Nontraumatic Emergency Radiology of the Thorax

Published on 12/06/2015 by admin

Filed under Radiology

Last modified 22/04/2025

Print this page

This article have been viewed 3990 times

CHAPTER 8 Nontraumatic Emergency Radiology of the Thorax

Scott White, Alejandra Duran-Mendicuti, Aaron Sodickson, Salvatore Viscomi, Michael Stella, M. Stephen Ledbetter

An emergency medicine physician once told a bright-eyed medical student on the first clinical shift:

All kinds of problems bring people in to see us with chest complaints. But just remember this—your job is to make sure nobody leaves the emergency department with any of only four entities undiagnosed, and those four entities are pneumothorax, aortic dissection, pulmonary embolism, and myocardial infarction. Now repeat that list to me four times (the medical student regurgitates the list four times). Good. Beyond these four entities, anything that’s causing their problem won’t cause you much of a problem.

While that is a rather spartan synopsis of nontraumatic thoracic pathology encountered in the emergency department (ED), the underlying tenet stands up well in the practical evaluation of ED patients presenting with chest pain and other symptoms. Accordingly, this chapter on nontraumatic emergency thoracic radiology is organized primarily around evaluation of these entities, beginning with a review of chest radiography technique, and followed by sections on the pleura/mediastinum, aortic diseases, pulmonary embolism, and acute coronary syndrome (ACS). Lung parenchymal findings are covered with less emphasis, but are readily available with greater detail in other texts, such as Thoracic Radiology: The Requisites.

Multidetector computed tomography (MDCT) technique deserves and receives special emphasis in this chapter because it has overtaken and continues to overtake other imaging modalities for chest evaluation in ED patients. Clinicians have become enamored with the rapid, highly accurate, and thereby efficient diagnostic capability CT provides. The CT industry and the radiology profession continue to further these advantages with faster, higher-resolution scanners allowing for greater clinical applications and increasingly refined diagnoses. Similarly, the pressure for greater diagnostic sensitivity and specificity in the ED is at odds with decreased imaging utilization. The volume of chest CT exams ordered by ED physicians will almost certainly continue to rise. Perhaps the only significant counterbalance to increased utilization is the concern for population radiation exposure and the associated risk of cancers induced by ionizing radiation. Radiologists should be familiar with and sensitive to these concerns, readily able to advise clinicians and protect patients by being good stewards of radiation safety.

CHEST RADIOGRAPHY

In a clinical environment where radiologists and ED physicians frequently have available “high-tech” CT scanners with dozens, even hundreds of detector rows, multiple radiograph sources, and electrocardiographic (EKG) gating, it is tempting to neglect or dismiss the value of the “low-tech” chest x-ray (CXR). However, neglecting the CXR and failing to develop the knowledge and skills needed for expert CXR interpretation is a mistake. Despite the advances in CT technology, the CXR remains the mainstay of first-line imaging for ED patients with chest complaints. The 2004 National Hospital Ambulatory Medical Survey (NHAMES) recorded 110.2 million ED visits, which resulted in 20.19 million CXRs performed. Indeed, a CXR is the most often ordered imaging exam as part of an ED patient visit. This compares to totals of approximately 22 million for all other x-rays combined, and 10.3 million for CT or magnetic resonance imaging (MRI) (the CT and MRI data were combined in the NHAMES report). Therefore, the emergency radiologist needs to have highly developed interpretive skills for chest radiography. Detailed instruction on CXR interpretation is well beyond the scope of this chapter, but many excellent sources are available for this, including the Requisites text on thoracic radiology.

One principle that serves radiologists well is the establishment of a systematic pattern through which the interpreter progresses as he or she reads a CXR. While naturally drawn to look at the lungs, and while evaluation of pulmonary parenchyma is paramount, the interpreter must unerringly include assessment of extrapulmonary anatomy on every CXR. This includes evaluation of anatomy below the diaphragm, of lung parenchyma posterior to the diaphragm, of the bony thorax, spine, shoulders, and chest wall soft tissues, of the heart contours, mediastinum, and pulmonary hila, of pleural surfaces and interfaces, and of complex areas such as the pulmonary apices. Establishing a pattern of evaluation that systematically includes all of the anatomy encountered on a CXR will serve to avoid unfortunate errors of CXR interpretation (Fig. 8-1).

Figure 8-1 Chest x-ray (CXR) of a patient presenting with cough and fever. A faint area of opacification within the left midlung proved to be pneumonia (black arrow). Additionally, a small left apical carcinoma was detected (white arrows). Due to multiple overlying structures, including the first ribs and the clavicles, the pulmonary apices are a “danger” area in terms of risk of missed findings on CXR. Close inspection of the apices is warranted on every CXR.

For evaluation of the lung parenchyma, CXR has been in use for nearly a century. Pulmonary pathology is detected primarily by changes in pulmonary parenchymal density, as either decreased parenchymal density (cystic lung diseases and pneumothorax) or increased parenchymal density (masses, fluid, infection, fibrosis, or atelectasis). Roughly 10.3% of ED visits fall within the category of respiratory diseases, with 3.1% due to upper respiratory infections, 1.7% due to asthma, and 1.4% due to pneumonia (3.4% in patients over 65 years old).

Pneumonia

CXR and chest imaging are generally of limited value in upper respiratory infections. On CXR, pneumonia manifests as areas of increased density within the pulmonary parenchyma. The increased density is usually focal but may be multifocal or diffuse. Depending on the causative organism and the health and immune status of the patient, the density may manifest as confluent parenchymal consolidation (where vessels are obscured within the density), as hazy ground glass density (where vessel margins can be detected within the density), or as coarse reticular or reticulonodular opacities. These densities may be marginated by pleural or fissural boundaries. If available, the use of both posteroanterior (PA) and lateral projections is often helpful in detecting, confirming, or localizing areas of increased parenchymal density (Figs. 8-2 and 8-3).

Figure 8-2 Chest x-ray of a patient with pneumonia. A, Posteroanterior view reveals a patchy opacity overlying the left lung base (large arrow). Note that the left cardiac border is preserved, indicating that the pneumonia is more likely in the left lower lobe. If it were lingular and abutting the left cardiac border, the left cardiac border would be obscured (the “silhouette” sign). B, The lateral view confirms left lower lobe location of the pneumonia, as indicated by increased density of the lower thoracic vertebral bodies (arrow). Normally, on a lateral chest x-ray, the thoracic vertebral bodies become increasingly lucent from superior to inferior. Increased density of lower thoracic vertebral bodies indicates the presence of radiodense material (in this case, a left lower lobe pneumonia) attenuating the x-ray beam.

Figure 8-3 Chest x-ray of another patient with pneumonia. A, Posteroanterior view reveals a faint opacity overlying the right lung base (arrows). The density does not abut the right heart or diaphragm, and thus the “silhouette” sign is of no help in determining middle versus lower lobe location. B, The lateral view reveals increased density overlying the heart (longer arrow), confined by the major fissure (shorter arrows), indicating right middle lobe location of the pneumonia. Note the vertebral bodies become more lucent from superior to inferior, indicating the absence of lower lobe processes overlying the vertebrae.

Heart Failure

There were 5.2 million ED visits for congestive heart failure (CHF) in 2004. At age 40, the lifetime risk of developing CHF for both men and women is one in five. Progression or exacerbation of CHF can cause dyspnea due to pulmonary vascular congestion or its more advanced form, pulmonary edema. Pulmonary edema can be classified as of hydrostatic origin (increased intravascular pressure, as in left ventricular failure or volume overload) or of permeability origin (increased permeability of the pulmonary alveolar-capillary membrane, as in acute respiratory distress syndrome [ARDS]). Although there is likely some overlap in the pathophysiology, the etiology of pulmonary edema cannot routinely be distinguished by CXR.

The sequence by which CHF leads to pulmonary edema and may manifest on CXR is familiar. Left ventricular failure results in increased pressure within the pulmonary venous system leading to pulmonary venous distention. The increased size of pulmonary veins on CXR, particularly in the upper lobes, manifests as “cephalization,” for which the determination can be relatively subjective at imaging. As pulmonary venous capacitance is exceeded, pulmonary vessels begin to leak transudative fluid into the pulmonary interstitium, including the septal, peribronchial, and perivascular spaces, and into the pleural space. The CXR correlates include septal lines (or “Kerley B” lines), peribronchial cuffing, indistinctness of vessel margins, and pleural effusions, respectively. With further progression, fluid begins to accumulate in alveolar spaces resulting in more confluent density that obscures vessel margins but in contrast accentuates the appearance of bronchial airways resulting in air bronchograms (Fig. 8-4).

Figure 8-4 Patient with cardiomegaly and recurrent episodes of congestive heart failure. A, Posteroanterior view obtained when the patient was otherwise healthy reveals grossly clear lungs and cardiomegaly. There is a normal paucity of lung markings at the periphery of the lungs near the pleural margins. B, A posteroanterior view obtained when the patient was in congestive heart failure with pulmonary edema reveals diffusely increased lung markings and faint opacification at the bases. C, A magnified view of the right lung base reveals prominent septal markings in the subpleural lung periphery, or “Kerley B” lines (arrow), as well as more consolidated density overlying the right lung base (asterisk).

COMPUTED TOMOGRAPHY

CT of the Pulmonary Parenchyma

Computed tomography allows for more refined analysis of the pulmonary parenchyma. Similar to CXR, pulmonary parenchymal pathology is detected by changes in parenchymal density and architecture. With section thicknesses of 2 mm or less, axial images allow for evaluation of pulmonary parenchymal architecture at the level of the secondary pulmonary lobule, the smallest unit of lung structure marginated by connective tissue septa. The secondary pulmonary lobule consists of small and terminal bronchioles (accompanied by terminal arterioles) along with their associated acini (usually 12 or fewer in number). This complex of terminal bronchovascular structures and their associated acini is confined by interlobular septae. The interlobular septae contain pulmonary veins and pulmonary lymphatics, the latter also found along the bronchovascular complex. Changes to the architecture of the secondary pulmonary lobule occur in a multitude of disease processes, and characterization of the architectural changes allows for refinement of differential diagnostic considerations. Of course, larger-order anatomy is clearly defined by CT as well.

Changes to the architecture and density of the secondary pulmonary lobule can be organized according to the structures involved, including peripherally located septal structures, centrally located bronchovascular structures, and more generalized abnormalities affecting both. A detailed instruction on CT of pulmonary parenchyma is beyond the scope of this chapter, but several excellent resources are available on this subject, including the Requisites text on thoracic radiology. However, a general overview of the differential considerations associated with architectural derangements described above is useful, and a brief synopsis is provided below.

Septal thickening may be smooth, nodular, or irregular. Differential considerations for smooth septal thickening include pulmonary edema (with septal thickening being the equivalent of “Kerley B” lines seen on CXR) or ARDS, and pneumonia or pulmonary hemorrhage (Fig. 8-5). Septal thickening seen with these entities may be associated with “ground glass” density, the so-called crazy paving pattern. The “crazy paving” pattern has also been described as characteristic of alveolar proteinosis. Nodular septal thickening may be seen in lymphangitic metastases as well as sarcoidosis and silicosis. Nodules seen in sarcoidosis are characterized as perilymphatic, and thus may be centrilobular or septal (lymphatics are associated with both the septae and bronchovascular structures). Pulmonary fibrosis may also result in septal thickening, which may be irregular in nature and may be associated with honeycombing or other common findings of fibrosis, such as traction bronchiectasis.

Figure 8-5 Patient with congestive heart failure and pulmonary edema. Thickened interlobular septae (arrows) define the borders of the intervening secondary pulmonary lobules. Additional findings that may be seen in congestive heart failure/pulmonary edema are shown, including centrilobular opacities (white oval), ground glass opacity (black oval), and pleural fluid seen posteriorly (single asterisk) and within the major fissure (double asterisks).

Centrilobular opacities or nodules generally imply abnormalities of the small airways or vascular diseases, and are usually related to bronchiolitis of varying etiologies. The combination of centrilobular opacities/nodules with dilated/opacified bronchioles has been termed the “tree-in-bud” sign (Fig. 8-6). In the ED, centrilobular opacities/nodules are usually seen in the setting of an infectious bronchiolitis that may be due to bacterial or atypical organisms. Pulmonary aspiration can give a similar appearance. Other differential considerations for centrilobular opacities/nodules include pulmonary edema and hemorrhage, vasculitis, bronchoalveolar cell carcinoma, acute or subacute hypersensitivity pneumonitis, noninfectious bronchiolitis (as can be seen in smoking-related interstitial lung disease/respiratory bronchiolitis), and panbronchiolitis.

Figure 8-6 Examples of centrilobular opacities (arrows) and the “tree-in-bud” sign (ovals) in a patient with a history of nonspecific interstitial lung disease who presented with an acute cough. Subsequent CT scan (images not provided) showed resolution of the centrilobular opacities, which were therefore presumed infectious (bronchiolitis or bronchopneumonia) in etiology.

More generalized (panlobular) processes may lead to either ground glass or more consolidated opacities. If ground glass opacities are detected, the differential diagnosis is assisted by identification of ancillary findings. Ground glass opacities in the presence of honeycombing or other findings of fibrosis should point toward differential considerations among fibrotic lung diseases. Ground glass opacities in the presence of septal thickening (“crazy paving” pattern) should raise differential considerations including pulmonary edema, pneumonias of varying etiologies, and pulmonary hemorrhage, as well as entities less commonly encountered in the ED, including hypersensitivity pneumonitis, ARDS, and alveolar proteinosis. If ground glass opacities are encountered in the absence of ancillary findings the differential considerations become even broader and include interstitial pneumonia, infection, and drug reaction, in addition to the entities listed above (Fig. 8-7). Differential considerations for more consolidated densities include entities such as pneumonias, pulmonary edema, ARDS, and pulmonary hemorrhage, as well as acute eosinophilic pneumonia, masses, and fibrotic processes. Familiarity with the concept of evaluating pulmonary parenchyma based on analysis of the structures of the secondary pulmonary lobule will serve emergency radiologists well in their approach to formulating differential diagnoses for parenchymal disease.

Figure 8-7 Example of ground glass opacities due to pulmonary edema in a postpartum patient with a history of preeclampsia, who presented with leg swelling and shortness of breath. Note the patchy hazy density that does not obscure the underlying vascular structures.

Chest CT Techniques and Protocols

This section outlines general principles guiding the selection of CT techniques and scan parameters. While there can be wide variability between scanner types and manufacturers in technological capabilities, nomenclature, and technical detail, the basic underlying considerations are applicable across these platforms. A working knowledge of these considerations can aid the radiologist in optimizing CT scan performance while remaining good stewards of radiation safety.

Patient Screening

Intravenous Contrast Material

All patients must be screened for allergy to intravenous contrast material (IVCM). While IVCM is a critical part of evaluation for pulmonary embolus and aortic dissection, it can be omitted if the clinical indication is primarily for abnormalities of the lungs. Although protocols vary by institution, all policies should include screening for a history of contrast allergy and risk factors for contrast-induced nephropathy.

Pregnancy

Women of child-bearing age should be screened, as concerns regarding the use of ionizing radiation during pregnancy must be weighed against medical necessity (e.g., pretest probability, urgency), exam performance characteristics (e.g., sensitivity, specificity), and alternative methods of imaging (e.g., MRI, ultrasound). Medically unnecessary exams should not be performed. Alternative imaging methods with similar exam performance characteristics should be employed when available. If CT or radiography is used, protocol modifications are frequently warranted during pregnancy, such as restricting to a single PA view chest radiograph with shielding of the abdomen, or elimination of pelvis and lower extremity scanning as part of CT evaluation for pulmonary embolus. Multiple-phase CT scanning, or so-called multi-pass CT protocols, should be altered to single-phase scanning unless multiple phases are critical to patient evaluation. The number of nondiagnostic scans should be minimized by taking proactive steps to optimize CT scan quality (e.g., patient coaching for breathing and motion, excellent IV placement), thereby reducing the need for repeat scanning.

Patient Preparation

Intravenous Line Placement

Whenever possible, contrast-enhanced chest CT studies should be performed through a 20-gauge or larger peripheral IV, preferably located in a right antecubital location. An 18- to 20-gauge size will easily allow for power-injected flow rates of 3 to 6 cc per second. Right-sided IV placement reduces streak artifact over the anterior superior mediastinum as compared with left-sided placement, due to elimination of dense IV contrast passing through the left brachiocephalic vein. Antecubital placement results in more reproducible contrast inflow characteristics and a reduced rate of infiltration as compared with more peripheral placement.

Heart Rate Control

Cardiac gated protocols (coronary CT angiography [CTA] for gated aortic dissection) generally require a regular cardiac rhythm. With current single-source 64-MDCT techniques, coronary CTA is best performed with heart rate controlled to 65 beats or less per minute. However, further advances such as quicker gantry rotation, greater numbers of detector rows, and/or dual x-ray source scanners are likely to lessen or obviate the current requirement for lower or regular heart rates.

Scan Protocol Considerations

Indication-specific CT Protocol Techniques

Routine Chest CT

This is often performed without IVCM to further assess a radiographic parenchymal abnormality (e.g., pulmonary nodule). However, IVCM is helpful for assessment of empyema or mediastinal or hilar abnormalities.

Pulmonary Embolus CTA

A typical protocol includes bolus detection on the main pulmonary artery, followed by a breathhold delay and scanning in the caudal to cranial direction. Most other thoracic CT protocols utilize a cranial to caudal scan direction. This scan direction is often reversed in pulmonary embolus CTA (CTPA) in order to minimize respiratory motion artifact in the lower lobes, where the majority of pulmonary emboli occur.

If a lower extremity venogram is not included, a high-quality CTPA may be performed with approximately 80 mL of IV contrast injected at 4 mL per second. If a lower extremity venogram is performed following CTPA, 100 to 125 mL of IVCM are instead needed, and images are acquired at intervals from the pelvis to the knees after an approximately 3-minute delay.

Upper Extremity Deep Venous Thrombosis or Superior Vena Cava Syndrome

A delayed venous phase scan should be performed if there is high clinical suspicion for upper extremity or central vein thrombosis. The delayed phase produces more uniform venous opacification and eliminates the mixing artifacts typically observed from dense IVCM injected through one arm mixing in the SVC with the unopacified blood arriving from the other arm (Fig. 8-8). Coronal reformations can greatly aid diagnosis.

Figure 8-8 Superior vena cava (SVC) thrombus following recent central line removal. A, Early phase axial image. Dense contrast in the SVC (curved arrow) produces streak artifact that partially obscures the SVC thrombus (arrow) and makes it difficult to differentiate streak artifact from mixing artifact with unopacified blood from the left brachiocephalic vein. B, Delayed (90 seconds) phase axial image eliminates streak and mixing artifacts.

Aortic Dissection CTA

Currently, full dissection CTA protocols may include:

• A noncontrast scan of the chest only.

• An arterial phase scan through the chest (+/- abdomen and pelvis) performed either with or without cardiac gating.

• A delayed phase scan through the chest, abdomen, and pelvis.

The initial noncontrast scan is used to assess for a thrombosed false lumen or isolated intramural hematoma, both of which will appear denser than the unopacified blood in the aortic lumen. If a noncontrast scan was not initially obtained (such as a pulmonary embolus CTA study), a 20- to 30-minute postcontrast delayed scan can be obtained, as IVCM will have adequately cleared the intravascular space. This permits the radiologist to troubleshoot the rare case in which a contrast-enhanced scan may raise suspicion for isolated intramural hematoma but is not sufficiently diagnostic for the radiologist.

The arterial phase scan is often performed with cardiac gating in order to reduce the pulsation artifact, which can mimic focal aortic dissection at the aortic root. However, this artifact has a characteristic appearance on axial and MPR images, and can typically be readily differentiated from aortic dissection (Fig. 8-9). If the radiologist is readily familiar with this diagnostic pitfall, a nongated arterial phase scan can be used so as to reduce radiation exposure, which is significantly greater on cardiac-gated scans. If cardiac gating is utilized, a regular heart rate is required, preferably 65 or fewer beats per minute. For younger ED patients (generally younger than 50 years old) with a low pretest probability of aortic dissection and a low incidence of significant atherosclerotic disease, we screen with a physician-monitored, nongated, contrast-enhanced chest CT only in order to reduce the radiation exposure from multiple phase protocols. If needed, additional scanning can be readily undertaken if results of this scan show anything other than a normal aorta. Very rarely is this necessary.

Figure 8-9 Pulsation artifact: typical aortic root pulsation artifact from a non-gated study, not to be confused with aortic dissection. A, A sharply defined band of low-density artifact (longer arrow) extends across the left side of the proximal ascending aorta. The artifact extends beyond the aortic margin to the right atrial appendage (shorter arrow), helping to confirm its artifactual nature. B, Sagittal reformation showing motion artifacts (arrows) arising in each cardiac cycle. C, Repeat study performed the following day with cardiac gating, eliminating the cardiac motion artifact.

The CT scan through the abdomen and pelvis is performed to assess distal extension of the intimomedial flap, involvement of visceral or iliac arteries, and end organ ischemia. Therefore, it should be prescribed as routine in high-risk patients or those with known aortic dissection presenting with acute symptoms. But the abdomen and pelvis can be scanned selectively in patients with low pretest probability. For those in whom dissection was identified by active physician monitoring of the screening chest CT, the scan can readily be continued through the abdomen and pelvis during the arterial phase with delayed imaging acquired as needed. This is a viable option for institutions with radiologists who can actively monitor CT acquisitions while the patient is on the gantry table. Although the 24/7 operations of the ED can make this challenging to implement during nonbusiness hours, greater organization around the practice of emergency radiology is likely to make this more feasible in the future.

Pneumomediastinum and/or Evaluation of Esophageal Injury

In the ED, CT is often used rather than fluoroscopy to assess for possible esophageal injury. The scan protocol should include an initial scan with neither oral nor intravenous contrast. Subsequently, the patient should take a few swallows of Gastrografin, and a repeat scan (with or without intravenous contrast, depending on indication) should be performed. The “scout” noncontrast scan is included in order to readily differentiate Gastrografin extravasation from soft tissue calcification or a small foreign body (e.g., a fish bone). Because both gas and contrast material extravasation are intrinsically high-contrast evaluations, low-dose techniques can be used to reduce radiation exposure.

Coronary CTA

There can be significant variability in CT scan technique depending on local CT scanner technology and institutional decisions about whether to perform a focused coronary CTA only or a “triple rule-out” that simultaneously assesses the aorta and pulmonary and coronary arteries. The numerous variables and technical details are beyond the scope of this chapter. Briefly, however, the study may be prescribed from a topogram or a low-dose unenhanced chest CT performed for calcium scoring. The CTA is then either initiated by automated bolus detection or, following a scan delay, derived from a test bolus. Cardiac gating is required, preferably with a heart rate equal to or less than 65 beats per minute. Tube current dose modulation schemes, if available, are recommended to reduce the associated radiation exposure by limiting tube current during the systolic portion of the cardiac cycle. Images are reconstructed at multiple intervals of the cardiac cycle to permit identification of image sets free of cardiac motion artifact. Numerous efforts in CT scanner technology are aimed at optimizing coronary CTA evaluation. As these technological advances are exploited through innovative clinical applications, CT protocols and scanning requirements for coronary CTA are likely to evolve in the short term.

THE PLEURA, PERICARDIUM, AND MEDIASTINUM

The Pleura

The pleura is a specialized tissue layer covering the lungs and the internal surface of the chest cavity. The visceral and parietal components form a continuous layer and create a potential cavity, the pleural space, which is generally visualized radiographically only when local or systemic pathology develops. Normally a small amount of fluid (15 to 20 mL) lies in each pleural cavity and lubricates the sliding lung surface during the respiratory cycle. Pleural fluid originates in the interstitium and is drawn away from the pleural space along a gradient toward the low-pressure pleural capillaries, ultimately draining into the pulmonary venous system. Pleural fluid is continuously secreted and absorbed; disruption of this equilibrium leads to effusion.

The pleural space is maintained below ambient pressure by chest wall recoil, the actions of gravity, and muscular contraction of the diaphragm. Loss of vacuum integrity leads to rapid collapse of the lung airspaces.

Pneumothorax

Pneumothorax is the presence of air within the pleural space. The etiologies of pneumothorax are multifactorial (Table 8-1). The most common imaging test performed to assess for pneumothorax is the chest radiograph, or CXR. Ideally, this is performed with upright patient positioning so that the air–pleural interface can be readily identified toward the apices. Expiratory imaging may accentuate the contrast between lung tissue and intrapleural gas. This method may be particularly helpful for smaller pneumothoraces, but is generally not required. The presence of pneumothorax is often more subtle on supine radiography, and small pneumothoraces can be easily overlooked.

Table 8-1 Etiologies of Pneumothorax

Etiology	Types
Trauma	Penetrating or blunt; barotrauma
Idiopathic	Primary or secondary
Iatrogenic	Following biopsy, surgery, central line placement, thoracentesis, mechanical ventilation, bronchoscopy
Infection	Pneumocystis carinii, necrotizing pneumonia
Neoplastic	Langerhans’ cell histiocytosis, sarcoma, lung cancer, various others
Congenital/Acquired	Emphysema, congenital bullae, lymphangioleiomyomatosis

Spontaneous pneumothorax occurs more frequently in smokers, asthmatics, and thin young men. The relative risk for male smokers is up to 20 times higher than for nonsmoking males. Risk increases in a linear relationship to the amount of smoking. Blebs are often found at thoracoscopy of patients presenting with spontaneous pneumothorax, and can be visualized by CT in up to 89% of patients. For this reason, chest CT is often ordered as a follow-on study to a CXR showing an unexpected spontaneous pneumothorax.

Primary spontaneous pneumothorax is differentiated from secondary spontaneous pneumothorax by the absence of underlying lung disease. Secondary spontaneous pneumothorax is due to underlying pleural or parenchymal disease, usually in the form of blebs or cysts. These may result from disease entities such as emphysema, Pneumocystis carinii pneumonia infection, or lymphangioleiomyomatosis. Spontaneous pneumothorax is sufficiently common as a component of the differential diagnosis for chest pain and dyspnea that a high index of suspicion should be maintained when evaluating routine chest radiographs for these indications. The frequency of central line placement in the ED is another reason to maintain a high index of suspicion for pneumothorax when evaluating all ED chest radiographs.

A tension pneumothorax occurs when air enters but cannot exit the pleural space. This can occur when the pleural defect allowing air passage into the pleural cavity mechanically functions like a ball-valve resulting in unidirectional air transit. The increasing intrapleural gas creates mass effect, and eventually can displace the mediastinum and compromise venous blood return to the heart. Progressive accumulation of air in the pleural cavity can lead to cardiovascular impairment and, if left untreated, death. Tension pneumothorax should be suspected when the trachea, heart, and/or mediastinum have been displaced from the midline, when the ipsilateral rib spaces have widened, and/or when the contralateral lung appears relatively compressed.

Several estimation methods for pneumothorax size are based on CXR measurements, but these are subject to wide interobserver variations in practice. A pragmatic rule of thumb is that a large pneumothorax is one that completely dehisces from the chest wall, while a small one does so only partially. This simple criterion identifies the majority of pneumothoraces requiring evacuation. If there is ambiguity as to whether an apical lucency represents intrapleural air or a large bulla, CT is generally definitive. Chest CT also permits volumetric or percent measurements of pneumothorax size that are more precise, although intervention usually occurs prior to CT evaluation.

Small pneumothoraces are usually followed radiographically to resolution. Failure of resolution and interval enlargement are common indications for invasive management. The treatment for a large pneumothorax (generally 15% or more of the hemithorax) is air aspiration or evacuation by intercostal chest tube or catheter. The radiologist should be familiar with the normal locations for appropriately inserted chest tubes. Intraparenchymal, extrathoracic, and interlobar placements should be communicated to the treating physician to allow for repositioning. The multiple sideports of the chest tube should all lie within the thoracic cavity. Increasing subcutaneous emphysema on repeat chest radiographs is a clue to the potential presence of an extrathoracic sideport and the need for repositioning or downsizing of the chest tube.

Additional radiographic signs associated with pneumothorax include the deep costophrenic sulcus sign seen on the supine radiograph. The “double diaphragm” sign is caused by nondependent air outlining the anteriordiaphragmatic attachment. Mimics of pneumothorax are multiple and may be due to overlying sheets, clothing, medical equipment, or skin folds. Repeat chest radiography after rearranging or removing the external structures is often adequate to exclude pneumothorax.

Recently, ultrasound (US) has gained attention as a rapid assessment tool to evaluate for pneumothorax in the ED and after chest interventions. The performance of US has been shown to be as good if not better than radiography, although it generally underperforms as a follow-up study. Additionally, greater variation in results should be expected due to differences in sonographer skill and experience. Traumatic pneumothorax is discussed in this text in Chapter 2, so only a brief comment is made here. The seal of the pleural space is easily broken by penetrating injury, and it is vulnerable because it is spread over a large surface area. Significant blunt trauma is estimated to cause pneumothorax in up to 30% to 40% of cases. It can lead to pneumothorax when broken ribs or other sharp structures expose the pleural space to atmospheric pressure, or when a preexisting bleb is ruptured. Fractured ribs on imaging should invoke a secondary search for pneumothorax. With positive pressure ventilation, barotrauma can lead to pneumothorax, particularly in patients with relatively noncompliant lung tissue (e.g., those with ARDS or interstitial lung disease). Barotrauma may also occur during sport diving or from nonpenetrating blast injuries resulting from explosive devices.

Pleural Effusion

Pleural effusion, defined as abnormally increased fluid within the pleural space, may be transudative or exudative in etiology. Transudative fluid is generally low in protein and typically develops in the setting of left heart failure, decreased oncotic pressure, or other systemic abnormality. Exudative fluid is often a consequence of breakdown of the barrier function of the pleura, leading to leakage of macromolecules and proteins. Less commonly encountered complex exudative pleural effusions include hemorrhagic fluid related to tuberculous infection, chylothorax related to thoracic duct injury, and eosinophilic effusions related to hypersensitivity (Table 8-2).

