Author + information
- Received January 16, 2013
- Revision received April 5, 2013
- Accepted April 12, 2013
- Published online October 1, 2013.
- Nuno Bettencourt, MD∗,†,‡∗ (, )
- Nuno Dias Ferreira, MD∗,
- Daniel Leite, MD∗,
- Mónica Carvalho∗,
- Wilson da Silva Ferreira∗,
- Andreas Schuster, MD, PhD§,
- Amedeo Chiribiri, MD, PhD†,
- Adelino Leite-Moreira, MD, PhD‡,
- José Silva-Cardoso, MD, PhD‡,
- Eike Nagel, MD, PhD† and
- Vasco Gama, MD∗
- ∗Cardiology Department, Centro Hospitalar de Gaia/Espinho, Vila Nova de Gaia, Portugal
- †Division of Imaging Sciences and Medical Engineering, King's College London, London, United Kingdom
- ‡Cardiovascular R&D Unit, Faculty of Medicine, University of Porto, Porto, Portugal
- §Georg-August-University and German Center for Cardiovascular Research (DZHK, Partner Site), Göttingen, Germany
- ↵∗Reprint requests and correspondence:
Dr. Nuno Bettencourt, Cardiology Department, Centro Hospitalar de Vila Nova de Gaia/Espinho EPE, Rua Conceição Fernandes, 4434-502 Vila Nova de Gaia, Portugal.
Objectives This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) for detection of ischemic scar and to test the additive value of CTDE as part of a comprehensive multidetector computed tomography (MDCT) stress–rest protocol including computed tomography perfusion (CTP) and computed tomography angiography (CTA) for the diagnosis of significant coronary artery disease (CAD).
Background CTDE has been recently described as a promising tool for noninvasive detection of myocardial scar, similarly to CMR-LGE techniques. Despite its theoretical potential as an adjunctive tool to improve MDCT accuracy for detection of CAD, its clinical performance has not been validated.
Methods One hundred five symptomatic patients with suspected CAD (age 62.0 ± 8.0 years, 67% men) underwent MDCT, CMR, and x-ray invasive coronary angiography. The MDCT protocol consisted of calcium scoring, stress CTP under adenosine 140 μg/kg/min, rest CTP + CTA, and a low-dose radiation prospective scan for detection of CTDE. CMR-LGE was used as the reference standard for assessment of scar. Functionally significant CAD was defined as the presence of ≥90% stenosis/occlusion or fractional flow reserve measurements ≤0.80 in vessels >2 mm.
Results CTDE had good accuracy (90%) for ischemic scar detection with low sensitivity (53%) but excellent specificity (98%). Positive and negative predictive values were 82% and 91%, respectively. On a patient-based model, MDCT protocol without integration of CTDE results had a sensitivity, specificity, and positive and negative predictive values of 90%, 81%, 80%, and 90%, respectively, for the detection of functionally significant CAD. Addition of CTDE results did not improve MDCT performance (90%, 77%, 77%, and 90%, respectively).
Conclusions CTDE has moderate accuracy for detection of ischemic scar in patients with suspected CAD. Integration of CTDE into a comprehensive MDCT protocol including stress–rest CTP and CTA does not improve MDCT accuracy for detection of significant CAD in intermediate-to-high pre-test probability populations.
Multidetector computed tomography (MDCT) coronary angiography represents the noninvasive gold standard for the assessment of the coronary arterial tree. It is particularly useful for the exclusion of coronary artery disease (CAD) in patients with intermediate/low pre-test probability, largely because of its high negative predictive value (1). However, in patients with higher pre-test probability, its performance is limited because the physiological significance of many lesions cannot be assessed (2). To overcome this limitation, MDCT stress–rest myocardial perfusion techniques (computed tomography perfusion [CTP]) have been described (3–6), and integrated protocols, providing both morphological (computed tomography angiography [CTA]) and functional (CTP) information in a single MDCT exam, have been tested (7–13).
MDCT ability to identify myocardial ischemic scar using computed tomography delayed enhancement (CTDE) has also been reported. This technique follows the same principles applied to cardiovascular magnetic resonance (CMR) late gadolinium enhancement (LGE) and could be particularly valuable as part of a MDCT protocol including CTA and CTP. The feasibility of performing such a comprehensive examination has already been shown, but CTDE's potential as an adjunctive tool to improve MDCT accuracy for the detection of CAD has never been validated (14).
