Comparison Study of Image Quality on Coronary Artery Visualization by pharmphresh31

VIEWS: 7 PAGES: 10

									Feb. 2007, Volume 4, No.2 (Serial No.27)                        Journal of US-China Medical Science, ISSN1548-6648, USA




     Comparison Study of Image Quality on Coronary Artery Visualization

           by Electron Beam Computed Tomography and 16-slice Spiral

                                           Computed Tomography

              ZHENG Min-wen, HUAN Yi, XU Jian, YANG Yong, GE Ya-li, SHI Ming-guo, PENG Yong
                  (Division of Radiology, Xijing Hospital, the Fourth Military Medical University, Xi’an 710032)


      Abstract: Objective Non-invasive coronary imaging holds a bright future in diagnosis of coronary artery
disease. However, clinical application is limited by cardiac motion artifacts, which cause incomplete
interpretability. The aim of this study is to compare the image quality of coronary artery visualization by electron
beam computed tomography (EBT) and 16-slice spiral computed tomography (16-slice spiral CT). Methods In 23
patients (18 male, 5 female, aged 49±4) with suspected obstructive coronary artery disease, ECG-gated EBT and
16-slice spiral CT angiography was respectively performed. Overall length of the visualized coronary arteries and
evaluable vessel length were measured. The image interpretability was classified as from grade A to C: grade A:
no cardiac artifacts; grade B: mild cardiac artifacts and grade C: significant cardiac artifacts. 3-dimension
reconstruction methods in BET and 16-slice spiral CT were compared by two experienced cardiovascular
radiologists. Results There was no significant difference in overall visualized length of the coronary arteries
evaluated by 16-slice spiral CT and EBT. 16-slice spiral CT significantly depicted longer diagnosable segments in
left anterior descending (length: 77.1±16.8 vs 59.3±22.1mm, P<0.05) and left circumflex coronary arteries (length:
65.8±19.5 vs 48.3±17.3mm, P<0.01) compared to EBT. The grade A and grade B percentage as well as the
percentage of diagnosable segments were higher measured by 16-slice spiral CT than that measured by EBT
(P<0.01). 16-slice spiral CT showed more grade C images of the coronary arteries than EBCT. Of all segments of
grade C determined by 16-slice spiral CT, most were middle part of right coronary arteries. However, grade C
determined by EBT were distal segments of left anterior descending and circumflex coronary arteries.
Non-calcification plaques within vascular wall were more easily shown by 16-slice spiral CT. The imaging quality
of reconstructed coronary arteries by conventional 3-dimensional reconstruction methods in 16-slice spiral CT
were better than that in EBT except cross section maximum intensity projection (MIP). Conclusion Due to its
higher spatial resolution and lower image noise, 16-slice spiral CT seems to be superior to EBT in visualization of
the coronary arteries although motion artifacts of right coronary artery were more frequent than in EBT. Both EBT
and 16-slice spiral CT enable non-invasive visualization of coronary arteries and evaluation of obstructive
coronary artery diseases.
      Key words: coronary angiography; coronary artery disease; computed tomography; computed X-ray

     In recent years, considerable progresses have been achieved in the field of non-invasive coronary imaging


  HUAN Yi (1961- ), female, Ph.D. supervisor, professor of Department of Radiology, Xijing Hospital, the Fourth Military Medical
University; research fields: CT and MR imaging diagnosis.


72
     Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                     and 16-slice Spiral Computed Tomography

with MRI, electronic beam computed tomography (EBT) and multislice spiral computed tomography (MSCT).
However, cardiac motion artifacts and calcification deposit in the coronary artery wall limits proper assessment of
images. Cardiac motion artifacts frequently occur in higher heart rates. Recently developed 16-, even 64- slice
spiral computed tomography decreases motion artifacts to a large extent due to thinner detector rows and
increased rotation speed. In this study, we compare the quality of coronary artery images generated with EBT and
16-slice spiral CT (or MSCT), and try to draw an outline of EBT and 16-slice spiral CT in clinical application.