Table 8-2 Etiologies of Pleural Collections

Etiology	Types
Trauma	Hemothorax, chylothorax
Infectious	Empyema, tuberculous effusion
Gestational/Fertility	Ovarian hyperstimulation
Inflammatory	Postpericardiotomy, eosinophilic
Drug induced	Multiple factors
Reactive	Ascitic, pancreatitis

Pleural fluid volume as low as 5 mL can be visualized on decubitus imaging. Approximately 200 mL is required to blunt the costophrenic sulcus on upright radiography. Small fluid collections can routinely be characterized on CT, US, and MRI. The CT appearance of pleural fluid varies depending on the density. Transudative effusions are near water density (i.e., approximately 0 Hounsfield units [HU]); hemorrhage and pus typically demonstrate increased density, often with a hematocrit effect in the case of hemorrhage. Low density on CT does not, however, exclude exudate. Long-standing infections may show complex septations or gas. US may help to characterize effusions by showing areas of loculation and septation, and is also commonly used for drainage guidance.

Pleural effusions may also be seen in the setting of malignancy, and occasionally in the setting of benign lesions, as seen in Meigs syndrome. Classically, Meigs syndrome is the presence of ascites and pleural effusion caused by benign ovarian fibroma, although other ovarian neoplasms are now commonly invoked as well. Meigs syndrome is uncommon and the pathophysiology is unclear. Most sources accept the explanation that the fluid is generated by the tumor itself and is often, but not exclusively, a transudate.

Hemothorax

Traumatic hemothorax can occur following blunt or penetrating trauma. Blunt trauma may cause direct venous or arterial rupture or may disrupt vessels secondary to rib fracture. Penetrating trauma is capable of mediastinal, intercostal, substernal, or diaphragmatic vascular injury with resultant intrapleural hemorrhage.

Intrapleural blood may manifest on the conventional radiograph as a localized collection or as diffuse relative opacification of one or both hemithoraces. Much smaller collections can be detected on CT with the added certainty of density characterization allowed by evaluation of HU in a region of interest. CT, particularly when performed during intravenous contrast administration, may identify active extravasation from arterial injury. Progressive enlargement of hemothorax is likely due to arterial injury, while venous hemorrhage is more likely to stabilize. Unevacuated hemothorax, particularly if large, can develop adhesions and fibrothorax, potentially restricting lung function.

Empyema

Empyema refers to abscess formation in the pleural space. Although usually secondary to adjacent pneumonia, it is also known to develop after hematogenous seeding, trauma, or iatrogenesis. In the setting of suspected empyema, chest CT should be performed with administration of IVCM. The presence of a “split pleura” sign is strongly suggestive of empyema. Several studies have shown improved health outcomes with early evacuation of empyema, often in conjunction with fibrinolytic therapy.

Chylothorax

Chylothorax is the development of effusion containing lymphatic fluid, most often secondary to thoracic duct injury or malignant invasion. The thoracic duct drains into the junction of the left axillary and internal jugular veins, with the majority of the course from the cisterna chyli lying on the right. Consequently, most chylous effusions are right-sided. A minority are bilateral. CT is helpful in diagnosis, as chylous fluid contains fats, which manifest as HU measurements less than zero.

Pericardium

The pericardium is an investing connective tissue layer covering the heart and proximal great vessels, separating the cardiac structures from the remainder of the middle mediastinum.

Pericardial Effusion

The etiologies of pericardial effusion are many and imaging is often nonspecific in this regard. Etiologies may include idiopathic, drug-induced, collagen vascular diseases, radiation-induced, malignancy, hypothyroidism, chylous effusion, infection (pericarditis), cardiac failure, and renal failure.

Hemopericardium

Hemopericardium is most often the result of acute blunt or penetrating trauma, from direct pericardial damage, myocardial contusion, or proximal aortic injury. Hemorrhage into the pericardial sac can rapidly lead to tamponade and circulatory collapse. In the absence of a history of trauma, type A aortic dissection, coronary artery aneurysm rupture, or postinfarct cardiac rupture must be considered. However, a dense pericardial effusion is not a sine qua non of hemopericardium or incipient tamponade. Radiographic and CT findings must be interpreted in conjunction with history and physical exam findings. Echocardiography should be performed if there are features causing concern for hemodynamic compromise (Fig. 8-10).

Figure 8-10 Axial CT images showing hemopericardium (white arrows) secondary to type A aortic dissection (black arrows indicate type A dissection flap within the ascending aorta; apparent duplication of the flap may be due to motion). Dissection extending into the descending thoracic aorta, with a hyperdense true lumen (curved arrow) and hypodense false lumen (asterisk), is noted.

Cardiac Tamponade

Cardiac tamponade is a potentially life-threatening restriction of cardiac function due to rapid influx of fluid or gas into the pericardial space, regardless of its source or nature.

Pneumopericardium

Pneumopericardium can result, uncommonly, from blunt trauma or penetrating injury, or may develop after positive pressure ventilation, often in association with pneumothorax. It has also been associated with illicit drug use, particularly crack cocaine inhalation, and is thought to be probably secondary to repeated Valsalva maneuvers. Although pneumopericardium generally resolves with withdrawal of the causative insult, tension pneumopericardium can develop due to a ball-valve mechanism and should be suspected if the cardiac contour is abnormally diminished, known as the “small heart” sign.

Malignancy

Metastatic involvement of the heart and pericardium is not unusual, occurring in 10% to 12% of all cancer patients and 30% of those with lung malignancy. The development of symptoms such as dyspnea, cough, and chest pain are potential clues to cardiac or pericardial spread, although these are quite nonspecific. Contrast-enhanced chest CT to exclude pulmonary embolism is a common initial evaluation for oncology patients presenting to the ED with new chest symptoms. Foreknowledge of a patient history of malignancy, particularly lung cancer, should prompt careful visual review of the heart and pericardium for alternative or additional explanation of symptoms.

Other pericardial diseases include pericarditis and constrictive pericarditis. Pericarditis may cause a pericardial effusion and/or pericardial thickening. The normal pericardial thickness varies from 1 to 3 mm, depending on the location and cardiac phase. Constrictive pericarditis may or may not be associated with pericardial thickening.

Mediastinum

Anatomy

The mediastinum is a space within the central chest cavity bounded laterally by the medial pleural surfaces, anteriorly by the sternum, and posteriorly by the spinal column. The inferior margin is bounded by the diaphragm, while the superior limit is usually arbitrarily defined by a line drawn horizontally from the manubriosternal joint to the T4 vertebra.

There are several ways to segment the mediastinum anatomically. The simplest method trisects the intrathoracic structures into anterior, middle, and posterior mediastinal compartments based on location relative to the heart.

Because the three main mediastinal compartments contain very different structures, localization of a lesion to one of the mediastinal compartments considerably aids the differential diagnosis for most abnormalities. The anterior compartment normally contains thymus (or thymic remnant), lymph nodes, and fat. During embryonic development, pluripotent germ cells migrate inferiorly through this region. The middle mediastinum contains the heart and pericardium, great vessels, tracheal bifurcation, phrenic nerves, and lymph nodes. The posterior mediastinum contains the esophagus, thoracic duct, vagus nerve and sympathetic chain branches, the azygos and hemiazygos veins, and the descending aorta.

Masses

The anatomic compartmentalization of the mediastinum described above assists the radiologist in formulating differential diagnostic considerations for mediastinal masses. Mediastinal masses may be detected as incidental findings in ED patients undergoing chest imaging, or they may be the cause of the patients’ presenting symptoms, such as chest pain, dysphagia or dyspnea, cough, hemoptysis, or SVC syndrome.

Pneumomediastinum

Typical causes of pneumomediastinum include trauma, asthma, drug use (Valsalva-induced), infection, and spontaneous pneumomediastinum. Traumatic causes (other than typical blunt or penetrating external trauma) may be the result of scope, tube, or line insertion (i.e., iatrogenic), positive pressure ventilation, or internal barotrauma related to vomiting (Boerhaave syndrome) or coughing against a closed glottis (Fig. 8-11). Pneumomediastinum generally requires no treatment apart from addressing the underlying cause.

Figure 8-11 CT showing spontaneous pneumomediastinum (black and white arrows) in an asthmatic patient.

Infection

Acute mediastinitis is a relatively rare entity, most commonly (90%) related to esophageal rupture. Other causes include tumor necrosis, spread of infection from adjacent compartments (e.g., descending cervical infection due to spread of oropharyngeal or deep neck infection along fascial planes), and infection secondary to procedures (e.g., sternotomy, mediastinoscopy). Signs and symptoms are nonspecific and include fever, tachycardia, and chest pain. Dysphagia is often noted in cases related to esophageal perforation. Conventional radiography may reveal tracking of air along the mediastinal planes, with variable extension into the soft tissues of the neck. CT lung windows are much more sensitive than radiography.

SVC Syndrome

SVC syndrome refers to the development of facial and upper extremity edema, swelling, and collateral venous engorgement secondary to superior vena cava obstruction. Associated findings include upper body erythema, confusion, chest pain, and hoarseness. In the developed world, most cases are related to malignant invasion of the SVC. Historically, tuberculous infection and syphilitic aneurysm were more common causes. Indwelling central catheters have come to play a larger role in recent years and can lead to SVC syndrome via stricture and thrombosis. Other causes of SVC syndrome include extrinsic compression by aneurysm or lymphadenopathy and fibrosing mediastinitis usually secondary to the inflammatory response elicited by infection with Histoplasmosis capsulatum. In younger patients, benign causes predominate. Malignancy is the most common cause over age 40.

Because SVC obstruction impairs venous drainage of the head and neck, cerebral and laryngeal edema may occur. For these reasons, the development of SVC syndrome is considered a medical emergency, although only a minority of patients are at risk of death from cerebral edema.

CXR may reveal a mediastinal abnormality in patients with clinically suspected SVC syndrome, such as a mediastinal mass or lymphadenopathy. Contrast-enhanced CT and MR angiography are the most reliable noninvasive methods of definitive diagnosis. CT is generally preferred as the initial study, owing to the prevalence of lung cancer and lymphoma as causative factors in the adult population and the ease and accuracy with which CT delineates these findings (Fig. 8-12).

Figure 8-12 SVC syndrome: axial and coronal contrast-enhanced CT. Multiple collaterals and dilatation of azygos and internal mammary veins (arrows) in a patient with severe SVC narrowing due to tumor mass encasement (oval).

THE THORACIC AORTA

Relevant CT Anatomy

On axial sections, the aorta is generally round and its diameter should gradually taper more distally as vessels branch from its lumen, although there are some notable exceptions that represent normal variants in its size and shape. It is important to be familiar with the appearance of these variants at imaging so that they are not mistaken for disease.

Normal variants of the aortic contour identified on CTA are best recognized on MPR or three-dimensional (3D) images; these include the aortic spindle, ductus diverticulum, branch vessel diverticula, and pseudocoarctation. The aortic spindle is a smooth circumferential bulge below the region of the isthmus, representing normal mild dilatation of the region of the posterior arch. The ductus diverticulum is a focal convex bulge along the anterior undersurface of the aortic isthmus. The obtuse angles it forms with the aorta can be used to help distinguish it from a post-traumatic pseudoaneurysm, which more characteristically forms acute angles with the aortic wall. Branch vessel diverticula show smooth obtuse margins with a branch vessel emanating from the apex of the infundibulum. These too can be mistaken for a post-traumatic pseudoaneurysm at a branch vessel origin. Recognition of these is aided by their tendency to occur in characteristic locations, such as the left subclavian artery and the third right intercostal artery. Pseudocoarctation results from elongation of the aortic arch, which characteristically results in kinking at the site where the aorta is tethered by the ligamentum arteriosum. Aortic wall thickness is best evaluated on axial CT images. Imaging of healthy adults has shown that the aortic wall thickness is less than 1 mm or imperceptible on axial CTA images.

Aneurysm

Thoracic aortic aneurysm is defined as an aortic size that is 50% greater than the expected aortic diameter. However, in practice, a 5-cm axial dimension is most often used since intervention is otherwise rarely considered in the asymptomatic patient. Distal aortic segments should generally be smaller in diameter than more proximal segments except for the previously noted anatomic variants. If this relationship is reversed, aortic monitoring for aneurysm development should be considered. CT can characterize thoracic aneurysms by their location and shape. Three fourths of aortic aneurysms are atherosclerotic in etiology, fusiform in shape, and located in the descending aorta. However, approximately 20% of atherosclerotic aneurysms are saccular, particularly in the arch and descending portions. Incidence of aortic rupture is related to aneurysm size. An important goal for imaging is to provide accurate measurements. Conventional catheter-based aortography underestimates true aortic diameter, as it reflects only the size of the patent lumen and does not include contributions from intraluminal thrombus or mural thickening. CT accurately evaluates aneurysm dimensions, the extent of intralumenal thrombus and mural thickening, the integrity of the aortic wall, and its relationship with contiguous structures. The chapter in this text on vascular emergencies contains a more detailed section on thoracic and abdominal aortic aneurysms and their complications. Thoracic aortic aneurysms carry a high incidence of concomitant abdominal aortic aneurysms such that screening of the abdomen should generally be undertaken with CT or US when thoracic aneurysm is detected.

Infectious Aortitis

An infected aorta, whether normal caliber or aneurysmal, can represent a diagnostic challenge, as patients may be asymptomatic until late stages or may present with nonspecific clinical symptoms. Nearly 50% of infected aneurysms occur in the thoracic or thoracoabdominal aorta, a significantly higher than expected rate given distribution of atherosclerotic aortic aneurysms. More than 90% show a saccular morphology, often lobulated. Early detection by CT can permit intervention before a rapidly progressive course with sepsis and/or rupture. CT reveals subtle periaortic edema, stranding, and fluid in the initial stages. Rim enhancement of periaortic soft tissues followed by disruption or loss of intimal calcifications often precedes aortic enlargement. Close follow-up CT is advised since development of large aneurysms and marked growth have been reported in short intervals. An infected aorta may also maintain a normal caliber. Periaortic gas and gas within the aortic wall are specific findings but are seen in less than 10% of cases. Lack of mural calcification within an aneurysm suggests a nonatherosclerotic etiology and therefore raises greater concern for infection. Nuclear scintigraphy with labeled leukocytes or gallium-67, when correlated with CT findings, can increase confidence in this diagnosis.

Syphilis, a now infrequent cause of infectious aortitis, has a typical course that can be assessed by CT. CT demonstrates enlargement of the aorta, which progresses to an aneurysm formation that most commonly involves the ascending portion or arch. Most syphilitic aneurysms are saccular, but about 25% are fusiform. The pattern of fine and pencil-line dystrophic calcification is characteristically seen. However, this finding is often obscured by thick, irregular, coarse calcifications of secondary atherosclerosis.

