Diagnosis Of Peripheral Pulmonary Embolism In Experimental And Clinical Research With Dual-Energy CT

PurposeTo evaluate the performance of dual-energy CT (DECT) based vascular iodine analysis for the detection of acute peripheral pulmonary embolism (PE) in a canine model with histopathological findings as the reference standard.Materials and MethodsThe study protocol was approved by our institutional animal committee. A total of16mongrel dogs weighing11to16kg were used for this study. The canines were randomly assigned to either an experimental group (n=12) or a control group (n=4). Anesthesia was induced with ketamine (Fujian Gutian Pharmaceutical Co., Ltd.), diazepam (Tianjin Lisheng pharmaceutical Limited by Share Ltd.), and atropine (Wuhan Dahua Weiye Pharmaceutical Chemical Co., Ltd.) via intramuscular injection and maintained with a mixture of propofol (AstraZeneca S.P.A. Viale Dell’Industria,3-20040Caponago MI), diazepam (Tianjin Lisheng pharmaceutical Limited by Share Ltd.), and atropine (Wuhan Dahua Weiye Pharmaceutical Chemical Co., Ltd.), administered via the right or left great saphenous vein. Blind placement of a24-gauge Angiocath into the right femoral vein was performed via the Seldinger technique and a5F custom-made catheter was inserted. Then, PE was induced in the experimental group by injection of autologous thrombi with a size of0.3×1cm. Thrombi were generated from15mL fresh blood and comminuted to generate appropriate fragment sizes to induce peripheral pulmonary emboli. In the control group, sham injections with saline were performed instead of thrombi. Post-injection digital subtraction pulmonary angiograms (DSA) were obtained on an Axiom Artis dTA unit after injection of20mL of Ultravist (300mg I/mL,2mL/kg; Bayer, Berlin, Germany) with a flow rate of4mL/s and a pressure of400psi. CT pulmonary angiography (CTPA) in DECT mode was performed and conventional CTPA images and DECT based vascular iodine studies were reconstructed. Two radiologists independently evaluated the number and location of PEs using conventional CTPA and DECT series and then combination of conventional CTPA with DECT based vascular iodine studies. Detailed histopathological examination of lung specimens served as reference standard. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of DECT and CTPA were calculated. Weighted κ values were computed to evaluate inter-modality and inter-reader agreement. All images were evaluated independently by two radiologists, who were unaware of pathologic findings. DECT based vascular iodine studies and CTPA image series were reviewed in random order and a minimum of two weeks apart in the same animal, to minimize reader recall. For CTPA and DECT, the presence, number, and location of pulmonary emboli were recorded on a per-clot basis. On CTPA, PE was defined as luminal filling defects or non-visualization of segmental pulmonary arteries compared to the contralateral side. Continuous filling defects that extended into branching vessels were regarded as a single embolus at the most proximal location. With DECT based vascular iodine analysis, pulmonary arteries color coded in red were regarded as positive for PE. Likewise, continuous pulmonary arteries color coded in red that extended into branching vessels were regarded as single PE, otherwise as two or more. When analyzing DECT based vascular iodine analysis series, we defined PE as color-coded red vessels or color-coded gray intraluminal filling defects because we found that many pulmonary vessels analyzed with the DECT based vascular iodine analysis software showed false-negative findings of intraluminal thrombus where virtual120kVp images could identify the partial filling defects. True negative was caculated on a per-lobe basis, that is, one or more emboli in the lung lode counted as one, none as zero.ResultsThree canines in the experimental group were excluded because of peri-procedural death (n=2) or poor image quality (n=1). Thus, data of four control dogs and nine experimental dogs were included into the final statistical analysis. In the control group, no emboli were detected, while in the experimental group, the histopathological results revealed237emboli in45lung lobes in9dogs. Of237emboli,11emboli were located in segmental pulmonary arteries,49in subsegmental pulmonary