Program Information
Evaluation of Metal Artifact Reduction (MAR) Software On Computed Tomography (CT) Images
V Huang*, K Kohli , BC Cancer Agency, Surrey, BC
Presentations
SU-E-I-13 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose:
A new commercially available metal artifact reduction (MAR) software in computed tomography (CT) imaging was evaluated with phantoms in the presence of metals. The goal was to assess the ability of the software to restore the CT number in the vicinity of the metals without impacting the image quality.
Methods:
A Catphan 504 was scanned with a GE Optima RT 580 CT scanner (GE Healthcare, Milwaukee, WI) and the images were reconstructed with and without the MAR software. Both datasets were analyzed with Image Owl QA software (Image Owl Inc, Greenwich, NY). CT number sensitometry, MTF, low contrast, uniformity, noise and spatial accuracy were compared for scans with and without MAR software. In addition, an in-house made phantom was scanned with and without a stainless steel insert at three different locations. The accuracy of the CT number and metal insert dimension were investigated as well.
Results:
Comparisons between scans with and without MAR algorithm on the Catphan phantom demonstrate similar results for image quality. However, noise was slightly higher for the MAR algorithm. Evaluation of the CT number at various locations of the in-house made phantom was also performed. The baseline HU, obtained from the scan without metal insert, was compared to scans with the stainless steel insert at 3 different locations. The HU difference between the baseline scan versus metal scan was improved when the MAR algorithm was applied. In addition, the physical diameter of the stainless steel rod was over-estimated by the MAR algorithm by 0.9 mm.
Conclusion:
This work indicates with the presence of metal in CT scans, the MAR algorithm is capable of providing a more accurate CT number without compromising the overall image quality. Future work will include the dosimetric impact on the MAR algorithm.
Contact Email: