|
Diagnosing Pulmonary Embolism
Although expert clinical diagnosis of pulmonary embolism is relatively accurate, little can be done for the patient without confirmation, and the traditional means to that end have been inconclusive, impractical, and hazardous. Greater reliability and the power to reveal other abnormalities have brought multidetector computed tomography to the diagnostic forefront, but it is not without its drawbacks.
By Sam Kini, MD, Douglas R. Lake, MD, Joseph J. Kavanagh, MD, MBA, James G. Ravenel, MD, U. Joseph Schoepf, MD, Philip Costello, MD, and Thomas Pope, MD
|
Dr. Kini is an associate professor of emergency medicine, Drs. Lake and Kavanaugh are radiology residents, Drs. Ravenel and Schoepf are associate professors of radiology, Dr. Costello is professor and chairman of radiology, and Dr. Pope is professor of radiology and orthopedics at the Medical University of South Carolina in Charleston. Dr. Pope is also a member of the EMERGENCY MEDICINE editorial board. |
Pulmonary embolism (PE) is a potentially fatal cause of chest pain and dyspnea that may be seen in the emergency department and may result in sudden death. It is not an uncommon presentation. The American Heart Association estimates that nearly 600,000 Americans develop PE annually, 25,000 are hospitalized each year with this diagnosis, and 60,000 succumb to the condition.
Pulmonary embolism may be difficult to diagnose, and is especially challenging in patients who have underlying heart or lung disease. Various tests are used to aid in the diagnosis, but most have limited reliability.
The box below lists the symptoms and signs that should prompt the emergency physician to consider PE in the differential diagnosis. It is important to note that the classic triad of hemoptysis, dyspnea, and chest pain is neither sensitive nor specific and occurs in fewer than 20% of patients in whom the diagnosis of PE is subsequently made. The second box below lists important risk factors for PE.
In this article, we will discuss the standard diagnostic tests and imaging modalities that have traditionally been used on patients with suspected PE. We will then focus on multidetector computed tomography (MDCT), which represents a significant breakthrough in the diagnostic approach to PE.
STANDARD DIAGNOSTIC TESTS
Once the diagnosis of PE has been considered, based on clinical presentation, existing risk factors, and a high index of suspicion, the emergency physician must choose the best test to confirm the diagnosis. Historically, routine tests such as a chest x-ray, arterial blood gas analysis, ECG, ventilation-perfusion (V/Q) scan of the lung, and Doppler ultrasound of the extremity for deep vein thrombosis (DVT) have been ordered in this setting. However, the interpretation of these tests is difficult and unreliable. Pulmonary angiography, of course, has always been the gold standard for confirming the diagnosis of PE, but for various reasons it may not always be practical in the emergency department.
It is possible for an experienced clinician to make a reasonably accurate diagnosis of PE without ordering any diagnostic tests. However, confirmation of that suspicion is necessary prior to initiating anticoagulation, consultation, and admission to the hospital. It is also important for physicians to realize that many of the standard tests are useful in ruling in the diagnosis of PE when the results are positive, and ruling out PE when the results are negative—but not both.
The D-dimer test is sensitive but interpretation can be problematic. When the test is negative in a low-risk patient, PE can be excluded. If the result is positive, the diagnosis of PE cannot be made without further confirmation using other, more reliable diagnostic tests such as MDCT or pulmonary angiography.
Furthermore, recent PIOPED II (Prospective Investigation of Pulmonary Embolism Diagnosis) re-commendations and diagnostic pathways also include discussions of the workup of low-, moderate-, and high-risk patients, depending on the risks associated with iodinated contrast material, radiation exposure, and the availability of diagnostic tests. Low- and moderate-risk patients as determined by the Wells or Geneva criteria should receive a D-dimer test; if the results are positive, most investigators advocate MDCT.
Chest x-rays are notoriously unremarkable in many patients with PE. At times, subtle clues such as a Westermark sign, dilated pulmonary vessels, or atelectasis with or without a small pleural effusion may be present. The finding of a triangular opacity adjacent to the costophrenic sulcus, the so-called Hampton hump, is highly suggestive of a pulmonary embolus but is rarely seen. The most common radiographic abnormality found in patients with PE is cardiomegaly.
An ECG is usually normal or may show nonspecific abnormalities such as sinus tachycardia or ST- or T-segment changes. It can show signs of right-heart strain such as negative T-waves in the precordial leads, right bundle-branch block, an S1Q3T3 pattern, or a QR wave in lead V1.
IMAGING MODALITIES
Traditionally utilized imaging tests such as pulmonary angiography and V/Q scanning are used less often than the standard tests just discussed for suspected PE. Angiography is an invasive procedure that is subject to complications; it is also fraught with poor interobserver reliability for subsegmental clots. The V/Q scan remains clinically unsatisfying for the following reasons: most scans are of intermediate or indeterminate probability (73% of all scans performed), the specificity of a low-probability exam is 10%, and there is poor interobserver correlation.
