Does the metal artifact reduction algorithm activation mode influence the magnitude of artifacts in CBCT images?

This study was conducted to assess the effectiveness of a metal artifact reduction (MAR) algorithm activated at different times during cone-beam computed tomography (CBCT) acquisition on the magnitude of artifacts generated by a zirconium implant.

Volumes were obtained with and without a zirconium implant in a human mandible, using the OP300 Maxio unit. Three modes were tested: without MAR, with MAR activated after acquisition, and with MAR activated before acquisition. Artifacts were assessed in terms of the standard deviation (SD) of gray values and the contrast-to-noise ratio (CNR) in 6 regions of interest with different distances (10 to 35 mm, from the nearest to the farthest) and angulations (70° to 135°) from the implant region.

In the acquisitions without MAR, the regions closer to the implant (10 and 15 mm) had a higher SD and lower CNR than the farther regions. When MAR was activated (before or after), SD values did not differ among the regions (P>0.05). The region closest to the implant presented a significantly lower CNR in the acquisitions without MAR than when MAR was activated after the acquisition; however, activating MAR before the acquisition did not yield significant differences from either of the other conditions.

Both modes of MAR activation were effective in decreasing the magnitude of CBCT artifacts, especially when the effects of the artifacts were more noticeable.