Nine years after it was first described, the use of CT to evaluate pulmonary embolism (PE) continues to generate both enthusiasm and controversy. There is tremendous appeal to an imaging modality that permits actual visualization of an embolus, unlike the more traditional imaging techniques for PE. Publications and abstracts have assessed several thousand PE patients imaged with CT and have found it promising for evaluating segmental and larger PE.^1-8 Technical advances, particularly the advent of multidetector-row CT (MDCT) technology, will greatly affect the role of CT in PE imaging.

Despite this work, however, no consensus has been reached on the overall accuracy of CT, a question that affects its cost-effectiveness.^9,10 Questions also remain about other aspects of the use of CT in PE, from appropriate patient selection to the clinical significance of a negative CT examination.

The advantages and shortcomings of conventional imaging for thromboembolic disease (chest x-ray, ventilation/ perfusion [V/Q] scan, lower extremity Doppler ultrasound, and pulmonary angiography) are well known. Chest x-ray findings, although usually abnormal, are most often nonspecific, and the role of chest radiography is largely to guide performance and interpretation of the V/Q scan. Nuclear scanning is highly sensitive for PE (98%) when anything other than a normal perfusion scan is considered; in this case, specificity is very low (~20%). Using high-probability scans alone to predict PE results in a sensitivity of only 41%.¹¹

Patients with intermediate-probability V/Q (and low-probability V/Q in the presence of continued clinical suspicion) require further imaging, such as lower extremity Doppler ultrasound or pulmonary angiography. It is believed that deep venous thrombosis of the lower extremities and pelvis is responsible for more than 90% of pulmonary emboli, but lower extremity ultrasound is positive in no more than half of patients with proven PE.

Pulmonary angiography, with a sensitivity of at least 95%, is still the gold standard but is used in only 15% of patients with an unresolved question of PE.¹² The risks of inappropriately treating-or not treating-thromboembolic disease are substantial.

Historical Development

Initial studies of helical CT for PE focused on the accuracy of helical CT compared with pulmonary angiography as the gold standard.^1,2 The sensitivity of helical CT for central emboli (segmental vessels and larger) was found to be 86% to 100%, but limitations in evaluating subsegmental clots were quickly recognized. A study conducted in 1995, for example, found a 25% sensitivity.² Subsequent work examined the relative merits of helical CT and V/Q scanning as the initial test for PE.^3,5 In these reports, CT continued to be accurate with respect to pulmonary angiography, and it was of particular value in diagnosing patients who had low- or intermediate-probability V/Q scans.

Another identified strength of helical CT relative to V/Q was the higher interobserver agreement noted for helical CT (k = 0.72 to 0.85) than for V/Q (k = 0.22 to 0.61).^5,13 A feature that has undoubtedly encouraged widespread use of CT rather than V/Q as an initial study is the high frequency of identifying an alternate diagnosis in up to 67% of patients who do not have PE.^14,15

Not every investigation, however, has been uniformly enthusiastic about CT.⁶ In comparing the results of various studies, it is essential to ascertain that optimal technique has been used in both performance and interpretation of the CT study.

Technique

If using a single-detector-row scanner, thin collimation of 2 to 3 mm with pitch 1.7 to 2 generally provides adequate resolution for detection of segmental and larger emboli. Dyspneic patients may be scanned using slightly thicker collimation to permit shorter breath-hold duration, but this will decrease detection of small emboli. Typical parameters for MDCT include a collimation of 1 to 2.5 mm and pitch 6. The volume of interest extends from the dome of the diaphragm (inferior pulmonary veins) up to the aortic arch, or approximately 10 to 12 cm along the z-axis. Patients are coached to hyperventilate two or three times before holding their breath in inspiration as the x-rays begin.

As most emboli go to the lower lobes, patients are scanned in a caudal-to-cranial direction, which helps ensure that dyspneic patients can hold their breath for at least the initial, lower lobe images. An injection rate of at least 3 cc/second should be used with 100 to 140 cc of full-strength nonionic contrast; many centers use higher flow rates.¹⁶ At higher injection rates, more dilute contrast may be used to avoid streak artifacts in the central veins. Contrast should be injected through an antecubital or more proximal vein.

The use of a standard injection delay is feasible in most cases; the value depends on flow rate but is usually between 15 and 20 seconds. A timing bolus may be of value if there is known severe pulmonary hypertension or circulatory delay. Images are reconstructed using overlapping intervals of 1 to 2 mm, unless 1-mm collimation is used. If cases are being filmed, those that are not clearly positive on film review must be viewed on a workstation to maximize accuracy.¹⁷

CT Interpretation

Thorough familiarity with pulmonary vascular anatomy is required to accurately interpret PE studies. Ideally, each of 20 segmental arteries, two pairs of which are generally fused, should be identified and contrast enhanced. The CT findings of acute PE are equivalent to angiographic findings: an intravascular filling defect associated with either complete vascular occlusion or tram-track streaming of contrast past a nonocclusive embolus (Figure 1). Multiple emboli are frequently seen in positive cases.

Chronic PE can be identified as peripheral thrombus with vascular recanalization (Figure 2). Standard mediastinal window and level settings often limit visualization of small thrombi within dense contrast; a wider window setting (~700) is helpful in this regard. In addition, it is important to be able to switch between the embolism window and the lung window, both to distinguish the segmental arteries from veins and to evaluate lung parenchymal findings such as infarcts that may raise a suspicion of embolism in that vascular territory.

Although it has not yet proved to be of general value, multiplanar reconstruction may occasionally aid in the definitive identification of an intravascular versus extravascular location of a possible filling defect (Figure 3).

Pitfalls

Evaluating PE CT involves a number of pitfalls. False-positive readings may result from hilar lymph nodes, hypodense vessel bifurcations, perivascular edema, or mucus-filled bronchi or hypodense pulmonary veins that are mistaken for arteries (Figure 4).

False-negative readings may result from technically suboptimal studies, including patient or respiratory motion, inadequate contrast density, image noise, or any other cause of inability to identify each segmental artery. This occurs in 5% to 10% of cases.¹⁸ False-negatives are also caused by failure to identify subsegmental emboli.

The question of subsegmental emboli deserves further attention, as it has been a focal point of controversy regarding CT in PE. The frequency of isolated subsegmental emboli varies among series, ranging from 6% to 36%.^11,19 The consequences of PE depend in part on the degree of pulmonary vascular occlusion, which is naturally less for isolated subsegmental clots than for those that are more central. But each patient’s cardiopulmonary status also plays a major role in determining the sequelae of emboli. Thus, patients with normal cardiopulmonary reserve may do well without treatment for subsegmental clots,²⁰ while patients with congestive heart failure or chronic obstructive pulmonary disease may require pulmonary angiography to completely exclude a small PE even after diagnostic-quality, negative CT and Doppler ultrasound.

Early experience with MDCT shows great promise for improved sensitivity for subsegmental emboli. MDCT permits true collimation as low as 1 to 1.25 mm, improving spatial resolution and reducing partial-volume effects. Research presented at the 2000 RSNA meeting by Ghaye and colleagues showed that subsegmental arteries were adequately depicted in 94% of patients with MDCT when 1.25-mm collimation was used and in only 62% of patients when single-slice helical CT with 2-mm collimation was used.²¹

At the same meeting, Schoepf and colleagues analyzed patients with proven PE and found that 1-mm collimation improved the demonstration of subsegmental emboli by 40% compared with 3-mm collimation, and by 14% compared with 2-mm collimation.²² Although large studies with MDCT have yet to be published, a recent report using dual-section CT showed an overall sensitivity of 90% compared with pulmonary arteriography, including the 6% of patients who had isolated subsegmental PE.²³ Among other attributes, MDCT provides more nearly isotropic voxels, which might validate the utility of multiplanar reformats for PE (Figure 5).²⁴

Continued Controversy

Despite nearly a decade of research, controversy remains regarding the overall role of CT in the workup of possible PE, particularly the prognostic significance of a negative CT. Some of this uncertainty relates to the wide range of sensitivities in published studies, which have varied from 53% to 100%.^1,6

Perhaps more realistically, a recent meta-analysis of 11 single-detector helical CT studies conducted between 1992 and 1999 showed a narrower range of 74% to 81% sensitivity for segmental and larger clots.²⁵ As the authors point out, most of those studies used 5-mm collimation; it is apparent that thinner collimation, especially using MDCT, will improve accuracy.