Duplex venous ultrasound of the lower extremities is advocated early in the evaluation of PE because it is relatively easy to perform and interpret, and a positive result obviates further testing because treatment of DVT and PE is similar. However, this test may miss more than half of all patients with PE, and it is relatively insensitive for diagnosing DVT in asymptomatic patients.
Magnetic resonance angiography (MRA) has not played a significant role in the evaluation of PE because of its long scan-acquisition time, motion artifact from respiration and cardiac activity, and limited spatial resolution. However, new ultra-fast MRA scanning techniques allowing acquisitions in 4 to 30 seconds are achieving diagnostic accuracy that approaches MDCT and may play a bigger role in evaluating patients with contrast allergy or renal insufficiency.
MDCT FOR PULMONARY EMBOLISM
Multidetector CT obtains images by spiraling down through the thorax (helical technique) and obtaining data from multiple (2, 4, 8, 16, 32, 64, or 128) detectors, depending on the generation of scanner used. The current generation of scanners allows the acquisition of a volume of data from the entire thorax in submillimeter resolution in less than 10 seconds. Images are obtained in a craniocaudal direction following a 4 ml/sec bolus injection of a contrast agent through an 18- or 20-gauge catheter, preferably in an antecubital vein. The images can then be reformatted in numerous planes for problem solving. Thin collimation technique at 1.25-mm-slice thickness or less has been shown to improve depiction of subsegmental vessels and improve interobserver agreement. However, thinner collimation increases image noise, and thus 2- to 2.5-mm collimation may be appropriate for larger patients who have a higher incidence of PE.
The diagnosis of acute PE is based on failure to enhance the entire lumen due to a filling defect. Frequently, the artery is enlarged compared to adjacent vessels. A partial filling defect may be surrounded by contrast material, producing the “lifesaver” sign (see image below) on images obtained perpendicular to the long axis of the vessel, or the “railway track” sign (see second image below) on longitudinal images of the vessel. Occasionally, a sharp vessel cutoff is detected.
 Lifesaver sign. Subsegmental left lower lobe pulmonary embolus (white circle) demonstrates the classic “lifesaver” sign of a dark thombus outlined by white contrast on axial images. The long axis of the embolus is perpendicular to the plane of scanning. |
 Railway track sign. This coronal reformatted image depicts a segmental left lower lobe embolus (white rectangle)—the so-called “railway track” sign. Note that the embolus is within the plane of imaging. |
Secondary signs of acute PE include peripheral wedge-shaped regions of ground-glass opacity or consolidation (CT Hampton hump), which occurs in 25% of cases of PE (see image below). Localized oligemia may also be seen with larger emboli, corresponding to the so-called CT Westermark sign. Larger areas of ground-glass opacification can be a secondary but quite nonspecific finding. Nonspecific abnormalities such as subsegmental atelectasis and small unilateral pleural effusions may also be demonstrated.
 CT Hampton hump. In this axial image, the black oval shows opacification of the superior segment of the right lower lobe—or a CT Hampton hump, representing infarct—accompanying a pulmonary embolus. |
Diagnoses other than PE may be made in up to 67% to 76% of patients thought to have PE, depending on the definition of “other abnormality.” This is a significant benefit with MDCT for the emergency department physician. Important diagnoses made or supported by CT include pneumonia, interstitial lung disease, aortic dissection, lung malignancy, and pleural disease. In the near future, MDCT scanners and dual-source CT scanners will be capable of excluding significant coronary artery stenosis, aortic dissection, and PE with one scan (the “triple-rule-out” scan).
After experiencing the initial embolic event, PE patients are at risk for circulatory collapse from right-heart failure. While right ventricular strain and failure are most optimally monitored with echocardiography, some morphologic abnormalities on CT may suggest the diagnosis. The right ventricular/left ventricular ratio measures diameters from the right ventricular free wall to the interventricular septum and from the interventricular septum to the lateral left ventricular myocardium. While this was first described using a four-chamber reconstructed view, standard transverse views provide similar ratios. The right ventricle/left ventricle ratio correlates relatively well with clinical severity. Significantly more adverse events occurred in patients with a ratio of more than 0.9. Deviation of the interventricular septum toward the left ventricle and contrast reflux into the hepatic veins also suggest right ventricular strain (see images below).
 CT assessment of prognosis. Note the relative enlargement of the right ventricle (solid black arrow) versus the left ventricle (dashed black arrow). The ratio is more than 0.9, which has been shown to correlate relatively well with the clinical severity of PE. Note too the relative straightening of the interventricular septum, which bows slightly into the left ventricle, also indicating right ventricular strain. |
 CT assessment of prognosis. This axial image through the level of the hepatic veins in the same patient demonstrates hepatic reflux (white arrows), further suggesting right ventricular strain. |
An additional examination that can be performed at the time of thoracic MDCT is indirect CT venography (ICTV). The majority of PIOPED II investigators preferred the combination of ICTV and thoracic MDCT.Between 120 to 180 seconds after contrast administration, an additional scan of the pelvis and lower extremities to the level of the popliteal veins occurs at the time of optimal venous opacification in a search for DVT.The additive value of ICTV to MDCT for PE ranges from 4% to 15%.In comparison with ultrasound, ICTV has a sensitivity of 89% to 97%, a specificity of 94% to 100%, a positive predictive value of 64%, and a negative predictive value of 98%, making it a useful adjunctive examination obviating the need for ultrasound in most cases.