                                                   METHODS

1. Study Population
     Twenty-three patients (ages 43-64, average 49±4, male 18, female 5) with suspected obstructive coronary
artery disease were performed with EBT, and subsequently MSCT between 3 days and 9 months (average 56±9
days) following EBT at the heart rates of 50-82/min. Five of twenty-three patients were performed percutaneous
coronary angiography (PCA), and three of them received coronary stent implantation before this study.
2. 16-slice Spiral CT
     MSCT scanning was performed with an Aquilion 16 CT scanner (Toshiba Medical Systems, Otawara, Japan).
Images were acquired in supine position gated with ECG in inspiratory breath hold, preceded by mild
hyperventilation. Initially, the position of the imaging volume and the contrast agent transit time were determined.
Then, the volume data set for coronary artery visualization was acquired in spiral mode. 16 parallel contiguous
slices were simultaneously acquired with 1.0mm collimation each (gantry rotation time 420 milliseconds/360°,
table feed 1mm/rotation, 120kV, tube current 300mA). The field of view was 180mm with an image matrix of 512
× 512 pixels. A bolus of 100ml contrast medium (Ultravist 300, Schering, Berlin, Germany) was intravenously
injected through cubital vein at 4ml/second, and the CT scan was initiated with a delay according to the previously
determined contrast medium transit time. During the scan, the patient’s ECG was digitized and continuously
recorded.
     The raw data set was then retrospectively reconstructed around 5 equidistant phases of the cardiac cycle
(40% to 80% of the R-R interval) use both the half-scan and the multi-segmental algorithm yielding data sets of
the entire heart from five different phases of the cardiac cycle.
3. Electron Beam Tomography
     EBT scanning was performed with a C-150 CT scanner (Imatron, San Francisco, CA, USA) in supine
position. The exact individual contrast transit time was determined by placing a region of interest in the ascending
aorta and evaluating a time-density curve. After determination of the heart position and measurement of the
contrast-agent transit time, axial cross-sections of the heart were acquired in inspiratory breath hold. Prospectively
triggered to the ECG at 40% of the R–R interval, one image was acquired after every heartbeat. Slice thickness
was 3mm, with a table feed of 2mm between two consecutive images. The tube current was 625mA at 130kV.
Images were reconstructed with a field of view of 150mm and a matrix of 512 × 512 pixels. A bolus of 150ml
contrast medium (Ultravist 300) was intravenously injected at a flow rate of 4ml/second through the cubital vein.
4. Data Postprocessing and Image Analysis
     All data sets were transferred from the CT to off-line image processing workstations (vitrea 2; vital, CA and
Accuview; Accuimage, CA, USA). In addition to the axial sectional images, curved multiplanar reformations
along the course of the 3 main coronary arteries were generated from each data set. All image data sets were


                                                                                                                   73
       Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                       and 16-slice Spiral Computed Tomography

