CHAPTER 97 Acute Pulmonary Thromboembolic Disease

Pulmonary embolism (PE) was clinically described in the early 1800s, and von Virchow first described the connection between venous thrombosis and PE.¹^,² In 1922, Wharton and Pierson reported the first radiographic description of PE,³ a commonly encountered entity whose diagnosis is clinically challenging because of its nonspecific symptoms.

Definition

Pulmonary thromboembolism is defined as the obstruction of the pulmonary artery or one of its branches leading to the lungs by a blood clot, usually from the leg, or foreign material causing sudden closure of the vessel. (Embolus is from the Greek embolos, meaning plug.) Emboli can also be composed of fat, air, or tumor tissue.

PE is a potentially fatal condition with substantial morbidity and mortality in untreated patients. The annual incidence of pulmonary embolism is estimated to be between more than 300,000 cases, resulting in approximately 50,000 to 100,000 deaths in the United States every year.⁴ Nearly 90% of PEs arise from lower extremity deep venous thrombosis (DVT). Diagnosis of pulmonary embolism remains a clinical challenge because of its nonspecific presentation. Mortality ranges from 3.5% to 15% and can be as high as 31% to 58% in the presence of shock.⁵^,⁶ Hence, prompt and accurate diagnosis is imperative because 65% of patients die within 1 hour of presentation and approximately 93% in the first 2.5 hours.⁷ Since its original description, imaging has played an important role in the diagnosis of PE. For many years, ventilation-perfusion (V/Q) scintigraphy has been the main imaging modality for the evaluation of patients with suspected PE. However, with the advent of and the widespread availability of faster computed tomography (CT) scanners, CT scanning has emerged as another important diagnostic test for the evaluation of not only PE but also DVT in select patients. Thus, familiarization with various aspects of imaging evaluation is of extreme importance for the interpreting physician.

Various imaging modalities are currently available to evaluate suspected PE in routine clinical practice. We will discuss current imaging options in the remainder of this chapter.

Etiology and Pathophysiology

Three primary influences predispose a patient to thrombus formation, which form the so-called Virchow triad: (1) endothelial injury; (2) stasis or turbulence of blood flow; and (3) blood hypercoagulability.^1,^2,⁸

More than 90% of all PEs arise from thrombi within the large deep veins of the legs, typically the popliteal vein and the larger veins central to it.^1,^2,⁴ The pathophysiologic consequences of thromboembolism in the lung largely depend on the cardiopulmonary status of the patient and on the size of the embolus, which in turn dictates the size of the occluded pulmonary artery.

PE has two important consequences: (1) an increase in pulmonary artery pressure produced by obstruction of flow, and possibly vasospasm caused by neurogenic mechanisms and/or release of mediators (such as thromboxane A2 and serotonin); and (2) ischemia of the downstream pulmonary parenchyma. Thus, occlusion of major vessels of more than 60% of the arterial bed suddenly increases pulmonary artery pressure, diminishes cardiac output, and causes right-sided heart failure (acute cor pulmonale) or even death. Usually, hypoxemia develops as a result of multiple mechanisms. If smaller vessels are occluded, the result is less catastrophic or the event may even be clinically silent.

MANIFESTATIONS

Clinical Presentation and Pretest Probability

Many patients with suspected PE present with nonspecific findings such as acute chest pain, and some are even asymptomatic. Some observers have suggested that because of the decline in tolerance for diagnostic uncertainty, diagnostic studies for the evaluation of PE may be overused in current practice. Pretest probability is an important concept, because the sensitivity, specificity, and predictive values of diagnostic imaging modalities are based on the pretest probability of PE, which is usually determined by clinical findings and laboratory test results.⁹^–¹¹ The Wells criteria (Table 97-1)¹² and Geneva score,¹³ which were devised and validated for clinical pretest probability, have been reported to have similar accuracy levels. Chagnon and colleagues¹⁴ have shown that these methods can be used to divide patients into low-, intermediate-, and high-probability groups for PE, although neither was accurate enough to diagnose PE in individual cases.

