Asian Cardiovasc Thorac Ann 2006;14:321-327
© 2006 Asia Publishing EXchange Ltd
Imaging of Calcified Coronary Arteries with Multislice Computed Tomography
Thomas Wittlinger, MD,
Ivo Martinovic, MD1,
Rainer Moosdorf, PhD1,
Anton Moritz, PhD
Department of Thoracic and Cardiovascular Surgery, University Hospital, Frankfurt/Main, Germany
1 Department of Heart Surgery, University Hospital, Marburg, Germany
For reprint information contact: Thomas Wittlinger, MD Tel: 49 69 630 183 315 Fax: 49 67 326 5370 Email: thomaswittlinger{at}t-online.de, Department of Thoracic and Cardiovascular Surgery, University Hospital, Theodor-Stern Kai 7, 60590 Frankfurt/Main, Germany.
 |
ABSTRACT
|
|---|
Reliable noninvasive detection of coronary artery disease is a prime goal in clinical cardiology. The aim of this study was to investigate the accuracy of multislice computed tomography in detecting coronary artery disease in correlation to the calcium score. Fifty patients with 61 stenoses > 50% and 41 occlusions underwent multislice computed tomography and conventional coronary angiography. Calcium scoring was calculated for the total coronary artery territory and patients were divided into 3 groups based on this score. Multislice computed tomography visualized 89% (365/500) of all coronary segments. The sensitivity and specificity for detection of stenoses > 50% or occlusion was 47%–92%, and 97%–100% for the calcium score. Forty of 500 segments were underestimated by multislice computed tomography, of which 39 were in the group with a calcium score > 400. Multislice computed tomography allows noninvasive angiographic evaluation of coronary artery disease with high diagnostic accuracy. However, the method strongly depends on the degree of vascular calcification and underestimates the degree of stenosis according to the calcium score. This new technology holds promise for noninvasive risk assessment in patients with known or suspected coronary artery disease.
|
INTRODUCTION
|
|---|
Early identification of coronary artery stenoses in asymptomatic patients can reduce morbidity and mortality. In 1998, 600,000 deaths due to coronary artery disease (CAD) were reported in Europe. Almost 50% of these patients died without any symptoms. Over 1 million diagnostic coronary angiograms are performed in Europe annually, only 50% of them with subsequent interventional therapy.1 Selective coronary angiography is the gold standard for assessing native coronary arteries and bypass grafts. However, as an invasive procedure, it carries a risk of 0.1% fatal and 0.96% severe complications.2 This shows the need for reliable non-invasive imaging tools for early diagnosis of CAD.
The recently developed multislice computed tomography (MSCT) is capable of rapid imaging of cardiac structures including the coronary arteries during a single breath-hold. The role of imaging has rapidly progressed from detecting calcification on non-enhanced scans to exact evaluation of stenoses even in small branch vessels.3 Noninvasive imaging methods such as electron-beam tomography (EBT) or magnetic resonance imaging have been evaluated for assessment of coronary artery patency, but they have significant limitations in reliable visualization of distal segments. The image resolution in conventional cardiac catheterization is in the range of 300 µm.4 Such high spatial resolution can, in principle, be achieved with MSCT. A great challenge is presented by the intrinsic periodic movements of the heart and the extrinsic respiratory movements that are superimposed over the cardiac movements. Both action components amount to 1–2 cm and, therefore, to 30 times the spatial resolution and approximately 5 or 10 times the vascular diameter. The displayed vascular continuity can also be interrupted by such action, simulating vascular occlusion. Turbulence evoked by calcified plaques may further complicate assessment of the degree of stenosis. Calcification of the coronary artery wall is a recognized marker of coronary atherosclerosis. Strong calcification of coronary vessels complicates assessment of the degree of the stenosis. Noncalcified atherosclerotic plaque often causes acute cardiac infarction by spontaneous rupture.
Electron-beam tomography and MSCT are particularly sensitive in detecting coronary calcification. The limitations of EBT are inadequate reproducibility, low sensitivity in noncalcified stenoses, and low 3-dimensional spatial resolution.4 In 1990, Agatston and colleagues5 managed to quantify coronary calcium in 584 patients and establish a correlation with coronary status. In 1999, Schmermund and colleagues6 developed a noninvasive EBT index through which patients with severe coronary disease could be identified by determining the Ca score of the left anterior descending artery (LAD) and the circumflex artery (CX). The significantly increased scan speed together with improved in-plane and z-resolution results in thinner collimated slice width and improved spatial and temporal resolution, which is necessary for computed tomography of the coronary arteries.7 The aim of this study was to investigate the accuracy of MSCT in detecting coronary artery stenosis according to the calcium score.