The aim of this study was to evaluate the diagnostic performance of low-radiation-dose CTDE for detection of ischemic scar using CMR-LGE as the standard and to test the additive value of CTDE as part of an comprehensive MDCT protocol, including CTA and stress–rest CTP, for the diagnosis of functionally significant CAD, using invasive x-ray coronary angiography (XA) with fractional flow reserve (FFR) evaluation as the reference standard.
We prospectively screened 176 consecutive patients with suspected CAD referred by general physicians to our hospital outpatient cardiology clinic from February 2010 to November 2011. Inclusion criteria were age >40 years, symptoms compatible with CAD, and at least 1 of the following: ≥2 risk factors or a positive/inconclusive treadmill test. Exclusion criteria included unstable clinical status, known CAD, valvular heart disease, atrial fibrillation, creatinine clearance ≤60 ml/min, and standard contraindications to CMR, contrast media, and adenosine. A total of 139 eligible patients were tested for exclusion criteria. Figure 1 summarizes the study flow and reasons for exclusions. Characteristics of the final population are summarized in Table 1.
After informed consent, patients were scheduled for CMR and MDCT in the week before XA. FFR was measured in all major patent epicardial coronary arteries with intermediate diameter stenoses (50% to 90%) as assessed by quantitative coronary angiography (QCA). CMR and MDCT results were fully blinded.
CMR was performed using established protocols on a 1.5-T Siemens Symphony Tim (Siemens, Erlangen, Germany) using a 12-channel receiver coil (15). Long- and short-axis cine images were obtained using a steady-state free precession breath-hold sequence for volumetric and functional analysis. LGE imaging was performed using a 2-dimensional phase-sensitive inversion-recovery breath-hold sequence ≥10 min after administration of contrast (0.2 mmol/kg). The entire volume of the heart was covered in 8-mm-thick short-axis projections with a gap of 2 mm between slices, and in standard long-axis cardiac planes.
MDCT comprehensive protocol
MDCT stress–rest protocol was performed as previously published, with the addition of a low-radiation scan for CTDE detection (Fig. 2) (8). All scans were performed using a Somatom Sensation 64 scanner (Siemens Medical Solutions, Forchheim, Germany) with no pre-test medication. The comprehensive MDCT protocol included 4 sequential acquisitions: calcium scoring, stress CTP, rest CTP, and CTDE. Table 2 summarizes the scan parameters of each portion of the protocol.
First, a low-dose prospective scan to assess coronary artery calcification was performed using established protocols (16). Then, adenosine infusion (140 μg·kg−1·min−1 for 3 to 6 min) was started, and a retrospectively gated scan was acquired during the first passage of contrast medium, using a bolus-tracking technique. Adenosine infusion was discontinued immediately after stress acquisition. Intravenous metoprolol targeting a heart rate ≤60 beats/min was administered in patients whose heart rate exceeded 65 beats/min 3 min after the suspension of the adenosine infusion. All patients received 0.5 mg of sublingual nitroglycerin before the rest scan. The latter was acquired 10 min after the first contrast injection, using prospective triggering. Timing and contrast administration were similar to the stress scan, using a test-bolus technique. To detect areas of CTDE, a fourth scan was performed 7.0 ± 0.3 min after the rest scan using prospective triggering at 65% of the RR interval. For this acquisition, a fixed tube voltage of 80 kV and a tube current of 160 mA were used, and collimation was increased to 1.2 mm. No additional contrast was administered.
Two blinded independent readers analyzed all CMR images. In cases of disagreement, a third reader adjudicated. Each of the 17 segments was classified on the basis of the presence and transmurality (subendocardial vs. transmural) of scar—defined as areas of myocardial enhancement using LGE imaging. Ischemic scar was assumed when subendocardial involvement was noted. Image quality and the degree of confidence in scar detection were classified independently using 4-class (0 to 3) scales: from poor to excellent and from very unconfident to very confident, respectively.
MDCT images were analyzed using dedicated software (Aquarius Intuition version 4.4.6, TeraRecon, Foster City, California) on dedicated workstations by 2 independent blinded readers. Each component of the MDCT scan was analyzed independently at a different time point of the study.