TABLE 97-1 Wells Clinical Decision Rule for Pulmonary Embolism

Variable	Points
Clinical signs and symptoms of deep vein thrombosis	3.0
Alternative diagnosis less likely than pulmonary embolism	3.0
Heart rate > 100/min	1.5
Immobilization (> 3 days) or surgery in the previous 4 weeks	1.5
Previous pulmonary embolism or deep vein thrombosis	1.5
Hemoptysis	1.0
Malignancy (receiving treatment; treated in the last 6 months or palliative)	1.0

Clinical probability of pulmonary embolism unlikely—4 points or less.

Clinical probability of pulmonary embolism likely—more than 4 points.

Adapted from Wells PS, Anderson DR, Rodger M, et al. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost 2000; 83:416-420.

Serum D-dimer levels can be used to gauge the presence of PE. D-dimer is a fibrin degradation product that increases with clot lysis, suggesting the presence of thrombosis. It has been shown that this test with levels above a threshold value of 0.5 mg/L has an excellent sensitivity of 95%, but the specificity is lower (38% to 83%)¹⁵ because of elevated levels in some conditions, such as active inflammation, infection, pregnancy, malignancy, and liver failure. Of all the methods available for measurement of D-dimer levels (Table 97-2),¹⁰ the VIDAS D-dimer assay (bioMérieux, Marcy l’Etoile, France), a rapid, quantitative, enzyme-linked immunosorbent assay (ELISA) procedure, has been reported to have a 96.4% sensitivity,¹⁶ thus ruling out PE in the absence of a calculation of clinical pretest probability.¹⁶^,¹⁷ However, this is not routine clinical practice. Several studies^10,^11,¹⁸ have demonstrated that a negative result of rapid quantitative ELISA D-dimer can be combined with the clinical probability score to exclude PE in up to 40% of patients, thereby eliminating further workup.

TABLE 97-2 Sensitivity and Specificity (%) of D-Dimer Based on Laboratory Technique

Laboratory Technique	Mean (Range)
Laboratory Technique	Sensitivity	Specificity
ELISA	95 (85-100)	44 (34-54)
Rapid quantitative ELISA	95 (83-100)	39 (28-51)
Rapid semiquantitative ELISA	93 (79-100)	36 (23-50)
Rapid qualitative ELISA	93 (74-100)	68 (50-87)
Quantitative latex agglutination	89 (81-98)	45 (36-53)
Semi quantitative latex agglutination	92 (79-100)	45 (31-59)
Whole blood agglutination	78 (64-92)	74 (60-88)

ELISA, enzyme-linked immunosorbent assay.

Adapted from Ginsberg JS, Wells PS, Kearon C, et al. Sensitivity and specificity of a rapid whole-blood assay for D-dimer in the diagnosis of pulmonary embolism. Ann Intern Med 1998; 129: 1006-1011.

Imaging Techniques and Findings

Chest Radiography

Initially, the chest radiography findings commonly are normal. However, in later stages, the radiograph may show findings that include a Westermark sign (dilation of pulmonary vessels and a sharp cutoff), atelectasis, a small pleural effusion, and an elevated diaphragm. Although chest radiographic findings may indicate an alternate diagnosis, this study alone is not sufficient to confirm the diagnosis of PE.

Chest radiographic findings of PE can be subtle, nonspecific, and infrequently present. Among the patients in a PIOPED study without any evidence of prior cardiopulmonary disease, 84% of patients with proven PE and 66% of those without PE had abnormal chest radiographs. In an actual clinical setting, certain findings may suggest a diagnosis of PE and direct further imaging.

The Westermark sign (Fig. 97-1)¹⁹ is a focal peripheral lucency beyond an occluded vessel accompanied by mild dilation of central pulmonary vessels, seen in 7% to 14% of cases of documented PE in the PIOPED study. It is a subtle finding caused by embolic obstruction or hypoxic vasoconstriction of pulmonary artery.

