Purpose: This study aims to make recommendations for fiducial marker selection that optimizes lung tumor tracking in SBRT patients. It is rare that all fiducials are tracked during treatment delivery and the selection is currently based on proximity to the GTV.
Methods: 10 patients were retrospectively reviewed. Each patient had 4 fiducials placed via bronchoscopy per manufacturer recommendations. The gross tumor volume (GTV) and each fiducial were contoured in each of the 10 phases from a simulation 4DCT. The center of mass (COM) was calculated for each fiducial and the GTV in every phase using a Matlab program with post-processing to eliminate noise. The centroid of every fiducial combination was found and compared to determine which centroid combination was closest to the GTV’s COM and which centroid’s motion was most like the GTV’s.
Results: 8 of the 10 patients had the centroid tracked that was not the centroid closest to the GTV’s COM and 1 patient had the centroid that matched the tumors motion the best tracked. The centroid that was closest to the GTV’s COM was closer than the tracked centroid on average by 8.5 mm and the centroid that kept the most consistent distance was more consistent on average by 0.3 mm.
Conclusion: The fiducials that are tracked during robotic SBRT delivery per usual clinical practice are not always the fiducials that are closest to the GTV’s COM, or that best correlate with GTV COM motion during breathing. An analysis of GTV and fiducial motion on 4DCT using the technique described here, along with machine data during clinical simulation on how well different fiducial combinations are tracked, can improve tumor tracking during SBRT delivery or inform the treating the physician on the expansions needed to create the PTV that would minimize the risk of marginal tumor recurrence.