The purpose of this study was to compare a combined dual-energy CT (DECT) and single-energy CT (SECT) metal artifact reduction technique with a SECT metal artifact reduction technique for detecting lesions near an arthroplasty in a phantom model. Two CT phantoms with a cobalt chromium sphere attached to a titanium rod, simulating an arthroplasty, within a background of soft-tissue attenuation containing spherical lesions (range, 10-20 mm) around the head and stem of different attenuations from the background (range of attenuation, 10-70 HU) were scanned with a single CT scanner individually (unilateral) and together (bilateral) with the following three dose-equivalent techniques: the currently used clinical protocol (140 kVp, 300 Reference mAs); 100 kVp; and DECT (100 kVp and 150 kVp with a tin filter). Three radiologists reviewed the datasets to identify lesions. Nonparametric AUC was estimated for each reader with each technique. Multireader ANOVA was performed to compare AUCs. Multiple-variable logistic regression analysis was used to identify factors affecting sensitivity and specificity. Accuracy was lower ( < 0.001) for the DECT 130-keV technique than for the 100-, 70-, and 140-kVp techniques. Sensitivity was higher with unilateral arthroplasties ( = 0.037), with greater contrast differences from background ( < 0.001), and with the SECT 100-kVp technique versus other techniques ( < 0.001). The difference in specificities of modalities was not statistically significant ( = 0.148). Combining DECT and SECT techniques does not provide additional benefits for lesion detection as opposed to using SECT alone.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.2214/AJR.19.22084 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!