Pulmonary embolism detection and characterization through quantitative iodine-based material decomposition images with spectral computed tomography imaging.

OBJECTIVES: To assess the diagnostic value of pulmonary embolism (PE) detection and characterization through quantitative iodine-based material decomposition images with spectral computed tomography (CT) imaging.
MATERIALS AND METHODS: Fifty-three patients underwent CT pulmonary angiography (CTPA) with spectral imaging mode with the simultaneous acquisition of 80 kVp and 140 kVp on a GE Discovery CT750HD scanner to generate monochromatic CTPA and material decomposition images. CTPA images were reviewed for the presence, localization, and degree (occlusive or nonocclusive) of PE. The iodine distribution in the lung parenchyma on the iodine-based material decomposition images was used to identify perfusion defects, which were then correlated to the CTPA findings. The iodine densities for the perfusion defects and the normal lung parenchyma were measured and statistically compared. Twelve PE patients underwent anticoagulation, and the iodine densities for the perfusion defects before and after the treatment were also measured and compared. The receiver operating characteristics curve was generated to assess the differential diagnostic performances of iodine density in distinguishing the presence or absence of PE and the occlusive or nonocclusive PE.
RESULTS: A total of 93 clots (51 occlusive and 42 nonocclusive) were found in 19 patients with lobar (26), segmental (54), or subsegmental (13) distribution. CTPA identified 88 clots initially and 5 more retrospectively with the help of iodine mapping. Thirty-three of 34 normal CTPA patients had symmetric iodine distribution. All occlusive clots and 11 nonocclusive clots showed clear evidence of iodine distribution defects. There was a significant difference for the iodine density among normal lung parenchyma (1.89 mg/mL [0.85-3.29 mg/mL]), nonocclusive perfusion defects (0.83 mg/mL [0.44-1.26 mg/mL]), and occlusive perfusion defects (0.27 mg/mL [0.00-0.62 mg/mL]) (P < 0.001). The iodine densities of perfusion defects before and after anticoagulation were significantly different (P < 0.001). Receiver operating characteristics analyses showed high discriminatory power for using the quantification of iodine density in distinguishing the presence or absence of PE and the occlusive or nonocclusive PE.
CONCLUSIONS: Spectral CT imaging generated both monochromatic CTPA images for morphologic analysis of PE and material decomposition images for quantitative depiction of pulmonary blood flow and perfusion defects. Quantification of iodine density may be used as a predictor in distinguishing the presence or absence of PE and the severity of PE.