Dual energy CT angiography: pros and cons of dual-energy metal artifact reduction algorithm in patients after endovascular aortic repair

Purpose: To evaluate the value of metal artifact reduction (MAR) post-processing and iodine MD images in fast kV-switching dual-energy computed tomography (DECT) in patients after endovascular aortic repair (EVAR). Materials and methods: Twenty-four consecutive EVAR patients (age 76 +/- 9 years, 7/24 (29%) with coils, 9/24 (37.5%) with 10 endoleaks) who underwent DECT angiography were included in this HIPAA-compliant, IRB-approved retrospective study. Monochromatic reconstructions included 55, 60, 65, 70, and 75 keV with and without MAR and iodine MD images. Near field, far field, and vessel artifacts were assessed subjectively (1 = none; 5 = severe) and objectively by measuring noise and contrast-to-noise ratio. Visibility of endoleak was evaluated (1 = optimal; 5 = not visible). Results: MARobjectively decreased artifacts from EVAR stents in the near field (60.7 +/- 25.4 HU vs. 70.1 +/- 34.2; p -.002) and subjectively increased near field (3.2 +/- 0.9 vs. 2.8 +/- 0.6; p <.001), far field (2.2 +/- 0.6 vs. 1.6 +/- 0.6; p < .001), and vessel (3.1 +/- 1.1 vs. 2.5 +/- 0.9; p <.001) artifacts. Near-field artifacts from coils were reduced by the MAR objectively (72.4 +/- 24.8 vs. 182.7 +/- 57.3 HU; p < .001) and subjectively (4.5 +/- 0.5 vs. 4.9 +/- 0.4; p < .02). CNR of standard reconstructions was optimal at 60 keV (38.3 +/- 16.8). Reconstructions without MAR and iodine MD images provided improved endoleak visualization in 6/10 (60%) of cases (median 1 for both) compared to MAR (median 3) (p < 0.001). However, MAR improved visualization in 1/10 (10%) cases due to endoleak location adjacent to a coil. Conclusion: DECT withMARreduced artifacts from coils and improved endoleak visualization in 1/10 (10%) cases due to location adjacent to a coil. However, MAR impaired endoleak visualization in 6/10 (60%) cases and should be reviewed combined with 60 keV standard reconstructions and iodine MD images.