Purpose: The clinical proton dose calculation algorithms using pencil-beam convolution superposition (PCS) and Monte Carlo (MC) methods in Eclipse™ TPS were validated in a phantom with the metallic port of a Natrelle-133 tissue expander (Allergan, PLC) placed between a silicone-filled prosthesis (Mentor, LLC) and solid water.
Methods: A proton spot-scanning field was generated to deliver a homogeneous dose to a target volume in the phantom. The metallic port was contoured based on geometrical dimensions provided by the manufacturer. The Hounsfield units (HUs) were assigned based on the corresponding stopping powers from lectures (silicone: 0.94, Titanium: 3.17, magnet: 5.50, solid water: 1.02) for PCS (PCS 15.6) algorithm. The materials used for MC (AcurosPT 13.7) calculation were selected (silicone: wax, Titanium: Aluminum, magnet: Brass) to have the closest stopping powers. The 2D planar doses at multiple depths close to the distal-dose fall off were measured using a MatriXX-PT detector (IBA-dosimetry, Germany) and compared with the calculations.
Results: The measured 2D planar doses at distal target edge (36.5 mm below the prosthesis) matched both PCS (35.4 mm) and MC (37.4 mm) calculations within 1 mm discrepancy in depth. The depth-doses collected from 2D planar doses behind the breast prosthesis were found to over-range in measurements by 1.0 mm and 0.0 mm versus PCS and MC calculations, respectively. However, a 5-8 mm over-range outside the target volume was observed when protons passed through the metallic port.
Conclusion: The assigned HU (-124) for silicone-filled breast prostheses with the institutional CT curve was experimentally validated for both PCS and MC algorithms. Limited by the lateral scattering in PCS algorithm and material selections for MC method, the disturbed doses caused by the high-Z metallic port cannot be predicted precisely. A two-field beam arrangement is suggested to average out and confine the uncertain doses inside the tissue expander.