PE is the third most common acute cardiovascular disease, after myocardial infarction and stroke (1). However, there is considerable uncertainty and confusion with regard to accurate diagnosis of this condition. The clinical signs and symptoms associated with PE are nonspecific, as are laboratory investigations, electrocardiograms, and chest radiographs. When PE occurs without infarction, the chest radiograph may be normal, or it may show any or all of the following: oligemia of the affected lung (the Westermark sign; see Fig. 2-21), increase in the size of the main pulmonary artery, elevation of the diaphragm, pleural effusion (usually small and unilateral), or discoid atelectasis. The chest radiograph is usually abnormal in patients with PE, however, with nonspecific subsegmental atelectasis being the most common abnormal finding (2). No chest radiographic sign is specific for pulmonary embolism or infarction, and the sensitivity of chest radiography for these conditions is poor. Even with a large pulmonary artery clot burden, the chest radiograph can be normal (3). The main role of the chest radiograph, therefore, is to exclude other diagnoses that might mimic PE clinically, such as pneumonia or pneumothorax. Because PE often goes undetected, the diagnosis of PE should be considered in any patient who presents with acute shortness of breath and pleuritic chest pain.

Deep venous thrombosis (DVT) originates most commonly in lower-extremity or pelvic veins, where they dislodge and propagate cranially into the pulmonary arterial tree. Radiologic studies used to diagnose thromboembolic disease include chest radiography, ventilation-perfusion (V/Q) scans, computed tomographic pulmonary angiography (CTPA), magnetic resonance imaging/magnetic resonance angiography (MRI/MRA), CT venography (CTV), MR venography, and lower-extremity ultrasound. Once the gold standard for diagnosing PE, catheter-based pulmonary angiography has largely been replaced by CTPA and is now used mainly when the results of CTPA and V/Q scanning are indeterminate and there is continued high clinical suspicion of PE. The ideal test to diagnose PE should be accurate, direct (objective), rapid, safe, readily available, and of reasonable cost. Because only approximately 30% of patients with clinically suspected PE have the disease (4), a diagnostic test that is able to provide information regarding the presence and significance of other chest disease would also be desirable. None of the common tests in use (other than CTPA) meet all or even most of these criteria. V/Q scintigraphy was the main imaging modality used in the evaluation of patients with suspected PE until the advent of multidetector CT scanning. A high-probability V/Q scan provides sufficient certainty to confirm the diagnosis of PE, while a normal or near normal scan reliably excludes the diagnosis. However, in the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study (4), indeterminate scans, which were present in 364 (39%) of 931 patients, showed a 30% incidence of PE, and low-probability scans (i.e., two thirds of V/Q scans in the PIOPED study) were not useful in establishing or excluding PE. In many institutions, CTPA has become the test of choice rather than V/Q scintigraphy or catheter-based pulmonary angiography. A suggested diagnostic algorithm for the evaluation of suspected PE is described in Table 17-1.

Recent studies have found the sensitivity of thin-section multidetector CTPA to be 96% to 100% and the specificity to be 89% to 98% for the detection of pulmonary emboli to the level of the subsegmental arteries (5,6). Characteristic findings of acute PE are: (a) partial central filling defect surrounded by a thin rim of contrast material, or (b) complete filling defect with obstruction of an entire vessel section (“vessel cutoff sign”) (Figs. 17-1, 17-2, 17-3, 17-4, 17-5). Pulmonary arteries that are completely obstructed by an acute embolus usually have an increased diameter (Figs. 17-6 and 17-7). Arteries peripheral to a central thrombus may or may not opacify. Central clot does not necessarily completely obstruct the distal flow of contrast. Although nonocclusive clot is depicted by CTPA, false-negative
scintigraphy in this setting is well known. Acute embolic obstruction of a large degree of the pulmonary circulation increases pulmonary vascular resistance, leading to acute PAH. CTPA findings suggesting this complication include right ventricular enlargement (right ventricle/left ventricle ratio >1) and straightening or leftward bowing of the interventricular septum (Fig. 17-8). Pitfalls to be aware of in diagnosing PE include lymph nodes; impacted bronchi (Figs. 17-9 and 17-10); respiratory motion; vessel bifurcation; unopacified pulmonary veins (Fig. 17-11); periarterial abnormalities (lymph node enlargement or infiltration of the axial interstitium by edema fluid, inflammation, or neoplasm); pulmonary artery catheters; and pulmonary artery sarcoma (7).

