Selected topics: emergency radiology
Computed tomography scan versus ventilation-perfusion lung scan in the detection of pulmonary embolism1

https://doi.org/10.1016/S0736-4679(01)00359-6Get rights and content

Abstract

This study compared the sensitivity and specificity of computed tomography (CT) scan and ventilation-perfusion (V-P) scan in detecting pulmonary embolism (PE) with pulmonary angiogram (AG) as the reference standard. Following a comprehensive search of the indexed medical literature, CT scan studies related to PE diagnosis were systematically evaluated to select those using AG as the reference standard and meeting specified methodologic criteria. Studies were further grouped by those reporting results for central PE findings only versus central and peripheral PE combined. A composite analysis of data derived from seven selected publications yielded sensitivity and specificity estimates for CT scan in detecting PE, which were statistically compared to the published results of a multi-center study reporting the sensitivity and specificity of the V-P scan with pulmonary AG as the reference standard. The calculated CT scan sensitivity was 77% for central PE only data and 81% for central and peripheral PE combined data, and the CT scan specificity was 91% and 98%, respectively. High-probability V-P scan sensitivity was 41% and specificity 97%; high- and intermediate-probability V-P scans combined yielded sensitivity 83% and specificity 52%. The sensitivity for PE detection was significantly greater for CT scan than for high-probability V-P scan; CT scan sensitivity was equivalent to V-P when high- and intermediate-probability scans were considered together. CT scan specificity for central and peripheral PE combined was equivalent to that of the high-probability V-P scan, but significantly greater than that of high- and intermediate-probability V-P scans considered together. Considering that only a small proportion of patients with suspected PE yield high-probability V-P scan results (which are usually indicative of PE), while as many as one-half of patients may yield intermediate-probability results (which are commonly not useful in PE diagnosis), our results suggest the CT scan may be an appropriate study for use by Emergency Physicians in the clinical evaluation of suspected PE.

Introduction

Pulmonary embolism (PE) is an important cause of patient morbidity and mortality, with an estimated 50,000–120,000 patient deaths per year 1, 2, 3, 4. The mortality rate reaches 30% for untreated PE; however, with treatment the mortality rate decreases to 2.5–10% 2, 5, 6. Classic symptoms for PE including pleuritic chest pain, hemoptysis, dyspnea, and tachypnea are present in many diseases of the chest resulting in a diagnostic challenge 7, 8, 9, 10. Because clinical presentations tend to be nonspecific and are so varied, diagnostic testing has a crucial role in the diagnosis of PE (11).

The ventilation-perfusion (V-P) scan has been widely used as the first-line radiologic examination in the diagnosis of suspected PE. In the mid-1980s, the multi-center Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study evaluated the diagnostic utility of the V-P scan for 731 patients with conclusive pulmonary angiogram (AG) results for PE (251 positive, 480 negative; Reference 12). PIOPED study findings revealed the utility of the V-P scan to be limited as summarized by the following: 1) high-probability results usually indicated PE (n = 116 with all but 14 positive for PE on AG), but only a small minority of study patients (16%) yielded high-probability results (sensitivity 41%, specificity 97%); 2) intermediate-probability results were found in 44% of study patients (n = 322) and were not helpful in establishing the diagnosis of PE (sensitivity 42%, specificity 45%); 3) only 5 of 55 patients with near-normal or normal V-P scans were positive for PE on AG (9%), making the diagnosis of PE very unlikely for near-normal or normal results (negative predictive value 98%), especially in the setting of low clinical likelihood for PE; and 4) the incidence of PE was ∼16% in the 238 patients with low-probability results (negative predictive value 84%; Reference 12).