FIGURE 97-1 Westermark sign on chest radiograph showing oligemia of the right lung fields in comparison to the left lung. It denotes focal peripheral lucency beyond an occluded vessel accompanied by mild dilation of central pulmonary vessels.

Another finding is enlarged central pulmonary vasculature, which can be easily missed. It may be caused by vessel distention by thrombus or by an acute rise in pulmonary arterial pressure secondary to distal emboli. A sausage-like configuration of focal enlargement of the right descending interlobar pulmonary artery caused by the physical presence of a clot has also been noted in many patients with acute PE and referred to as the knuckle sign. Hampton hump ²⁰ is classically referred to as a conical peripheral opacity pointing toward the hilum. These are multiple, subpleural lower lobe infarcts seen as ill-defined opacities without air bronchography (Figs. 97-2 and 97-3).

FIGURE 97-2 Classic case of pulmonary embolism with Hampton hump sign showing subpleural infarct opacities on radiograph (A) and corresponding axial CT scans (B, C).

(Courtesy of Dr. Jeffrey P. Kanne, Cleveland Clinic, Cleveland, Ohio.)

FIGURE 97-3 A, B, Axial CT scans of 66-year-old patient with acute right upper lobe pulmonary embolism. B, The arrow points toward a developing infarct in the posterior segment of right upper lobe.

Focal parenchymal abnormalities are the most commonly encountered lesions on chest radiographs in PIOPED series patients. These particularly include subsegmental atelectasis (Fleischner lines),²¹ seen as linear opacities in lung bases; these are transient in nature and thought to be caused by mucus plugging, hypoventilation, or distant airway closure. Focal air space consolidation represents true pulmonary infarction, with ischemic necrosis or pulmonary hemorrhage without infarction occurring in 10% to 60% of patients. Small, unilateral, pleural effusions and diaphragmatic elevation are seen in a large number of PE patients.

In summary, the chest radiograph is often normal in patients with suspected PE. The sensitivity (33%) and specificity (59%) are low. The major role is in exclusion of conditions such as pulmonary edema, pneumonia, or pneumothorax, whose symptoms may overlap with those of PE. In addition, chest radiographs are used for correlation of the interpretation of V/Q scintigraphy results.

Lower and Upper Extremity Ultrasonography

As noted, 90% of PEs originate from thrombosis within the deep venous system of the lower extremities and are mostly clinically silent; other sources are the pelvic, renal, and upper extremity veins.²² Significant morbidity and mortality are associated with undiagnosed DVT; thus, methods such as contrast venography (CV), magnetic resonance venography (MRV), and lower extremity ultrasonography are commonly used for its detection. Real-time gray scale, continuous wave and pulsed Doppler, and color Doppler imaging, along with ancillary techniques such as the Valsalva maneuver and manual blood flow augmentation, are used to image the lower extremity veins.

Acute thrombus is often anechoic with variable echogenicity, making compression ultrasonography an important tool for assessment. The diagnosis is established by lack of venous compression caused by intraluminal thrombus. Lack of appropriate response to a Valsalva maneuver also indicates thrombosis of the central veins outside the field of view. The reliability of a negative compression ultrasonography examination is high.

Spectral Doppler analysis is particularly useful for vessels that cannot be visualized directly, such as the inferior vena cava ( IVC), superior vena cava (SVC), and brachiocephalic veins. Normal patent vessels show respiratory phasicity, whereas a monophasic waveform suggests venous obstruction.

Color Doppler imaging is useful to identify deeper veins such as the iliac veins and superficial femoral veins, for which direct venous compression is not possible, or for obese or swollen extremity patients. Thrombosis is identified as the absence of color flow within a vessel lumen at baseline and with augmentation; the sensitivity and specificity are 95% and 98%, respectively, for femoropopliteal DVT.²²

Upper extremity thrombosis is screened using spectral and color Doppler techniques for veins inaccessible to direct compression. The incidence of PE with upper extremity DVT may be as high as 12% to 15%, making these diagnostic tests highly significant.

