Dual-energy CT angiography of the lung in patients with suspected pulmonary embolism: initial results

Abstract

To evaluate the feasibility of dual-energy CT angiography (CTA) of the lung in patients with suspected pulmonary embolism (PE). 24 patients with suspected PE were examined with a single-acquisition, dual-energy CTA protocol (A-system: 140 kV/65 mAsref, B-system: 80kV/190 mAsref) on a dual-source CT system. Lung perfusion was visualized by color-coding voxels containing iodine and air using dedicated dual-energy postprocessing software. Perfusion defects were classified by two blinded radiologists as being consistent or non-consistent with PE. Subjective image quality of perfusion maps and CTA was rated using a 5-point scale (1: excellent, 5: poor). The reading of a third independent radiologist served as the standard of reference for the diagnosis of PE. In all patients with PE (n=4), perfusion defects classified as being consistent with PE were identified in lung areas affected by PE. Both readers did not record perfusion defects classified as being consistent with PE in any of the patients without PE. Thus, on a per patient basis the sensitivity and specificity for the assessment of PE was 100% for both readers. On a per segment basis the sensitivity and specificity ranged between 60-66.7% and 99.5-99.8%. The interobserver agreement was good (k= 0.81). Perfusion defects rated as non-consistent with PE were most frequently caused by streak artifacts from dense contrast material in the great thoracic vessels. The median score of the image quality of both the perfusion maps and CTA was 2. In conclusion, dual-energy CTA of pulmonary embolism is feasible and allows the assessment of perfusion defects caused by pulmonary embolism. Further optimization of the injection protocol is required to reduce artifacts from dense contrast material.