50
Integrated Cardiothoracic Imaging with Computed Tomography Lucas L. Geyer, MD1,2 Justin R. Silverman, BA1 Aleksander W. Krazinski, BA1 Pal Suranyi, MD, PhD1 James G. Ravenel, MD1 Stefan Wirth, MD, PhD, MSc2 Philip Costello, MD1 U. Joseph Schoepf, MD1 1 Department of Radiology and Radiological Science, Medical
University of South Carolina, Charleston, South Carolina 2 Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital, Munich, Germany
Address for correspondence U. Joseph Schoepf, MD, Heart and Vascular Center, Medical University of South Carolina, Ashley River Tower, MSC 226, 25 Courtenay Dr, Charleston, SC 29401 (e-mail: [email protected]).
Abstract Keywords
► ► ► ►
chest imaging heart disease lung diseases cardiopulmonary comorbidities ► computed tomography angiography ► dual-source computed tomography
The respiratory and the cardiovascular systems are intimately connected. Because of the high degree of morphological and functional interaction, pathophysiological processes in one compartment are likely to induce adaptive changes in the other. Computed tomography (CT) plays a central role in the diagnostic work up of both thoracic and cardiac disorders. Historically, these two systems have been evaluated separately; however, CT technology has evolved remarkably over recent decades. Up-to-date advanced imaging strategies allow for a combined assessment of the cardiopulmonary unit. Besides improved techniques of electrocardiogram (ECG)-synchronization for obtaining both morphological and functional information, latest advances of dualsource CT (DSCT) have shown great promise for even more comprehensive integrated cardiothoracic imaging.
Diseases of the respiratory and the cardiovascular system are leading causes of morbidity and mortality in the Western Hemisphere (1, 2). Chest computed tomography (CT) plays a central role in the diagnostic work up of known or suspected thoracic pathology. Because of their inherent interrelationship, pathophysiological changes in the pulmonary system are likely to induce adaptations in the intimately connected cardiovascular system, and vice versa.1 Therefore, evaluation for concomitant cardiovascular disorders should be part of the interpretation of any chest CT. As thoracic CT imaging is usually not performed in sync with cardiac activity (non-ECG synchronized data acquisition), the simultaneous assessment of the heart is often limited by cardiac motion. However, CT technology has substantially evolved over the last decades, often allowing a degree of cardiac assessment even with routine chest CT protocols. Depending on the clinical scenario, there also exist more advanced image acquisition strate-
Issue Theme Thoracic Imaging; Guest Editor, Martine Remy-Jardin, MD, PhD
gies that integrate ECG-synchronization for detailed cardiac and coronary imaging (►Fig. 1). The purpose of this contribution, thus, is (1) to summarize the technical developments of CT imaging which enabled integrated cardiopulmonary imaging, (2) to discuss appropriate scan protocols, and (3) to provide context through a review of common cardiopulmonary disorders.
Technical Evolution of Cardiothoracic Imaging Reducing Cardiac Motion Artifact In the early years of thoracic CT imaging, image interpretation was frequently hampered by respiratory or cardiac motion artifacts. Whereas respiratory motion induces extensive blurring, cardiac motion leads to more subtle changes by transmitting minor pulsations to adjacent mediastinal struc-
Copyright © 2014 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0033-1363451. ISSN 1069-3424.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Semin Respir Crit Care Med 2014;35:50–63.
Geyer et al.
Fig. 1 Techniques of electrocardiography (ECG)-synchronization. The width of the gray bars reflects the duration of data acquisition in relation to the cardiac cycle; the amplitude of the gray bars reflects the tube current applied during the scan. (a) Retrospective ECG-gating: image data are continuously acquired over several full cardiac cycles with a slow table movement and the use of a constant tube voltage. This method provides both functional and morphological information. (b) Retrospective ECG-gating with tube current modulation: In contrast to (a), tube current is reduced outside of the target phase (in this case: 50–90%). (c) Prospective ECG-pulsing: Image data are sequentially acquired over an extended period during the cardiac cycle. Besides the application of full tube current during a target phase, the tube current is reduced during the remaining data acquisition window. Consequently, functional and morphological information can be obtained. (d) Prospective ECG-triggering: In contrast to (c), image data are sequentially (step-and-shoot) acquired targeted on a specific phase during the cardiac cycle (diastole or systole). This method provides mainly morphological information. (e) High-pitch dual-source CT technique: This prospectively ECG-triggered scan mode is useful to acquire motion-free images during a single heart beat because of an ultra fast table speed.