CTPA findings diagnostic of chronic PE include mural thrombus (adherent to the arterial wall), which may or may not be calcified (Fig. 17-12); webs; stenosis or strictures of the arteries (Fig. 17-13); and a central "dot" of contrast surrounded by circumferential thrombus, which is indicative of recanalization. Ancillary findings include mosaic perfusion with decreased caliber of vessels in the hypoattenuated areas of lung (Fig. 17-14), enlarged pulmonary arteries and right ventricle (Figs. 17-15 and 17-16), and enlarged bronchial arteries (Figs. 17-3 and 17-17). CTPA, like conventional angiography, usually enables distinction between acute and chronic PE; this is not possible with scintigraphy.

The clinical significance of small emboli is unclear, but data suggest that small, untreated clots in patients without impaired cardiopulmonary reserve may not be associated with poor outcome (8). Several investigations have found that the negative predictive value of CTPA ≥97% (9), suggesting that anticoagulants can be safely withheld when CTPA is normal and of good diagnostic quality. In patients without concomitant cardiopulmonary disease, no difference in the incidence of recurrent PE between treated and untreated patients with small clots has been noted (10). However, in patients with limited cardiopulmonary reserve, such small emboli may be fatal. Isolated subsegmental clot on single-detector CT is very unusual, and the risk of anticoagulation may exceed the risk of morbidity and mortality from the suspected clot in this setting (11).

Major advantages of CTPA over V/Q scintigraphy to investigate patients suspected of acute PE include (a) direct visualization of emboli on CTPA; (b) evaluation of the lung parenchyma and mediastinum, which may provide an alternate diagnosis; and (c) capability of acquiring a CTV study without







additional contrast (Fig. 17-18). Investigators have shown that CTPA provides an alternative diagnosis (e.g., pneumonia, pneumothorax, pleural effusion, pericarditis, aortic dissection, aortic aneurysm, congestive heart failure, rib fracture, lung nodules or mass, mediastinal mass or air, gallstones, chronic obstructive pulmonary disease) in up to two thirds of patients with an initial suspicion of PE (8). Limitations of CTPA include patients with an allergy to contrast material, impaired renal function, the inability to lie supine, the inability to be transported to the CT scanner, or inadequate intravenous access. Other limitations of CTPA include motion artifact caused by inability of the patient to hold his or her breath or by adjacent cardiac motion, poor contrast bolus enhancement, image noise in large patients, partial volume averaging (Fig. 17-19), parenchymal disease, and streak artifact from lines and tubes or dense contrast material.

PE and DVT are different manifestations of the same clinical disease. One advantage of CTPA is the ability to add CTV, from the iliac crest to the tibial plateau, to detect DVT in the legs
and pelvis (Figs. 17-20 and 17-21). Both CTPA and CTV can be accomplished with the same bolus of contrast agent. Unlike lower-extremity ultrasound, CTV can image the external and internal iliac veins. Venous thrombosis can also occur in the upper extremities and in the thorax and can be detected on CTPA (Fig. 17-22).

The D-dimer assay, a test that detects one of the products of fibrin breakdown in the blood, is an important rapid initial test for DVT and PE. Recent studies show that the enzyme-linked immunosorbent assay D-dimer test can accurately rule out DVT and PE in the vast majority of cases (11). However, this test can be falsely positive in postoperative patients, patients on anticoagulation, and patients with recent trauma.

MRI is useful in the evaluation of suspected PE when patients are allergic to iodinated contrast medium. Because it does not involve ionizing radiation, it is also advantageous in children and pregnant women.

Anticoagulant therapy must be considered for DVT as well as for PE; therefore, ultrasound of the deep venous system should have a primary screening role in patients suspected of PE. Ultrasound imaging has the advantages of being readily available and noninvasive. If ultrasound is negative for DVT, depending on the degree of clinical suspicion, further evaluation is generally obtained with a V/Q scan or CTPA.

The diagnostic feature of PE on a V/Q scan is a perfusion defect in a region of normally ventilated lung - the so-called "mismatched perfusion defect." Interpretation of V/Q scans is based on a comparison of the V/Q images and the chest radiograph, which gives rise to a report of "normal" or of low, intermediate, or high probability of PE (4). An abnormal V/Q scan indicating a low probability for recent PE is one in which the individual perfusion defects are smaller than 25% of a segment, regardless of the chest radiographic and ventilation scan appearances; are matched on the ventilation scan; or are accompanied by larger chest radiographic abnormalities. A high-probability scan is one in which there are two or more perfusion defects that are not matched by corresponding ventilation defects or chest radiographic abnormalities, including at least one of segmental or larger size. In the appropriate clinical setting, a high-probability V/Q scan indicates a probability of PE exceeding 90%. An intermediate-probability V/Q scan, also described as an indeterminate scan, is an abnormal scan that does not fit into the low- or high-probability categories. It includes those with perfusion defects that, although matched, correspond in size and shape to an area of opacity on the chest radiograph (and, therefore, may represent infarction or pneumonia) or with perfusion defects in areas of severe obstructive lung disease, pulmonary edema, or pleural effusion.