Data are also emerging regarding the clinical outcomes of patients with negative CT scans who do not receive anticoagulation. Three such studies showed subsequent PE in five of 388 patients, or 1.2%, compared with eight of 294, or 2.7%, whose original V/Q was very low or low probability and who did not receive anticoagulation.^4,14,26 At the Mayo Clinic, 951 patients had an incidence of PE after negative CT of approximately 1.2% at three months’ follow-up, a value not dissimilar to that after negative pulmonary arteriography.¹⁸

Opinions vary regarding imaging algorithms for PE and the role of CT. Factors to consider include the patient’s cardiopulmonary status, allergy to contrast, and breath-hold capability; the availability of the procedure, including technical and interpretive expertise; and cost. Various centers use CT for all patients, for inpatients, or only in cases of equivocal V/Q scans. It is clear that, whichever CT technology is being used, a negative CT will never conclusively exclude the possibility of thromboembolic disease. No matter where it is positioned in the algorithm, CT must remain one of several tests used in the evaluation of these patients.²⁷

Evaluation for deep vein thrombosis (DVT) is essential in all patients with a negative CT, and much work has focused on the combination of pulmonary CT arteriography and lower extremity CT venography.²⁸ Patients at our hospital with signs and symptoms of both PE and DVT initially undergo duplex ultrasound. In patients without leg symptoms, we use the chest x-ray as a primary factor in determining which lung imaging test to use first (Figure 6). Patients with nontrivial abnormalities on the chest x-ray, who are more likely to have nondiagnostic V/Q scans,²⁹ undergo CT. If the CT is negative, lower extremity ultrasound is indicated. Patients with essentially normal chest x-rays undergo V/Q scanning; those with intermediate- or low-probability results with continued clinical suspicion undergo CT. Pulmonary angiography retains a role in equivocal cases and for those in whom the detection of small emboli is essential.

