preview images with 5% or 50-msec incremental steps of the R-R interval through the cardiac cycle [14–16]. Although these methods are relatively simple, there may be some limitations in the selection of the appropriate phase for end-systole and end-diastole. A direct comparison of ventricular size on the axial image series only at the midventricular level is not easy and is quite subjective. This is because the position of the heart is continuously changing during contraction. On the basis of this concept, the phase-selection method was designed for end-systole and end-diastole in reference to ECG.
In previous reports, a reconstruction method based on ECG has usually been described for evaluating coronary artery disease on cardiac MDCT [11, 17, 18]. No reports, to our knowledge, have defined an appropriate phase-selection method based on ECG for the measurement of end-systole and end-diastole volumes on cardiac MDCT examination. The results derived from the ECG-based LV volume and functional assessment show that the two-phase reconstruction technique could be a relatively simple method for evaluating LV volumes and function with cardiac MDCT. The results measured by the two-phase reconstruction method based on ECG were similar to those measured by the multiphase reconstruction method of cardiac MDCT, which consisted of 20 phases of an axial image series in 5% intervals throughout the cardiac cycle and required a great deal of time to handle the image data sets for each case. In this study, however, only 5–10 min were needed to reconstruct the two data sets based on ECG for end-systole and end-diastole and to analyze LV volumes and function from the same data sets.
Even though the LV volumes measured by cardiac MDCT using the two-phase reconstruction method were similar to those measured by echocardiography, cardiac MDCT yielded an overestimation of the ejection fraction of 2.9% ± 8.7% (mean ± SD) compared with echocardiography. Dirksen et al. [19] reported that the mean LV ejection fraction was slightly lower with cardiac MDCT than with echocardiography but that this difference was not statistically significant. This discrepancy cannot be explained simply, but a potential cause might be interobserver variability. In our study, we found an interobserver variability of 5.8% for the LV end-diastolic volume and 10.5% for the LV end-systolic volume with cardiac MDCT using the two-phase reconstruction method. These data are similar to those from another study that reported an interobserver variability of 6.8% for the LV end-diastolic volume and 8.5% for the LV end-systolic volume using the volumetric MDCT measurements [14].
Although cardiac MDCT can be used as a tool to evaluate ventricular volume and function, its routine use has some limitations. Cardiac MDCT requires contrast medium to discriminate the ventricular cavity, and the temporal resolution is inferior to that of either electron beam CT or MRI [14]. The use of a contrast medium causes a volume load and might influence LV ejection fraction in cardiac MDCT. Furthermore, cardiac MDCT requires a temporal resolution of approximately 20 msec to completely avoid any motion artifact in imaging the heart [20]. Therefore, the limited temporal resolution used in this study might be insufficient to acquire precise end-systolic volume. A stair-step artifact may interrupt the ventricular volume evaluation, particularly in the systolic phase. Another limitation is radiation exposure to the patient. The total radiation dose was measured to be approximately 4–6 mSv depending on the scanning range and the patient's body weight. Considering the need to administer contrast medium, the radiation exposure to the patient, and the limited temporal resolution, using only cardiac MDCT for analyzing ventricular volume and function does not appear to be reasonable now.
Our study has some limitations that will be overcome in the future. First, the number of patients in this study is too small to generalize the results in sicker patients or patients with arrhythmia. Second, although the results of cardiac MDCT are compared with those of echocardiography, further studies will be needed in the future to compare the results obtained on cardiac MDCT with those obtained on MRI, which is accepted as a better gold standard. Finally, there were no results comparing the differences between the two-phase reconstruction method and the method described in previous reports [14–16], in which end-systole and end-diastole images were selected after reviewing the serial preview images. However, this study has the advantage of offering an objective reference for selecting the appropriate image sets for end-systole and end-diastole measurements without visually comparing the sizes of the LV cavity on the serial preview images.
In conclusion, cardiac MDCT allows ventricular volume and function to be assessed and depicts direct images of the coronary artery in patients with suspected coronary artery disease. The two-phase reconstruction method based on ECG using cardiac MDCT is relatively simple and fast for evaluating LV volumes and function. In addition, it can provide objective reconstructive points for the end-systole and end-diastole phases without the need to review the serial preview images.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