In summary, color flow Doppler imaging and compression ultrasonography have a high sensitivity (89% to 100%) and specificity (89% to 100%) for the detection of proximal DVT in symptomatic patients. However, compression ultrasonography has a low sensitivity (38%) and a low positive predictive value (26%) in patients without symptoms of DVT. Patients with positive findings for DVT can be anticoagulated irrespective of their V/Q scan results; other patients must have more invasive investigations performed to rule out PE definitively.

Ventilation-Perfusion Scintigraphy

V/Q scanning of the lungs is an important diagnostic modality for establishing the diagnosis of PE. The PIOPED classification scheme allows for a more meaningful interpretation of the V/Q scan to treat patients with anticoagulation. The scans should be interpreted primarily as a diagnostic or nondiagnostic pattern, indicating whether the patient has or does not have a high likelihood of having PE.

A normal perfusion scan excludes a diagnosis of PE, with its negative predictive value close to 100%.²³^–²⁵ It is a safe, relatively inexpensive, and widely available test that allows the acquisition of single-breath, equilibrium, and washout images. A high-probability V/Q scan (Figs. 97-4 and 97-5), in combination with high-probability clinical findings, corresponded to PE in 96% of patients, whereas a low-probability clinical assessment showed PE in only 4% of patients in a major study.⁹

FIGURE 97-4 Anterior projection of a Tc 99m diethylenetriamine pentaacetic acid (DTPA) aerosol image demonstrates relatively homogeneous ventilation.

(Courtesy of Dr. Donald Neumann, Cleveland Clinic, Cleveland, Ohio.)

FIGURE 97-5 Left posterior oblique (A) and right lateral planar (B) views of Tc 99m macroaggregated albumin pulmonary perfusion scans show multiple large perfusion defects compatible with a high-probability scan in a patient with pulmonary embolism.

(Courtesy of Dr. Donald Neumann, Cleveland Clinic, Cleveland, Ohio.)

In PIOPED-1, a V/Q scan in patients with a normal chest radiograph was diagnostic in 52% of patients with suspected PE,²⁶ and one study has shown it to be diagnostic in 91% of patients.²⁷ This PIOPED group retrospectively analyzed perfusion scans alone and found the results to be equivalent to those of the V/Q technique.²⁸ In the PISA-PED trial,²⁹ only perfusion scans were obtained. These were classified according to the shape of defects, yielding sensitivity and specificity of 86% and 93%, respectively, for abnormal perfusion scans compatible with PE, with the result being dichotomous. These studies indicate that the ventilation scan can be eliminated, thus reducing cost and radiation dose. It can also be used as a preferred alternative for patients unable to undergo CT angiography.

It can be used particularly for reproductive female patients whose chest radiograph is likely to be normal and in whom the breast irradiation dose from CT angiography can be minimized using the perfusion scan only.⁹^,³⁰ Another significant use is follow-up of proven PE patients undergoing anticoagulation.

Ventilation-Perfusion Single Photon Emission Computed Tomography

V/Q single photon emission computed tomography (SPECT) is a newer modality that may improve both sensitivity and specificity of the V/Q scan (Fig. 97-6). Reinartz and associates³¹ analyzed 83 patients who underwent four-slice spiral CT, V/Q planar, and V/Q SPECT studies, giving sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of 97%, 91%, 98%, and 90%, respectively, for SPECT techniques. This study used aerosolized technetium 99m (Tc 99m) for the ventilation portion of the V/Q scan instead of the radioisotope xenon 133. Tc 99m is five times smaller in diameter and has a 20% efficiency of pulmonary deposition in comparison to 2% for xenon 133, thus helping improve the results.³²^,³³ Unlike PIOPED, this study considered all the mismatched results as PE, regardless of the size. This technique has the potential to improve the interpretation of V/Q scanning by incorporating computerized interpretations.