Aortic Dissection

Aortic dissection occurs when intravascular blood breaches the intima and dissects within the media of the aortic wall. Imaging is critical for establishing this diagnosis and for guiding medical or surgical intervention. Chest radiography continues to play a role in the initial assessment of patients with suspected aortic dissection, principally because it provides readily available diagnostic information that helps to exclude other differential considerations (e.g., pneumothorax). Its value is otherwise limited, since a normal CXR cannot exclude, nor can an abnormal CXR confirm, aortic dissection. Findings of aortic and/or mediastinal widening lack specificity and require further evaluation with cross-sectional imaging, while up to 25% of dissection cases will appear normal at radiography. Patients for whom radiography provides insufficient alternate explanation of their symptoms and for whom aortic dissection remains a diagnostic consideration must undergo cross-sectional imaging to exclude dissection.

CT, MRI, and transesophageal echo (TEE) have all been utilized for the diagnosis of acute dissection. CT compares well with MRI and TEE. Although all three modalities were reported to have 100% sensitivity, specificity was highest for CT at 100%. Aortography has a sensitivity of only 88% with false negative diagnoses often related to thrombosed or faint opacification of the false lumen, equal opacification of true and false lumen, unusual intimal tears, and intimal tears proximal to the catheter tip. When CT findings were directly compared with surgical findings, Yoshida found 100% accuracy, sensitivity, and specificity for CT. CT was also effective in identifying the entry site of the intimal tear, as well as determining whether there was pericardial effusion or aortic arch involvement, with 95% sensitivity and 100% specificity as confirmed with surgical results. The advantage of CT in aortic arch involvement has also been confirmed by Sommer, who found a sensitivity/specificity of 93%/97% for CT, compared with 60%/85% for TEE and 67%/88% for MRI.

The principal criterion for CT diagnosis of aortic dissection is the presence of an intimomedial flap separating the true lumen from the false lumen. After intravenous contrast administration, the false lumen may opacify completely, partially, or, if thrombosed, not at all. An intimal flap was identified in 70% of conventional CT studies but has been reported in 100% of volumetric CT studies. Secondary findings, which are less specific but may be helpful in equivocal cases, include displacement of intimal calcifications toward the lumen, aortic widening, and mediastinal and pleural hemorrhage. Calcification of neointimized mural thrombus can occasionally mimic displaced intimal calcifications and lead to false positive diagnosis if not recognized.

CT reliably classifies dissections as either type A or type B (Fig. 8-13; see also Fig. 8-10), the principal criterion for determining surgical (type A) versus medical (type B) intervention. In addition, CTA has been shown to reliably differentiate the true from false lumen. Distinguishing the true from false lumen had been less important with conventional surgical therapy but has since become critical for endovascular management. The false lumen is usually located in the right anterolateral position of the ascending aorta, the superior aspect of the arch, and the left lateral position of the descending aorta. However, the configuration of aortic dissection can be quite variable. The most reliable imaging criteria are identifying the continuity of the true lumen with the undissected portion of the normal aorta either proximally or distally, and identifying the false lumen’s termination into a blind sac. Establishing continuity can be more difficult for dissections that involve the aortic root, originate at the aortic arch, or involve the entire aorta. Lepage found that an acute angle between the dissection flap and outer wall on an axial CT image (the “beak” sign) and a larger luminal cross-sectional area were the two most useful indicators of the false lumen for both acute and chronic dissections as they were both sensitive and specific. The false lumen can be filled with either contrast-enhanced blood or thrombus. “Cobwebs,” which represent thin radiolucent filling defects attached to the aortic wall, are also highly specific for the false lumen but are seen infrequently. Outer wall calcifications are useful for indicating the true lumen in acute dissection but are unreliable in chronic dissection since neointima within a false lumen may calcify. While the direction of flap curvature is not a useful indicator, direct visualization of intimomedial rupture, seen as intimal discontinuity, evaginating intraluminal flap, and communication of the two lumina, reliably identifies the direction of entrance tear and differentiates true from false lumen. Identification of the tear site is important since both surgical and endovascular treatment aim to occlude the tear and induce thrombosis of the false lumen. In patients with aortic arch involvement, one lumen often spirals around the other, resulting in the appearance of three lumens on an axial CT image. The inner lumen invariably represents the true lumen. Enhanced CT typically shows either isodensity or hyperdensity of the true lumen (see Fig. 8-13).

Figure 8-13 Aortic dissection. A, Diagram of Stanford aortic dissection classification. Type A dissections involve the ascending aorta, while type B dissections involve the descending aorta only. Although no longer in common clinical use, the older DeBakey classification scheme is included for reference. B, Image is of the patient shown in Figure 8-10. Axial image at the level of the aortic arch shows an intimal flap (arrow). The false lumen is hypoattenuating as a result of slow flow (F). C, Axial image at a level just above the arch shows involvement of the great vessels (arrows). D, 3D image with volume rendering clearly reveals extent of dissection (arrows).

Periodic imaging of patients with dissection is important regardless of prior surgery. A notable advantage of CT is the detection and characterization of thoracic and abdominal complications. Irregularity of the aortic wall, extravasation of contrast, and hyperattenuating pleural collections may indicate aortic rupture. Retrograde and anterograde extension of the dissection can be readily recognized when compared with prior studies. Chronic dissections should be closely monitored for aneurysmal formation, which occurs in about one third of cases. These aneurysms are prone to rupture and are the most frequent cause of late death.

Type A complications include intrapericardial rupture causing tamponade and great vessel or coronary artery involvement; CT can readily identify these (see Fig. 8-13). With ECG-gated MDCT, coronary and cerebral perfusion can be evaluated using the same contrast bolus. Moreover, ECG-gated MDCT permits dynamic imaging of the left ventricular outflow tract and assessment of aortic valve regurgitation. Further imaging to include an abdominal CTA may also be useful as a predictor of postoperative death in acute aortic dissection. The number of abdominal organs demonstrating decreased parenchymal enhancement was shown to be a strong factor in outcome for patients with acute aortic dissection, whereas the number of dissected branches was not a factor in postoperative death.

Acute Intramural Hematoma

Intramural hematoma (IMH) constitutes 10% to 20% of acute aortic syndromes. Acute IMH may be the result of rupture of the vasa vasorum with hypertension proposed as the most frequent risk factor. The clinical presentation of patients with acute IMH is quite similar to other acute aortic syndromes. Therefore, CT has a major role in the initial diagnosis of IMH and in the detection of early (e.g., ulceration, saccular aneurysm, dissection) and late complications (e.g., fusiform aneurysm, rupture).

CT is the study of choice for making the initial diagnosis and has a reported accuracy of 100%. In the acute stage, unenhanced axial images reveal a hyperdense, smooth crescentic collection located eccentrically in the aortic wall, often without significant narrowing of the adjacent aortic lumen (Fig. 8-14). CT shows the IMH to have a constant circumferential relationship with the aortic wall rather than the spiraling configuration of aortic dissection. Internal displacement of intimal calcifications toward the lumen is typically seen. The ascending aorta and proximal descending aorta are most frequently involved. CTA images show no enhancement within the area of the IMH, and typically there is no branch occlusion.

Figure 8-14 Intramural hematoma (IMH). A, Axial noncontrast image shows an acute IMH as a hyperdense crescentic collection in the distal arch (curved arrow) with internal displacement of intimal calcifications (arrow). B, Axial CT angiography image shows the decreased conspicuity of the arch hematoma (arrow) on contrast-enhanced study. C, Sagittal CT angiography image in a patient presenting concomitantly with an IMH and ulcer-like projection (arrow) in the distal descending thoracic aorta.

It is important to recognize the potential pitfalls in the diagnosis of IMH. If an unenhanced CT is not obtained, contrast in the aortic lumen can diminish conspicuity and even obscure detection of more subtle IMH on enhanced CTA images (see Fig. 8-14). IMH is not to be confused with mural hematoma, which is often associated with atheromatous plaque or aneurysm and is typically irregular and low in attenuation. The distinction is best appreciated on noncontrast axial CT images where calcifications are frequently a marker of the intima. Mural thrombus layers over the calcified intima, whereas IMH is subintimal in location.

While there has been controversy regarding the prognosis and management of IMH, serial CT scans best identify resolution, progression, and/or complications. CT has documented favorable responses with medical treatment with complete resolution of the hematoma and low complication rates. Other studies have found these lesions to be more serious than dissection and recommend surgical replacement of the aorta to prevent a high rate of early rupture. For example, a multicenter study demonstrated a 50% complication rate within 30 days of initial presentation consisting of contained rupture, dissection, and aneurysm. The discrepancy seems to be partly related to patient presentation. In patients with acute chest pain, complications are more common, but a benign course is likely in lesions seen incidentally in patients imaged by CT for nonvascular reasons.

IMH is classified as type A or B using the same criteria as the Stanford dissection classification. Location is important in determining prognosis. Type A IMH has been shown more likely to progress to complications such as ulcer, dissection, and aneurysm, whereas IMH localized in the descending thoracic aorta without an associated ulcer demonstrates less frequent complications on CT follow-up. In addition, a maximum aortic diameter greater than 5 cm on the initial CT imaging has been shown to be an independent predictor of progression of type A IMH.

IMH can occur with or without any other aortic abnormality. While most intramural hematomas resolve on follow-up CT, follow-up imaging is still important in the acute stages to assess for the development of ulcers (see Fig. 8-14). Ulcers can lead to saccular aneurysms and, less frequently, to overt dissection. Expansion of the hematoma may be a sign of impending rupture and is readily identified by CT. Although many cases of IMH not associated with ulcer have a stable early radiographic and clinical course, long-term follow-up may be indicated since fusiform aneurysm formation can be a common late complication in these patients.

Ulcerlike Diseases

There is some controversy in the literature regarding nomenclature and clinical significance of aortic ulcers identified radiographically. Stanson first described a penetrating atherosclerotic ulcer (PAU) as a distinct entity from dissection and aneurysm, and defined it as an ulcerating atherosclerotic lesion that penetrates the internal elastic lamina allowing for hematoma formation within the media of the aortic wall. The principal CT characteristics (Fig. 8-15) seen in nearly all patients include advanced atherosclerotic disease of the thoracic aorta and a focal ulcer with adjacent subintimal hematoma beneath a frequently calcified and inwardly displaced intima. Clinical presentation varies from asymptomatic to acute chest pain similar to aortic dissection. Location can be helpful, since most cases of PAU are identified in the middle and distal thirds of the descending thoracic aorta, compared with type B dissection, which involves the proximal descending thoracic aorta.

Figure 8-15 Penetrating atherosclerotic ulcer with acute hematoma. A, Axial CT angiography image reveals a focal outpouching of an aortic ulcer (curved arrow) with adjacent acute hematoma (arrow). B, Sagittal reformation shows relationship of ulcer (curved arrow) with intimal calcifications (arrows) and the typical location and extent of atherosclerosis.

Ulcerlike lesions that occur in the setting of IMH (see Fig. 8-14) are likely distinct from penetrating atherosclerotic ulcers and tend to have a more aggressive course. CT identifies these ulcerlike projections (ULPs) along with or following the resolution of acute IMH in patients typically presenting with chest pain. As opposed to PAUs, ULPs are located predominantly in the ascending aorta, distal aortic arch, and proximal descending thoracic aorta.

Regardless of ulcer classification, clinical presentation is the most important factor in patient management. Incidentally identified ulcers have shown a generally benign course, while those initially diagnosed in symptomatic patients show an unpredictable course with frequent clinical and radiographic progression. Because neither the presence nor absence of subsequent clinical symptoms in an initially symptomatic patient is a reliable predictor of clinical course, serial CT is recommended to document resolution.

While multiple studies have reported similar initial CT characteristics during the acute stage of the ulcer regardless of clinical outcome, serial CT findings determine progression and help determine whether patients should be managed conservatively or surgically. For example, interval change in ulcer size and morphology can be helpful in prognosis. Ulcers that demonstrate a diameter of greater than 20 mm and depth of greater than 10 mm have a very high risk of progression. In the early follow-up period, CT identifies frequent progression of ascending aorta and arch ULPs to saccular aneurysm formation, rupture, and dissection. Aneurysms that develop from incorporation of the ulcer into the aortic lumen are further characterized by a faster than average growth rate. Overt aortic dissection generally occurs within the first month of follow-up, and a 38% rupture rate has been reported during the initial hospital admission. Finally, the presence and interval increase of pleural effusions on serial CT have been found to be an independent factor for clinical and radiographic progression.

Postoperative Thoracic Aorta

Grafts, and, more recently, stents, can be used for thoracic aortic surgical repair. Two widely performed techniques utilize prosthetic grafts or homografts for thoracic aortic repair or replacement. In the interposition graft technique, there is total excision of the diseased segment of native aorta with graft anastomoses to the proximal and distal excision sites. The inclusion graft technique implies closure of the remaining diseased aorta around the graft, thus creating a potential space between the graft and the native aortic wall that can contain thrombus, flowing blood, or both.

Normal Graft Postoperative Findings

Quint conducted a study on 235 CT studies on 114 patients to describe the normal postoperative appearance of thoracic aortic interposition grafts on serial CT studies. Low-density material surrounding or adjacent to the aortic graft (including anterior mediastinum for ascending grafts and extrapleural for descending grafts) was noted in a large number of patients for months to several years after surgery. The volume of material decreased over time in many patients, while in others it remained unchanged, possibly representing hematoma that evolved into fibrous tissue. The presence of such material on postoperative CT studies should be confirmed to be of low density and should not be mistaken for evidence of leak or infection.

In aortic graft interposition, felt strips are used to buttress the anastomosis in patients with dissection or other disease of the aortic wall. Felt pledgets are often used to repair the bypass cannulation site in the native aorta and the air evacuation needle site in the graft. Potential pitfalls can occur with felt strips and felt pledgets, as these high-density structures simulate contrast material or contained leak. Knowledge of the operative material used and routine initial noncontrast CT can assist in avoiding these potential pitfalls. Furthermore, the high density of the felt strips helps to readily detect the anastomotic site on CT.

Another potential pitfall lies in the variable appearances of coronary artery anastomoses after aortic arch reconstruction. The coronary artery grafts may be anastomosed with a button of native aortic root, which when conspicuous can simulate a pseudoaneurysm at the proximal graft anastomosis. Total arch replacement with re-implantation of the head and neck vessels into the graft can also result in complex appearances on enhanced CT. In cases of type A dissection repair, small outpouchings at the anastomotic site are identified and develop when there is a perfused false lumen distal to the anastomotic site. Another postoperative appearance that should be recognized occurs when a stage 1 “elephant trunk” technique is performed to replace the ascending aorta and arch in patients with diffuse aneurysmal processes. The presence of a free segment of graft material projecting into the proximal descending aorta identifies patients who may require subsequent repair of the descending aorta.

Rofsky described the normal postoperative CT and MRI appearance in 34 patients after undergoing continuous-suture graft-inclusion technique for repair of aortic aneurysms and dissection involving the ascending aorta. Perigraft thickening was noted in 33% of CT studies, best seen on axial images. The perigraft thickening was symmetric with a concentric distribution around the graft. Thrombi within and outside the wrap were detected in 24% of the cases with CT. Mass effect on the graft was seen in 13% of the CT studies, mostly in cases with coexistent perigraft flow. There was no evidence of clinical compromise on the basis of follow-up examination. In patients with aortic dissection repair, 40% of CT studies demonstrated an intimal flap distal to the graft, and 40% of the studies showed extension of the dissection to the great vessels.

In aneurysm repair, the length of the graft does not always match the extent of the aneurysm, with persistent dilatation of the remaining native aorta. With chronic dissection repair, a double-channel aorta distal to the site of repair is an expected finding and is deliberately preserved to ensure perfusion of those organs dependent on the false lumen for their blood supply.

Graft Complications

CT can identify important early mediastinal complications related to the grafting procedure, such as mediastinal abscess, anastomotic dehiscence, or postoperative bleeding. Delayed complications detected by CT include anastomotic dehiscence and leakage, pseudoaneurysm formation (Fig. 8-16), infection, and new or progressing disease of the native aorta. Paravalvular leaks can also be seen after aortic valve repair.

Figure 8-16 Postoperative complications from type A repair. A, ECG-gated axial CT angiography image reveals outpouching of contrast (curved arrow) consistent with a pseudoaneurysm in the ascending aorta that was not apparent on non-gated images. A dissection flap in the descending aorta is noted (arrow). B, Coronal reformatted image better displays the pseudoaneurysm (arrow).

CT is recognized as the most sensitive method of detecting leak. A leak is confidently diagnosed when contrast extravasation is identified. To avoid pitfalls, an unenhanced scan should be performed in postoperative patients to obtain an unequivocal baseline appearance of postoperative repair for comparison with the contrast-enhanced CTA images. Leak can originate at both the proximal and distal ends of the graft, referred to as perigraft leaks. Leakage at the proximal end of the graft seam predisposes to aortic rupture and underscores the importance of adequate follow-up imaging and early diagnosis.

Pseudoaneurysm formation is a major complication of the graft-interposition techniques resulting from partial dehiscence of one of the suture lines and is not surrounded by any layer of the native aorta. CT is a sensitive method for the diagnosis of anastomotic pseudoaneurysm after aortic reconstruction. A pseudoaneurysm is identified as an outpouching of the graft contour and is usually partially thrombosed. Perigraft flow seen after graft inclusion technique can be considered a “protected” pseudoaneurysm. However, the presence of perigraft flow after interposition techniques is more ominous since there is no jacket of native aorta to confine the flow of blood.

CT should be the first imaging modality performed when there is concern for aortic graft infection due to its high sensitivity and specificity. However, it can be difficult to distinguish normal postoperative findings from signs of infection in the early postoperative period. It is rare to find perigraft air beyond 1 week after surgery, but many do not consider it pathognomonic of graft infection until 4 to 7 weeks after surgery. Other authors suggest it is definitely abnormal beyond 2 weeks. After the early postoperative period, axial images show a ring of fat density around the graft. Perigraft fluid that persists beyond 6 to 12 weeks after surgery should be considered highly suspicious of infection.

Thoracic Aortic Stent-Grafts

The main goals of endovascular repair, as demonstrated by CT, include absence of flow in the aneurysm sac and cessation of aneurysmal growth. In the immediate post-treatment phase of thoracic aneurysms, slight expansion of the aneurysm diameter, presumably related to inflammatory changes, has been reported in up to 30% of patients. However, aneurysm diameter and volume decrease in 48% to 67%, remain unchanged in 22%, and mildly increase in 11% to 22% in the absence of detectable endoleak at 1-year follow-up. After the 1-year follow-up period, diameter and sac volume remain constant in this group. CT volumetric measurements have been shown to be more sensitive and less variable when compared with maximal diameter measurements. Cross-sectional diameter may not accurately reflect all the changes in the shape of thoracic aortic aneurysms, which can have irregular fusiform or bizarre saccular configurations.

In successful endovascular treatment of dissection, CTA immediately demonstrates an obliteration of the entry tear, excluding flow through the intimal tear and redirecting flow exclusively into the true lumen. Follow-up CT demonstrates an interval decrease in the mean diameter of the false lumen and increase in the caliber of the true lumen. The false lumen also demonstrates partial or complete thrombosis.

Stent-Graft Complications

CTA can accurately identify and classify endoleaks. Five types of endoleaks described include type I, flow around the graft attachment sites (proximal or distal); type II, retrograde flow from side branches into the aneurysmal sac; type III, graft disruption or graft malfunction; type IV, graft leaking or porosity; and type V, endotensive enlargement of the aneurysm sac without a visualized leak. An initial unenhanced scan should be performed to avoid mistaking high-density mineralization or postoperative materials for contrast leakage. A late acquisition is recommended, particularly for detecting type II endoleaks, since retrograde filling of the sac may be apparent only on the delayed phase. In addition, very narrow windows are recommended to detect subtle endoleaks. Interval increase in aneurysm diameter or sac volume can be an indirect sign of an endoleak. Axial CTA images can then confirm the presence and indicate the cause of the endoleak based on the configuration and localization of the leakage in relation to the stent-graft. Type I endoleaks are readily detectable (Fig. 8-17) and can cause a substantial increase in aneurysm size and lead to rupture if undetected or untreated. However, in the early postoperative period, CTA can detect very small type I endoleaks that have a tendency to seal spontaneously. Serial CTA can be used to monitor these cases for persistent leak and associated interval increases in aneurysm dimensions.

Figure 8-17 Type I endoleak after endovascular repair of an enlarging penetrating aortic ulcer. A, Axial CT angiography image shows extravasation of contrast adjacent to proximal edge of stent (arrow). B, Sagittal reformatted image shows focal area of high density within sac (arrow) not present on precontrast images, consistent with leakage.

Type I endoleaks in thoracic aneurysms have been reported to occur when the proximal neck length is less than 2 cm, emphasizing the importance of accurate initial CT dimensional criteria. Fattori reports that late endoleaks nearly always result from dilatation of the aortic neck, which should be measured on each follow-up CT scan to anticipate potential endoleak. There is usually an interval increase in the proximal and distal neck diameter in the first year of follow-up, but no further increases generally occur afterward. Stent migration is a less common cause of type I leak and is best depicted on orthogonal images.

Type II endoleaks result from an incompletely thrombosed aneurysm sac with retrograde perfusion by a feeding vessel, typically an intercostal artery. Serial CT usually shows spontaneous sealing or no increase in aneurysm dimensions. However, continued CT surveillance is generally recommended because of the potential risk of rapid expansion or rupture.

Thrombosis of the stent-graft can complicate endovascular repair of the aorta. CT identifies partial thrombi within the aortic stent-graft as nonenhancing peripheral circular or semicircular filling defects. Short-interval CT follow-up is warranted, since prognosis varies from interval resolution to complete thrombosis. Shortening in the craniocaudal axis of the treated aneurysm, and, less commonly, stent migration, can result in kinking of the unsupported portion of the graft. MIP and MPR images are superior to axial sections in demonstrating a kink and planning repair.

Additional potential fatal complications identified by CT are dissection and pseudoaneurysm formation. Introduction of the delivery system can cause iatrogenic intimal injury and a resulting false lumen. In endovascular repair of type B dissection, extension of the dissection to the ascending aorta, originating from the proximal uncovered part of the stent-graft, is readily detected by CTA. Pseudoaneurysms at the anastomotic site are rarely seen with endografts but, when present, are often detected far from the free edge of the stent-graft, possibly related to hydraulic stresses or intimal defects associated with deployment.

Assessment of Aortic Coarctation after Surgery

CTA imaging assesses all clinically relevant anatomic features after surgical treatment of aortic coarctation. When directly compared with the imaging findings of catheter angiography and correlated with pressure measurements, CTA was found to be reliable enough to replace catheter angiography in the postoperative care of patients with coarctation. CTA detects subtle findings of stenosis, aneurysms, pouches, and intimal flaps before any clinical complications are manifested (Fig. 8-18). Intramural calcification at the former coarctation site marks the site at risk for potential future dissection and atherosclerotic disease. Detection of such calcification is an advantage of CT over MRI, given the susceptibility artifact usually encountered with the latter. CTA also provides the necessary information for those patients requiring further invasive procedures. Narrowing on 3D images should be further evaluated by systolic pressure gradients directed in the region of morphologic abnormality, thereby decreasing contrast dose and number of projections needed.

Figure 8-18 Post-coarctation repair. A, Axial CT angiography image in a patient presenting with blood pressure discrepancy between the upper and lower extremities after surgical repair revealed a focal bulge in the descending thoracic aorta (arrow). B, Parasagittal reformatted image more clearly reveals a postoperative pseudoaneurysm (arrow). C, Example of 3D volume rendering after endovascular stent placement in another patient shows no complication.

IMAGING EVALUATION OF PULMONARY EMBOLISM

Clinical Considerations

Pulmonary embolism (PE) is the third most common acute cardiovascular disease after myocardial infarction and stroke, and is associated with a high mortality rate. The risk of death is particularly high during the acute phase but then decreases over time. Two-week mortality or in-hospital clinical worsening in patients with undiagnosed pulmonary embolism ranges between 25% and 38% versus less than 1% to 20% in the diagnosed and treated patients. Therefore, mortality due to PE can be reduced by prompt diagnosis and treatment in the ED as well as early prognostic stratification and identification of patients at high risk for hemodynamic deterioration.

The difficulty of diagnosing acute PE based on clinical and laboratory findings alone is why imaging studies play such a key role in the diagnostic algorithm for PE evaluation. After a brief review of the epidemiology and pathophysiology of this disease and discussion of relevant clinical findings, the different imaging modalities commonly employed for evaluation are described. Emphasis is given to CT pulmonary angiography, given its clear advantages in the ED setting. The rapid examination time afforded by CT, its widespread availability, and, perhaps most importantly, its accuracy in evaluation of PE and a host of additional thoracic abnormalities all make it the imaging exam of choice.

Epidemiology

The incidence of venous thromboembolism (both deep venous thrombosis and pulmonary embolism) increases exponentially with age for both men and women. Incidence rates are somewhat higher in women of childbearing age, whereas, after the age of 45 years, incidence rates are generally higher for men. Venous thromboembolism (VTE) recurs frequently; about 30% of patients develop recurrence within the next 10 years. The hazard of recurrence is highest within the first 6 to 12 months.

Independent risk factors include patient age, surgery, trauma, hospital or nursing home confinement, active malignancy with or without concurrent chemotherapy, central vein catheterization or transvenous pacemaker, prior superficial vein thrombosis, varicose veins, and neurological disease with extremity paresis. Among women, additional risk factors include oral contraceptive use, hormone therapy, pregnancy, and the postpartum period.

Pathophysiology

Thrombi present in the deep veins of the upper and, more commonly, lower extremities may dislodge and embolize to the pulmonary arteries. The resultant pulmonary arterial obstruction increases the pulmonary vascular resistance and induce, the release of vasoactive agents such as serotonin by platelets. The alveolar dead space is increased and impairs gas exchange. As right ventricular afterload and pressure increase, right ventricle failure may result. Paradoxical embolism can occur in the presence of a patent foramen ovale or atrial septal defect. Pulmonary emboli are often multiple and have a predilection for the lower lobes, likely related to increased pulmonary blood flow to the lung bases. Only a minority of pulmonary emboli result in lung infarction because of the bronchial arterial collateral circulation.

Clinical Findings

Objective assessment of the clinical probability of PE should be made prior to imaging. Examples of indices used include the Wells and Geneva scores. The PIOPED II investigators have published recommendations regarding the diagnostic pathways for PE. A normal D-dimer in a patient with a low or moderate probability of PE by clinical assessment requires no further testing. If there is high probability by clinical assessment, further testing is needed regardless of the D-dimer results. If the pulmonary CT angiography is negative and CT venography was not performed or was technically inadequa te, venous US or MR venography is recommended. If the pulmonary CT angiography and CT venography are negative, other options include serial venous US, pulmonary digital subtraction angiography, and pulmonary scintigraphy. If the D-dimer is abnormal, the majority of the PIOPED investigators preferred the combination of CT pulmonary angiography and CT venography. However, patient radiation dose is always a concern; thus, in younger and female patients, venous US should be considered as an alternative to CT venography. In patients with severe allergy to iodinated contrast material, pulmonary scintigraphy may be a useful alternative. Other options include serial venous US examinations, CT pulmonary angiography with gadolinium, or MR pulmonary angiography with gadolinium. In patients with impaired renal function, venous US is recommended. If negative, then pulmonary scintigraphy should be performed. In pregnant patients, although often positive, D dimer testing should still be performed. If the D dimer is positive, venous US is recommended. If the US is negative, most of the PIOPED II investigators recommend pulmonary scintigraphy over CT pulmonary angiography, although some studies indicate that the absorbed dose to the fetus is less with CT pulmonary angiography than with a perfusion scan.

Imaging Techniques

The following paragraphs discuss the imaging techniques for pulmonary embolism (Box 8-1).

Box 8-1 Pulmonary Imaging Techniques

1. Chest radiograph (CXR)

2. Ventilation-perfusion scan (VQ)

3. CT pulmonary angiography (CTPA) and CT venography (CTV)

4. MR pulmonary angiography (MRPA) and MR venography (MRV)

5. Conventional pulmonary angiography (PAgram)

Chest Radiograph

Although the chest radiograph is a poor predictor of PE, it remains one of the initial steps in the work-up of ED patients presenting with chest pain and for whom there is clinical suspicion for PE. It can be diagnostic of other causes for patients’ presenting symptoms mimicking PE such as pneumothorax, pneumonia, and pulmonary edema. Chest radiography plays an essential role if a ventilation-perfusion scan is chosen as the imaging modality to confirm or exclude PE.

In the observations from the PIOPED study, the most common radiographic findings in patients with PE were atelectasis and/or parenchymal areas of increased opacity in the lower lung zones and pleural effusions. A negative chest radiograph had a negative predictive value of only 74%. Although infrequently observed, some CXR findings have been described as more suggestive of PE. The Hampton sign refers to a well-defined pleural-based area of increased opacity with a convex medial border, and relates to pulmonary infarction. It can help to differentiate pulmonary infarct from pleural thickening or effusion (Fig. 8-19). The Westermark sign refers to an area of oligemia distal to a large vessel that is occluded by PE. In the PIOPED study, oligemia was relatively uncommon but a highly specific radiographic finding for prediction of PE at angiography. The Fleischner sign refers to a prominent central artery, caused either by pulmonary hypertension secondary to peripheral emboli or by distention of the vessel by a large clot (Fig. 8-20). However, the PIOPED study concluded that there were no differences between patients with PE and those without PE in regard to most chest radiographic findings. Thus, the chest radiograph cannot exclude or confirm the presence of PE. It is most useful for detecting alternate chest pathologies and as a correlate exam for a ventilation-perfusion exam.

Figure 8-19 Atelectasis versus infarcts (Hampton sign). A, Posteroanterior view of the chest shows right base peripheral opacities (arrow) and a small right pleural effusion. B, Contrast-enhanced CT reveals segmental and subsegmental filling defects (arrow) in the right lower lobe.