tures or lung parenchyma (►Fig. 2). Historically, common diagnostic pitfalls included artifactual intimal flaps of the ascending aorta mimicking aortic dissection2 or distortion of pericardial lung segments compromising the accurate assessment of small pulmonary nodules, vessels, or subtle parenchymal abnormalities.3 In general, cardiac motion artifacts typically have affected thin-section chest CT by doubling the borders of thoracic structures, particularly in the lower lung lobes.4–6 Ritchie et al6 postulated that an acquisition time of less than 19.1 milliseconds would be required to suppress cardiac motion artifacts at all heart rates. Later studies indicated that a heart rate-independent temporal resolution
of less than 100 milliseconds can eliminate motion artifacts during diastole at normal resting heart rates. However, a temporal resolution of less than 50 milliseconds might still be required for motion-free imaging when studies are acquired during systole or at very high heart rates (above 100 bpm).7–9 Initial attempts were made to address these limitations by shortening acquisition times and introducing ECG-synchronized acquisitions using electron beam CT or single slice CT.10,11 In this context, Schoepf et al11 has shown a significant reduction of cardiac motion artifacts that might mimic pulmonary diseases by using an ECG-gated thin-section protocol. However, although the above mentioned artifacts could be Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
51
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Integrated Cardiothoracic Imaging with CT
Integrated Cardiothoracic Imaging with CT
Geyer et al.
Fig. 2 Reduction of cardiac motion artifacts by using electrocardiography (ECG)-synchronization. Double contours along the mediastinal and cardiac borders caused by cardiac motion in non-ECG-synchronized computed tomography images ([a] white arrows) can be suppressed by using ECG-synchronization techniques in cardiothoracic CT (b). In addition, ECG-synchronization (c) reduces blurring of vascular calcifications in nonECG-gated chest CT ([d] black arrow).
reduced by these techniques, coverage of the entire thorax was not possible within a reasonable examination time. A key step toward motion-reduced CT imaging of the entire chest and dedicated cardiac CT imaging was the introduction of multidetector-row CT (MDCT) systems. The first MDCT generation with simultaneous 4-slice data acquisition provided subsecond gantry rotation speed and an enlarged volume coverage in the z-direction compared with single slice CT.10 Besides its impact on cardiac imaging, 4-row MDCT improved the evaluation of the pulmonary artery tree on segmental and subsegmental levels for pulmonary embolism (PE) diagnosis.12,13 Sequential, prospectively ECG-triggered scanning (step-and-shoot) allowed for cardiac motion artifact reduction at temporal resolutions of 250 milliseconds. Although scan range and speed were improved compared with single slice CT, protocols for integrated cardiothoracic imaging were performed with a 2.5-mm section thickness and were limited to a scan range of 150 mm; the average scan duration was 25 second.14 In contrast, retrospective ECG-gating as an alternative means of ECG-synchronized CT image acquisition provided a higher spatial (section thickness, 1 mm) and temporal resolution (up to 125 milliseconds). However, this approach was associated with low-pitch (i.e., slow table speed per rotation) spiral acquisition resulting in a limited z-coverage of 100 mm and an extended scan time of 30 second.15 In addition, radiation doses tended to be relatively high because of the slow scan speeds involved with data oversampling required for retrospectively ECG-gated image reconstruction. Regarding CT of the entire thorax (approximately 300 mm), motion-reduced, thin-section imaging within a single breath hold was feasible by using modified reconstruction algorithms but only for a specific diagnostic focus, for example, aortic dissection or coronary bypass grafts.16 Truly, motion-free data acquisition were still confined to cardiac CT. Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
Integrating Morphology and Function in Whole Chest CT Imaging In subsequent years, CT technology evolved rapidly and therefore, more advanced cardiothoracic scan protocols were enabled by the advent of 16-slice CT.17,18 Besides increased temporal and spatial resolution as well as faster coverage in the z-direction, ECG-synchronized data acquisition of the whole chest allowed for simultaneous cardiac function assessment. Initially, two different approaches were introduced: a monophasic protocol with measurement of global cardiac function during a single chest CT scan,19 and dual phase CT data acquisition using the highest possible spatial resolution for both the heart and the thoracic organs.20 Using a single CT scan, a necessary adaptation of the scan parameters was the use of a 16 1.5 mm collimation to ensure practicable breath hold duration below 30 second.19 Besides the reliable evaluation of cardiac function, motionsuppressed imaging of the lung parenchyma and the mediastinal vessels became possible. The image quality was sufficient to analyze gross pathologies of the coronary arteries,21 but it did not allow for a detailed assessment. In addition, the capabilities of 16-row MDCT systems permitted the first ECGsynchronized CT protocols for a comprehensive single scan assessment of both the heart and the cardiothoracic vessels in patients with acute chest pain.22 With the introduction of 64-row MDCT scanners in 2004, routine diagnostic cardiothoracic imaging became feasible through single breath hold scans of the entire chest with submillimeter collimation.23–25 Salem et al25 have investigated a cohort of 133 patients with bronchopulmonary or pulmonary vascular diseases obtaining functional parameters of the heart and morphological information of the entire thorax. The chest was examined within a single CT scan implementing an ECG-pulsing technique resulting in a remarkably high rate (92%) of studies with diagnostic image quality. Turning the focus from cardiopulmonary to
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
52
Geyer et al.
