Click here to


Are you sure ?

Yes, do it No, cancel

Metal Artifact Reduction Algorithms for CT in the Trauma Setting: Do They Take Too Long?

S Rose*, M Lipford, N Stabo, C Bartels, M Lubner, T Szczykutowicz, University Wisconsin-Madison, Madison, WI


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: We quantify the increase in reconstruction time when employing metal artifact reduction (MAR) and analyze the impact on the duration of trauma exams at our institution.

Methods: Our dataset consists of 127 patient exams performed on 5 CT scanners. We quantified reconstruction times using the Content Time DICOM tag of the first and last images. With and without MAR timing comparisons were performed with cervical-spine reconstructions of 1.25mm slice thickness and 0.625mm spacing. Reconstruction time per axial slice was calculated to account for differences in scan range.
Exam duration, defined here as time between first acquisition and completion of last reformat, was calculated using the Acquisition and Content Time DICOM tags. Durations were compared between exams in which MAR was employed in one or more reconstructions sent to PACS and exams in which MAR was never employed. A single trauma exam often fulfills multiple orders. To control for this, comparisons were limited to exams fulfilling the same number of orders.

Results: Median reconstruction time increased by 342% when MAR was employed, from 12s to 53s (p<0.001). Median reconstruction time per axial slice similarly increased by 350%, from 0.039s to 0.173s (p<0.001). Despite the increase in reconstruction time, we did not observe an increase in duration for exams employing MAR. MAR containing exam duration was 30.2 minutes versus non MAR duration of 29.9 minutes (p=0.65).

Conclusion: MAR can increase reconstruction time by as much as 350%, but we did not observe a corresponding increase in exam duration in the setting of trauma. We believe this is attributable to the short duration of reconstruction relative to total exam time, the fact that trauma exam durations are highly variable, and that technologists perform other tasks while images reconstruct.

Funding Support, Disclosures, and Conflict of Interest: T. Szczykutowicz and C. Bartels supply CT protocols to GE Healthcare under a licensing agreement. T. Szczykutowicz has an equipment grant with GE HealthCare, consults for GE and iMALOGIX, and is on the medical advisory board of iMALOGIX. M. Lubner has received prior research funding from Phillips Healthcare and Ethicon.


CT, Reconstruction, Quality Assurance


IM- CT: Metal artifact reduction

Contact Email