| ObjectiveAcute pulmonary embolism (PE) is a common, often undiagnosed andpotentially fatal disease because of without special symptoms and reliable laboratorytests. A conclusive imaging test such as pulmonary CT angiography (CTPA), a planarventilation/perfusion (V/Q) scans or V/Q SPECT scintigraphy is needed to confirm orexclude the diagnosis of PE. SPECT/CT can achieve Q-SPECT scintigraphy,V/Q-SPECT scintigraphy, plain CT scan and CTPA based on single imagingexamination. Meanwhile, the fusion imaging obtained through further processingworkstations, theoretically, which may enhance their strengths and overcome theirweaknesses. There is, however, less evidence concerning comparisons betweenQ-SPECT and CTPA and Q-SPECT/CT and Q-SPECT/CTPA. Although severalstudies have been carried out about diagnostic ability of the above imaging tests, theresults of evaluations is still controversial owing to the limitations of clinical research,which includes lack of gold standard from pathology, drawbacks of retrospectivestudies, a high-dose radiation etc. The aim of our study was to compare, throughestablishing a PE rabbit model whose final diagnosis was confirmed by pathology, thediagnostic ability of Q-SPECT, CTPA, Q-SPECT/CT, and Q-SPECT/CTPA obtainedsimultaneously using a combined SPECT/CT scanner, in order to offer anexperimental reference and theoretical foundations for clinicians selecting suitableimaging test to diagnose PE.Methods (1)A total of30New Zealand white rabbits with an average body weight of2.5kg, the PE model was made by injecting Gelfoam into the femoral veins of rabbitsafter confirming lungs were normal by plain CT scan.(2)21experimental rabbits underwent SPECT/CT scans, from which Q-SPECT,plain CT scan and CTPA were created, and Q-SPECT/CT, and Q-SPECT/CTPAfusion images were then obtained and evaluated.99mTc-labelled macroaggregatedhuman albumin (99mTc-MAA，74~111MBq) was used for Q-SPECT and Ultravist forcontrast agent(300mgI／m1，4ml/kg).(3)After SPECT/CT scans, rabbits were sacrificed with an overdose ofpentobarbital. The lungs and heart of each rabbit were subsequently removed en bloc,and fixation was performed using a1:10dilution of formaldehyde.3to5days afterfixation, serial sectioning of the lungs was performed at1-mm intervals. Pathologicdetermination of locations and numbers of lung lobes with PE were recorded.(4) All images were evaluated by consensus of two experienced nuclearradiologists; each was blinded to pathologic findings. The presence of PE atQ-SPECT was defined as one or more perfusion defects. Pulmonary artery emboli atCTPA were defined as the finding of luminal filling defects or nonvisualisation ofsegmental pulmonary arteries compared with the contralateral side. The presence ofPE at Q-SPECT/CT was defined as the presence of perfusion defects at Q-SPECT or adiscovery of perfusion defects at Q-SPECT through plain CT imaging. Pulmonaryartery emboli at Q-SPECT/CTPA were defined as a positive result at Q-SPECT orCTPA, and as the absence of PE when being negative at both Q-SPECT and CTPA.The locations and numbers of lung lobes with PE were recorded.(5) Using the pathologic results as the standard of reference, the per lobesensitivity, specificity, accuracy, positive predictive value, and negative predictivevalue for detection of pulmonary emboli with Q-SPECT, CTPA, Q-SPECT/CT andQ-SPECT/CTPA methods were determined. McNemar’s tests were performed tocompare the accuracy of these different imaging methods. Kappa analysis was used todetermine the agreement between two nuclear radiologists. Kappa values <0.20wereinterpreted as poor agreement,0.21to0.40as fair agreement,0.41to0.75as moderate agreement,0.76to1.0as good agreement. P values <.05regarded asstatistically significant. Statistical analysis was performed using SPSS version19.0.Results(1)On the basis of the results of histopathology, pulmonary emboli were detectedin at least one lobe in19of the21rabbits. Overall, histologically confirmed emboliwere present in a total of26pulmonary lobes and absent in79lobes.(2) Q-SPECT scintigraphic measurements yielded14true-positive results,74true-negative results, along with5false-positive results and11false-negativeresults,these results corresponded to sensitivities of53.9%[95%CI:33.4%-73.4%] andspecificities of93.7%[95%CI:85.8%-97.9%] and accuracy of83.8%[95%CI:76.8%-90.9%] and positive predictive value of73.7%[95%CI:48.8%-90.9%] and negativepredictive value of86.1%[95%CI:76.9%-92.6%].(3) CTPA measurements yielded19true-positive results,76true-negative results,along with3false-positive results and7false-negative results,these resultscorresponded to sensitivities of73.1%[95%CI:52.2%-88.4%] and specificities of96.3%[95%CI:89.6%-99.2%] and accuracy of90.5%[95%CI:84.9%-96.1%] andpositive predictive value of86.4%[95%CI:65.1%-97.1%] and negative predictivevalue of91.6%[95%CI:83.4%-96.5%].(4) Q-SPECT/CT measurements yielded20true-positive results,74true-negative results, along with5false-positive results and6false-negativeresults,these results corresponded to sensitivities of76.9%[95%CI:56.4%-91.0%] andspecificities of93.7%[95%CI:85.8%-97.9%] and accuracy of89.5%[95%CI:83.7%-95.4%] and positive predictive value of80.0%[95%CI:59.3%-93.2%] and negativepredictive value of92.5%[95%CI:84.4%-97.2%].(5) Q-SPECT/CTPA measurements yielded23true-positive results,72true-negative results, along with7false-positive results and3false-negativeresults,these results corresponded to sensitivities of88.5%[95%CI:69.9%-97.6%] andspecificities of91.1%[95%CI:82.6%-96.4%] and accuracy of90.5%[95%CI:85.3%-96.1%] and positive predictive value of76.7%[95%CI:57.7%-90.1%] and negative predictive value of96.0%[95%CI:88.8%-99.2%].(6) McNemar’s tests showed that assessments based on Q-SPECT/CT andQ-SPECT/CTPA had higher diagnostic sensitivity for the detection of pulmonaryemboli than Q-SPECT (P＜.005)， Q-SPECT/CT had higher diagnostic accuracythan Q-SPECT (P<.005).There were no statistically significant differences betweenthe diagnostic accuracy of other different imaging methods. There were also nostatistically significant differences between positive or negative predictive value of allimaging tests.(7) Kappa values of Q-SPECT, CTPA, Q-SPECT/CT and Q-SPECT/CTPA forreaders1and2were0.902、0.915、0.973、0.884。Conclusions(1) Q-SPECT/CT imaging provides good sensitivity, specificity, which has adiagnostic performance corresponded roughly to Q-SPECT/CTPA, CTPA, but betterdiagnosibility than Q-SPECT. Meanwhile, using Q-SPECT/CT imaging, the diagnosisof two readers has the best consistency.(2) Q-SPECT/CTPA has an great sensitivity, specificity, which has a greaterdiagnostic performance than Q-SPECT, but the accuracy almost as well as CTPA.(3) Q-SPECT has an diagnostic performance corresponded roughly to CTPA. |
PDF Full Text Request