Fig. 3 Triple-rule-out computed tomography in a 66-year-old male with history of human immunodeficiency virus presenting with chest pain, ST elevation, and increased troponin. Coronary computed tomography angiogram (CTA) shows occlusion of the proximal left anterior descending artery (a), confirmed by coronary catheterization (b). Myocardial assessment shows corresponding hypodensity (c) and hypokinesis (functional images are not displayed) of the entire LAD territory, consistent with myocardial infarction. In addition, bilateral symmetric ground-glass opacities and scattered cysts (d, e) can be identified in the lung being suspicious for an atypical infectious process and most likely coexistence of pulmonary edema.
cardiovascular, the proximal and midcoronary segments were assessable using the same scan protocol even without administration of β-blockers in patients with a heart rate below 80 bpm.24 Furthermore, CT is increasingly used in the context of triaging patients with equivocal chest pain presenting to the emergency department. So-called “triple-ruleout” protocols allow for the safe exclusion of common causes of acute chest pain, such as PE, aortic dissection, and significant coronary artery disease (CAD), within a single examination (►Fig. 3).26
Resolving Remaining Limitations of Integrated Cardiopulmonary CT Imaging Although 64-row MDCT brought ECG-synchronized cardiopulmonary CT into clinical routine, several sources of image quality impairment remained relevant: respiratory motion artifacts with longer scan durations, cardiac motion artifacts due to limited temporal resolution, and artifacts related to the ECG-pulsing algorithm, for example, nonoptimal ratios between pitch and heart rate (nonheart-rate-adapted pitch). In particular, for patients with respiratory diseases, the prolonged scan duration—and therefore breath hold—of retrospectively ECG-gated CT scans ran the risk of incurring higher rates of respiratory motion artifacts. However, in the study by Salem et al,25 the average scan duration was 18
second and the vast majority of patients were able to hold their breath throughout the image acquisition; relevant respiratory motion artifacts occurred in only 3% of scans. One can expect this percentage to be further reduced with more recent CT systems having 128, 256, and 320 detector rows and further decreased scan durations.27 In a development apart from the industry trend of ever increasing detector row numbers, a dual-source CT (DSCT) system was introduced in 2005 with two X-ray tubes offset by 90 degrees28; this was followed by the release of a second generation DSCT in 2009.29 The most important benefit of DSCT is its high physical temporal resolution of 83 or 75 milliseconds, respective of generation, at all heart rates,28,30 which greatly enhances the robustness of image acquisition vis-à-vis cardiac motion artifacts. In this context, Johnson et al31 have demonstrated the advantage of DSCT for preserving diagnostic image quality in cardiothoracic imaging even in patients with high heart rates. Another advantage of DSCT is its capability of scanning in an ECG-synchronized highpitch mode (i.e., high table speed per rotation), which eliminates motion artifacts, reduces radiation, and provides adequate visualization of the cardiothoracic organs, including mediastinal vessels32 and the coronary arteries,33,34 all within a fraction of a heart beat during a single breath hold (►Fig. 4). Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
53
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Integrated Cardiothoracic Imaging with CT
Integrated Cardiothoracic Imaging with CT
Geyer et al.
Fig. 4 High-pitch chest computed tomographic (CT) angiography in a 70-year-old male. Chest CT angiography reveals a type A aortic dissection with extension of the flap into both common carotid arteries ([a] black arrows), as well as bilateral large pleural effusions with adjacent atelectases (b). In addition, scattered coronary artery calcifications can be identified (c).
Expected Benefits of Integrated Cardiac Assessment in Thoracic Imaging In light of the developments illustrated above, the integration of cardiac and coronary imaging into thoracic CT has the potential to broaden the spectrum of clinical indications for imaging patients with respiratory disorders. Although preserving the image quality that is required to evaluate the primary disease, modern CT scanners have the capability to provide complementary information on the entire cardiothoracic system. Accordingly, the implementation of ECG-synchronized CT protocols might be indicated if additional information about cardiac structure and/or function could be beneficial: (1) the evaluation of thoracic disorders and their cardiac effects—in particular, on right ventricular (RV) function—or (2) the evaluation of thoracic consequences caused by cardiac disorders.35
Scan Protocols In the context of a stepwise diagnostic approach to thoracic disease, chest radiography is often the primary imaging modality used for the initial evaluation of patients with suspected cardiopulmonary disorders. However, plain radiographs are usually followed by a thoracic CT scan providing for a more sophisticated and comprehensive evaluation of chest Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