inspected concerning the presence of motion artifacts by one investigator experienced in cardiac and coronary
imaging. Two blinded radiologists independently evaluated the MSCT and EBT scans by using 9-segment model
including left coronary artery (LCA) - left main segment (LMS); proximal, middle, and distal segments of left
anterior descending artery (LAD); proximal and distal segments of left circumflex artery (LCx); and right
coronary artery (RCA) - proximal, middle, and distal segments of RCA. Enhanced scan images were
reconstructed with the methods of curved multiplanar reformations (CPR), multiplanar reformations (MPR) and
maximum intensity projection (MIP). The volume rendering (VR) was utilized for reconstruction of images in
some cases. Quantitative analysis is comprised of the following measurements: (1) the overall length of the
visualized segment of each coronary artery. (2) The evaluable vessel length depicted free of motion artifacts of
each coronary artery. For doing steps 1 and 2, a path was drawn exactly along the course of the concerned vessels
or segments on the curved multiplanar reconstructions. The vessel length was automatically read by machine. (3)
The image interpretability was classified as from grade A to C (grade A is a vessel segment contour uninterrupted
and sharp; grade B is a vessel segment contour slightly interrupted and not clearly delineated but still can be
evaluated; grade C is a vessel segment contour obviously interrupted and severely affected by artifacts). Then, the
percentages of image interpretabilities were calculated from grade A to C in MSCT and EBT scans. (4)
Comparison of atherosclerotic coronary artery plaque visualization between MSCT and EBT scans. Of all 9
patients underwent both MSCT and EBT within 5 days to 23 days, noncalcified plaques were detected in 6
patients. Every plaque in same segment was observed according to MSCT and EBT reconstruction imaging
respectively, and then compared.
5. Statistical Studies
     Statistical analysis was performed using a PC-based software program (SPSS 10.0). Mean and standard
deviations were calculated for all variables. After testing for equality of variances, the t test for paired samples,
and unpaired samples, respectively, as well as the χ2 test for ratio samples were used to compare MSCT and EBT
parameters. P<0.05 was considered to represent a statistically significant result.

                                                         RESULTS

1. The Evaluable and Diagnosable Lengths of Coronary Artery Images
     There was no significant difference in overall visualized length of the coronary arteries evaluated by MSCT
and EBT. 16-slice MSCT significantly depicted longer diagnosable segments in left anterior descending (length:
77.1±16.8 vs 59.3±22.1mm, P<0.05) and left circumflex coronary arteries (length: 65.8±19.5 vs 48.3±17.3mm,
P<0.01) compared to EBT. (Table 1)
     Table 1 Comparison of evaluable and diagnosable lengths of coronary arteries by MSCT and EBCT (mm). (P<0.05,
     significant; P<0.01, very significant) 16-slice MSCT significantly depicted longer diagnosable segments in left anterior
       descending (length: 77.1±16.8 vs 59.3±22.1mm, P<0.05) and left circumflex coronary arteries (length: 65.8±19.5 vs
                                             48.3±17.3mm, P<0.01) compared to EBCT.
      Coronary            Evaluable lengths (mm)                 P             Diagnosable Lengths (mm)               P
      Arteries          MSCT                  EBT              Value           MSCT                EBT              Value
        LM             16.4±6.1             16.4±5.9          >0.05           16.4±6.1            16.4±5.9         >0.05
        LAD           124.3±25.8           120.5±24.0         >0.05          77.1±16.8            59.3±22.1        <0.05
        LCx           88.7±19.5            79.5±18.3          >0.05          65.8±19.5            48.3±17.3        <0.01
        RCA           147.3±19.8           145.7±19.0         >0.05          117.6±23.0          121.6±20.8        >0.05



74
     Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                     and 16-slice Spiral Computed Tomography

2. Quality Comparison of Coronary Artery Imagines by MSCT and ESCT
      Both in MSCT and, to a lesser extent, in EBT, coronary visualization was affected by remaining motion
artifacts. In MSCT, a significantly smaller proportion of the coronary arteries was visualized free of motion
artifact than in EBT. The grades of 207 coronary artery segments in 23 patients were showed as Table 2. In MSCT
scans, 82.2% (74/90) of grade a segments located at proximal segment of coronary artery and LM. 47.1% (16/34)
of grade C segments were respiratory artifacts in the middle of RCA (Figure 1). In EBCT scans, 83.9% (52/62) of
grade C located at LAD and distal LCx (Figure 2). The percentage ratio of grade A and B in MSCT and EBCT
scans (the ratio of diagnosable segments) were respectively 83.6% (173/207) and 70.0% (145/207) (P<0.01).

