The Effectiveness of Using Fewer Implanted Fiducial Markers for Prostate Target Alignment

Abstract

Due to patient comfort, CT and MRI imaging artifacts, and dose shadowing effect for proton therapy, the use of fewer metal objects is preferred. In this study, we evaluated the alignment accuracies by using fewer (one or two) fiducial markers and the impact of fiducial locations on the final result. Three gold fiducial markers were implanted in 10 patients near the apex, left-mid gland, and the base (superior) of the prostate for daily image-guided IMRT on a CT-on-rails system. Patients received 3 CT scans per week in treatment position. The acquired CT images are used to determine the couch correction needed to align the prostate using one, any combinations of two, or all three fiducial markers. The prostate target was also aligned relative to the original treatment planning CT with the aid of an in-house CT-to-CT soft tissue and bony registration methods. The bony registration method simulates patient setup without any implanted fiducials. Residual shifts for the pelvic bone and any combinations of fiducial alignment methods were compared relative to the CT-CT soft tissue alignment method. A time-trend analysis using the linear-regression method was also used to determine if there was a significant time-trend migration. A total of 256 CT scans and 1280 alignments were performed. Residual systematic shifts when using a single superior marker were 1.0 mm, 0.43 mm, and 0.08 mm in the anterior-posterior (AP), superior-inferior (SI), and right-left (RL) directions, respectively. For the mid-gland marker, the systematic shifts were 1.0 mm, 0.48 mm, and 0.07 mm, respectively. For the apex marker, the systematic shifts were 0.04 mm, 0.15 mm, and 0.22 mm, respectively. The residual systematic shifts for the bone alignments were 0.21 mm, 0.08 mm, and 0.05 mm, respectively in the AP, SI, and RL directions. To measure the setup reproducibility, the random residual shifts in one standard deviation (1SD) were 1.49 mm in AP, 1.26 mm in SI, and 0.69 mm in RL for the superior markers; and 1.33 mm in AP, 1.55 mm in SI, and 0.73 mm in RL for the mid marker; and 1.9 mm in AP, 1.07 mm in SI, and 0.7 mm in RL for the apex marker. If bone alignment was used, the 1SD random residual shifts were much larger at 2.8 mm, 2.4 mm, and 0.67 mm in AP, SI, and RL directions, respectively. If we use all three fiducial makers, the systematic shifts were 0.8 mm in AP, 0.25 mm in SI, and 0.12 mm in RL. The random residual shifts are 1.13 mm in AP, 1.0 mm in SI, and 0.5 mm in RL. The best combination of two-fiducial alignment was the superior and apex fiducials: the 1SD was 1.22 in AP, 0.97 mm in SI, and 0.61 mm in RL. Two out of the 10 patients had the mid marker migrated from its original position by approximately 2.3 mm near the end of treatment. One patient had all three markers migrated 2.8 mm. Another patient had a migration of approximately 4.1 mm in the sup marker and 2.4 mm in the mid marker. As a group, the use of single fiducial marker may be acceptable and better than the bone alignment. However, there were significant individual variations in using single fiducials due to organ deformation and migration over the course of treatment. The use of two fiducials can significantly reduce the impact of organ deformation and migration effect. In particular, the combination of the apex and superior fiducial markers gave the best performance rival to the use of all three fiducial markers.