(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: This study seeks to identify the ideal locations for gold fiducial marker placement to optimize tumor tracking.
Methods: Patients receiving robotic SBRT for lung tumors received 4D-CT imaging. 19 patients were included in the study. The patients usually had 4 fiducials placed with a minimum of 3 fiducials being placed. The Gross Tumor Volume (GTV) and each fiducial was contoured, with post-processing to eliminate noise. The center of mass was calculated for every fiducial and the GTV in 10 phases of breathing. Centroids of different fiducial combinations were compared with one another to determine the combination that best correlated with the GTV center of mass. The 3 centroid combinations tracked were the centroid from all fiducials, from the two fiducials closest to the GTV, and from the two fiducials that individually tracked the best. A two-mean t-test was used to compare the results.
Results: The centroid of the two fiducials closest to the GTV tracks best. On average this centroid was closest to the GTV (mean=1.89 cm). When comparing the total centroid (mean= 2.29 cm) with the centroid from the two best individually tracking fiducials (mean=2.26 cm), the difference was not statistically significant. However, when comparing the two closest fiducials’ centroid with both other combinations, statistical significance was observed. The two closest fiducials tracked better than the centroid from all fiducials (p=0.061). The two closest fiducials centroid tracked better than the two individually best tracking fiducials (p=0.006).
Conclusion: The data would suggest that it is better to place fewer fiducials as close to the tumor as possible, within the limits of how well the system can distinguish individual fiducials. More research is needed to determine optimal fiducial placement around a tumor. Currently it is preferred that at least three fiducials are tracked so rotational movement can potentially be tracked.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
