6 mo) were prospectively recruited to undergo late phase DECT for characterization of their MI.Computed tomography(CT) examinations were performed using a dual source CT system(64-slice Definition or 128-slice Definition FLASH,Siemens Healthcare) with initial first pass and 10 min late phase image acquisitions.Using the 17-segment model,regional systolic function was analyzed using first pass CT as normal or abnormal(hypokinetic,akinetic,dyskinetic).Regions with abnormal systolic function were identified as infarct segments.Late phase DE scans were reconstructed into:140 kVp,100 kVp,mixed(120 kVp) images and iodine-only datasets.Using the same 17-segment model,each dataset was evaluated for possible(grade 2) or definite(grade 3) late phase myocardial enhancement abnormalities.Logistic regression for correlated data was used to compare reconstructions in terms of the accuracy for detecting infarct segments using late myocardial hyperenhancement scores.RESULTS:All patients reported prior history of documented myocardial infarction,with most occurring more than 5 years prior(n = 18;75% of cohort).Fiftyfive of 408(13%) segments demonstrated abnormal wall motion and were classified as infarct.The remaining 353 segments were classified as non-infarcted segments.A total of 1692 segments were analyzed for late phase enhancement abnormalities,with 91(5.5%) segments not interpretable due to artifact.Combined grades 2 and 3 compared to grade 3 only enhancement abnormalities demonstrated significantly higher sensitivity and similar specificity for detection of infarct segments for all reconstructions evaluated.Evaluation of different voltage acquisitions demonstrated the highest diagnostic performance for the 100 kVp reconstruction which had higher diagnostic accuracy(87%;95%CI:80%-90%),sensitivity(86%-93%;95%CI:54%-78%) and specificity(90%;95%CI:86%-93%) compared to the other reconstructions.For sensitivity,there were significant differences noted between 100 kVp vs 140 kVp(P">
海量資源,盡在掌握