Fast Monte Carlo simulation on a voxelized human phantom deformed to a patient

Area of Science:

• Medical Physics
• Radiotherapy Dosimetry
• Computational Imaging

Background:

• Accurate absorbed dose simulation is crucial for effective radiotherapy planning.
• Traditional methods often rely on calibration curves and can be computationally intensive.
• Patient-specific dosimetry requires methods that can adapt to individual anatomy.

Purpose of the Study:

• To present a novel method for fast absorbed dose simulation directly from patient CT scans.
• To eliminate the need for explicit calibration curves in dose calculation.
• To enable patient-specific dosimetric studies using deformable phantom technology.

Main Methods:

• Geometrical deformation of a standard voxelized human phantom to match patient CT images.
• Utilizing the inherent material composition and density data of the phantom.
• Comparison with Monte Carlo (MC) simulations using PENELOPE and a commercial collapsed cone convolution algorithm.

Main Results:

• Absorbed dose calculations showed a maximum distance to agreement of 2 mm compared to reference MC simulations.
• The method demonstrated validity across various patient anatomies and random density phantoms.
• Fast simulations achieved 2% statistical uncertainty in 10 minutes for a 6 MV photon beam.

Conclusions:

• The deformed phantom method enables accurate, patient-specific absorbed dose calculations directly from CT data.
• This approach facilitates dosimetric studies and organ-specific information retrieval.
• It validates the use of condensed-history algorithms like PENFAST for efficient absorbed dose estimation in clinical settings.