                         Table 2   Percentages of grades in coronary artery imagines in MSCT
                            and EBCT scans. (P<0.05, significant; P<0.01, very significant)
             Quality                     MSCT                          EBT                     P Value
             Grade A                 43.5% (90/207)               31.9% (66/207)               <0.05
             Grade B                 40.1% (83/207)               38.1% (79/207)               >0.05
             Grade C                 16.4% (34/207)               30.0% (62/207)               <0.01

3. Visualization of Plaque and Stenosis
     Of 6 patients, 18 atherosclerotic plaques were observed in coronary arteries with MSCT. Of them, there are 3
calcified plaques, 3 non-calcified plaques and 13 mixed plaques (7 in LAD, 2 in RCA and 4 in LCx). 14 plaques
were observed by EBCT in 23 patients. Compared to MSCT scans, 4 plaques were not detected by EBCT. 2 of
them were non-calcified plaques in superior wall and inferior wall of LAD. However, the plaques at the same
position were detected by MSCT (Figure 3, and 4). Other 2 of undetected plaques in EBT scans were artery wall
thickness, which were well observed in MSCT scans.
4. Analysis of 3-dimentional Reconstructions by Independent Radiologists
     With volume rendering (VR) reconstruction, coronary artery imagines generated by MSCT showed better
quality of 3-dimentional coronary arteries in length, width and contours than EBCT, especially in LAD, LM and
LCx (Figure 5, 6). Reconstructed with curved multiplanar reformations (CPR) for MSCT and EBCT scans,
coronary artery imagines also showed better quality in MSCT scans compared to EBCT scans, particularly in
LAD and LCx. Unlike reconstructions of EBCT scans, artifacts were observed in RCA in reconstructions of
MSCT scans. Multiplanar reformations (MPR) reconstruction of MSCT scans showed location of atherosclerotic
plaques and lumen stenosis through analysis of vessel vertical section imagines (Figure 7). MPR reconstruction
of EBCT scans did not show any plaques or stenosis for the same arteries (Figure 8). However, maximum
intensity projection (MIP) reconstruction of EBCT showed similar imagines qualities to that of MSCT.

                                                 DISCUSSION

1. Comparison of Length in Visualized and Diagnosable Segment
     Due to higher spatial resolution and lower contrast-to-noise ratio, MSCT generated better quality of coronary
artery imagines than EBT. Excluding showing similar length in LM, MSCT also depicted longer visualization
segments in most of coronary arteries than EBT, particularly in distal segments of LAD and LCx. This different
was not statistically significant. Our results are coincidence with Achenbach’s observations in their study with
4-slice spiral computed tomography and EBT[1]. Furthermore, MSCT really depicted the longer diagnosable
segments in LAD and LCx than EBT, especially in LCx (P<xxx). Generally, both in MSCT and, to a lesser extent,


                                                                                                               75
     Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                     and 16-slice Spiral Computed Tomography