pathology. Thoracic CT scan protocols come in a variety of flavors which may be custom-tailored to the underlying clinical indication and technical capabilities of the CT system, as well as institutional practices and experience. In patients with suspicion of noncardiovascular chest disorders, the routinely acquired thoracic CT study typically is a non-ECGsynchronized image acquisition. However, the latest developments in CT technology—as mentioned above—increasingly enable the integration of ECG-synchronization into routine chest CT protocols (►Table 1).36,37 Historically, the use of ECG-synchronization ordinarily involved the penalty of increased radiation exposure; hence, these techniques were typically reserved for dedicated evaluation of cardiac pathology or cardiothoracic disorders that clearly required the concomitant analysis of heart morphology and/or function. The decision of whether or not to use ECG-synchronization for cardiothoracic CT image acquisition still requires careful consideration of the expected benefits from such an approach. Despite recent years’ drastic decreases in overall radiation exposure from ECG-synchronized CT acquisitions, a chest CT study performed with traditional retrospective ECG-gating will ordinarily still result in a relatively higher radiation dose than a routine, non-ECG-synchronized acquisition. However, this acquisition technique would provide concomitant evaluation of cardiac function and may thus be beneficial and still desirable in certain clinical scenarios. When only information on cardiac anatomy is sought, with cardiac function being of limited interest, there exist several ECG-synchronization strategies, recently rediscovered or newly devised, that enable motion-free imaging of the chest at a radiation dose that is equal to standard non-ECG-synchronized thoracic CT studies—thus, feasible without the penalty of higher radiation exposure. Using an anthropomorphic phantom, Sommer et al38 compared the radiation exposure resulting from a standard non-ECG-synchronized chest CT scan (2.68 mSv), a retrospectively ECG-gated acute chest pain “triple-rule-out” protocol (19.27 mSv), and a high-pitch chest pain protocol (2.65 mSv) on a DSCT system. Transferring these scan parameters into clinical practice, the authors estimated mean effective radiation dose equivalents of 4.40 0.83 mSv (standard non-ECGsynchronized protocol), 4.08 0.81 mSv (high-pitch protocol), and 20.4 5.3 mSv (retrospectively ECG-gated protocol) using relatively high conversion factors of 0.0180–0.0188 mSv/(mGy cm) (standard conversion factor, 0.014 mSv/ [mGy cm]). However, the resulting estimated radiation dose depends on several factors such as the desired image quality, scanner technique, and conversion factor. For instance, d’Agostino et al39 investigated 105 patients with pulmonary or vascular disease, using a retrospectively ECGgated 64-slice MDCT protocol of the entire thorax to obtain both morphological and functional information. An estimated mean effective dose of 4.95 1.59 mSv was achieved by adjusting scan parameters (e.g., tube current-time product, 200 mAs; for comparison, typical range in cardiac CT: 500– 900 mAs) and integrating tube current modulation algorithms based on both patient size and ECG-tracing (ECG pulsing technique). Their suggested CT protocol for combined
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
54
Integrated Cardiothoracic Imaging with CT
Geyer et al.
55
Table 1 Examples of scan protocols for integrated (cardio-)thoracic CT imaging Wide detector CTa
DSCT Cardiac function 1
Triple-rule-out
Triple-rule-out
1
1
Tube voltage (kV)
Triple-rule-out
Chest
Heart
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120
120
Tube current (modulated)
320 mAs/rot
400 mAs/rot
320 mAs/rot
250 mA
400 mA
Collimation
2 64 0.6 mm
2 64 0.6 mm
2 64 0.6 mm
320 0.5 mm
320 0.5 mm
Pitch
–
0.2–0.5 (HR adapted)
3.2
–
–
Rotation time (ms)
0.28
0.28
0.28
0.35
0.35
Total CM volume (mL)
60–80
6 mL per second of scan duration (> 50 mL)
120
90
Flow rate (mL/s)
6.0
6.0
5.0
3.0/5.0/5.0
Injection protocol
1. Pure CM 2. 50 mL mixture 30/70% 3. 30 mL saline solution
1. Pure CM 2. 50 mL mixture 30/70% 3. 30 mL saline solution
1. Pure CM 2. 100 mL saline solution
Bolus timing
Test bolusc
Test bolusc
Test bolusc
Bolus tracking
ROI
Ascending aorta
Ascending aorta
Ascending aorta
Pulmonary trunk
Delay (s)
Peak time þ 4 s
Peak time
Peak time
Scan range
Carina–diaphragm
Lung apex–diaphragm
Lung apex–diaphragm
2
Heart
1. 45 mL pure CM 2. 45 mL pure CM 3. 30 mL saline solution
6 (threshold 200 HU)
6 (after chest scan)
Scan direction
Craniocaudal
Craniocaudal
Craniocaudal
ECG synchronization
Prospective, dual-step pulsing
Retrospective; ECG-pulsing window width according to HR
Prospective; single heart beat
–
Prospective; 1–2 heart beats
Phase of the RR-cycle with full tube current
65–75% (HR 75 bpm) 35–45% (HR > 75bpm)
70% (HR 70 bpm) 30–70% (HR > 70 bpm)
75%
–
70–80% (HR 65 bpm) 30–80% (HR > 65 bpm)
Pulsing window of the RR-cycle with reduced tube current
30–90% (20% tube current)
70–30% (4% tube current)
–
Scan mode
Sequential
Helical
Helical
Scan duration (s)
ca. 10–15 (depending on HR)
ca. 17–22 (depending on HR)
Integrated Cardiothoracic Imaging with Computed Tomography Lucas L. Geyer, MD1,2 Justin R. Silverman, BA1 Aleksander W. Krazinski, BA1 Pal Suranyi, MD, PhD1 James G. Ravenel, MD1 Stefan Wirth, MD, PhD, MSc2 Philip Costello, MD1 U. Joseph Schoepf, MD1 1 Department of Radiology and Radiological Science, Medical
University of South Carolina, Charleston, South Carolina 2 Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital, Munich, Germany
Address for correspondence U. Joseph Schoepf, MD, Heart and Vascular Center, Medical University of South Carolina, Ashley River Tower, MSC 226, 25 Courtenay Dr, Charleston, SC 29401 (e-mail: [email protected]).