arteries,177in fifth-order-or more distal-pulmonary arteries. Reader1detected87emboli using conventional CTPA and266emboli with DECT based vascular iodine analysis. Reader2identified85emboli using conventional CTPA and246emboli with DECT based vascular iodine analysis. With the histopathological results as the reference standard, readers1and2correctly identified166/284and165/279emboli in45lung lobes in9dogs with conventional CTPA versus DECT based vascular iodine analysis. Readers1and2incorrectly diagnosed153/95and152/65emboli with conventional CTPA versus DECT based vascular iodine analysis. These performances correspond to sensitivities of36.1%and86.1%, specificities of98.8%and56.3%, PPVs of98.8%and76.7%, and NPVs of34.5%,70.8%for reader1. Reader2had sensitivities of35.9%and88.2%, specificities of100%and68.4%, PPVs of100%and84.9%, and NPVs of34.5%and74.1%. Overall, both readers combined had sensitivities of39.7%and94.9%, specificities of96.4%and71.4%, PPVs of96.9%and87.9%, and NPVs of35.9%and87.0%using conventional CTPA and DECT based vascular iodine analysis, respectively. When findings at CTPA and DECT based vascular iodine analysis were combined for PE detection, sensitivity, specificity, PPV, and NPV were93.1%,76.9%,89.4%,84.2%, respectively, which were higher than those of conventional CTPA alone (P<0.001). The paired χ2test (McNemar test) showed that DECT based vascular iodine analysis had higher sensitivity, higher NPV, and lower specificity than conventional CTPA (both P< 0.001) for both readers. Kappa statistics were also calculated to quantify the inter-reader and inter-modality agreement for the detection of PE using CTPA and DECT based vascular iodine analysis. Kappa values of0.65and0.70, each with a P<0.001, were found for DECT based vascular iodine analysis and CTPA, respectively. A κ value of0.553(P<0.001) resulted in moderate inter-modality agreement.ConclusionWith histopathological findings as the reference standard, DECT based vascular iodine analysis improves the sensitivity for detecting peripheral PE compared with CTPA, albeit at the price of decreased specificity and PPV. Of note, however, is the observation that the combination of conventional CTPA with DECT based vascular iodine analysis preserved the sensitivity of the latter method while to some degree restoring the specificity of conventional CTPA to the overall analysis. Accordingly, our results suggest that neither method should be used in isolation, but rather in a combined approach to add incremental overall accuracy to the CT-based diagnosis of peripheral PE. PurposeTo evaluate the value of dual energy CT with application of Lung Vessels in diagnosing pulmonary embolism (PE) of children.Materials and Method47patients with nephrotic syndrome and aged≤18years old (32male,15female, mean age15years old). Two radiologists detected the presence or absence of emboli and counted the clots on a per-patient and per-lobe basis with Lung Vessels, respectively. From the raw spiral projection data of both tubes of DECT, images were automatically reconstructed into three image reconstruction sets-80kVp,140kVp, and fused CTPA images with30%of attenuation information from the80kVp data and70%from the140kVp data. These latter fused series closely resemble conventional single-energy120kVp CTPA data and were used for morphological analysis of lung vessels. Section thickness was0.75mm, with a reconstruction interval of0.50mm (67%overlap). All images were transferred to a commercially available workstation (Syngo MMWP VE23A; Siemens) equipped with the DECT post-processing software package (VA21) including the "Lung Vessels" software application. In arteries with a diameter of about2mm or less, the CT-value at the vessel center measured with a typical clinical, single-energy CT scanner ordinarily is systematically reduced by partial volume effects. The "Lung Vessels" application is designed to overcome this limitation by highlighting iodine within small peripheral pulmonary vessels, thus enabling DECT based vascular iodine analysis. The results are generated based on the spectral analysis of80kVp and140kVp data and are displayed as color-coded multiplanar reformations and three-dimensional volume-renderings, in which vessels containing iodine are color coded in blue, whereas soft tissue or