in EBT, coronary visualization was affected by remaining motion artifacts[2]. In MSCT, a significantly smaller
proportion of the coronary arteries was visualized free of motion artifact than in EBT[3]. These resulted in the
lower visualized and diagnosable length in distal segments of LAD and LCx in EBT than in MSCT. For larger
coronary arteries such as LM and RCA, EBT generated the same quality imagines as MSCT due to quicker
scanning speed (0.1s/imagine). In Chinese people, EBT even depicted longer RCA because of RCA priority in
Chinese.
2. Comparison of Quality
     Our study showed higher rate of grade A (P<0.05), but lower rate of grade C (P<0.01) in MSCT than EBT.
This indicated that the vessel imagine quality in MSCT was better than in EBT. The higher spatial, temporal and
density resolution in MSCT leads to the better quality of vessel imagines, especially in small vessels such as distal
segments of LAD and Lcx. Meanwhile, EBT was limited by lower spatial resolution for clearly depicting distal
segments of LAD and LCx. In our study, 52 of 62 un-diagnosable segments of grade c (83.9%) mostly located in
the distal segments of LAD and LCx. Investigators reported that heart rate negatively related to the vessel imagine
quality. That means that fast heart causes more motion artifacts in coronary artery imagines. MSCT images were
more frequently affected by motion artifacts than EBT[4, 5] which has a fast scanning speed (0.1s). About half of
grade c (47.1%) in MSCT were involved in the motion artifacts, especially in RCA because RCA located at the
free wall of right ventricle and has a larger motion extent[6]. However, motion artifacts can be partly diminished in
MSCT because MSCT scans can be reconstructed to display the good imagines through multi-temporal selections.
Furthermore, grade A and B segments in MSCT were more than that in EBT, which provided more diagnosable
segments compared to EBT.
3. Comparison of 3-dimentional Reconstruction and Plaque Imagines
     Both MSCT and EBT un-invasively showed coronary stenosis and plaques[7]. 16-slice CT has significantly
higher spatial, temporal and density resolution than EBT and acquires data with 16-slice CT has significantly
higher spatial, temporal and density resolution than EBT and acquires data with equal voxel acquisition. This
advantage promises a better quality of reconstructed coronary artery imagines with MPR and CPR, which it is
easy to visualize the superior and inferior plaques in parallel and vertical sections and evaluate the degree of
stenosis. Reportedly, 16-slice spiral CT has higher diagnosability for coronary stenosis with stenosis degree >50%
of lumen and lower heart rates (<65bpm), which the sensibility, specificity, positive predictive value and negative
predictive value were respectively 95.7%, 95.8%, 86.3% and 98.8%[8]. Some of them, the sensibility and
specificity were up to 98%[7]. Accurate diagnostic rates were the highest in LAD and LM, but lowest in LCx[9].
Due to insufficient spatial resolution, it is difficult for EBT. Diagnosis with EBT scans mainly depends upon
original axial imagines and axial or oblique MIP imagines. Of them, MPR imagines were not good enough to read,
CPR imagines were significantly affected by echelon artifacts, and axial position imagines often can not depicted
superior and inferior plaques in longitudinal vessels. In this study, two pieces of plaques in superior and inferior
wall were detected by MSCT but not EBT. Another advantage of MSCT over EBT is vessel probe software which
can automatically visualizes target vessel with double-centerline curved multiplanar reconstruction (CPR).The
software helped to shorten reconstruction time and displayed lesions easy to skip. There were 2 thin plaques
attached to lateral wall of vessel which was not displayed by EBT with the manually drawn centerline but shown
by MSCT with automatic vessel probe technique. The contour of heart and coronary vessels on fake color volume
rendering (VR) were much clear in MSCT than in EBT. Although there is not statistic analysis reported, the
MSCT performs better than EBT in visualizing the coronary wall as well as the lumen because of its higher

76
      Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                      and 16-slice Spiral Computed Tomography

temporal resolution and spatial resolution.
4. Future Directions
      Both MSCT and EBT are non-invasive coronary angiography. EBT has advantages of fast scanning speed
and less of motion artifacts. Because of insufficient spatial resolution, EBT can not clearly show small vessels
such as distal segments of LAX and LCx. EBT is also limited by high cost. Compared to EBT, MSCT has higher
spatial, temporal and density resolution, and acquires better quality of coronary artery imagines, especially for
LAD and LCx. However, the imagines in distal segments are still not clear enough for clinical diagnosis.
Furthermore, motion artifacts in RCA also indicate that scanning speed has already limited MSCT application.
The newly developed 64-slice or more slice spiral CT will strength this area and will generates good quality of
coronary artery imagines. EBT, particularly MSCT holds a promising future in non-invasive coronary angiography.
Although EBT or MSCT can not fully substitute of coronary angiography in diagnosis of coronary artery disease,
it has been approved that MSCT and EBT are valuable in screening diagnosis of coronary artery disease.