Abstract Keywords
► ► ► ►
chest imaging heart disease lung diseases cardiopulmonary comorbidities ► computed tomography angiography ► dual-source computed tomography
The respiratory and the cardiovascular systems are intimately connected. Because of the high degree of morphological and functional interaction, pathophysiological processes in one compartment are likely to induce adaptive changes in the other. Computed tomography (CT) plays a central role in the diagnostic work up of both thoracic and cardiac disorders. Historically, these two systems have been evaluated separately; however, CT technology has evolved remarkably over recent decades. Up-to-date advanced imaging strategies allow for a combined assessment of the cardiopulmonary unit. Besides improved techniques of electrocardiogram (ECG)-synchronization for obtaining both morphological and functional information, latest advances of dualsource CT (DSCT) have shown great promise for even more comprehensive integrated cardiothoracic imaging.
Diseases of the respiratory and the cardiovascular system are leading causes of morbidity and mortality in the Western Hemisphere (1, 2). Chest computed tomography (CT) plays a central role in the diagnostic work up of known or suspected thoracic pathology. Because of their inherent interrelationship, pathophysiological changes in the pulmonary system are likely to induce adaptations in the intimately connected cardiovascular system, and vice versa.1 Therefore, evaluation for concomitant cardiovascular disorders should be part of the interpretation of any chest CT. As thoracic CT imaging is usually not performed in sync with cardiac activity (non-ECG synchronized data acquisition), the simultaneous assessment of the heart is often limited by cardiac motion. However, CT technology has substantially evolved over the last decades, often allowing a degree of cardiac assessment even with routine chest CT protocols. Depending on the clinical scenario, there also exist more advanced image acquisition strate-
Issue Theme Thoracic Imaging; Guest Editor, Martine Remy-Jardin, MD, PhD
gies that integrate ECG-synchronization for detailed cardiac and coronary imaging (►Fig. 1). The purpose of this contribution, thus, is (1) to summarize the technical developments of CT imaging which enabled integrated cardiopulmonary imaging, (2) to discuss appropriate scan protocols, and (3) to provide context through a review of common cardiopulmonary disorders.
Technical Evolution of Cardiothoracic Imaging Reducing Cardiac Motion Artifact In the early years of thoracic CT imaging, image interpretation was frequently hampered by respiratory or cardiac motion artifacts. Whereas respiratory motion induces extensive blurring, cardiac motion leads to more subtle changes by transmitting minor pulsations to adjacent mediastinal struc-
Copyright © 2014 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0033-1363451. ISSN 1069-3424.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Semin Respir Crit Care Med 2014;35:50–63.
Geyer et al.
Fig. 1 Techniques of electrocardiography (ECG)-synchronization. The width of the gray bars reflects the duration of data acquisition in relation to the cardiac cycle; the amplitude of the gray bars reflects the tube current applied during the scan. (a) Retrospective ECG-gating: image data are continuously acquired over several full cardiac cycles with a slow table movement and the use of a constant tube voltage. This method provides both functional and morphological information. (b) Retrospective ECG-gating with tube current modulation: In contrast to (a), tube current is reduced outside of the target phase (in this case: 50–90%). (c) Prospective ECG-pulsing: Image data are sequentially acquired over an extended period during the cardiac cycle. Besides the application of full tube current during a target phase, the tube current is reduced during the remaining data acquisition window. Consequently, functional and morphological information can be obtained. (d) Prospective ECG-triggering: In contrast to (c), image data are sequentially (step-and-shoot) acquired targeted on a specific phase during the cardiac cycle (diastole or systole). This method provides mainly morphological information. (e) High-pitch dual-source CT technique: This prospectively ECG-triggered scan mode is useful to acquire motion-free images during a single heart beat because of an ultra fast table speed.