Figure 8-20 Fleischner sign. A, Portable film showing an asymmetrically enlarged right pulmonary artery (arrow). B, Contrast-enhanced CT shows bilateral pulmonary emboli with large clot in the right pulmonary artery (arrow).

Ventilation-Perfusion Scan

Ventilation-perfusion (VQ) scan is often the second modality of choice when pulmonary embolism is suspected. Data from the PIOPED study showed that although it has a very high sensitivity (98%), the specificity is very low (10%). Technetium 99m macroaggregate albumin (MAA) is the radiopharmaceutical agent used for pulmonary perfusion, and it is injected intravenously with the patient in a supine position. If the perfusion images and the chest film are normal, the study is of low probability and the exam may be concluded. Otherwise, ventilation imaging is performed with xenon 133. The scintigraphic hallmark of pulmonary embolism is a perfusion defect corresponding to a bronchopulmonary segment that displays normal ventilation and with no correlating abnormality in the chest film. The modified PIOPED criteria are used for interpretation, determining if the study is normal, and determining if it is of low, intermediate, or high probability (Box 8-2). This subject is discussed in depth in Nuclear Medicine: The Requisites and in Chapter 12.

Box 8-2 Revised PIOPED Ventilation-Perfusion Scan Criteria

HIGH PROBABILITY (≥80%)

≥2 large mismatched segmental perfusion defects or the arithmetic equivalent in moderate or large plus moderate defects ^∗

INTERMEDIATE PROBABILITY (20%–70%)

One moderate to two large mismatched segmental perfusion defects or the arithmetic equivalent in moderate or large plus moderate defects ^∗

Single matched ventilation-perfusion defect with clear chest radiograph ^†

Difficult to categorize as low or high, or not described as low or high

LOW PROBABILITY (≤19%)

Nonsegmental perfusion defects (e.g., cardiomegaly, enlarged aorta, enlarged hila, elevated diaphragm)

Any perfusion defect with a substantially larger chest radiographic abnormality

Perfusion defects matched by ventilation abnormality ^† provided that there is (1) a clear chest radiograph and (2) some areas of normal perfusion in the lungs

Any number of small perfusion defects with a normal radiograph

From Gottschalk A, Sostman HD, Coleman RE Ventilation-perfusion scintigraphy in the PIOPED study. Part II: Evaluation of the scintigraphic criteria and interpretations. J Nucl Med 34:1119–1126, 1993. Reprinted with permission.

NORMAL

No perfusion defects or perfusion outlines exactly the shape of the lungs seen on the chest radiograph (note that hilar and aortic impressions may be seen and the chest radiograph and/or ventilation study may be abnormal)

^∗ Two large mismatched perfusion defects are borderline for “high probability.” Individual readers may correctly interpret individual scans with this pattern as “high probability.” In general, it is recommended that more than this degree of mismatch be present for the “high probability” category.

^† Very extensive matched defects can be categorized as “low probability.” Single VQ matches are borderline for “low probability” and thus should be categorized as “intermediate” in most circumstances by most readers, although individual readers may correctly interpret individual scans with this pattern as “low probability.”

CT Pulmonary Angiography and CT Venography

CT Pulmonary Angiography

In routine practice, CTPA performed on MDCT scanners has become the imaging modality of choice when the diagnosis of PE is suspected. Sensitivities ranging from 53% to 100% and specificities ranging from 83% to 100% have been reported, although these wide ranges are explained mostly by technological improvements over time, most notably the advent of multidetector-row technology for image acquisition, but also the use of picture archival communication system (PACS) for image interpretation. Today, CTPA is considered the “gold standard” for evaluation of PE. The negative predictive value of a normal CTPA study is high, close to 98%, regardless of whether multidetector-row technology was used or underlying lung disease was present. It is therefore a very valuable tool for the evaluation of patients presenting to the ED in whom clinical suspicion for PE warrants imaging.

The benefits of CTPA include evaluation of the lung parenchyma, mediastinal structures, and chest wall, along with direct visualization of pulmonary emboli. Thus, in addition to high accuracy for pulmonary emboli diagnosis, CTPA can accurately detect numerous alternative or concurrent pathologies such as aortic dissection, pneumothorax, pneumonia, or malignancy. For ED patients who often present with nonspecific chest complaints, this is one of the most compelling benefits of CT, the ability to broadly assess and exclude not just one, but most diagnoses under clinical consideration.

Technical Factors in CTPA

When interpreting a CTPA examination, the quality of the study should be assessed and reported, as it reflects diagnostic confidence. Adequate contrast opacification is essential at all levels of the pulmonary arterial tree (Fig. 8-21). If the entirety of the pulmonary arterial tree is not well opacified, one cannot state to which level of the pulmonary arterial branches one can confidently evaluate for the presence of pulmonary embolism.

Figure 8-21 Poor opacification of the pulmonary arteries due to delayed imaging relative to the contrast bolus enhancement peak. A, The aorta is much more densely opacified than the pulmonary artery. Thrombus in the right main pulmonary artery (arrow) is difficult to discern. B, Using a narrower window improves visualization of the right main pulmonary thrombus (arrow at left) as well as a thrombus within the anterior left upper lobe segmental artery (arrow at right).

A careful and systematic evaluation at the main, segmental, and subsegmental pulmonary arterial branches is required. Volumetric acquisition and multidetector-row technology allow a relatively facile review of hundreds of images by scrolling or cine mode. Appropriate windowing is necessary to optimize emboli detection. With a window width too narrow, small nonocclusive emboli can be obscured. Brink and colleagues suggest a window width equal to the measured mean attenuation of the enhanced main pulmonary artery plus two standard deviations and a window level equal to one half of this value. Varying the window width and level until the pulmonary valve is visualized within the contrast-opacified pulmonary artery is another means of appropriate window/level selection. These modified window settings can increase the conspicuity of artifacts caused by image noise and flow. Evaluation of the pulmonary vessels and bronchi in lung window settings is useful for confirming motion artifact.

Diagnostic Criteria for CTPA

The following CTPA findings are used as diagnostic criteria for acute pulmonary embolism:

1. Complete arterial occlusion by thrombus with lack of enhancement of the lumen, plus or minus enlargement of the vessel (Fig. 8-22).

2. Partial filling defect of the vessel surrounded by contrast, the “polo mint” sign if the image is acquired perpendicular to the long axis of the vessel and the “railway track” sign if the image is acquired parallel to the long axis of the vessel (Fig. 8-23).

3. Peripheral intraluminal filling defect that forms acute angles with the arterial wall (Figs. 8-24 and 8-25).

Figure 8-22 Acute pulmonary embolism. A right lower lobe segmental artery reveals complete occlusion by thrombus (arrow at left). This vessel is enlarged compared with the contralateral left lower lobe segmental artery (arrow at right), which also contains a large but nonoccluding thrombus.

Figure 8-23 Acute pulmonary embolism. A, Bilateral lower lobe emboli exhibit the “polo mint” sign (curved arrows). Partial filling defects within the vessels are surrounded by contrast. More occlusive thrombi are seen in nearby arteries (arrows). B, Oblique coronal reformatted image of right lower lobe pulmonary embolism showing the “railway track” sign (curved arrows), the longitudinal axis equivalent of the “polo mint” sign. Left-sided emboli are also seen (arrows). C, Oblique axial reformatted image showing a “saddle” embolus, right and left main pulmonary emboli straddling the bifurcation (arrows).

Figure 8-24 Acute pulmonary embolism. Axial contrast-enhanced CT image showing an eccentrically positioned partial filling defect in the left interlobar artery (arrow), forming acute angles with the arterial wall.

Figure 8-25 Acute pulmonary embolism. Axial CT image shows massive clot in the left main pulmonary artery (arrow).

Ancillary findings of acute PE, although not specific, include peripheral wedge-shaped areas suggestive of lung infarct and linear bands.

The CTPA findings used as diagnostic criteria for chronic PE include:

1. Complete occlusion of the vessel (by thrombus) with vessel diameter smaller than adjacent patent vessels.

2. Peripheral, crescent-shaped intralumenal defect that forms obtuse angles with the vessel wall.

3. Contrast-opacified thick-walled vessels (due to recanalization), with affected vessels that may be smaller than adjacent nondiseased vessels.

4. Web or flap within a contrast-opacified vessel (Fig. 8-26).

5. Extensive bronchial or other systemic collateral vessels.

6. Accompanying mosaic pattern of perfusion of the lung.

7. Calcification within eccentric vessel thickening.

Figure 8-26 Chronic puolmonary embolism. A, Axial contrast-enhanced CT image shows a linear filling defect in the right interlobar artery (arrows) consistent with a web in a patient with prior pulmonary embolism. B, Coronal reformation displays the extension of the web (arrow).

CTPA Interpretation Errors

The interpreting radiologist has to be aware of several artifacts that can occur with CTPA. These potential pitfalls have been divided into three categories: technical, anatomic, and physiological. Knowledge of these pitfalls will improve the radiologist’s performance and confidence in the interpretation of CTPA.

The quality of arterial enhancement in CTPA depends highly on the physiologic conditions under which the study is performed. Abnormalities of venous inflow, extrapulmonary shunts, and intrapulmonary shunts can adversely influence pulmonary arterial enhancement. Another physiological artifact that may occur is the transient interruption of the contrast column in the pulmonary arteries, a result of increased venous return of unopacified blood from the inferior vena cava. This occurs when negative intrathoracic pressure is generated by inspiratory effort immediately prior to imaging. A variable amount of unopacified blood enters the right atrium, diluting and potentially even disrupting the contrast column. The column of unopacified blood that flows through the pulmonary arteries can simulate pulmonary emboli. Clues to help identify this artifact include lack of opacification in multiple vessels at the same level bilaterally without vascular expansion, and the presence of unopacified blood in the right heart chambers on the preceding images followed by the presence of unopacified blood in the left atrium, ventricle, and aorta on later images. The most common causes for a nondiagnostic CTPA are poor contrast bolus, respiratory motion, and graininess of the images owing to patients’ body habitus.

Prognostic Factors for CTPA

Most late deaths in patients diagnosed with PE are due to underlying diseases. However, the main cause of death within 30 days is right ventricular failure. It is therefore important to identify patients at high risk of right ventricular failure in order to establish the appropriate treatment. Thrombolysis, catheter intervention, and surgical embolectomy may be performed as adjuncts to anticoagulation and may reverse right ventricular failure and reduce the risk of recurrence and death. With newer-generation scanners, standardized cardiac views can be obtained when performing CTPA. Right ventricle enlargement on a four-chamber view in the setting of acute PE correlates with right ventricle dysfunction by echocardiography (Fig. 8-27). Right ventricle enlargement is present when the ratio of the diameter of the right ventricle to the diameter of the left ventricle is greater than 0.9. Other cardiovascular parameters seen on CTPA have been evaluated as predictors of mortality in patients with severe PE. These include findings such as reflux of contrast into the inferior vena cava, convex leftward bowing of the interventricular septum, and increased diameter of the azygos vein, superior vena cava, and aorta.

Figure 8-27 Right heart strain in patient with bilateral pulmonary emboli. A, Axial CT image showing right ventricle enlargement (RV). Note also straightening or leftward bowing of the interventricular septum (S), with resultant diminutive-appearing left ventricular cavity (LV). B, Axial image reveals reflux of contrast into the inferior vena cava (IVC) and hepatic veins (arrows).

CT Venography

Limited lower extremity CT venography (CTV) provides direct imaging of the inferior vena cava, pelvic veins, and lower extremity veins immediately after CTPA with the same contrast bolus. Technical limitations are not uncommon and include poor venous enhancement and streak artifacts from orthopedic hardware. The reported sensitivity and specificity using MDCT are 100% and 97%, respectively.

The drawbacks of combining CTV with CTPA include the higher volume of contrast required to produce adequate vein opacification and the additional radiation exposure to ovaries and testes. Radiation exposure to the gonads can be reduced by starting to scan at the acetabulum. Because of these issues, the value of combining CTV with CTPA is still debated.

MR Pulmonary Angiography and MR Venography

MR pulmonary angiography (MRPA) and MR venography (MRV) are considered second-line diagnostic tools (particularly in the ED) in the evaluation of PE. Practical considerations include their higher cost, limited availability, longer examination times, and reduced performance as compared with MDCT. As technology improves, MR imaging may play a greater role in the evaluation of patients with venous thromboembolic disease, particularly those patients with allergies to iodinated contrast material.

MR Pulmonary Angiography

The routine use of MRPA in the evaluation of pulmonary embolism has been limited by technical and practical factors. Image degradation from respiratory and cardiac motion is common. High spatial resolution is necessary because of the small diameter of the branch vessels of the pulmonary arterial tree. High temporal resolution is required to produce arterial-phase-only images, thus avoiding the pulmonary venous enhancement that can obscure the evaluation of the arteries. MRPA studies are performed during suspended respiration. The timing of the acquisition with respect to the gadolinium-based contrast injection can be crucial if one is to capture the pulmonary arterial phase.

Conventional 3D-gadolinium-enhanced MR angiography (MRA) can be performed by acquiring a single coronal 3D MRA with a large field of view (FOV), encompassing both lungs, using a single dose of gadolinium chelate. To avoid wraparound artifact, the FOV must be large enough to cover the two lungs, along with the arms and shoulders. Another approach is to perform two separate sagittal acquisitions, one for each lung, with smaller FOV and higher spatial resolution during two separate breath-holds and contrast injections. This method has the disadvantage of requiring two separate acquisitions rather than one, and, furthermore, sagittal volumes may not optimally cover the central pulmonary circulation.

Another option for data acquisition is time-resolved MR angiography, an approach that eliminates the need for bolus timing. Using very fast acquisition methods, a time series of 3D images is acquired. One of these 3D datasets is expected to capture the arterial enhancement, while a later dataset captures venous enhancement. This method has higher temporal resolution and lower spatial resolution as compared with conventional 3D-gadolinium-enhanced MRA. Although the spatial resolution of time-resolved 3D MR angiography is less than that of CT pulmonary angiography, confident diagnoses can be made at the main, lobar, and segmental levels.

Emboli are detected as intraluminal filling defects or vascular cutoffs, just as in CTPA (Fig. 8-28). Gupta and colleagues emphasized the need to perform and review overlapping, small, subvolume targeted MIP images, as pulmonary emboli often lodge where abrupt changes in vessel diameter are found; this typically occurs where vessels branch. It may be difficult on single thin coronal sections to tell the difference between a normal bifurcation and a small embolus, and the review of targeted MIP images may help with this distinction. Because there is often some degree of unwanted pulmonary venous opacification, it is useful to obtain larger subvolume MIP images so that vessels can be followed to the hilum, thus allowing the differentiation of arteries and veins. Major drawbacks for pulmonary MRA include cardiac and respiratory motion artifacts, losses of signal in the presence of complex blood flow patterns, and magnetic susceptibility effects from the air-containing lungs.