The PIOPED investigators further reported the negative predictive value at 95% when low-probability V-P scan was considered in conjunction with low clinical likelihood; the clinician’s assessment of PE likelihood was recorded before the scan (described as prior probability), but how such was determined by clinicians was not explained (12). However, other studies have questioned the utility of low-probability V-P scans in patients with concomitant clinical problems such as cardiopulmonary disease, abnormal chest X-ray studies, and inadequate cardiopulmonary reserve 13, 14, 15, 16, 17, 18, 19, 20, 21. The PIOPED investigators also reported 25–30% differences among expert observers in classifying low- versus intermediate-probability V-P scans.

The literature indicates the diagnostic utility of the V-P scan to be limited to the small proportion of patients with high-probability results or to those with normal/near-normal or low-probability scans in conjunction with low clinical likelihood. Unfortunately, this leaves a large group of patients with intermediate-probability results for whom the V-P scan is not diagnostically useful. Nevertheless, the risk of missed PE may result in death; alternatively, treating patients unnecessarily for PE exposes them to the risk of life-threatening complications from anticoagulation, which have been reported to be 5% at 1 year (22). Therefore, an alternative to the V-P scan would be helpful for Emergency Physicians faced with cases of possible or suspected PE.

Computed tomography (CT) lung scanning has been used clinically and studied for evaluation of suspected PE. Supporters of the CT lung scan argue that it is faster and more accurate than the V-P scan and often provides an alternative diagnosis 23, 24. Critics argue that a CT scan may miss small subsegmental emboli, that it exposes the patient to an i.v. contrast dye load, and there is insufficient support in the medical literature for its use (25). The accuracy of CT scan readings based on training and experience of radiologists also has been questioned 26, 27.

We are not aware of a randomized, prospective clinical trial that compares CT lung scan and V-P lung scan with pulmonary AG as the reference standard. Pulmonary AG is the gold standard for detection of PE with both sensitivity and specificity >95%; however, the invasiveness of AG and the morbidity (6%) and mortality (0.5%) have limited its use 28, 29, 30. Consequently, some studies have compared CT scan to V-P scan results (i.e., high-probability V-P scan as the reference standard), others have compared CT scan to clinical outcome or autopsy results, and still others have provided results for only a subset of study patients undergoing AG (e.g., when the V-P scan is equivocal). Much of the clinical research on this topic varies in other respects, such as patient selection criteria, whether readers are blinded to comparison study results, and whether peripheral as well as central PE detection is addressed (31).

The objective of our study was to compare the utility of the CT scan and V-P scan in the detection of PE. We systematically searched the indexed literature for studies comparing CT lung scans to pulmonary AG that could be used as a basis for comparison to the PIOPED study data. Our aim was to perform a meta-analysis, calculating sensitivity, specificity, and accuracy of the CT scan and to compare these results to those reported for the V-P scan in the PIOPED study.

Section snippets

Literature search

Studies related to the detection of PE by use of the CT scan were identified via computerized search of English-language medical literature via PubMED; search criteria included the Medical Subject Headings pulmonary embolism, computed tomography (CT), and ventilation perfusion scan. Helical and spiral CT scan descriptors were equally considered in this search. The search was conducted for studies published between January 1980 and June 2000 and were supplemented by individually searching

Results

The full review of 18 publications resulted in initial exclusion of four based on retrospective methodology, for studying only those patients with negative CT scan findings, or for failing to include AG as the reference standard 33, 34, 35, 36. Three additional studies were excluded for failure to blind test readers and to employ AG as the primary reference standard 37, 38, 39. No studies were excluded based on criterion #5; studies using the descriptor spiral CT scan, helical CT scan, and

Discussion

Our literature search revealed a number of publications related to the use of CT scan in the detection of PE, but an absence of a prospective, randomized clinical trial comparing CT scan to V-P scan. Also, relatively few studies employed pulmonary AG as the reference standard for most or all patients; several studies instead employed AG only for those with negative CT scan or discordant CT scan and V-P scan results, which introduces selection bias. Other studies were found to use surrogate

Conclusion

The current standard of practice using the V-P scan for the initial evaluation of PE leaves much to be desired. The literature indicates limited utility of the V-P scan to those patients with high-probability results and normal or low-probability scans in conjunction with low clinical likelihood for PE. Unfortunately, these patients comprised less than 30% in the PIOPED study, pointing to the problem frequently encountered by Emergency Physicians faced with many patients presenting with

Acknowledgements

We thank Melissa Schlenker, Senior Research Assistant at York Hospital, for her assistance in preparing this manuscript.