tures or lung parenchyma (►Fig. 2). Historically, common diagnostic pitfalls included artifactual intimal flaps of the ascending aorta mimicking aortic dissection2 or distortion of pericardial lung segments compromising the accurate assessment of small pulmonary nodules, vessels, or subtle parenchymal abnormalities.3 In general, cardiac motion artifacts typically have affected thin-section chest CT by doubling the borders of thoracic structures, particularly in the lower lung lobes.4–6 Ritchie et al6 postulated that an acquisition time of less than 19.1 milliseconds would be required to suppress cardiac motion artifacts at all heart rates. Later studies indicated that a heart rate-independent temporal resolution
of less than 100 milliseconds can eliminate motion artifacts during diastole at normal resting heart rates. However, a temporal resolution of less than 50 milliseconds might still be required for motion-free imaging when studies are acquired during systole or at very high heart rates (above 100 bpm).7–9 Initial attempts were made to address these limitations by shortening acquisition times and introducing ECG-synchronized acquisitions using electron beam CT or single slice CT.10,11 In this context, Schoepf et al11 has shown a significant reduction of cardiac motion artifacts that might mimic pulmonary diseases by using an ECG-gated thin-section protocol. However, although the above mentioned artifacts could be Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
51
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Integrated Cardiothoracic Imaging with CT
Integrated Cardiothoracic Imaging with CT
Geyer et al.
Fig. 2 Reduction of cardiac motion artifacts by using electrocardiography (ECG)-synchronization. Double contours along the mediastinal and cardiac borders caused by cardiac motion in non-ECG-synchronized computed tomography images ([a] white arrows) can be suppressed by using ECG-synchronization techniques in cardiothoracic CT (b). In addition, ECG-synchronization (c) reduces blurring of vascular calcifications in nonECG-gated chest CT ([d] black arrow).
reduced by these techniques, coverage of the entire thorax was not possible within a reasonable examination time. A key step toward motion-reduced CT imaging of the entire chest and dedicated cardiac CT imaging was the introduction of multidetector-row CT (MDCT) systems. The first MDCT generation with simultaneous 4-slice data acquisition provided subsecond gantry rotation speed and an enlarged volume coverage in the z-direction compared with single slice CT.10 Besides its impact on cardiac imaging, 4-row MDCT improved the evaluation of the pulmonary artery tree on segmental and subsegmental levels for pulmonary embolism (PE) diagnosis.12,13 Sequential, prospectively ECG-triggered scanning (step-and-shoot) allowed for cardiac motion artifact reduction at temporal resolutions of 250 milliseconds. Although scan range and speed were improved compared with single slice CT, protocols for integrated cardiothoracic imaging were performed with a 2.5-mm section thickness and were limited to a scan range of 150 mm; the average scan duration was 25 second.14 In contrast, retrospective ECG-gating as an alternative means of ECG-synchronized CT image acquisition provided a higher spatial (section thickness, 1 mm) and temporal resolution (up to 125 milliseconds). However, this approach was associated with low-pitch (i.e., slow table speed per rotation) spiral acquisition resulting in a limited z-coverage of 100 mm and an extended scan time of 30 second.15 In addition, radiation doses tended to be relatively high because of the slow scan speeds involved with data oversampling required for retrospectively ECG-gated image reconstruction. Regarding CT of the entire thorax (approximately 300 mm), motion-reduced, thin-section imaging within a single breath hold was feasible by using modified reconstruction algorithms but only for a specific diagnostic focus, for example, aortic dissection or coronary bypass grafts.16 Truly, motion-free data acquisition were still confined to cardiac CT. Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
Integrating Morphology and Function in Whole Chest CT Imaging In subsequent years, CT technology evolved rapidly and therefore, more advanced cardiothoracic scan protocols were enabled by the advent of 16-slice CT.17,18 Besides increased temporal and spatial resolution as well as faster coverage in the z-direction, ECG-synchronized data acquisition of the whole chest allowed for simultaneous cardiac function assessment. Initially, two different approaches were introduced: a monophasic protocol with measurement of global cardiac function during a single chest CT scan,19 and dual phase CT data acquisition using the highest possible spatial resolution for both the heart and the thoracic organs.20 Using a single CT scan, a necessary adaptation of the scan parameters was the use of a 16 1.5 mm collimation to ensure practicable breath hold duration below 30 second.19 Besides the reliable evaluation of cardiac function, motionsuppressed imaging of the lung parenchyma and the mediastinal vessels became possible. The image quality was sufficient to analyze gross pathologies of the coronary arteries,21 but it did not allow for a detailed assessment. In addition, the capabilities of 16-row MDCT systems permitted the first ECGsynchronized CT protocols for a comprehensive single scan assessment of both the heart and the cardiothoracic vessels in patients with acute chest pain.22 With the introduction of 64-row MDCT scanners in 2004, routine diagnostic cardiothoracic imaging became feasible through single breath hold scans of the entire chest with submillimeter collimation.23–25 Salem et al25 have investigated a cohort of 133 patients with bronchopulmonary or pulmonary vascular diseases obtaining functional parameters of the heart and morphological information of the entire thorax. The chest was examined within a single CT scan implementing an ECG-pulsing technique resulting in a remarkably high rate (92%) of studies with diagnostic image quality. Turning the focus from cardiopulmonary to
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
52
Geyer et al.