Figure 8-28 Acute pulmonary embolism (arrow). A 56-year-old man presented with chest pain and shortness of breath. Axial oblique maximum intensity projection image of 3D dynamic contrast-enhanced MR acquisition.

MR Venography

MRV, like CTV, is clearly superior to venous US in the task of delineating the inferior vena cava and pelvic veins. MRV is significantly more sensitive and accurate than US in the detection of lower extremity thrombosis.

MRV can be performed without intravenous contrast agents, using time-of-flight (TOF) MRA techniques, although gadolinium-enhanced techniques tend to provide faster acquisitions, better signal-to-noise ratio, and greater accuracy in states of slow flow or tortuous venous anatomy. Gadolinium-enhanced MRV is performed as two large FOV coronal two-dimensional (2D) or 3D gradient-echo volumes encompassing deep veins of the abdomen and pelvis, and then the thighs. The reported sensitivity and specificity of the MR venogram in the pelvis and thighs are 100% and 95%, respectively. In the calf, the sensitivity and specificity reported are lower, noted to be 87% and 97%, respectively. The combination of MRA and MRV in a single exam stands up as a strong diagnostic modality in the evaluation of thromboembolic disease.

Conventional Pulmonary Angiography (PAgram)

At most institutions, percutaneous, catheter-based pulmonary angiography has been replaced by CTPA as the standard method of reference for the evaluation of PE. A PAgram is associated with minimal but definite morbidity and mortality, and, thus, many clinicians are reluctant to refer patients for this procedure even when there is diagnostic uncertainty. In the emergency setting, the greater accessibility of MDCT at all times of day, coupled with the added value of a complete assessment of the entire thorax, which includes most relevant differential considerations other than PE, has firmly positioned it as the modality of choice. The criteria for the diagnosis of PE at PAgram are similar to those at CTPA and include the findings of an intravascular filling defect or an abrupt vessel cutoff.

CORONARY CT ANGIOGRAPHY IN THE ED

Evaluation of Chest Pain for Exclusion of Acute Coronary Syndrome

For patients presenting to the ED with chest pain, medical imaging has been established as a primary means of delineating life-threatening entities such as acute aortic syndromes, pulmonary embolism, and pneumothorax. This section defines the emerging role of coronary CT angiography (CCTA) in evaluating ED patients suffering with chest pain potentially due to acute coronary syndrome (ACS), and provides a review of CT-based coronary artery anatomy, anomalies, and findings in coronary disease.

Acute Coronary Syndrome

The American Heart Association defines coronary heart disease (CHD) as a category that includes acute myocardial infarction, other acute ischemic (coronary) heart disease, angina pectoris, atherosclerotic cardiovascular disease, and all other forms of chronic ischemic heart disease. The term acute coronary syndrome describes patients who present with either acute myocardial infarction (MI) or unstable angina (UA). Our understanding of coronary artery disease (CAD) continues to evolve.

Conventional, catheter-based coronary angiography allows identification of significantly narrowed coronary arteries, or stenotic “culprit lesions,” to which targeted revascularization therapy can be directed. The need for revascularization therapy of a particular culprit lesion should be supported by evidence of myocardial ischemia (using ECG, cardiac biomarkers, perfusion imaging, stress testing, etc.). Treating culprit lesions without supporting evidence of ischemia or symptoms has been coined the “occulo-stenotic reflex.”

More recently, the concept of the “vulnerable plaque” has gained favor; this refers to a soft, inflammatory atherosclerotic plaque with a thin or discontinuous fibrous cap. These are at risk for rupture or thrombosis, which often leads to coronary luminal compromise or occlusion. Due to positive (outward) remodeling, vulnerable plaques may not be stenotic and may therefore be occult when evaluated by conventional angiography since this visualizes only the patent vessel lumen. This concept, too, is in evolution in that vulnerable plaques are likely numerous rather than solitary, and in that disrupted, unstable, and/or ulcerated plaques extend beyond the culprit lesion. Thus, the concept of a vulnerable plaque may in fact extend further to the concept of a “vulnerable artery” or even a “vulnerable patient.”

With the advent of ECG-gated multidetector (and now multisource) CCTA, noninvasive visualization and analysis of the coronary arterial tree/system have become feasible. CCTA has already shown high accuracy in defining significant (greater than 50%) coronary stenoses, as well as the ability to detect soft and calcific plaques. It shows great promise as an ED triage tool that can help identify and discriminate patients with ACS from those with chest pain of other etiologies.

Epidemiology

It is estimated that 15.8 million Americans aged 20 or older suffer from CHD. In 2007 an estimated 700,000 Americans had new coronary attack and about 500,000 will have a recurrent attack. It is estimated that an additional 175,000 silent first heart attacks occur each year, and the subsequent lifetime risk of developing CHD after age 40 is 49% for men and 32% for women.

From 2002 to 2004, the Centers for Disease Control recorded more than 110 million ED visits in the United States, of which approximately 30 million had a primary diagnosis of “injury.” Of the remaining 80 million, roughly 7.3 million (approximately 10%) had a primary diagnosis of chest pain (nonspecific), heart disease, or heart attack or other ischemic heart disease. Acute chest pain prompts 6 million patients annually to undergo an ED evaluation in order to exclude ACS.

Clinical Considerations

Triage of patients presenting with acute chest pain is a common task performed in the ED. Clinical findings, EKG changes, and cardiac biomarkers (troponin or CK-MB) are the current standard of care tools used to exclude ACS in these patients. Unfortunately, triage decisions based on these indices are often ineffective. Normal cardiac biomarkers on ED presentation do not exclude ACS, and serial testing at intervals requiring up to 6 to 10 hours of hospital time is necessary to exclude MI. Furthermore, reported rates of missed ACS have ranged from 2% to 8%. Missed ACS carries a twofold increase in mortality and accounts for 20% of ED malpractice dollar losses.

Concern for missing ACS in ED patients with chest pain often prompts admission, or at least prolonged observation and testing (delayed cardiac biomarkers, stress testing, or perfusion imaging) in the ED or an observation unit, of approximately 2.8 million such patients per year in the United States, at a cost of $6 to $12 billion annually. Up to 60% of these patients do not have an ACS and would have been eligible for early discharge, thus highlighting the need for a rapid and more accurate triage tool for this patient population.

A recent report of ED trends and projections shows the importance of achieving this goal. The decade between 1994 and 2004 saw an increase of 18% in ED visits and a 12% decline in hospitals offering ED services. EDs in the United States are overcrowded, with 40% to 50% reportedly pushed beyond capacity between 2003 and 2004, and nearly 60% of patients waiting 30 minutes or more to be seen by an ED physician. Changes in process, technology, and staffing are being considered to improve ED throughput, but diagnostic imaging will surely continue to play a key role toward achieving this goal.

CCTA is showing promise as a tool that could rapidly exclude ACS in low- or intermediate-risk ED patients presenting with chest pain. CCTA has proven effective in visualizing calcific and soft atherosclerotic plaques as well as in identifying significant (greater than 50%) coronary stenoses. Moreover, CCTA has shown a very high negative predictive value in excluding significant CAD and stenoses in normal or insignificantly diseased coronary arteries.

Because of its high negative predictive value, CCTA is well suited to rapidly and safely exclude CAD in low or intermediate/indeterminate-risk patients presenting to the ED with nonspecific chest pain. Several pilot studies on relatively small numbers of patients have shown good performance of CCTA for this application, with a negative CCTA accurately indicating the absence of ACS as well as an extremely low risk of near-term cardiac events following ED discharge. While CCTA is not currently in widespread clinical practice in the management of ED patients with chest pain, innovator sites have begun to implement this technique, and the impending significance of CCTA in this emerging application warrants review in this chapter.

Coronary Artery Anatomy

The coronary arteries are the first branches off the ascending aorta, located immediately above the aortic valve. Between the aortic valve and the sinotubular ridge are three focal outpouchings called the sinuses of Valsalva. Normally, the right coronary artery (RCA) ostium arises from the right sinus of Valsalva, and the left main coronary artery (LMA) ostium arises from the left sinus of Valsalva. The posterior or non-coronary sinus normally has no coronary ostium.

The word coronary is derived from the Latin coronalis, which means “of, relating to, or resembling a crown.” The coronary arteries can be envisioned as a set of two crossing rings which encircle (or “crown”) the heart, with the rings defined by the intersecting atrioventricular sulcus and interventricular septum (Fig. 8-29). Respectively, the RCA and left circumflex (LCX) arteries mirror one another and course within the right and left hemi-rings of the atrioventricular sulcus, while the left anterior descending (LAD) and posterior descending (PDA) arteries course within the anterior and posterior hemi-rings formed over the interventricular septum (Figs. 8-30 and 8-31).

Figure 8-29 Crossing ring diagrams of the coronary arteries. A, The architectural skeleton of the heart and coronary arteries can be envisioned as a set of “crossing rings,” with the atrioventricular sulci (AVS) and interventricular septum (IVS) defining the planes within which the crossing rings (and coronary arteries) lie. B, As viewed from the front, the heart is “tilted” within the thorax, or oriented such that the IVS points inferolaterally toward the left. C, When viewed from the front, the heart is rotated about the IVS long axis such that the right atrium and right ventricle (RA and RV) are positioned more anteriorly relative to the left atrium and ventricle (LA and LV). D, Finally, viewed inferiorly, as in a standard axial CT image, the RV is positioned relatively anterior, the LA relatively posterior, and the LV and RA toward the patient’s left and right, respectively. The right coronary artery (RCA) and left circumflex arteries (LCX) course within the right and left atrioventricular sulci, respectively. The left anterior descending (LAD) and the posterior descending arteries (PDA) course over the anterior and posterior aspects of the interventricular septum, respectively. The posterior junction of the AVS and IVS is termed the “crux,” and is the point where the RCA (in a right dominant circulation) takes a downward turn to supply the PDA.

Figure 8-30 Coronary arteries applied to the crossing ring diagram. A, Diagram of the coronary arterial tree (right dominant circulation) superimposed on a frontal view of the “crossing ring” AVS and IVS architectural skeleton. The RCA arises from the aortic right sinus of Valsalva and courses within the right AVS to the crux, where it gives off the PDA branch (which travels under the inferior margin of the IVS) as well as posterolateral branches (PLA, which supply the inferior and inferolateral left ventricle walls). The acute marginal (AM1, AM2, etc.) branches off the RCA supply the right ventricle walls. The LCX arises from the left main coronary artery (LMA) and mirrors the RCA as it courses within the left AVS. It supplies obtuse marginal (OM1, OM2, etc.) branches, which perfuse the lateral wall of the left ventricle. The LMA arises from the left sinus of Valsalva. The LAD (essentially a continuation of the LMA) turns to course over the anterior aspect of the IVS and usually continues over the cardiac apex. The LAD supplies diagonal branches (D1, D2, D3, etc.), which perfuse the anterior, and anterolateral, walls of the left ventricle. B, 3D reformatted image of CCTA superimposed on a frontal view of the “crossing ring” architectural skeleton. Vessels are labeled as in Figure 30A. Diagonal 1 is very small and not labeled, seen just above D2. Vessel labeled “S” is a large septal perforator arising from the LAD. The three OM branches are prominent, and the LCX is not seen beyond the OM3 origin. In a left dominant circulation, the LCX would continue to the crux to supply the PLA and PDA branches.

Figure 8-31 Axial view of the coronary arteries. 3D reformatted image of CCTA as viewed from the inferior aspect of a patient, as in a standard axial CT, superimposed on a diagram of the architectural skeleton of the AVS and IVS. The RCA arises from the right sinus of Valsalva at approximately the 10–12 o’clock position, and the left main coronary artery (LMA) arises from the left sinus of Valsalva generally at the 2–5 o’clock position. The RCA and LCX course within the right and left AVS, respectively, and the LAD and PDA course over the anterior and inferior borders of the IVS, respectively. On an axial image, the diagonal branches (D1, D2, etc.) extend toward the patient’s left from the LAD, and the septal perforators (S) extend toward the patient’s right from the LAD.

“Coronary dominance” describes which artery supplies the PDA as well as the posterolateral left ventricular branch (PLA) and the artery to the atrioventricular (AV) node. In right coronary dominance (approximately 85% of the population), the RCA supplies the PDA, along with the left posterolateral branches and AV nodal artery. In left coronary dominance (approximately 8%), these vessels are supplied by the LCX. In co-dominant coronary circulation (approximately 7%), both the LCX and RCA contribute to these vessels.

RCA

After arising from the right (or anterior) sinus of Valsalva, the RCA courses deep to the right atrial appendage within the right AV groove, following the AV groove until it intersects with the posterior interventricular septum, at the “crux” of the heart (see Fig. 8-29). For the purposes of clinical description of lesion location, the RCA is often divided into proximal (ostium to first main right ventricular branch), mid (first main branch to the acute margin), and distal (acute margin to the crux). Usually, the first branch off the RCA is the conus branch; however, the conus branch often (23% to 51%) has a separate ostium off the right sinus of Valsalva. The conus branch courses anteriorly in an arc over the conus or infundibulum and may form a collateral bridge to LAD branches (the “arc of Vieussens”), thus offering protection in cases of proximal RCA or LMA/LAD disease. The sinoatrial (SA) nodal branch/artery most commonly arises from the proximal RCA (60%) but often originates from the LCX (40%). This branch extends posteriorly along the interatrial groove and supplies branches to both atria as well as the SA node. The right ventricular, or acute marginal, branches emerge at right or acute angles from the RCA and vary in size and number as they supply the right ventricular myocardium. In right dominant circulations, the RCA continues in the atrioventricular sulcus until it intersects with the posterior interventricular septum at the crux, where the RCA gives off the PDA branch that courses along the posterior interventricular septum, supplying the lower third of the interventricular septum, as well as PLA branches (which supply the inferior wall of the left ventricle and the interventricular septum) and the AV nodal branch (which extends cephalad from the crux to the AV node).

Figure 8-32 Normal coronary artery anatomy, as visualized by CCTA. A, Axial maximum intensity projection (MIP) slab image showing the courses of the proximal coronary arteries. B, Oblique coronal MIP reformatted image of the RCA. Focal hyperdensities within the RCA (arrows) are foci of calcific atherosclerosis. The proximal LMA is also visualized. C, Oblique axial MIP reformatted image of the LAD. A small septal perforator (S) and diagonal branches (D) arise from the LAD. Nearby coronary veins (V) are also visualized. D and E, Volume-rendered reformatted images of the coronary arteries.

LMA

The LMA arises from the left sinus of Valsalva and follows a short (usually 1 to 2 cm) course between the pulmonary trunk anteriorly and left atrial appendage posteriorly (Fig. 8-32). It normally bifurcates into the LAD and LCX arteries but may trifurcate with a third intermediate artery arising between the LAD and LCX. This intermediate artery has various names, including “ramus intermediate” or “ramus medians,” “median artery,” “left diagonal artery,” and “straight left ventricular artery,” and supplies different portions of the anterolateral left ventricular wall.

