Value of pulmonary perfusion/ventilation imaging and 64-slice computed tomographic pulmonary angiography for acute pulmonary embolism in rabbits

Luo Yan,Zhang Qing,Hu Yun,Su Hongyuan,Yu Feiying
DOI: https://doi.org/10.16016/j.1000-5404.2010.02.015
2010-01-01
Abstract:Objective To evaluate the sensibilities and specificities of diagnosis for acute pulmonary embolism with perfusion/ventilation imaging(V/Q) or 64-slice computer tomography pulmonary angiography(SCTPA).Methods Fourteen experimental rabbits were randomly divided into 2 groups,that is,experimental group(n=12),and control group(n=2).The rabbits of experimental group were further equally randomized into 4 subgroups,2-hour group,24-hour group,3-day group and 7-day group.The animal model of acute pulmonary embolism were prepared by jugular vena catheter using sponges as emboli.The rabbits of control group received an injection of normal saline via catheter.Pulmonary perfusion/ventilation imaging and pulmonary angiography were performed after the operation,then the rabbits were killed at the corresponding time points and the pulmonary tissue were pathological inspected to search and record the numbers and location of emboli.The control group rabbits were killed in 24 h and 3 d respectively for pathological observation.Results One rabbit of experimental group died during the operation,and the left 11 of 12 rabbits were made as acute pulmonary embolism model successfully.Totally 165 pulmonary arteries were anatomized after imaging diagnosis.Forty-three emboli were found and 39 pulmonary embolisms were made.V/Q imaging detected 36 segmental pulmonary arteries embolism,and another 10 embolisms were missed and 7 embolisms were false.SCTPA diagnosed 32 embolism arteries,another 13 embolisms were missed and 6 embolisms were misdiagnosed.The total sensibilities,specificities and accuracies were 74.4%,94.4% and 89.7% respectively for V/Q imaging,and 66.7%,95.2% and 88.5%,respectively for SCTPA in diagnosis of pulmonary artery embolism.Conclusion The sensitivity and accuracy for diagnosing pulmonary embolism with pulmonary perfusion/ventilation imaging are superior than 64-slices computer tomographic pulmonary angiography,but inferior in the specificity.However,there is no statistical difference.
What problem does this paper attempt to address?