Coronary calcification was calculated using an effective slice thickness of 3 mm with a detection threshold of 130 HU and reported as the mean Agatston score. For CTA and CTP analysis, both stress and rest acquisitions were evaluated using sets of 11 retrospective phases from the stress scan, and a single-phase (65%) reconstruction from the rest scan (8). Soft (Siemens-B25f) and very smooth (Siemens-B10f) frequency filters were used for CTA and CTP reconstructions, respectively. Resulting CTA datasets were analyzed for detection of CAD according to the 17-segment modified American Heart Association classification (17). On the basis of the information obtained from both stress and rest reconstructions, each segment was graded, according to stenoses: 1 = normal, 2 = <50%, 3 = 50% to 70%, 4 = ≥70%/occlusion, 5 = uninterpretable. Analysis of CTP was performed according to the standard 17-segment model, using 10-mm-thick multiplanar reformat planes (short-axis and 2, 3, and 4 chambers) (18). Stress images were analyzed as cines, integrating perfusion with regional wall motion information and compared with rest.
For CTDE analysis, a 65% single-phase reconstruction was obtained using a very smooth filter (Siemens-B10f) and a slice thickness of 1.2 mm. Reading was performed using the same multiplanar reformat planes and reporting model (18). It was typically initiated using 10-mm-thick average intensity projections, and set, by protocol, to narrow window and level settings (W300, L150). The reading physician was allowed to adjust these display settings after an initial exploratory reading, and to change slice thickness and projections as needed. Each of the 17 segments was classified based on the presence and transmurality (subendocardial vs. transmural) of CTDE, defined as areas of myocardial enhancement when compared with remote regions of the myocardium. Readers were asked to determine whether ischemic scar was present and to classify image quality and degree of confidence, as described for CMR. Interobserver disagreements were resolved by consensus. Signal-to-noise ratio (mean signal intensity/standard deviation of signal intensity) of the CTDE scan was estimated using a 10-mm2 region of interest in the left ventricle.
MDCT Radiation Exposure Estimation
Effective radiation dose exposure for each component of the MDCT comprehensive protocol (calcium score, stress CTA/CTP, rest CTA/CTP, and CTDE) was calculated by the method proposed by the European Working Group for Guidelines on Quality Criteria in CT: product of the chest coefficient (0.014) and the dose-length product obtained during each scan (19).
XA and FFR assessment
XA was performed according to standard techniques. Excluding the left main, arteries with a caliber >2 mm and intermediate stenosis (diameter stenosis ≥50% and <90%) as assessed by QCA (Siemens Leonardo XP, Siemens AG, Munich, Germany, and IC3D v188.8.131.52A software, Paieon Medical, Rosh Ha'ayin, Israel) were evaluated using pressure wire (PressureWire Certus, St. Jude Medical, St. Paul, Minnesota). FFR was determined by RadiAnalyzer (St. Jude Medical) under steady-state hyperemia, obtained with intravenous adenosine (140 μg/kg/min) infusion over 3 to 6 min. Arteries were recorded as having significant flow-limiting disease if they had stenoses ≥90% (≥50% in the left main stem) or had a FFR value ≤0.80 in vessels >2 mm.
CTDE per-patient performance for the detection of ischemic scar was tested against CMR-LGE as the reference standard. The potential value of integrating CTDE with the combined CTA + CTP analysis was also tested as an additive tool to increase MDCT performance for the detection of functionally significant CAD as assessed by XA + FFR. The diagnostic performances of each component of MDCT (CTA, CTP, and CTDE), as well as the integration of CTA + CTP (integrated protocol) and CTA + CTP + CTDE (comprehensive protocol) were compared against XA + FFR as the reference standard. Ninety-five percent confidence intervals (CI) were calculated based on the binominal distribution. “Nonevaluable” coronary segments in CTA were coded as being positive for CAD when CTA alone was considered; in the integrated protocol, they were classified as negative or positive for functionally significant CAD, according to the CTP results of their territory. In the comprehensive protocol, these vessels were coded positive if scar was detected in the corresponding territory; if not, they were coded according to the CTP results, similar to the integrated protocol (Fig. 3). CMR-LGE and CTDE intraobserver and interobserver agreements were tested using Cohen's kappa statistic.
All data are described as means and standard deviations for continuous variables and as percentages for categorical variables. Differences in continuous variables were assessed using Student paired t tests. The area under the receiver-operating characteristic (ROC) curve was calculated for all diagnostic-testing strategies using XA + FFR assessment as the gold standard. A p value <0.05 was considered significant. Data analysis was performed using SPSS analysis software (Release 17, SPSS, Chicago, Illinois).
The study protocol was approved by the local research ethics committee and complies with the Declaration of Helsinki. Written informed consent was obtained from all participants.
The final population consisted of 105 symptomatic individuals (age 62 ± 8 years; 67% men) with an intermediate or high pre-test probability according to the modified Diamond-Forrester score (20).
All CMR and MDCT scans were performed within 9.0 ± 7.7 days before XA, and all patients completed the study protocol without adverse effects.
Myocardial LGE was visualized in 24 patients; 17 of those had an ischemic pattern (Fig. 4). In 7, however, an intramural/subepicardial pattern was detected. Intraobserver and interobserver agreements for ischemic scar detection were very good (kappa = 0.77 and 0.66, respectively).
Among the 105 CTA examinations, 33 (31%) had at least 1 nonevaluable segment—usually because of the presence of extensive calcification. Among the patients who had fully interpretable scans, 10 had no atherosclerotic disease, 24 had mild disease (<50% stenosis), and 38 had significant stenosis (≥50%). When the nonevaluable segments were considered to represent disease, 71 (68%) patients were categorized as having significant CAD. CTP defects were identified in 38 patients (36%) and in 53 (17%) vascular territories.
Myocardial CTDE was described in 13 patients (12%), including 2 with a nonischemic pattern. Frequently, areas of CTDE had the same density as the blood pool, appearing in the short-axis plane as complete or subendocardial interruptions of the circular shape of the myocardium (Fig. 4). Of the 11 patients with an ischemic CTDE pattern, 9 had evidence of perfusion defects on rest CTP, whereas 2 were normal. The CTDE image quality of most of the scans was classified as good (57) or excellent (27), and only 1 scan was classified as poor. Readers felt confident or very confident in the vast majority of cases (62 or 31 cases, respectively); moderately low confidence was reported in 12 cases. Compared with the scans of patients weighing ≤80 kg, CTDE scans of patients >80 kg had a lower signal-to-noise ratio (5.3 ± 2.23 vs. 7.3 ± 2.65; p = 0.003), worse image quality (1.6 ± 0.61 vs. 1.9 ± 0.64; p = 0.03), and lower reported confidence (1.8 ± 0.52 vs. 2.04 ± 0.51; p = 0.02).
Mean radiation exposure of the entire MDCT protocol was 5.5 ± 0.95 mSv (3.9 to 9.9 mSv). The CTDE scan was responsible for only 0.50 ± 0.10 mSv of radiation exposure. Estimated effective radiation exposure for each component of the MDCT protocol is described in Table 2. Good reproducibility of CTDE analysis was found, with very good intraobserver (kappa = 0.78) and interobserver agreement (kappa = 0.76).
XA and FFR results
Seventy patients had some degree of coronary stenosis on visual analysis and were evaluated using QCA: of those, 59 had stenosis ≥50%, including 20 with total occlusions. Nineteen patients with intermediate stenosis in vessels >2 mm were evaluated by FFR. Using this approach, a total of 48 patients were classified as having functionally significant CAD: single-vessel disease was seen in 24 patients, 16 had double-vessel disease, and 8 had triple-vessel disease. Left main disease was found in 5 of these patients.
CTDE performance for ischemic scar detection using CMR-LGE as a reference
CTDE accurately detected ischemic scar in 9 of 17 patients identified by CMR. On a per-patient level, CTDE had good accuracy (90%) for ischemic scar detection with low sensitivity (53%) but excellent specificity (98%). CTDE performed better in patients weighting ≤80 kg and was not significantly affected by heart rate during acquisition (Table 3).
MDCT performances for detection of functionally significant CAD are summarized in Table 4. Isolated CTA analysis had an excellent sensitivity and negative predictive value (100%). However, specificity and positive predictive value were low (60% and 68%, respectively). CTP, conversely, had higher specificity (93%), at a cost of lower sensitivity (71%). The integration of data from CTA and CTP resulted in a sensitivity of 90% and specificity of 81%. Addition of scar information as detected by CTDE (CTA + CTP + CTDE) (Fig. 3) did not improve overall MDCT accuracy (sensitivity 90%, specificity 77%).
The ROC analysis for the prediction of functionally significant CAD is represented in Figure 5. The areas under the curve for significant CAD detection were 0.80 (95% CI: 0.71 to 0.89) for CTA alone, 0.82 (95% CI: 0.73 to 0.91) for CTP alone, 0.58 (95% CI: 0.47 to 0.69) for CTDE alone, 0.85 (95% CI: 0.77 to 0.93) for the CTA + CTP integrated protocol, and 0.83 for the comprehensive protocol, including analysis of CTDE (95% CI: 0.75 to 0.92).
The main findings of our study are that: 1) low-radiation CTDE performed immediately after a stress–rest MDCT protocol is capable of scar detection with reasonable accuracy but low sensitivity; and 2) the addition of CTDE to a stress–rest CTA + CTP integrated protocol does not improve the global accuracy of MDCT for the detection of functionally significant CAD in patients with intermediate-to-high pre-test probability.
We and others have previously shown that integration of CTP with CTA improved diagnostic accuracy of MDCT in patients with intermediate-to-high pre-test probability, mainly because of an increased specificity in heavily calcified coronary arteries (7–12). In this study, we added CTDE analysis to our integrated CTA + CTP protocol, aiming to test its potential as an additive tool for the diagnosis of CAD. A similar approach has been proposed for CMR using LGE to improve the accuracy of stress perfusion, but currently, this algorithm is seldom used (21).
In our study, ischemic scar was documented in a significant proportion of patients using CMR-LGE; however, CTDE was only able to detect about one-half of those cases. Furthermore, the addition of CTDE to the integrated protocol including CTA and CTP did not improve MDCT performance for the detection of functionally significant CAD; this occurred because the majority of patients with ischemic scar also had reversible defects compatible with ischemia (already detected as positives) and because some patients had ischemic scar with normal coronary arteries. Inclusion of CTDE into the comprehensive MDCT protocol had a limited, but positive, impact because ischemic scar was detected in 2 patients with normal arteries and no perfusion defect on rest CTP. However, the potential advantage of a better prognostic and therapeutic management has to be balanced against the extra time needed for the CTDE scan and radiation exposure.
Only 1 previous study analyzed CTDE in the context of a comprehensive MDCT protocol, including CTA and stress/rest CTP: the final population of 34 patients was selected from those who underwent a nuclear stress test and XA within 3 months, including a significant proportion of patients with known CAD (14). Integration of CTDE did not improve MDCT ability to detect CAD, as defined by stenosis ≥70%. Differently from Blankstein et al. (14), we only included intermediate/high pre-test probability symptomatic patients without known history of CAD and used FFR as the reference standard. Additionally, CMR-LGE was used as the standard to individually evaluate CTDE performance in scar detection. The entire MDCT comprehensive protocol, including CTA, CTP, and CTDE, was completed with an effective radiation exposure that represents less than one-half of the exposure usually reported for nuclear studies. The CTDE acquisition represents a very small proportion of that value, being responsible for an average radiation exposure of as low as 0.5 ± 0.1 mSv.
CMR has the unique capability of scar detection using LGE techniques, which allow an optimized contrast between scar and myocardium. Iodinated contrast agents used in MDCT have similar kinetics to the gadolinium chelates used in CMR (22). Thus, images acquired 4 to 30 min after contrast administration may show regional hyperenhancement, corresponding to areas of myocardial scar, similar to CMR-LGE (23–25). Despite no consensus regarding the optimal protocol for contrast administration and timing of CTDE imaging, our methods are in line with previous studies. Contrast doses of 120 to 150 ml or 2 ml/kg and delay times ranging from 5 to 15 min after contrast material injection have been described (25,26). Differently from CMR-LGE, where the contrast may be optimized using specific inversion times selected to null the normal myocardium, CTDE has to rely on different tones of gray, corresponding to different Hounsfield units, according to the different degrees of radio-opacity of the tissues. CTDE areas do not appear as highly contrasted bright zones surrounded by dark myocardium. In fact, in our population, CTDE was most commonly detected as segmental areas where the myocardium could no longer be defined and differentiated from the gray area corresponding to the blood pool. Visual detection of those areas may be difficult, especially when slight subendocardial involvement is present; this may explain the relatively low sensitivity when compared to CMR. Furthermore, it has been suggested that the use of a low-dose approach by setting tube voltage to 80 kV, despite the intrinsic advantage of better signal-to-noise and contrast-to-noise ratios (27–30), may be associated with an insufficient image quality, especially in obese patients (30,31). In this context, Habis et al. (32) proposed a strategy of tube voltage adjustment according to weight (80 kV for patients ≤80 kg and 100 kV for patients weighting >80 kg). Our results seem to support this strategy because image quality, readers' confidence, and CTDE diagnostic performance for detection of scar were lower in patients weighting >80 kg.