Fig. 3 Triple-rule-out computed tomography in a 66-year-old male with history of human immunodeficiency virus presenting with chest pain, ST elevation, and increased troponin. Coronary computed tomography angiogram (CTA) shows occlusion of the proximal left anterior descending artery (a), confirmed by coronary catheterization (b). Myocardial assessment shows corresponding hypodensity (c) and hypokinesis (functional images are not displayed) of the entire LAD territory, consistent with myocardial infarction. In addition, bilateral symmetric ground-glass opacities and scattered cysts (d, e) can be identified in the lung being suspicious for an atypical infectious process and most likely coexistence of pulmonary edema.
cardiovascular, the proximal and midcoronary segments were assessable using the same scan protocol even without administration of β-blockers in patients with a heart rate below 80 bpm.24 Furthermore, CT is increasingly used in the context of triaging patients with equivocal chest pain presenting to the emergency department. So-called “triple-ruleout” protocols allow for the safe exclusion of common causes of acute chest pain, such as PE, aortic dissection, and significant coronary artery disease (CAD), within a single examination (►Fig. 3).26
Resolving Remaining Limitations of Integrated Cardiopulmonary CT Imaging Although 64-row MDCT brought ECG-synchronized cardiopulmonary CT into clinical routine, several sources of image quality impairment remained relevant: respiratory motion artifacts with longer scan durations, cardiac motion artifacts due to limited temporal resolution, and artifacts related to the ECG-pulsing algorithm, for example, nonoptimal ratios between pitch and heart rate (nonheart-rate-adapted pitch). In particular, for patients with respiratory diseases, the prolonged scan duration—and therefore breath hold—of retrospectively ECG-gated CT scans ran the risk of incurring higher rates of respiratory motion artifacts. However, in the study by Salem et al,25 the average scan duration was 18
second and the vast majority of patients were able to hold their breath throughout the image acquisition; relevant respiratory motion artifacts occurred in only 3% of scans. One can expect this percentage to be further reduced with more recent CT systems having 128, 256, and 320 detector rows and further decreased scan durations.27 In a development apart from the industry trend of ever increasing detector row numbers, a dual-source CT (DSCT) system was introduced in 2005 with two X-ray tubes offset by 90 degrees28; this was followed by the release of a second generation DSCT in 2009.29 The most important benefit of DSCT is its high physical temporal resolution of 83 or 75 milliseconds, respective of generation, at all heart rates,28,30 which greatly enhances the robustness of image acquisition vis-à-vis cardiac motion artifacts. In this context, Johnson et al31 have demonstrated the advantage of DSCT for preserving diagnostic image quality in cardiothoracic imaging even in patients with high heart rates. Another advantage of DSCT is its capability of scanning in an ECG-synchronized highpitch mode (i.e., high table speed per rotation), which eliminates motion artifacts, reduces radiation, and provides adequate visualization of the cardiothoracic organs, including mediastinal vessels32 and the coronary arteries,33,34 all within a fraction of a heart beat during a single breath hold (►Fig. 4). Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
53
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Integrated Cardiothoracic Imaging with CT
Integrated Cardiothoracic Imaging with CT
Geyer et al.
Fig. 4 High-pitch chest computed tomographic (CT) angiography in a 70-year-old male. Chest CT angiography reveals a type A aortic dissection with extension of the flap into both common carotid arteries ([a] black arrows), as well as bilateral large pleural effusions with adjacent atelectases (b). In addition, scattered coronary artery calcifications can be identified (c).
Expected Benefits of Integrated Cardiac Assessment in Thoracic Imaging In light of the developments illustrated above, the integration of cardiac and coronary imaging into thoracic CT has the potential to broaden the spectrum of clinical indications for imaging patients with respiratory disorders. Although preserving the image quality that is required to evaluate the primary disease, modern CT scanners have the capability to provide complementary information on the entire cardiothoracic system. Accordingly, the implementation of ECG-synchronized CT protocols might be indicated if additional information about cardiac structure and/or function could be beneficial: (1) the evaluation of thoracic disorders and their cardiac effects—in particular, on right ventricular (RV) function—or (2) the evaluation of thoracic consequences caused by cardiac disorders.35
Scan Protocols In the context of a stepwise diagnostic approach to thoracic disease, chest radiography is often the primary imaging modality used for the initial evaluation of patients with suspected cardiopulmonary disorders. However, plain radiographs are usually followed by a thoracic CT scan providing for a more sophisticated and comprehensive evaluation of chest Seminars in Respiratory and Critical Care Medicine
