To investigate dosimetric errors in proton therapy treatment planning due to titanium implants, and to determine how these affect postoperative passively scattered proton therapy for chordoma patients with orthopedic hardware.
The presence of titanium hardware near the tumor may affect the dosimetric accuracy of proton therapy. Artifacts in the computed tomography (CT) scan can cause errors in the proton stopping powers used for dose calculation, which are derived from CT numbers. Also, clinical dose calculation algorithms may not accurately simulate proton beam transport through the implants, which have very different properties as compared to human tissue. The authors first evaluated the impact of these two main issues. Dose errors introduced by metal artifacts were studied using phantoms with and without titanium inserts, and patient scans on which a metal artifact reduction method was applied. Pencil-beam dose calculations were compared to models of nuclear interactions in titanium and Monte Carlo simulations. Then, to assess the overall impact on treatment plans for chordoma, the authors compared the original clinical treatment plans to recalculated dose distributions employing both metal artifact reduction and Monte Carlo methods.
Dose recalculations of clinical proton fields showed that metal artifacts cause range errors up to 6 mm distal to regions affected by CT artifacts. Monte Carlo simulations revealed dose differences >10% in the high-dose area, and range differences up to 10 mm. Since these errors are mostly local in nature, the large number of fields limits the impact on target coverage in the chordoma treatment plans to a small decrease of dose homogeneity.
In the presence of titanium implants, CT metal artifacts and the approximations of pencil-beam dose calculations cause considerable errors in proton dose calculation. The spatial distribution of the errors however limits the overall impact on passively scattered proton therapy for chordoma.
The authors would like to thank Tom Madden and Hanne Kooy, Ph.D., for providing the implementation of the pencil-beam dose calculation algorithm, and Jan Schümann, Ph.D., for the support of the TOPAS Monte Carlo code. Thanks to Juliane Daartz, Ph.D., for helpful feedback on the paper. This work was supported by the Federal Share of program income earned by Massachusetts General Hospital on C06-CA059267, Proton Therapy Research and Treatment Center.
II. METHODS AND MATERIALS
II.A. Study design
II.A.1. CT metal artifacts
II.A.2. Beam transport through titanium
II.A.3. Evaluation of chordoma treatment plans
II.B. CT scans
II.B.1. Phantom scan
II.B.2. Patient scans
II.C. Treatment planning
II.C.1. Phantom for metal artifacts study
II.C.2. Phantom for dose calculation study
II.C.3. Patient treatment plans
II.D. Pencil-beam dose calculation algorithm
II.E. Monte Carlo simulations
II.E.1. Treatment head
II.E.2. Patient anatomy
II.E.3. Physics models
II.E.4. Physics model validation
II.F. Proton–nuclear interactions with titanium
III.A. Impact of metal artifacts
III.A.1. Phantom study
III.A.2. Patient study
III.B. Impact of dose calculation method
III.B.1. Geant4 physics model validation
III.B.2. Proton-nuclear interactions with titanium
III.B.3. Phantom study
III.B.4. Patient treatment plans
III.C. Overall impact on chordoma treatment plans
IV. DISCUSSION AND CONCLUSIONS
Data & Media loading...
Article metrics loading...
Full text loading...