In this single-center study, only symptomatic patients without known CAD and an intermediate-to-high pre-test probability were included. A small percentage of patients had to be excluded because of contraindications, such as renal dysfunction or arrhythmia. As such, these results may not apply to other groups of patients referred for CTA. FFR was only measured in vessels with intermediate stenosis on QCA assessment. Stenoses with QCA <50% were assumed to be irrelevant, and stenoses ≥90% were considered functionally significant. Although this was performed to avoid potential iatrogenic complications, and reflects current clinical practice in many centers, it may lead to a small bias. Another limitation concerning study design is the performance of 2 contrasted scans a few minutes apart: this assures clinical applicability but is associated with some drawbacks related to contrast redistribution and optimization of heart rate for the CTA scan. Furthermore, CTDE was tested using a very low radiation protocol without the use of any additional contrast. Despite the appeal of this approach, because it could be easily implemented in a comprehensive MDCT protocol without increase of patient risks, it is possible that CTDE could be optimized using different parameters—namely higher tube current and current intensity or dedicated injections of higher doses of contrast—as the ideal CTDE protocol is still under research. Post-processing software may also be an important tool for optimizing scar detection using MDCT. Nevertheless, our study shows that CTDE integration into a MDCT comprehensive protocol—using commercially standard and widely available software and hardware—is feasible and may inform about the presence of scar in patients with suspected CAD.
The authors acknowledge financial support from the Department of Health via the National Institute for Health Research comprehensive Biomedical Research Centre award to Guy's and St Thomas' NHS Foundation Trust in partnership with King's College London and King's College Hospital NHS Foundation Trust. Dr. Bettencourt was funded by Fundação para a Ciência e Tecnologia, Portugal, under grant number SFRH/BD/45989/2008 and received grant support from the Portuguese Society of Cardiology and the European Society of Cardiology. Dr. Schuster acknowledges grant support from the British Heart Foundation (FS/10/029/28253; RE/08/003) and the Biomedical Research Centre (BRC-CTF 196). Dr. Chiribiri was funded by the Wellcome Trust and Engineering and Physical Sciences Research Council under grant number WT 088641/Z/09/Z; he has received grant support from Philips Healthcare. Dr. Nagel has received grant support from Philips Healthcare and Bayer Healthcare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- cardiac magnetic resonance
- computed tomography angiography
- computed tomography delayed enhancement
- computed tomography perfusion
- fractional flow reserve
- late gadolinium enhancement
- multidetector computed tomography
- quantitative coronary angiography
- receiver-operating characteristic
- x-ray invasive coronary angiography
- Received January 16, 2013.
- Revision received April 5, 2013.
- Accepted April 12, 2013.
- 2013 American College of Cardiology Foundation
- Meijboom W.B.,
- Van Mieghem C.A.G.,
- van Pelt N.,
- et al.
- George R.T.,
- Arbab-Zadeh A.,
- Miller J.M.,
- et al.
- Feuchtner G.,
- Goetti R.,
- Plass A.,
- et al.
- Ko S.M.,
- Choi J.W.,
- Hwang H.K.,
- Song M.G.,
- Shin J.K.,
- Chee H.K.
- Ko B.S.,
- Cameron J.D.,
- Leung M.,
- et al.
- Bettencourt N.,
- Chiribiri A.,
- Schuster A.,
- et al.
- Blankstein R.,
- Shturman L.D.,
- Rogers I.S.,
- et al.
- Cerqueira M.D.,
- Weissman N.J.,
- Dilsizian V.,
- et al.
- ↵Bongartz G, Golding SJ, Jurik AG, et al. European Guidelines for Multislice Computed Tomography. Funded by the European Commission. Contract number FIGM-CT2000-20078-CT-TIP. March 2004. Available at: http://www.msct.eu/CT_Quality_Criteria.htm. Accesed November 14, 2012.
- Klem I.,
- Heitner J.F.,
- Shah D.J.,
- et al.
- Gerber B.L.,
- Belge B.,
- Legros G.J.,
- et al.
- Rodriguez-Granillo G.A.,
- Rosales M.A.,
- Baum S.,
- et al.
- Mahnken A.H.,
- Koos R.,
- Katoh M.,
- et al.
- Habis M.,
- Capderou A.,
- Ghostine S.,
- et al.