The use of MDCT for detecting PE is not without drawbacks.Paramount among these is radiation dose considerations. Typical radiation doses for routine MDCT studies for PE average 5.7 mSv, which is approximately double the average aggregate background radiation dose per year (3.0 mSv).Current estimates are that an effective dose of 5 mSv (500 mrem) corresponds to a risk of developing a fatal cancer of 2.5 per 10,000.Young female patients are often scanned for PE, and few realize a 5 mSv thoracic MDCT for PE imparts a radiation dose to the female breast equivalent to 10 to 25 two-view mammograms and up to as many as 100 to 400 chest radiographs. The potential latent carcinogenic effects of this dose are not clearly known, but a retrospective study of 1030 women with scoliosis who underwent multiple thoracic spine radiographic examinations as young girls revealed a nearly two-fold statistically increased risk for incident breast cancer.
CAUSES OF MISDIAGNOSIS
A variety of problems may reduce the utility of MDCT. These include patient-related, technical, anatomic, and pathologic factors. One common patient-related factor is a respiratory-motion “stair-step” artifact of alternating bands of low and high attenuation along the course of a vessel. Technical factors include streak artifact, which results from dense contrast bolus descending in the superior vena cava, and window-setting artifact (see images below). Window-setting artifact is best avoided with PE-specific window and level settings of 700 and 100 Hounsfeld units, respectively.
Window-level artifact. Two axial images at the level of the interventricular septum in the same patient are shown at different window and level settings, demonstrating the importance of these settings. The first is set at a mediastinal window with a level setting of 400/40, which is commonly used to detect PE. However, the embolus (this image, white circle) is difficult to identify with confidence. But with a PE-specific window and a level setting of 700/100, the embolus becomes apparent (see image below).
|
The most common anatomic and pathologic problems encountered include partial-volume averaging effects from lymph nodes and mucous plugging, respectively. Nearly all of these artifacts are best avoided by reviewing images reconstructed at the thinnest section possible—at most 2.5 mm—and in multiple planes, such as coronal or sagittal reconstructions. A 5-mm section thickness is generally considered inadequate for reliable exclusion of PE.
INCREASED UTILIZATION
The excellent performance of MDCT (100% sensitivity, 89% specificity, 91% accuracy)—and anticipated improvement with 16- to 64-detector row units and upcoming dual-source CT scanners will likely continue to fuel the trend of utilization of MDCT for the detection of PE. Furthermore, a recent meta-analysis showed the clinical validity of a negative MDCT scan to exclude PE is similar to that reported for conventional pulmonary angiography. The low rate of PE after negative MDCT (1% to 1.7%) should also support its increased usage.
back to top
Suggested Reading
Abcarian PW, et al.: Role of a quantitative D-dimer assay in determining the need for CT angiography of acute pulmonary embolism. AJR Am J Roentgenol 182(6): 1377, 2004.
Buller HR, et al.: Antithrombotic therapy for venous thromboembolic disease: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 126(3 Suppl): 401S, 2004.
Ghaye B, et al.: Can CT pulmonary angiography allow assessment of severity and prognosis in patients presenting with pulmonary embolism? What the radiologist needs to know. Radiographics 26(1): 23, 2006.
Goldhaber SZ: Multislice computed tomography for pulmonary embolism—a technological marvel. N Engl J Med 352(17): 1812, 2005.
Perrier A, et al: Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med 352(17): 1760, 2005.
Prologo JD and Glauser J: Variable diagnostic approach to suspected pulmonary embolism in the ED of a major academic tertiary care center. Am J Emerg Med 20(1):5, 2002.
Prologo JD, et al: CT pulmonary angiography: a comparative analysis of the utilization patterns in emergency department and hospitalized patients between 1998 and 2003. AJR Am J Roentgenol 183(4): 1093, 2004.
Quiroz R, et al.: Clinical validity of a negative computed tomography scan in patients with suspected pulmonary embolism—a systematic review. JAMA 293(16): 2012, 2005.
Quiroz R, et al.: Right ventricular enlargement on chest computed tomography: prognostic role in acute pulmonary embolism. Circulation 109(20): 2401, 2004.
Stein PD, et al.: Diagnostic pathways in acute pulmonary embolism: recommendations of the PIOPED II investigators. Radiology 242(1): 15, 2007.
Washington L, et al.: CT for thromboembolic disease. Radiol Clin North Am 40(4): 751, 2002.
Wells PS, 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 83(3): 416, 2000.
Wittram C, et al.: CT angiography of pulmonary embolism: diagnostic criteria and causes of misdiagnosis. Radiographics 24(5): 1219, 2004. |
|
