Encrypted login | home

Program Information

Impact of a Metal Artifact Reduction Technique On Delineation of GTV Contours in RTP for Head and Neck Cancer

no image available
N Nagami

N Nagami*, H Arimura , M Ohishi , T Imaizumi , S Tokumaru , K Uba , A Hiroki , H Irie , Kyushu University, Fukuoka City, Fukuoka Prefecture


SU-I-GPD-J-77 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To investigate the impact of a metal artifact reduction technique on delineation of gross tumor volume (GTV) contours in radiation treatment planning (RTP) for head and neck cancer.

Methods: A metal artifact reduction (MAR) technique (Single Energy Metal Artifact Reduction, SEMAR, TOSHIBA) could reduce the metal artifacts in planning computed tomography (CT) images, which were reconstructed by interpolating artifact voxels in sinograms. A GTV was simulated using a piece of protein - fat conjugated composition. Head CT images without and with metal artifacts were acquired as planning CT images by scanning a head phantom without and with putting crowns (gold-palladium alloy) on teeth, respectively. The planning CT images with the metal artifacts were processed with and without the MAR technique. The image acquired without crowns was considered a reference image without metal artifacts. Six radiation oncologists delineated three contours of the target on the reference image, CT images with and without MAR. Since PTV D90, which was defined as the minimum dose delivered to 90% of PTV, would be affected by the uncertainties of GTV contours due to the metal artifacts, D90 was employed for evaluating the impacts of the MAR technique on delineation of GTV contours.

Results: The average D90 for contours on the reference image was larger than that for those on the image without MAR (p < 0.05), and also that for those on the image with MAR was greater than that for those on the image without MAR (p < 0.05). On the other hand, there was no statistically significant difference with respect to D90 between contours on the reference image and the image with MAR (p > 0.05).

Conclusion: The MAR technique could improve the accuracy of delineation GTV contours on planning CT images.

Contact Email: