Detection of pulmonary embolism comparison between dual energy CT and MR angiography in a rabbit model.

RATIONALE AND OBJECTIVES

The purpose of this study was to compare the efficacy of dual-energy computed tomography (DECT), time-resolved magnetic resonance (MR) imaging (MRI) perfusion measurements, and high-resolution MR angiography (MRA) for the detection of pulmonary embolism (PE) in a rabbit model.

MATERIALS AND METHODS

Two hours after Gelfoam (n = 16) or saline (n = 2) injection into the femoral vein, 18 rabbits were first imaged using dual-energy computed tomographic pulmonary angiography (CTPA), from which blood flow imaging (BFI) were produced. Next, the rabbits underwent time-resolved MR perfusion measurements and MRA using a 3.0-T scanner. Two radiologists who were blinded to histopathologic findings independently evaluated the results of CTPA, BFI, time-resolved MR perfusion, and MRA for each rabbit and recorded the locations and number of pulmonary clots on a per lobe basis. Immediately after MRI examination, pathologic determination of the locations and numbers of lung lobes with PE was recorded. Histopathologic results served as the reference standard to determine the sensitivity and specificity of DECT (BFI and CTPA) and MRI (time-resolved MR perfusion and high-resolution MRA) for the detection of PE.

RESULTS

The induction of PE was successful in all animals, but three rabbits were excluded because of death prior to MRI procedures (caused by complications during the embolization procedure); only data from the remaining 15 rabbits were included in the final analysis. Histopathology demonstrated 24 lobes with clots and 51 lobes without clots. CTPA, BFI, and MRI correctly identified PE in 23, 23, and 20 lobes, respectively, and the absence of emboli in 48, 46, and 46 lobes, respectively; these results corresponded to sensitivities of 95.8%, 95.8%, and 83.3% and specificities of 94.1%, 90.2%, and 90.2% for reader 1 and sensitivities of 91.7%, 91.7%, and 87.5% and specificities of 90.2%, 88.2%, and 88.2% for reader 2 for CTPA, BFI, and MRI, respectively. Good or excellent intermodality and interreader agreement among CTPA, BFI, and MRI were found.

CONCLUSIONS

DECT can simultaneously provide high-resolution pulmonary artery and lung iodine mapping with slightly high diagnostic accuracy for the detection of PE compared to MRI in an experimental rabbit model of PE. This improvement was not statistically significant given the study sample size.