Vol. 35
No. 1/2014
pathology. Thoracic CT scan protocols come in a variety of flavors which may be custom-tailored to the underlying clinical indication and technical capabilities of the CT system, as well as institutional practices and experience. In patients with suspicion of noncardiovascular chest disorders, the routinely acquired thoracic CT study typically is a non-ECGsynchronized image acquisition. However, the latest developments in CT technology—as mentioned above—increasingly enable the integration of ECG-synchronization into routine chest CT protocols (►Table 1).36,37 Historically, the use of ECG-synchronization ordinarily involved the penalty of increased radiation exposure; hence, these techniques were typically reserved for dedicated evaluation of cardiac pathology or cardiothoracic disorders that clearly required the concomitant analysis of heart morphology and/or function. The decision of whether or not to use ECG-synchronization for cardiothoracic CT image acquisition still requires careful consideration of the expected benefits from such an approach. Despite recent years’ drastic decreases in overall radiation exposure from ECG-synchronized CT acquisitions, a chest CT study performed with traditional retrospective ECG-gating will ordinarily still result in a relatively higher radiation dose than a routine, non-ECG-synchronized acquisition. However, this acquisition technique would provide concomitant evaluation of cardiac function and may thus be beneficial and still desirable in certain clinical scenarios. When only information on cardiac anatomy is sought, with cardiac function being of limited interest, there exist several ECG-synchronization strategies, recently rediscovered or newly devised, that enable motion-free imaging of the chest at a radiation dose that is equal to standard non-ECG-synchronized thoracic CT studies—thus, feasible without the penalty of higher radiation exposure. Using an anthropomorphic phantom, Sommer et al38 compared the radiation exposure resulting from a standard non-ECG-synchronized chest CT scan (2.68 mSv), a retrospectively ECG-gated acute chest pain “triple-rule-out” protocol (19.27 mSv), and a high-pitch chest pain protocol (2.65 mSv) on a DSCT system. Transferring these scan parameters into clinical practice, the authors estimated mean effective radiation dose equivalents of 4.40 0.83 mSv (standard non-ECGsynchronized protocol), 4.08 0.81 mSv (high-pitch protocol), and 20.4 5.3 mSv (retrospectively ECG-gated protocol) using relatively high conversion factors of 0.0180–0.0188 mSv/(mGy cm) (standard conversion factor, 0.014 mSv/ [mGy cm]). However, the resulting estimated radiation dose depends on several factors such as the desired image quality, scanner technique, and conversion factor. For instance, d’Agostino et al39 investigated 105 patients with pulmonary or vascular disease, using a retrospectively ECGgated 64-slice MDCT protocol of the entire thorax to obtain both morphological and functional information. An estimated mean effective dose of 4.95 1.59 mSv was achieved by adjusting scan parameters (e.g., tube current-time product, 200 mAs; for comparison, typical range in cardiac CT: 500– 900 mAs) and integrating tube current modulation algorithms based on both patient size and ECG-tracing (ECG pulsing technique). Their suggested CT protocol for combined
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
54
Integrated Cardiothoracic Imaging with CT
Geyer et al.
55
Table 1 Examples of scan protocols for integrated (cardio-)thoracic CT imaging Wide detector CTa
DSCT Cardiac function 1
Triple-rule-out
Triple-rule-out
1
1
Tube voltage (kV)
Triple-rule-out
Chest
Heart
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120 (BMI > 25 kg/m) 100 (BMI < 25 kg/m)
120
120
Tube current (modulated)
320 mAs/rot
400 mAs/rot
320 mAs/rot
250 mA
400 mA
Collimation
2 64 0.6 mm
2 64 0.6 mm
2 64 0.6 mm
320 0.5 mm
320 0.5 mm
Pitch
–
0.2–0.5 (HR adapted)
3.2
–
–
Rotation time (ms)
0.28
0.28
0.28
0.35
0.35
Total CM volume (mL)
60–80
6 mL per second of scan duration (> 50 mL)
120
90
Flow rate (mL/s)
6.0
6.0
5.0
3.0/5.0/5.0
Injection protocol
1. Pure CM 2. 50 mL mixture 30/70% 3. 30 mL saline solution
1. Pure CM 2. 50 mL mixture 30/70% 3. 30 mL saline solution
1. Pure CM 2. 100 mL saline solution
Bolus timing
Test bolusc
Test bolusc
Test bolusc
Bolus tracking
ROI
Ascending aorta
Ascending aorta
Ascending aorta
Pulmonary trunk
Delay (s)
Peak time þ 4 s
Peak time
Peak time
Scan range
Carina–diaphragm
Lung apex–diaphragm
Lung apex–diaphragm
2
Heart
1. 45 mL pure CM 2. 45 mL pure CM 3. 30 mL saline solution
6 (threshold 200 HU)
6 (after chest scan)
Scan direction
Craniocaudal
Craniocaudal
Craniocaudal
ECG synchronization
Prospective, dual-step pulsing
Retrospective; ECG-pulsing window width according to HR
Prospective; single heart beat
–
Prospective; 1–2 heart beats
Phase of the RR-cycle with full tube current
65–75% (HR 75 bpm) 35–45% (HR > 75bpm)
70% (HR 70 bpm) 30–70% (HR > 70 bpm)
75%
–
70–80% (HR 65 bpm) 30–80% (HR > 65 bpm)
Pulsing window of the RR-cycle with reduced tube current
30–90% (20% tube current)
70–30% (4% tube current)
–
Scan mode
Sequential
Helical
Helical
Scan duration (s)
ca. 10–15 (depending on HR)
ca. 17–22 (depending on HR)
