Special Feature

New Diagnostic Approaches to Pulmonary Embolism

By Stephanie B. Abbuhl, MD, FACEP, and Raquel M. Schears, MD, MPH

Pulmonary embolism (pe) continues to be a difficult diagnostic challenge for emergency medicine physicians. The lack of a reliable, noninvasive method for detection, coupled with nonspecfic patient presentations and examination findings, results in the majority of PE cases being diagnosed postmortem. It is estimated that 650,000 cases of PE occur annually in the United States, with PE as the primary cause of death or as a significant contributor in 200,000 deaths, making this entity the third most frequent cause of death overall in hospitalized patients.1

The current standard of care in the initial ED evaluation of patients with suspected PE is radionuclide ventilation-perfusion (VQ) scanning of the lung. Unfortunately, VQ scans confirm or exclude PE in less than one- third of all patients with PE.2 The remainder of patients fall into a non-diagnostic category usually requiring further testing to make a diagnosis.

While pulmonary angiography has been the gold standard for the diagnosis of PE, it is far from the ideal test. It is an invasive test associated with a 0.5% mortality risk and a serious morbidity risk of about 1%.2 In addition, angiography requires intravenous contrast, is not always available, requires significant radiologic expertise, and can be associated with combined technical and professional fees as high as $8500.

There have been recent advances in spiral CT, MRI, and D-dimer testing that suggest exciting new roles for these modalities in the ED evaluation of PE. The advantages and disadvantages of these new approaches will be reviewed.

D-Dimer Assays

Clot formation anywhere in the body activates the endogenous fibrinolytic system. When cross-linked fibrin is broken down by plasma, D-dimers can be measured by either an enzyme-linked immunosorbent-assay (ELISA) or a latex agglutination test. The quantitative ELISA test has been the most promising, with sensitivities for thromboembolic disease reported as high as 97-100% (CI 92-100).3 Unfortunately, the standard ELISA test has been time consuming compared to the rapid, but less sensitive, latex agglutination assay. Recently, however, a rapid ELISA assay has been developed that may expand the role of D-dimer testing in the ED.3 The biggest disadvantage to the D-dimer assay is its lack of specificity (ranging from 15% to 70%).4,5 D-dimer levels can be elevated due to numerous causes, including surgery, infection, cancer, and liver disease.

Because D-dimer results appear less sensitive but non-specific for venous thromboembolic disease, a positive D-dimer level is of little value, but a negative test is potentially useful to exclude disease (high negative predictive value). In a recent outcome study by Ginsberg and colleagues, 1177 consecutive patients with suspected PE underwent an evaluation that included an assessment of pretest probability, a rapid bedside D-dimer test, VQ scan, and bilateral ultrasonography of the lower extremities.6 All patients were followed-up for three months to assess the adequacy of the testing strategy using reasonable definitions of PE and deep vein thrombosis (DVT). Overall, the D-dimer assay showed a sensitivity of 85% with a specificity of 68%. However, in a subgroup analysis, patients with low pretest probability and a normal D-dimer test had a negative predictive value of 99% and this combination occurred in 44% of the study sample. Ginsberg et al also looked at the subgroup of patients with a non-diagnostic VQ scan (low or intermediate probability) and a normal D-dimer test and found this combination had a negative predictive value of 97%.

In another important, recent study, D-dimer levels were assayed by a rapid ELISA test in 918 consecutive patients undergoing PE evaluation.7 A normal D-dimer concentration (< 500 mcg/L) ruled out venous thromboembolism in 31% of the cohort with a negative predictive value of 99%. We are cautiously optimistic that there may be a role for D-dimers in the ED, particularly in the subgroup of patients with a low pretest probability or in other groups when combined with additional tests.

Spiral CT

Advances in CT technology have made spiral CT an exciting, alternative modality for evaluating PE. Spiral CT allows continuous volumetric image data acquisition during a single breath hold. If scanning is timed appropriately after contrast injection, there is excellent visualization of the pulmonary arteries and an embolism will appear as a filling defect in the center or periphery of the artery. In more peripheral (segmental and subsegmental) arteries, a clot can completely occlude or even enlarge an artery,8 making it more difficult to visualize or distinguish from bronchial lymph nodes.

A primary advantage of spiral CT is that it images many important chest structures and other potential diagnoses can be assessed. In many hospitals, spiral CT is more readily available than VQ scanning on a 24-hour basis. Finally, CT is cost competitive with other modalities; the charge for a spiral CT is often less than that for a VQ scan.

Several studies have compared spiral CT with angiography, and most have shown an overall sensitivity in the 85-100% range with a specificity in the 78-100% range.9-11 Many reports have also shown spiral CT to have a much higher sensitivity for central clot than for peripheral clot. In one recent study, the overall sensitivity was 86% but when subsegmental emboli were included, the sensitivity fell to 63%.11 However, the clinical significance of isolated subsegmental emboli is not at all clear. Isolated subsegmental emboli occurred in only 5.6% of patients in the PIOPED study and the inter-observer agreement for detecting them on pulmonary angiography was only 66%.2 More importantly, it may be that small peripheral emboli in patients with good cardiopulmonary reserve are of little consequence when there is no evidence of proximal DVT in serial ultrasound studies of the lower extremities. This important concept of assessing for the risk of recurrent emboli, in addition to looking for evidence of significant emboli in the lungs, is the rationale behind the success of outcome studies that have tested the strategy of withholding anticoagulation in patients with a nondiagnostic VQ scan and a negative bilateral lower extremity ultrasound.12,13 We are optimistic that large scale outcome studies combining spiral CT with ultrasound may show this strategy to be equally or more effective than algorithms using VQ scanning.

There are potential disadvantages to spiral CT that must be considered. First, for patients who are allergic to intravenous contrast or who have renal insufficiency, spiral CT may not be the test of choice. Second, a spiral CT study to evaluate PE requires interpretation by an experienced radiologist. Third, no outcome studies have been published to date that have evaluated a testing strategy using CT alone or in combination with ultrasound without VQ scanning. One study used spiral CT instead of pulmonary angiography in 164 patients in an algorithm where the patients had already had an intermediate probability VQ scan and a negative bilateral duplex ultrasound.14 In the 112 patients who had a negative CT, anticoagulation was withheld and the incidence of venous thromboembolism was 5.4% at three months follow-up. There is a pressing need for large, prospective studies to determine the outcome of patients in whom anticoagulation is withheld after a negative CT scan (with or without other studies such as ultrasound or D-dimer testing), but primarily replacing VQ scans early in the algorithm.

Magnetic Resonance Angiography (MRA)

Nonangiographic MRI has the ability to demonstrate PE as intravascular filling defects on cross-sectional images. However, early results were poor due to interpretation problems including those caused by respiratory-motion artifact, lack of contrast between in-plane blood flow and an embolus, and signal intensity loss peripherally. Fortunately, there have been several key innovations in MR hardware. It is now possible to produce fast, three-dimensional images, which, when combined with dynamic gadolinium enhancement, allow high-resolution angiography to be performed in a single arrested breath.

Meaney and associates reported the results of a small prospective trial that compared gadolinium-enhanced, three-dimensional magnetic resonance angiography (MRA) to the gold standard of conventional pulmonary angiography in 30 patients with suspected PE.15 Enrolled patients received both studies, which were reviewed independently by three radiologists blinded to the study hypothesis. The diagnostic criterion for PE was either the presence of an intravascular filling defect or a trailing embolus sign. In eight patients with emboli proven by pulmonary angiography, all five lobar and 16 of 17 segmental emboli were identified by the MRA technique. The three sets of readings demonstrated sensitivities between 75% and 100% and specificities between 95% and 100%.

More recently, Gupta and colleagues demonstrated less positive results than those of Meaney et al comparing standard pulmonary digital subtraction angiography to three-dimensional MRA in diagnosing PE.16 Though the studies were designed similarly, differences in sensitivities and specificities occurred because the Gupta study attempted to extend the diagnostic ability of the MRA technique to detect subsegmental emboli. The results reaffirmed the accuracy of three-dimensional pulmonary MRA to depict lobar and segmental emboli, but also demonstrated the technique was unable to depict 80% of subsegmental emboli.

The controversy regarding the clinical significance of subsegmental emboli is the same for MRA as for CT scan, and has been discussed above. Certainly, continued advances in pulmonary MRA techniques will improve detection of subsegmental emboli. However, what to do with this information clinically will depend on large, prospective, clinical outcome trials.

Visualizing clots in the pulmonary vasculature using MRA has a few significant advantages over CT. First, with the use of non-iodinated contrast agents such as gadolinium, which are not nephrotoxic, MRA can be utilized in patients with iodinated contrast allergy and renal insufficiency. Second, MRA avoids ionizing radiation and can be safely used in pregnant patients. Third, MR venography of the legs and pelvis for evaluation of DVT can be combined with pulmonary vascular imaging to provide an assessment of clot burden at several sites within a single examination in under one hour.