|
PATIENTS AND METHODS
|
|---|
During one year, 50 patients (36 male, 14 female) with atypical chest pain underwent both MSCT and coronary angiography. Inclusion criteria were a pretest intermediate likelihood of CAD, inconclusive electrocardiographic (EKG) findings, and more than one risk factor in combination with atypical chest pain. Informed consent was obtained from all patients. The mean interval between coronary angiography and MSCT was 7 days (range, 1–13 days). The mean age of the patients was 63.4 years. Mean heart rate was 66 beats·min–1, a beta blocker (esmolol 50 µg·min–1·kg–1) was given to all patients with a heart rate > 70 beats·min–1.
Patients were examined with an MSCT-scanner with retrospective EKG-gating that acquires 4 channels of data at a rotation time of 500 msec (Somatom Plus 4 VolumeZoom, Siemens Medical Systems, Forchheim, Germany). A native scan with a 3 mm slice thickness, 4 x 1.5 mm collimation, and an increment of 1.5 mm of the heart in the craniocaudal direction followed a thoracic topogram. All patients underwent Ca score assessment, whereby they were classified into 3 groups with scores of 0–400 (grade 1), 400–800 (grade 2) and > 800 (grade 3). Circulation time was determined by a test bolus of 20 mL nonionic contrast medium at a flow rate of 3 mL·sec–1. All examinations were performed during deep inspiration. The heart and native coronary arteries were depicted by the first contrast medium series (120 kv, 400 mAs, collimation 4 x 1 mm, slice thickness 1.5 mm, 130 mL nonionic contrast medium with a flow of 3.0 mL·sec–1), and a scanning area from the coronary outlet to the cardiac apex, beginning with the superior thoracic aperture, in the direction of the diaphragm. If no contraindication existed, a short-term beta blocker (esmolol) was applied to improve image quality. Examination time was approximately 30 seconds, with rotation time of 0.42 sec/360°. The use of a 512 x 512 matrix and a field of view of 21 cm resulted in a pixel size of 0.29 mm2. Table feed speed was 6.6 mm·sec–1 (5 mm/360°).
Raw data were calculated with retrospective EKG triggering in end-systole and end-diastole. Depending on heart rate, the basic reconstruction algorithm was either a segmented (heart rate < 65 beats·min–1) or a biphasic (heart rate > 65 beats·min–1) algorithm. The segmented algorithm permitted a temporal resolution of up to 250 msec, whereby total slice data were obtained during one heart action (RR-interval). By addition of data gained from 2 RR-intervals, the biphasic algorithm permitted a temporal resolution of up to 125 msec. All axial slices were reconstructed with a thickness of 1.25 mm, increment of 0.6 mm, and kernel B 30, corresponding to a medium-soft tissue kernel. The corresponding reconstruction interval was ascertained individually for each patient and vessel. Optimal reconstruction time was selected after making 3 reconstructions per data record, at intervals of 50 msec each. Data were digitally stored and assessed off-line. Assessment of axial slices followed with a 512 x 512 matrix on a Siemens workstation. Depending on the constellation of findings, multiplanar reconstruction, maximal intensity projection, and volume rendering were made for individual patients and vessel segments.
Two reviewers blinded to the results of the conventional coronary angiography and all clinical information evaluated the MSCT scans in a joint reading manner. Each segment was classified as either interpretable or non-interpretable, according to image quality. In the remaining segments considered patent by the observers, the vessels were screened for stenotic lesions: patent, 50%–70%, or > 90% narrowing of the luminal diameter. Vessel wall calcification was assessed visually, and quantitatively determined on an off-line workstation (Leonardo, Siemens), based on the standard built-in algorithm using an Agatston score equivalent adapted for MSCT. The diagnostic accuracy was calculated using sensitivity and specificity tables. The 95% confidence intervals for the sensitivity and specificity were calculated in correlation with the Geigy scientific tables.8 All calculations were performed with SAS/Stat version 8.2 software (SAS Institute, Cary, NC, USA).
|
RESULTS
|
|---|
The distribution of stenotic lesions is given in Table 1
. Of the 500 segments assessed, 365 (89%) were correctly evaluated (Figure 1, Tables 2 and 3). Of the 90 non-interpretable segments (Table 2), 69 were within the territory of the distal right coronary artery (15 stenoses in segment 4) and CX (37 stenoses in segment 12) as well as in the first diagonal branch of the LAD (17 stenoses in segment 9). Multislice computed tomography overestimated 5 of the 365 segments, and a > 90% stenosis in segment 5 was assessed as 50%. In the remaining 5 cases, there was deviation by a single grade. Multislice computed tomography underestimated the findings in 40 of 365 cases (Table 2). In 32 cases, there was underestimation of a single grade. Four cases were underestimated by 2 or 3 grades (Table 3).