vessels with low or no iodine content are color coded in red. All images were evaluated independently by two radiologists with5and2year of experience in the interpretation of DECT, respectively, who were unaware of pathologic findings. DECT based vascular iodine studies and CTPA image series were reviewed in random order and a minimum of two weeks apart in the same animal, to minimize reader recall. For CTPA and DECT, the presence, number, and location of pulmonary emboli were recorded on a per-animal, a per-lung lobe and number of clots basis. On CTPA, PE was defined as luminal filling defects or non-visualization of segmental pulmonary arteries compared to the contralateral side. Continuous filling defects that extended into branching vessels were regarded as a single embolus at the most proximal location. With DECT based vascular iodine analysis, pulmonary arteries color coded in red were regarded as positive for PE. Likewise, continuous pulmonary arteries color coded in red that extended into branching vessels were regarded as single PE, otherwise as two or more. When analyzing DECT based vascular iodine analysis series, we defined PE as color-coded red vessels or color-coded gray intraluminal filling defects because we found that many pulmonary vessels analyzed with the DECT based vascular iodine analysis software showed false-negative findings of intraluminal thrombus where virtual120kVp images could identify the partial filling defects. First of all, we analyzed the presence of PE on a-patient basis, and no PE counted0, if only one clot was found counted1. Then on per-lobe basis, we counted number of clots in left upper lobe, left lower lobe, right upper lobe, right middle lobe and right lower lobe pulmonary arteries with DECT Lung Vessels. One or more emboli in one lung lobe counted as1, otherwise counted0. At last, on number of clots basis, the number of clots from both sides lobar pulmonary arteries extending to subsegmental pulmonary arteries were counted, true negative findings were regarded according to way on per-lobe basis. Analysis of combination of DECT based vascular iodine studies with CTPA image series were reviewed a minimum of one apart in the same patient, to minimize reader recall, and the third experienced radiologist was included when disparity occurred. PE on CTPA was regarded as filling defects in pulmonary arteries. With conventional CT pulmonary angiography (CTPA) as reference standard, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) were computed to detect PE for two readers with Lung Vessels application. Inter-reader agreement was also calculated.ResultsMean CT vale of attenuation in pulmonary arteries was356.4HU in these young patients, of which3patients (6.4%,3/47) were moderate enhanced in pulmonary arteries, and others (93.6%,44/47) were all good enhanced. PE incidence was27.6%(13/47) in13children with nephritic syndrome (51clots found in35lung lobes) with CTPA in our study, among them3patients were diagnosed as isolated peripheral PE, and10patients were multiple PE. We detected51pulmonary emboli and35lobes with PE in13patients with PE with CTPA. Two readers analyzed PE on a per-patient, a per-lung lobe, and a per-clot basis with Lung Vessels. On a per-patient basis, there had30and26patients with PE for readers1and2with Lung Vessels, corresponding to sensitivity, specificity, PPV, and NPV of100%,64.7%,52%,100%and100%,76.5%,61.9%,100%of readers1and2, respectively. On a per-lung lobe basis, readers1and2detected74and63clots with Lung Vessels, respectively, corresponding to sensitivity, specificity, PPV, and NPV of94.3%,79.5%,52.4%,98.8%and97.1%,85.5%,55%,99.4%of readers1and2, respectively. On a per-clot basis,134and105clots were detected with dual energy CT Lung Vessels application, corresponding to sensitivity, specificity, PPV, and NPV for DECT of96.1%,64.9%,36.3%,98.9%and94.1%,79%,45.7%,98.3%of readers1and2, respectively. Excellent or good inter-reader agreement was observed on a per-patient, per-lobe and per-clot basis (kappa value=0.828,0.859,0.695; all P <0.001).Conclusion Dual energy CT with Lung Vessels application has a high sensitivity and negative predictive value for PE detection in children. More clots were detected with Lung Vessels application than conventional CTPA, which deserved further investigation.