Helical CT in the Diagnosis of Pulmonary Embolism
Abstract & Commentary
Synopsis: There are limited data supporting the use of helical CT in the diagnosis of pulmonary embolism.
Source: Rathbun SW, et al. Sensitivity and specificity of helical computed tomography in the diagnosis of pulmonary embolism: A systematic review. Ann Intern Med 2000; 132:227-232.
To determine the sensitivity and specificity of helical computed tomography (CT) for the diagnosis of pulmonary embolism and to determine the safety of withholding anticoagulant therapy in those with helical CTs negative for pulmonary embolism, this review paper searched the MEDLINE database for prospective studies evaluating the use of helical CT for the diagnosis of pulmonary embolism. Inclusion criteria were the study of consecutive patients suspected of pulmonary embolism with and without the disease, a broad spectrum of patient characteristics, performance of helical CT and pulmonary angiography or another reference test in all patients, and independent interpretation of the CT scan and pulmonary angiogram or other reference test. While no study met all criteria for adequate evaluation of sensitivity and specificity of helical CT, the studies reviewed showed sensitivity ranging from 53-100% and specificity ranging from 81-100%. In no study were patients with negative helical CT scans who had anticoagulant therapy withheld tested for venous thromboembolism.
Rathbun and associates drew four conclusions from their review:
1. No study on helical CT for detection of pulmonary embolism published before October 1999 met all methodological criteria for adequate evaluation of sensitivity and specificity of helical CT.
2. There is a wide range of sensitivity (53-100%) for helical CT detection of pulmonary embolism.
3. Several studies suggest considerable interobserver variation in the sensitivity and specificity of helical CT detection of pulmonary embolism.
4. No study on helical CT for detection of pulmonary embolism published before October 1999 adequately evaluates the safety of withholding anticoagulant therapy in patients with negative helical CT studies.
Comment by Jeffrey S. Klein, MD
This study illustrates that despite strong support in the radiology literature for its use, helical CT remains at a relatively early stage in its development as a tool for the detection of pulmonary emboli as compared to ventilation perfusion scintigraphy (V/Q scan) and pulmonary angiography. Despite the inherent limitations in V/Q scans and significant inter- and intraobserver variability in its interpretation, there exist, in combination with evaluation of lower extremities for deep venous thrombosis with ultrasound or impedance plethysmography, good outcome data for these exams while similar data for helical CT remain limited.
It should be noted that although this paper reviewed studies published up to October 1999, there is a study published in July 1999 (Lomis NN, et al. J Vasc Interv Radiol 1999;10:707-712) reporting encouraging outcome data for patients with negative helical CT scans who were followed without anticoagulation. As multidetector row helical CT becomes more widely available, and visualization of first-order subsegmental pulmonary arteries becomes the norm, the issue of sensitivity and specificity of helical CT as compared to pulmonary angiography (itself an imperfect test that is infrequently used even when indicated) should eventually be put to rest. In addition, outcome data should eventually support the efficacy of helical CT pulmonary angiography, likely in combination with simultaneous CT venography of the lower extremities or ultrasound examination of the lower extremities, in the evaluation of venous thromboembolic disease. However, until such data become available, use of helical CT by pulmonary medicine physicians at least at some institutions will remain limited to selected cases where V/Q scans are unavailable or nondiagnostic and pulmonary angiography cannot be performed.
The published sensitivity of helical CT for pulmonary embolism is:
a. less than 30%.
b. less than 50%.