Diagnostic yield of CT thorax angiography in patients suspected of pulmonary embolism: independent predictors and protocol adherence

OBJECTIVES: To determine the diagnostic yield of computed tomography scanning of the pulmonary arteries (CTPA) in our centre and factors associated with it. Differences between specialties as well as adherence to protocol were investigated. METHODS: All patients receiving a first CTPA for pulmonary embolism (PE) in 2010 were included. Data about relevant clinical information and the requesting specialty were retrospectively obtained. Differences in diagnostic yield were tested using a chi-squared test. Independent predictors were identified with multivariate logistic regression. RESULTS: PE on CTPA was found in 224 of the 974 patients (23 %). Between specialties, diagnostic yield varied from 19.5 to 23.9 % (p = 0.20). Independent predictors of diagnostic yield were: age, sex, D-dimer, cough, dyspnea, cardiac history, chronic obstructive pulmonary disease (COPD), atelectasis/consolidation, intrapulmonary mass and/or interstitial pulmonary disease on CT. Wells scores were poorly documented (n = 127, 13.0 %). Poor adherence to protocol was also shown by a high amount of unnecessary D-dimer values with a high Wells-score (35 of 58; 58.6 %). CONCLUSIONS: The diagnostic yield of CTPA in this study was relatively high in comparison with other studies (6.7–31 %). Better adherence to protocol might improve the diagnostic yield further. A prospective study could confirm the independent predictors found in this study. TEACHING POINTS: • Pulmonary embolism is potentially life-threatening and requires quick and reliable diagnosis. • Computed tomography of the pulmonary arteries (CTPA) provides this reliable diagnosis. • Several independent predictors of diagnostic yield of CTPA for pulmonary embolism were identified. • Diagnostic yield of CTPA did not differ between requesting specialties in our Hospital. • Better protocol adherence could improve the diagnostic yield of CTPA for pulmonary embolism.