Background: There is no noninvasive method to rule out pulmonary embolism when the clinical suspicion for pulmonary embolism is high. We did a prospective observational study to determine the negative predictive value of spiral computed tomography (CT) in this situation.

Methods: We performed spiral CT scans of the thorax in consecutive patients with high clinical suspicion of pulmonary embolism with intermediate or low probability ventilation-perfusion scans. Patients with negative or indeterminate spiral CT results had conventional angiography at the discretion of the attending physician. Only patients with positive spiral CT results or positive conventional angiograms were treated. All patients were observed for 6 months for evidence of venous thromboembolic disease. Clinical outcome without treatment or the results of conventional angiography were used as reference standards. False-negative results were defined as a negative spiral CT with a positive conventional angiogram or any diagnosis of venous thromboembolism within 6 months.

Results: Among the 103 patients who were studied, spiral CT scans were positive in 22 patients, indeterminate in 10 patients, and negative in 71 patients. Twenty-seven (26%) patients had pulmonary embolism by clinical outcome, including 3 of the 71 patients with negative spiral CT scans and 2 of the 10 patients with indeterminate scans. A negative spiral CT result had a likelihood ratio of 0.12 (95% confidence interval [CI]: 0.04 to 0.35) with a negative predictive value of 96% (95% CI: 88% to 99%). Using conventional angiography only as the reference standard, a negative spiral CT result had a likelihood ratio of 0.08 (95% CI: 0.02 to 0.31) and a negative predictive value of 93% (95% CI: 77% to 98%).

Conclusions: Spiral CT has a high negative predictive value for pulmonary embolism and may replace conventional angiography in the workup of pulmonary embolism. Patients with indeterminate spiral CT results should be considered for conventional angiography.