LAD

The LAD arises from the LMA and courses around the left side of the pulmonary trunk toward the proximal interventricular septum where it takes a roughly 90-degree downward turn and courses over the interventricular septum toward the apex (see Fig. 8-32). This downward turn over the proximal interventricular septum has clinical significance, as this is the point where the LAD becomes amenable to surgical bypass. The second diagonal branch often originates near this downward turn. By convention, the LAD is divided into three segments: proximal (LAD origin to first septal perforator), middle (first septal perforator to downward turn along interventricular sulcus/origin of second diagonal), and distal (second diagonal origin to vessel end). The LAD provides two major types of branches: the septal perforators, which generally arise at 90-degree angles from the LAD and penetrate into the interventricular septum to supply the upper two thirds of the septum, and the diagonal (D) branches, which arise at acute angles from the LAD to course over and supply the anterior/anterolateral wall of the left ventricle. The number of diagonal arteries varies, and they are named in order of their origin from the LAD (D1, D2, etc.). The LAD usually extends over the apex to supply the distal inferior wall and interventricular septum, but the PDA may supply these regions as well, and this distal LAD-PDA confluence is a source of potential collateralization in cases of RCA or LAD disease.

LCX

The LCX arises from the LMA, and its course mirrors that of the RCA as it courses under the left atrial appendage to follow the left atrioventricular sulcus toward the crux. The major branches off the LCX are the marginal (or obtuse marginal) branches which vary in size and number. The obtuse marginal (OM) branches overlie and supply the lateral wall of the left ventricle and are labeled according to the order of their origin off the LCX (OM1, OM2, etc.). For clinical description, the LCX is divided into three segments: proximal (from LCX ostium to first major obtuse marginal branch [OM1]), mid (between OM1 and OM2), and distal (vessel distal to OM2). In a right dominant circulation the LCX will terminate short of the crux, whereas in co-dominant and left dominant circulations the LCX will make respectively greater contributions to the PDA, PLA, and AV nodal circulations.

Coronary Artery Anomalies

Coronary artery anomalies are a heterogeneous constellation of anatomic variants (Fig. 8-33). While relatively uncommon (0.46% to 5.6% of normal population), certain forms carry elevated risk, making it important for the radiologist to detect and accurately define these variants. In particular, coronary anomalies cause up to 19% to 30% of sudden deaths in athletes. CCTA is superior to conventional angiography in delineating the ostial origin and proximal path of anomalous coronary arteries. Anomalies may be classified by the presence or absence of a shunt, or they may be classified as “malignant” or “nonmalignant” depending on their course and propensity to cause clinical symptoms or sudden death.

Figure 8-33 Diagrams of coronary artery variants, as they would be visualized on axial CT images. A, Diagram of normal proximal coronary artery anatomy. Normally, the RCA arises from the right coronary sinus (Rt), and the LMA (which supplies the LCX and LAD) arises from the left coronary sinus (Lt) of the aorta (Ao). B–F, Examples of malignant “interarterial” anatomic variants. These variants portend increased risk for clinical symptoms or sudden cardiac death. The RCA may arise from the left sinus of Valsalva and course between the aorta and pulmonary artery (PA) (B) to achieve its normal right AVS position. C is an axial source image from a patient with this variant. Note the acute angle take-off and narrow ostial area of the anomalous RCA (curved arrow) as compared with the normal LMA. Position of the anomalous interarterial RCA between the aorta and pulmonary artery also results in increased risk of RCA compression, especially during exercise. Conversely, either the LMA (D) or the LAD (E) may arise from the right sinus of Valsalva and follow an interarterial course. Coronary anomalies resulting in a shunt (F) include coronary atrioventricular (AV) fistulas and coronary artery origins from the PA. These variants predispose to ischemia or symptoms/sudden death and are thus considered malignant. F is a diagram of an AV fistula, with dilated RCA draining to the right atrium (RA), rather than draining normally into the coronary sinus and left atrium.

Anomalies not resulting in a shunt include anomalous coronary artery origins (origin from another coronary artery, another sinus of Valsalva or above the sinotubular ridge, and independent branch origins from the sinus of Valsalva), myocardial bridges, congenital aneurysms, and hypoplasia/atresia. Anomalies resulting in a shunt include artery origin from the pulmonary artery (called Bland-White-Garland syndrome for the LMA) and fistulas.

Malignant variants (associated with an increased risk of MI and sudden death) primarily consist of anomalous artery origins that result in an interarterial course, with the anomalous artery coursing between the aorta and pulmonary artery (see Fig. 8-33). Several factors contribute to jeopardizing the anomalous vessel in this position, including compression of the vessel between the aorta and pulmonary artery, especially during exercise, and acute-angle take-off and narrow ostial area of the proximal vessel resulting in diminished flow or occlusion. The most common interarterial variant is the RCA arising from the left sinus of Valsalva and coursing between the aorta and pulmonary artery to reach the right AV sulcus, although an LMA or LAD arising from the right sinus of Valsalva with an interarterial course may pose a higher risk to the patient. Anomalous coronary artery origin from the pulmonary artery is also considered malignant and is associated with myocardial ischemia and sudden death in early childhood.

Myocardial bridging occurs when a coronary artery (usually the middle third of the LAD) deviates from a normal epicardial course to reside under superficial myocardial fibers for a short distance before reemerging to a normal epicardial position (Fig. 8-34). The prevalence and clinical significance of myocardial bridging are nebulous. While angiographic series report prevalence between 0.5% and 2.5%, pathologic series have reported prevalence between 15% and 85%. Generally thought to be benign, any specific case of myocardial bridging may have clinical significance, and cases have been associated with angina, myocardial ischemia and infarction, and sudden death. CCTA is effective at diagnosing myocardial bridging, due to its ability to visualize the myocardium as well as coronary arteries. The ability to image in diastolic and systolic phases may help gauge the clinical significance by defining the severity of lumenal compromise during systole. Interestingly, the bridged segment is often free of atherosclerotic disease. Because of this potential for clinical significance, myocardial bridging warrants reporting when encountered.

Figure 8-34 Example of myocardial bridging of the LAD. Long axis maximum intensity projection (top left), long axis curved multiplanar reformatted images (two top right images), and cross-section reformatted images (bottom row of three images) of the LAD. The proximal LAD is in its normal position within the epicardial fat (e) overlying the interventricular septum. However, the LAD deviates from its normal epicardial course to course under superficial myocardial fibers (M) for a short distance before reemerging to assume its normal epicardial course. Bottom row of cross-section reformatted images shows a “bridge” of myocardial fibers (M) overlying the LAD at the level of the myocardial bridging. The myocardial fibers separate the LAD from the epicardial fat (e).

Coronary artery fistulas are abnormal communications between a coronary artery and another vascular structure such as the right ventricle (45%), the right atrium (25%), a pulmonary artery (15%), or the superior vena cava, left atrium, or left ventricle (less than 10%) (see Fig. 8-33). The involved coronary is often dilated and tortuous, and the anomalous connection may result in a left-to-right shunt, a hemodynamic steal phenomenon, and hypoperfusion or ischemia of the myocardium normally perfused by the involved vessel.

Image Evaluation

A potentially very important triage tool in the ED, the function of CCTA is to detect or exclude the presence of CAD in a patient with chest pain. Therefore, familiarity with the CCTA findings in CAD is required. CCTA detects CAD primarily by defining two findings: coronary artery stenoses and mural atherosclerotic plaque. These two findings permit the radiologist to exclude or confirm CAD in ED patients. Ancillary evaluation depends on scanning technique and software availability, and includes evaluation of left ventricular wall motion and function (which bestows important physiologic/functional correlation with coronary artery findings) and possibly evaluation of myocardial perfusion defects (scar/ischemia imaging) or delayed enhancement (scar/viability imaging). This section focuses on coronary stenosis and mural atherosclerotic plaque.

Coronary Artery Stenosis

As neither CCTA nor conventional angiography can measure the hemodynamic effect of a particular lesion, the degree of stenosis is used as a surrogate to determine lesion significance. The CCTA threshold most often used to define a significant stenosis is 50% or greater luminal narrowing, using a nearby “normal” segment of the vessel as a reference. A 50% stenosis by CCTA corresponds to approximately 70% stenosis on conventional planimetric angiography.

While acquiring an ECG-gated CCTA exam, images are obtained in each of (generally) 10 to 20 phases throughout the cardiac cycle. Images from the (usually diastolic) phase in which the least coronary motion artifact is present are then utilized for image analysis. To evaluate CCTA images, review of the axial images as well as MPR and MIP images is performed (Figs. 8-35 and 8-36). Although reformatted images are valuable in defining and delineating stenoses, the axial images remain the cornerstone of the evaluation, as virtually all pathologies can be identified on the axial images. Once a stenosis is detected on the axial images, further evaluation of the lesion is best performed with two long-axis reformats through the vessel. Generally, 3- to 5-mm thin-slab MIP images are also useful for evaluating lesions in the long axis, and interactive evaluation is facilitated with sliding or rotating thin-slab MIP images. The interpreter should be aware that when using MIP images, stenoses may be masked or obscured by volume averaging.

Figure 8-35 Examples of coronary artery stenosis. A, An oblique axial maximum intensity projection reformatted image of the LAD reveals disease beyond the origins of septal (S) and diagonal (D) branches. A significant appearing stenosis is noted (curved arrow). B, An oblique long axis reformatted image through a diagonal artery in another patient shows a severe stenosis (between the two long arrows). Areas of apparent stenosis or occlusion are artifactual, related to volume averaging with adjacent epicardial fat (small arrows). C, Volume-rendered reconstruction of the diagonal artery stenosis (between long arrows) seen in B. D, An example of software-assisted quantitative assessment of coronary stenosis. Middle and right images are orthogonal long-axis curved multiplanar reformatted images through the vessel (LAD in this case), and the two left images are short-axis cross-sectional reformatted images through the vessel. Double-headed arrows indicate the levels of the cross-section reformatted images relative to the long-axis images. The software can automatically detect the “edge” of the coronary artery (based on a threshold of attenuation or attenuation difference) and draw a line at the detected borders. With the vessel borders defined, the diameters and cross-sectional areas at any particular point can be calculated, and with these values the severity of stenosis can be quantified.

Figure 8-36 Coronary artery atherosclerotic plaque, as visualized by CCTA. A, A diffusely diseased LAD reveals hyperdense calcific (C) atherosclerotic plaque as well as mixed (M) “soft” and calcific atherosclerotic plaque. The mixed plaque appears to be causing stenosis. B, Magnified image of A, showing components of the mixed plaque. Calcific components (C) appear as hyperdense relative to the opacified vessel, while the “soft” components (Soft) are hypodense. The mixed plaque is causing stenosis of the vessel.

Once detected, a stenosis can be visually measured, or estimated, using a nearby “normal” segment of the vessel as a reference, or software-assisted quantitative assessment of the stenosis may be employed (see Fig. 8-35). Comparison of the visual estimate/measurement method versus quantitative software-assisted assessment of stenoses has revealed that software-assisted assessment provides higher accuracy for about the same user time required for visual estimation. However, while 64-slice CCTA has shown high accuracy in detecting 50% or greater stenoses, the ability to provide exact, quantitative measures of stenosis severity is hampered by current limits in spatial resolution. Because of this, the severity of stenoses is often classified in quartiles (0% to 25%, 26% to 50%, 51% to 75%, 76% or greater) or tertiles (0% to 30%, 31% to 70%, 71% to 99%), or some other variant, depending on local radiologist/cardiologist preferences.

Despite this relative lack of quantitative accuracy, CCTA has proven highly accurate in detecting 50% or greater stenoses, and has shown high negative predictive values (generally 95% to 100%) in excluding significant coronary disease. Since the ED chest pain patient population evaluated by CCTA should have a low prevalence of disease (low or intermediate ACS risk patient with normal cardiac biomarkers and EKGs), the majority of CCTA scans done for ED patients should be essentially normal, which should facilitate more rapid image evaluation/interpretation and thus more rapid discharge or appropriate treatment for non-ACS ED patients with chest pain.

Atherosclerotic Plaque Imaging

The presence of coronary artery plaque increases ACS risk even in the absence of coronary artery stenoses. While higher total volume or burden of atherosclerotic plaque carries higher ACS risk, different types of plaque also impart different levels of ACS risk, as plaque vulnerability varies depending on plaque composition. Histologically, atherosclerotic plaque consists of numerous components including calcification (generally calcium hydroxyapatite), fibrous tissue, inflammatory cells (predominantly macrophages and monocytes), smooth muscle, lipid components (lipid-laden macrophages, necrotic lipid cores, etc.), and hemorrhagic foci.

Both major and minor histologic criteria for defining plaque vulnerability have been described. Major histologic criteria include active inflammation, a thin cap with a large lipid-necrotic core, endothelial denudation with platelet aggregation, fissured plaques, and stenosis larger than 90%. Minor histologic criteria include superficial calcified nodules, glistening yellow plaque (seen at angioscopy), intraplaque hemorrhage, endothelial dysfunction, and positive (outward) remodeling.

For coronary plaque analysis, CCTA has inherent limits in spatial and contrast resolution that prevent discernment of many of the components of plaque. However, calcium is well defined by CCTA and can be distinguished from “soft” plaque components (see Fig. 8-36). Coronary calcium scoring, or quantification of calcific coronary atherosclerotic plaque utilizing electron beam and MDCT, has been in use for over a decade. Generally, a threshold of 130 Hounsfield units is used to define calcified atherosclerotic plaque. “Soft,” or noncalcified, plaque can be discerned from calcified plaque on CCTA in this manner. Noncalcified soft plaque is heterogeneous in composition, consisting of various quantities of components as described above. On CCTA, soft coronary artery plaque is visualized as a structure that can be clearly assignable to the vessel wall (in at least two views) with densities less than the intraluminal contrast (see Fig. 8-36). On CCTA, plaque is generally classified as calcified, noncalcified, or mixed.

Whereas CCTA has only limited ability to define the specific composition of noncalcified plaque, lower CT density generally indicates a greater lipid concentration, and, in fact, larger lipid pools or lipid cores (larger than 2 mm) may be visualized within the plaque of larger coronary arteries as hypodense spots demonstrating a density of at least 20 Hounsfield units less than the average surrounding noncalcified plaque tissue. Thus, despite inherent limitations, CCTA may be able to identify some of the major and minor plaque vulnerability criteria, including larger lipid-necrotic cores, stenoses greater than 90%, superficial calcified nodules, and positive (outward) remodeling.

As regards calcium scoring, both electronic beam CT and gated MDCT have been shown to be reliable methods of quantifying coronary calcified plaque. However, with CCTA (using intravascular ultrasound as a gold standard), plaque volume per vessel tends to be underestimated for mixed and noncalcified plaque, and overestimated for calcified plaque. Despite these limits, it remains true that an individual’s ACS risk will increase with both increasing total plaque volume/burden and increasing plaque vulnerability, as defined by the characteristics described above. While the relevance and clinical significance of CCTA-detected nonstenotic coronary plaque in ED chest pain patients is not currently well defined, knowledge of the presence, type, and general burden of coronary plaque might in the future aid in further refining treatment pathways by better stratifying patient risk for ED physicians treating chest pain (and potential ACS) patients.