There are several disadvantages to MRA. MR scanning takes somewhat longer than CT scanning, and poses some difficulties due to remote patient monitoring, especially for acutely ill and unstable patients. Also, the ability to scan patients with morbid obesity, claustrophobia, or metallic implants is limited. The cost-effectiveness of MRA for the diagnosis of PE remains unknown. At many institutions, MRA is associated with significantly higher charges than for either CT of VQ scan. MRA is not widely available and requires expertise that is not routinely available 24 hours/day. Finally, there are no prospective outcome trials to evaluate a clinical strategy where patients with negative MRAs are sent home without anticoagulation.


It is very likely that our approach to the diagnosis of PE will change substantially in the next five years. We are optimistic that spiral CT has the potential to replace VQ scanning as the initial diagnostic test in many PE algorithms. MRI appears to have equal or greater potential as an effective test, especially by providing information about clot in both the lower extremities and the lungs. However, the cost, lack of availability, and expertise required may prevent MRA from assuming a primary role in most EDs. Rapid ELISA D-dimer testing has already been introduced in some ED settings and may be used cautiously in patients with a low pretest probability for thromboembolic disease, or in other patients in combination with additional noninvasive tests. (Dr. Schears is Assistant Professor of Emergency Medicine, Department of Emergency Medicine, Hospital of the University of Pennsylvania.)


1. Dalen JE, et al. Natural history of PE. Prog Cardiovasc Dis 1975;17:257-270.

2. Prospective Investigation of Pulmonary Embolism Diagnosis Investigators. Value of the ventilation-perfusion scan in acute pulmonary embolism: Results of the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). JAMA 1990;263:2753-2759.

3. Bounameaux H, et al. D-dimer testing in suspected venous thromboembolism: An update. Q J Med 1997;90:437-442.

4. Ginsburg JS, et al. D-dimer in patients with clinically suspected pulmonary embolism. Chest 1993;104:1679-1684.

5. Goldhaber SZ, et al. Quantative plasma D-dimer levels among patients undergoing pulmonary angiography for suspected pulmonary embolism. JAMA 1993;270:2819-2822.

6. Ginsburg JS, et al. Sensitivity and specificity for a rapid whole-blood assay for D-dimer in the diagnosis of pulmonary embolism. Ann Intern Med 1998;128:1006-1011.

7. Perrier A, et al. Non-invasive diagnosis of venous thromboembolism in outpatients. Lancet 1999;353:190-195.

8. Holbert JM, et al. Role of spiral computed tomography in the diagnosis of pulmonary embolism in the emergency department. Ann Emerg Med 1999;33:520-528.

9. Remy-Jardin M, et al. Diagnosis of pulmonary embolism with spiral CT: Comparison with pulmonary angiography and scintigraphy. Radiology 1996;200:699-706.

10. Goodman LR, et al. Detection of pulmonary embolism in patients with unresolved clinical and scintigraphy diagnosis: Helical CT versus angiography. Am J Roentgenol 1995;164:1369-1374.

11. Mayo JR, et al. Pulmonary embolism: Prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 1997;205:447-452.

12. Hull RD, et al. A noninvasive strategy for the treatment of patients with suspected pulmonary embolism. Arch Intern Med 1994;154:289-297.

13. Wells PS, et al. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med 1998;128:997-1005.

14. Ferretti GR, et al. Acute pulmonary embolism: Role of helical CT in 164 patients with intermediate probability at ventilation-perfusion scintigraphy and normal results at duplex US of the legs. Radiology 1997;205:453-458.

15. Meaney JF, et al. Diagnosis of pulmonary embolism with magnetic resonance angiography. N Engl J Med 1997;336:1422-1427.

16. Gupta A, et al. Acute pulmonary embolism: Diagnosis with MR angiography. Radiology 1999;210:353-359.

All of the following are true concerning the diagnosis of PE, except:

a. a positive D-dimer level is a very specific test with a high positive predictive value to rule-in PE.

b. both spiral CT and MRI have a much higher sensitivity for central clot than for subsegmental (peripheral) clot.

c. the clinical significance of isolated subsegmental emboli is currently debated.

d. one advantage of MRI is the ability to look for clot in both the lungs and in the lower extremities in a single examination.

Which of the following is a correct pairing of diagnostic test with its related significant limitation in the diagnosis of pulmonary embolism?

a. MRA and high fetal risk

b. spiral CT and risk of nephrotoxicity from intravenous contrast

c. D-dimer assay and patient consent

d. spiral CT and prolonged scanning time