There was a clear correlation between diagnostic accuracy and Ca grade. The accuracy of MSCT decreased with rising calcium load; 96.7% of coronary segments in patients with a grade 1 score agreed with the MSCT, while underestimation of the findings occurred in only 2.7% and overestimation in 0.8%. Patients with a grade 2 score agreed in 82% of cases, 14.8% were underestimated, and 2.3% were overestimated. In patients with Ca grade 3, 64% of cases agreed, 24% were underestimated, and 1.8% were overestimated (Table 4). With increased calcium load in the vessels, a clear reduction in diagnostic reliability of the method became evident. With good image quality, agreement was 96.25% with only 1.1% overestimated and 2.6% underestimated. With medium image quality, agreement was 80.8% with 1% of cases overestimated and 18.3% underestimated. Diagnostic reliability decreased with limited image quality, with agreement between MSCT and coronary angiography of only 61.5%; 2.6% were overestimated and 35.9% were underestimated. In our opinion, heart rate plays a decisive role, as most of the non-interpretable findings were in patients with heart rates > 75 beats·min–1 (Table 5).
There were 102 high-grade stenoses and occlusions, 30 stenoses of 70%–90%, 31 stenoses of > 90%, and 41 coronary occlusions. From the 30 cases with stenoses of 70%–90%, 28 (93%) were correctly assessed; of 31 stenoses > 90%, 25 (81%) were correctly assessed. Thirty-five (85%) of 41 occlusions were confirmed by MSCT. Within the area of the proximal right coronary artery (segments 1 and 2), 13 of 17 (76%) were correctly assessed, as well as 16 of 19 (84%) stenoses and occlusions of the proximal LAD (segments 5 and 6). The results in the area of the proximal CX were surprising as 13 of 16 (82%) stenoses were acceptably identified. All stenoses of 70%–90% and 7 of 9 with > 90% stenoses as well as all occlusions in the medial LAD were recognized (Table 6). The unchanged high specificity may be explained by the low number of inconspicuous vessel segments in the Ca grade 3 group. The sensitivity and specificity for the detection of > 50% stenosis was 92%/100% (grade 1), 80%/100% (grade 2) and 47/97% (grade 3), respectively (Figure 2, Table 7).
|
DISCUSSION
|
|---|
Multislice computed tomography sensitively identifies calcified plaques, but the extent and site of calcification often does not equate with the degree of stenosis. In our study, 365 (89%) of the 500 coronary segments were assessed correctly. Sixty-nine of 90 non-assessable segments were located within the distal coronary territory (segments 4, 9, 12), which we consider to have a clinically insignificant effect on noninvasive coronary screening. Coronary angiography will continue to be the gold standard for vessels with a small caliber. This is especially due to reduced financial resources in public health systems as, in our opinion, noninvasive coronary diagnostics are mainly suitable for risk evaluation, and differentiated coronary visualization for planning effective therapy will continue to be carried out invasively. In our study, there was 89% agreement in stenosis grading between coronary angiography and MSCT. The good results in the proximal range were surprising as 81% of stenoses were accurately recognized. This contradicts earlier MSCT results where the CX was difficult to visualize due to the slanted outlet from the left main stem.
There was 89% sensitivity for MSCT detection of stenosis. If deviations of one grade are included, the sensitivity was 97.8%. A clear relationship was seen between Ca grading and diagnostic accuracy; the reliability of MSCT decreased with increasing calcium load (Figure 3). These results confirm the high diagnostic reliability of this noninvasive method in patients without a history of CAD, and they endorse its use in routine diagnosis. In patients with coronary calcification, MSCT is limited to noninvasive follow-up. In planning interventional procedures, coronary angiography is still necessary. The sensitivity of MSCT in detecting stenosis depended on image quality. The majority of patients with moderate or adequate image quality had a heart rate < 70 beats·min–1. This highlights the need for negative chronotropic medication, and for cooperation between radiologists and cardiologists.