Only 15% or fewer of thromboemboli cause pulmonary infarction (12). It is not known why some emboli cause infarction and others do not, but it is likely a result of compromise of both the pulmonary and bronchial arterial circulation. This is most likely to occur with peripheral emboli and in patients with left heart failure or circulatory shock (13). It is known that bronchial circulation alone can sustain the lung parenchyma without infarction occurring (14).

Pulmonary infarction results in airspace opacities that are usually multifocal and predominantly in the lower lungs. They usually appear within 12 to 24 hours after the embolic event. The opacities are classically peripheral, with a triangular or rounded shape (thus the term Hampton hump), and they are always in contact with the pleural surfaces (Figs. 17-23 and 17-24). The apex of the opacity is directed toward the lung hilum. Occasionally, lobar opacity resembling pneumonia can occur. Air bronchograms are rarely present. It is important to note that the opacities can represent a combination of pulmonary hemorrhage and atelectasis without infarction, in which case clearing occurs within a week. Infarction takes several months to resolve, often with residual scarring (Fig. 17-25). As infarcts resolve, they melt away "like an ice cube" (giving rise
to the so-called "melting ice cube sign"; Fig. 2-16). The opacity clears from the periphery first, whereas in pneumonia the opacity clears homogeneously, both centrally and peripherally at the same time. Cavitation can occur within infarcts but is rare without coexisting infection, either secondary infection of an infarct or a result of septic emboli or vasculitis.

PAH is defined as pulmonary artery pressures above the normal systolic value of 30 mm Hg or above the mean value of 18 mm Hg. There are numerous causes of PAH (Table 17-2), which is classically categorized as either precapillary or postcapillary. Regardless of the etiology, the radiologic features are similar and include enlargement of the central pulmonary arteries and narrowing or "pruning" of the peripheral pulmonary artery branches (Fig. 17-26). Right ventricular enlargement is often appreciated on the lateral chest radiograph. However,
substantial PAH may be present in patients with normal chest radiographs. CT more accurately depicts the size of the pulmonary arteries and cardiac chambers. As a general rule, PAH is present when the main pulmonary artery diameter exceeds that of the ascending aorta or is 29 mm or more in diameter (15) (Fig. 17-27). In long-standing and severe PAH, the enlarged central pulmonary arteries may develop thrombus and peripheral calcification. This is most often seen in patients with Eisenmenger physiology, a condition characterized by a reversal in the direction of a long-standing severe left-to-right shunt (i.e., atrial septal defect, ventricular septal defect, patent ductus arteriosus) (Fig. 17-28). In postcapillary and some precapillary disorders, changes of pulmonary venous hypertension may be seen. The most salient finding is cephalization of pulmonary vasculature, which represents recruitment of upper lobe vasculature secondary to a diversion of blood flow. Pericardial effusions, usually small to moderate in size, are commonly associated with PAH.

Primary pulmonary artery sarcomas are exceedingly rare. They involve the central pulmonary arteries and often completely occlude the involved vessel. The appearance on CT can mimic massive acute PE. However, when complete occlusion and expansion of the main, left, or right pulmonary arteries is seen, tumor should be considered. Even so, it is more common that this appearance is caused by metastatic tumor or invasion of the pulmonary artery by adjacent mediastinal or central bronchogenic carcinoma than by primary pulmonary artery sarcoma.

Tumors most commonly known to embolize through the pulmonary arterial circulation include bronchioloalveolar carcinoma; carcinomas of the breast, kidney, stomach, liver, and prostate; and choriocarcinoma (16). The characteristic CT features are dilatation and beading of peripheral pulmonary arteries
(Fig. 17-29). The appearance can resemble the "tree-in-bud" appearance of small airway disease (17). Maximum-intensity-projection images can be helpful in showing continuity of the distal vessels with more central pulmonary vessels in cases of tumor emboli. Many more cases of tumor emboli are discovered at autopsy than antemortem.