References

1. Remy-Jardin M, Remy J, Wattinne L, Giraud F. Central pulmonary thromboembolism: diagnosis with spiral volumetric CT with the single-breath-hold technique-comparison with pulmonary angiography. Radiology 1992;185:381-387.
2. Goodman LR, Curtin JJ, Mewissen MW, et al. Detection of pulmonary embolism in patients with unresolved clinical and scintigraphic diagnosis: helical CT versus angiography. AJR 1995;164:1369-1374.
3. Remy-Jardin M, Remy J, Deschildre F, et al. Diagnosis of pulmonary embolism with spiral CT: comparison with pulmonary angiography and scintigraphy. Radiology 1996;200:699-706.
4. Ferretti GR, Bosson JL, Buffaz PD, 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.
5. Mayo JR, Remy-Jardin M, Muller NL, et al. Pulmonary embolism: prospective comparison of spiral CT and ventilation-perfusion scintigraphy. Radiology 1997;205:447-452.
6. Drucker EA, Rivits SM, Shepard JO, et al. Acute pulmonary embolism: assessment of helical CT for diagnosis. Radiology 1998;209:235-241.
7. Herold CJ, Hahne J, Ghaye B, et al. Prospective evaluation of pulmonary embolism: diagnostic performance of spiral CT angiography in the ESTIPEP trial. Radiology 1999;213(P):126.
8. Van Strijen MJ, De Moyne W, Kieft GJ, Bloem JL. Diagnosis of pulmonary embolism with spiral CT: a prospective cohort study in 617 consecutive patients. Radiology 1999;213(P): 127.
9. Van Erkel AR, van Rossum AB, Bloem JL, et al. Spiral CT angiography for suspected pulmonary embolism: a cost-effectiveness analysis. Radiology 1996;201:29-36.
10. Paterson DI, Schwartzman K. Strategies incorporating spiral CT for the diagnosis of acute pulmonary embolism: a cost-effectiveness analysis. Chest 2001;119:1791-1800.
11. PIOPED 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.
12. Henschke CI, Mateescu I, Yankelevits DF. Changing practice patterns in the work-up of pulmonary embolism. Chest 1995;107:940-945.
13. Blachere H, Latabe V, Montaudon M, et al. Pulmonary embolism revealed on helical CT angiography: comparison with ventilation-perfusion radionuclide lung scanning. AJR 2000;174: 1041-1047.
14. Garg K, Slieler H, Welsh CH, et al. Clinical validity of helical CT being interpreted as negative for pulmonary embolism: implication for patient treatment. AJR 1999;172:1627-1631.
15. Kim KI, Muller NL, Mayo JR. Clinically suspected pulmonary embolism: utility of spiral CT. Radiology 1999;210:693-697.
16. Prokop M, Schaefer-Prokop CM, Hahne J, Herold CJ. Image quality and diagnostic accuracy of CTA for acute pulmonary embolism: comparison of various protocols for contrast administration. Radiology 1999;213(P):472.
17. Gosselin MV, Rubin GD, Leung AN, et al. Unsuspected pulmonary embolism: prospective detection on routine helical CT scans. Radiology 1998;208:208-215.
18. Ryu JH, Swensen SJ, Olson EJ, Pellikka PJ. Diagnosis of pulmonary embolism with use of computed tomographic angiography. Mayo Clin Proc 2001;76:59-65.
19. Oser RF, Zuckerman DA, Gutierrez FR, Brink JA. Anatomic distribution of pulmonary emboli at pulmonary angiography: implications for cross-sectional imaging. Radiology 1996;199:31-35.
20. Stein PD, Henry JW, Relyea B. Untreated patients with pulmonary embolism. Outcome, clinical, and laboratory assessment. Chest 1995;107:931-935.
21. Ghaye B, Szapiro D, Delannoy V, et al. How far does multidetector CT allow the analysis of peripheral pulmonary arteries? A comparative study with thin collimation single detector CT. Radiology 2000;217(P):508.
22. Schoepf UO, Holzknecht NG, Helmberger TK, et al. Segmental and subsegmental pulmonary embolism with thin-slice multi-detector array spiral CT (MDCT). Radiology 2000;217(P):509.
23. Qanadli SD, Hajjam ME, Mesurolle B, et al. Pulmonary embolism detection: prospective evaluation of dual-section helical CT versus selective pulmonary angiography in 157 patients. Radiology 2000;217: 447-455.
24. Simon M. Boiselle PM, Choi JR, et al. Paddle-wheel CT display of pulmonary arteries and other lung structures: a new imaging approach. AJR, 2001;177:195-198.
25. Harvey RT, Gefter WB, Hrung JM, Langlotz CP. Accuracy of CT angiography versus pulmonary angiography in the diagnosis of acute pulmonary embolism: evaluation of the literature with summary ROC curve analysis. Academic Radiology 2000;7:786-797.
26. Goodman LR, Lipchik RJ, Kuzo RS, et al. Subsequent pulmonary embolism: risk after a negative helical CT pulmonary angiogram-prospective comparison with scintigraphy. Radiology 2000;215:535-542.
27. Perrier A, Howarth N, Didier D, et al. Performance of helical computed tomography in unselected outpatients with suspected pulmonary embolism. Ann Intern Med 2001;135:88-97.
28. Loud PA, Katz DS, Bruce DA, et al. Deep venous thrombosis with suspected pulmonary embolism: detection with combined CT venography and pulmonary angiography. Radiology 2001;219:498-502.
29. Goldberg SN, Palmer EL, Scott JA, Fisher R. Pulmonary embolism: prediction of the usefulness of initial ventilation-perfusion scanning with chest radiographic findings. Radiology 1994;193:801-805.