                                                       REFERENCES
[1]   Achenbach, S., Giesler, T., Ropers, D., et al. Comparison of Image Quality in Contrast-enhanced Coronary-artery Visualization
      by Electron Beam Tomography and Retrospectively Electrocardiogram-gated Multislice Spiral Computed Tomography. Invest
      Radiol, 2003, 38(2): 119-128.
[2]   Lembcke, A., Wiese, T.H., Schnorr, J., et al. Image Quality of Noninvasive Coronary Angiography Using Multislice Spiral
      Computed Tomography and Electron-beam Computed Tomography: Intraindividual Comparison in an Animal Model. Invest
      Radiol, 2004, 39(6): 357-364.
[3]   Zheng, M.W., Sun, L.J., Huan, Y., et al. EBT in Evaluation of Coronary Artery Stenosis. J Chin Radiol, 2004, 12(2): 101-104.
      (in Chinese)
[4]   Cademartiri, F., Runza, G., Marano, R., et al. Diagnostic Accuracy of 16-row Multislice CT Angiography in the Evaluation of
      Coronary Segments. Radiol Med (Torino), 2005, 109(1-2): 91-97.
[5]   Lv, B., Dai, R.P., Jiang, S.L. Analysis of Coronary Artery Imagine Quality Acquired by Enhanced EBT. Acta Chin Med Acad,
      2002, 24: 348.
[6]   Achenbach, S., Giesler, T., Ropers, D., et al. Detection of Coronary Artery Stenoses by Contrast Enhanced, Retrospectively
      Electrocardiographically-gated Multislice Spiral Computed Tomography. Circulation, 2001, 103(5): 2535-2538.
[7]   Nieman, K., van der Lugt, A., Pattynama, P.M., et al. Noninvasive Visualization of Atherosclerotic Plaque with Electron Beam
      and Multislice Spiral Computed Tomography. J Interv Cardiol, 2003, 16(2): 123-128.
[8]   Kuettner, A., Trabold, T., Schroeder, S., et al. Noninvasive Detection of Coronary Lesions Using 16-detector Multislice Spiral
      Computed Tomography Technology: Initial Clinical Results. J Am Coll Cardiol, 2004, 44(6): 1230-1237.
[9]   Matsuo, S., Nakamura, Y., Matsumoto, T., et al. Visual Assessment of Coronary Artery Stenosis with
      Electrocardiographically-gated Multislice Computed Tomography. Int J Cardiovasc Imaging, 2004, 20(1): 61-66.


                                                                                                          (Edited by Jane Chen)




                                                                                                                                 77
     Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                     and 16-slice Spiral Computed Tomography




          Fig. 1    Motion artifacts in middle RCA generated by MSCT-CPR reconstruction.↑: motion artifacts.




           Fig. 2    LAD imagines generated by EBT-CPR reconstruction. Distal and middle segments of LAD
                                       were not clear enough for clinical diagnosis.




78
Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                and 16-slice Spiral Computed Tomography




     Fig. 3 LAD of CPR reconstruction in 16-slice spiral CT clearly showed non-calcified plaque located
                                       in inferior wall of proximal LAD.




    Fig. 4   In the same patient, LCx of EBT-CPR reconstruction did not show plaque at the same location,
                              axial MIP reconstruction also did not show plaque.




                                                                                                            79
         Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                         and 16-slice Spiral Computed Tomography




     Fig. 5   Heart and coronary artery imagines reconstructed with VR in 16-slice spiral. LM, Full length of LAD and LAD
                                      branches and proximal RCA were clearly observed.




Fig. 6    In the same patient, Heart and coronary artery imagines reconstructed with VR in EBT. RCA was clearly showed up.
           However, LM and proximal LAD were not clearly displayed, and middle and distal LAD were not displayed.




80
    Comparison Study of Image Quality on Coronary Artery Visualization by Electron Beam Computed Tomography
                                    and 16-slice Spiral Computed Tomography




 Fig. 7    Longitudinal and vertical sections of proximal LAD reconstructed with MRP in 16-slice spiral CT. Non-calcified
                plaque in internal and inferior wall was clearly displayed. The lumen stenosis was about 30%.




Fig. 8    In the same patient, EBT-MPR reconstructed vessel imagines in the same location with cross section. The imagines
                                        were not good enough to display any plaque.




                                                                                                                            81

								
To top