Most studies have been performed with EBT which has the advantage of a very short examination time of 100 msec; however, the slice thickness of 3 mm is a disadvantage. So far, EBT studies on graft occlusions are limited by the small number of cases. In two studies, the sensitivity of EBT in detecting arterial bypass occlusions was 80% and 95%, and 92% and 100% for vein graft occlusions.9 Multislice computed tomography has been developed as an alternative to EBT, mainly due to lower costs and a higher degree of implementation by radiology departments. In contrast to EBT, MSCT features a longer exposure time of 250 msec while simultaneously acquiring 4 slices per heart beat, and the effective slice thickness is superior to that of EBT. This is reflected in improved resolution of 9 line pairs/cm compared to 6 line pairs/cm for EBT.10 Electron-beam tomography can be performed in patients with a heart rate > 70 beats·min–1 without the need for chronotropic medication. Leber and colleagues10 examined 101 patients with coronary stenoses > 50% and limited these to a 9-step segment model where the peripheral coronary segments were excluded; 77% of segments were interpretable with EBT and 82% of segments were interpretable with MSCT. Of the stenoses with 50% luminal narrowing, 76% were identified by EBT and 82% by MSCT. Exclusion of high-grade stenoses was possible with 93% reliability by EBT and 96% by MSCT.10 In 102 patients, Kopp and colleagues6 reached a sensitivity of 86% to 93%, as well as specificity of 98% to 99%. According to Achenbach,11 sensitivity of 4-line technology is between 72% and 93%, and specificity ranges from 84% to 98%. However, 32% of all coronary segments had to be excluded due to low image quality. Heart rate is decisive for image quality; sensitivity decreases from 91% to 33% with an increase of heart rate from < 60 to > 79 beats·min–1.12 Our study confirmed this trend as the lowest accuracy was obtained in arrhythmic patients with heart rates > 75 beats·min–1. Therefore, this patient group should not be examined by MSCT, as radiation dose and diagnostic information are inconsistent.
Several studies have confirmed the strong correlation between Ca score, CAD characteristics, and degree of stenosis.9 The high negative predictive value of the method should be highlighted here: if no calcium plaques are assessed by MSCT, the risk of coronary disease is approximately 6%.5 Segmental identification of even high-grade coronary calcification does not permit valid conclusions about the existence and severity of stenoses.13 Breen and colleagues4 evaluated 100 patients with EBT and coronary angiography. The sensitivity and specificity for detecting calcified deposits in patients with significant stenosis were 100% and 47%. The largest reported series of 710 patients with symptomatic CAD showed a sensitivity of 95%, and specificity of 44% in detecting calcification as an indicator of > 50% stenosis.14 The absence of calcification implies no angiographically significant coronary stenosis; however, it does not exclude unstable atherosclerotic plaques. Plaques in heavily calcified segments were not considered at increased risk of rupture and thus no increase in cardiac risk.15,16 On the other hand, new studies have proved that so-called "soft plaques" are more significant in potential for spontaneous rupture and coronary risk. With a combination of unenhanced and contrast-enhanced CT angiography, stenoses are better assessed within coronary vessels. Unlike coronary angiography, MSCT is not just a matter of endoluminal vessel imaging; adjacent wall structures are also depicted. In this way, plaques containing lipids and mixed fibrous plaques can be assessed. These results correlate well with intravascular ultrasound scans.17
Multislice computed tomography should not be implemented as a single modality as sensitivity, specificity, and spatial resolution still remain lower than those of coronary angiography. It might be indicated in patients with an intermediate likelihood of CAD, in symptomatic patients after coronary angioplasty, following a noninvasive stress test, and in high-risk individuals. The implementation of 16/64-line technology should eliminate the limitations of low resolution in the future. Certain invasive procedures may be replaced by MSCT if significant stenoses in vessel sections accessible to revascularization can be reliably eliminated. This has to be verified by large randomized studies, which so far have not been performed for 4- or 16-slice MSCT. Severe calcification is one of the main reasons for non-diagnostic images of coronary arteries acquired by 4-slice systems. Early reports described how minor calcifications impaired image quality because lumen obstruction could not be sufficiently visualized.18,19 The spatial resolution of 4-slice systems provides assessment of structures > 0.9 to 1 mm. With 16-slice techniques, maximal spatial resolution of 0.5 x 0.5 x 0.6 mm is possible.3 This seems to be sufficient to assess the lumen of the relevant coronary segments. The introduction of sub-millimeter detector rows is expected to improve the assessment of severely calcified and stenosed small distal branches. Multislice computed tomography has seen continued and rapid technical development since its inception. Higher spatial resolution, contrast definition, and advances in motion correction have led the way to its routine use in evaluating CAD. The ongoing advances with 16- and 64-slice MSCT may further increase its potential role in cardiac risk stratification.20
