Risk-benefit analysis of pulmonary CT angiography in patients with suspected pulmonary embolus.

OBJECTIVE The objective of our study was to estimate the mortality benefit-to-risk ratio of pulmonary CT angiography (CTA) by setting (ambulatory [emergency department or outpatient] or inpatient), age, and sex. MATERIALS AND METHODS A retrospective evaluation of 1424 consecutive pulmonary CTA examinations was performed and the following information was recorded: examination setting, patient age, patient sex, pulmonary CTA interpretation for pulmonary embolus (PE), and CT radiation exposure (dose-length product). We estimated mortality benefit of pulmonary CTA by multiplying the rate of positive pulmonary CTA examinations by published estimates of mortality of untreated PE in ambulatory and inpatient settings. We estimated the lifetime attributable risk of cancer mortality due to radiation from pulmonary CTA by calculating the estimated effective dose and using sex-specific polynomial equations derived from the Biological Effects of Ionizing Radiation VII report. We calculated benefit-to-risk ratios by dividing the mortality benefit of preventing a fatal PE by the mortality risk of a radiation-induced cancer. RESULTS Pulmonary CTA diagnosed PE in 188 of 1424 patients (13.2%). Both inpatients (101/723, 14.0%) and emergency department patients (74/509, 14.5%) had significantly higher rates of PE than outpatients (13/192 [6.8%]). Males received significantly (p = 0.02451) higher radiation dose (9.7 mSv) than females (8.4 mSv), but males had a significantly (p < 0.0001) lower lifetime attributable risk of cancer mortality than females. Assuming an untreated PE mortality rate of 5% for ambulatory patients and 30% for inpatients, the benefit-to-risk ratio ranged from 25 for ambulatory patients to 187 for inpatients. Ambulatory women had the lowest benefit-to-risk ratio. CONCLUSION The benefit-to-risk ratio of pulmonary CTA in patients with suspected PE ranges from 25 to 187 and can be increased by optimizing the radiation dose.