References (72)

  • T.J Russi et al.

    Clinical utility of computed tomography in the diagnosis of pulmonary embolism

    Clinical Imaging

    (1997)
  • P.B Fontanarosa

    Radiologic contrast-induced renal failure

    Emerg Med Clin North Am

    (1988)
  • C.I Henschke et al.

    Changing practice patterns in the workup of pulmonary embolism

    Chest

    (1995)
  • J Dalen et al.

    Natural history of pulmonary embolism

    Prog Cardiovasc Disease

    (1975)
  • K.M Moser

    Pulmonary embolism

  • Prevention of venous thrombosis and pulmonary embolism

    JAMA

    (1986)
  • J.L Carson et al.

    The clinical course of pulmonary embolism

    N Engl J Med

    (1992)
  • H Daisley

    Pulmonary embolism as a case of death

    West Indian Med J

    (1990)
  • P.M Kemp et al.

    Pulmonary embolismis the clinical history a useful adjunct to aid the interpretation of the equivocal lung scan?

    Clin Nucl Med

    (1996)
  • R.D Hull et al.

    Pulmonary embolism in outpatients with pleuritic chest pain

    Arch Intern Med

    (1988)
  • Value of the ventilation/perfusion scan in acute pulmonary embolismresults of the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED)

    JAMA

    (1990)
  • R.D Hull et al.

    The low-probability lung scan

    Arch Intern Med

    (1995)
  • D.F Worsley et al.

    A detailed evaluation of patients with acute pulmonary embolism and low- or very-low-probability lung scan interpretations

    Arch Intern Med

    (1994)
  • S.N Goldberg et al.

    Pleural effusion and ventilation/perfusion scan interpretation for acute pulmonary embolus

    J Nucl Med

    (1996)
  • R Smith et al.

    Clinical outcomes of patients with suspected pulmonary embolism and low-probability aerosol-perfusion scintigrams

    Radiology

    (1987)
  • D Kahn et al.

    Clinical outcome of patients with a “low probability” of pulmonary embolism on ventilation-perfusion lung scan

    Arch Intern Med

    (1989)
  • J.G Rajendran et al.

    Review of 6 month mortality follwing low-probability lung scans

    Arch Intern Med

    (1999)
  • H.W Gray et al.

    Pulmonary embolism exclusiona practical approach to low probability using the PIOPED data

    Eur J Nucl Med

    (1998)
  • G.W Petty et al.

    Complications of long term anticoagulation

    Ann Neurol

    (1988)
  • P.D Stein et al.

    The role of newer diagnostic techniques in the diagnosis of pulmonary embolism

    Curr Opin Pulm Med

    (1999)
  • E.E Coche et al.

    Acute pulmonary embolismancillary findings at spiral CT

    Radiology

    (1998)
  • Opinions regarding the diagnosis and management of venous thromboembolic disease

    Chest

    (1998)
  • E.A Drucker et al.

    Acute pulmonary embolismassessment of helical CT for diagnosis

    Radiology

    (1998)
  • S.M Bates et al.

    Helical computed tomography and the diagnosis of pulmonary embolism

    Ann Intern Med

    (2000)
  • H Blachere et al.

    Pulmonary embolism revealed on helical CT angiographycomparison with ventilation-perfusion radionuclide lung scanning

    AJR

    (2000)
  • P.D Stein et al.

    Complications and validity of pulmonary angiography in acute pulmonary embolism

    Circulation

    (1992)
  • Cited by (0)

    1

    Selected Topics: Emergency Radiology is coordinated by Jack Keene, md, of Emergency Treatment Associates, Poughskeepsie, New York

    View full text