|
REFERENCES
|
|---|
- Windecker S, Maier-Rudolph W, Bonzel T, Heyndrickx G, Lablanche JM, Morice MC, et al. Interventional cardiology in Europe 1995. Working Group Coronary Circulation of the European Society of Cardiology. Eur Heart J 1999;20:484–95.[Abstract/Free Full Text]
- Bach R, Espinola-Klein C, Ozbek C, Jung F, Schieffer H. Complications after coronary angiography and balloon dilatation. Dtsch Med Wochenschr 1993;118:1669–76.[Medline]
- Kuettner A, Trabold T, Schroeder S, Feyer A, Beck T, Brueckner A, et al. Noninvasive detection of coronary lesions using 16-detector multislice spiral computed tomography technology: initial clinical results. J Am Coll Cardiol 2004;44:1230–7.[Abstract/Free Full Text]
- Breen JF, Sheedy PF 2nd, Schwartz RS, Stanson AW, Kaufmann RB, Moll PP, et al. Coronary artery calcification detected with ultrafast CT as an indication of coronary artery disease. Radiology 1992;185:435–9.[Abstract/Free Full Text]
- Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827–32.[Abstract]
- Kopp AF, Schroeder S, Kuettner A, Baumbach A, Georg C, Kuzo R, et al. Non-invasive coronary angiography with high resolution multidetector-row computed tomography. Results in 102 patients. Eur Heart J 2002;23:1714–25.[Abstract/Free Full Text]
- Schmermund A, Denktas AE, Rumberger JA, Christian TF, Sheedy PF 2nd, Bailey KR, et al. Independent and incremental value of coronary artery calcium for predicting the extent of angiographic coronary artery disease: comparison with cardiac risk factors and radionuclide perfusion imaging. J Am Coll Cardiol 1999;34:777–86.[Abstract/Free Full Text]
- Geigy scientific tables. Vol. 2, statistical tables. 8th ed. Basle: Ciba-Geigy, 1985:224.
- Stanford W. Coronary artery calcification as an indicator of preclinical coronary artery disease. Radiographics 1999;19:1409–19.[Free Full Text]
- Leber AW, Knez A, Becker C, Becker A, White C, Thilo C, et al. Non-invasive intravenous coronary angiography using electron beam tomography and multislice computed tomography. Heart 2003;89:633–9.[Abstract/Free Full Text]
- Achenbach S. Clinical Use of Multi-Slice CT Coronary Angiography. Herz 2003;28:119–25.[Medline]
- Mennicke M, Giesler T, Ropers D, Baum U, Ulzheimer S, Wenkel E, et al. Influence of heart rate on image quality and detection of coronary stenoses with multislice spiral CT. Biomed Tech (Berl) 2002;47 Suppl 1 Pt 2:782–5.
- Richardson PD, Davies MJ, Born GV. Influence of plaque configuration and stress distribution on fissuring of coronary atherosclerotic plaques. Lancet 1989;2(8669):941–4.[Medline]
- Budoff MJ, Georgiou D, Brody A, Agatston AS, Kennedy J, Wolfkiel C, et al. Ultrafast computed tomography as a diagnostic modality in the detection of coronary artery disease: a multicenter study. Circulation 1996;93:898–904.[Abstract/Free Full Text]
- Wexler L, Brundage B, Crouse J, Detrano R, Fuster V, Maddahi J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications. A statement for health professionals from the American Heart Association Writing Group. Circulation 1996;94:1175–92.[Free Full Text]
- Cheng GC, Loree HM, Kamm RD, Fishbein MC, Lee RT. Distribution of circumferential stress in ruptured and stable atherosclerotic lesions. A structural analysis with histopathological correlation. Circulation 1993;87:1179–87.[Abstract/Free Full Text]
- Kopp AF, Schroeder S, Baumbach A, Kuettner A, Georg C, Ohnesorge B, et al. Non-invasive characterisation of coronary lesion morphology and composition by multislice CT: first results in comparison with intracoronary ultrasound. Eur Radiol 2001;11:1607–11.[Medline]
- Achenbach S, Giesler T, Ropers D, Ulzheimer S, Derlien H, Schulte C, et al. Detection of coronary artery stenoses by contrast-enhanced, retrospectively electrocardiographically-gated, multislice spiral computed tomography. Circulation 2001;103:2535–8.[Abstract/Free Full Text]
- Becker CR, Ohnesorge BM, Schoepf UJ, Reiser MF. Current development of cardiac imaging with multidetector-row CT. Eur J Radiol 2000;36:97–103.[Medline]
- Jakobs TF, Becker CR, Wintersperger BJ, Herzog P, Ohnesorge B, Flohr T, et al. CT angiography of the coronary arteries with a 16-row spiral tomography. Effect of spatial resolution on image quality. Radiologe 2002;42:733–8.[Medline]
TOP
ABSTRACT
INTRODUCTION
