Technical note: Beamlet-free optimization for Monte-Carlo-based treatment planning in proton therapy

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• Proton therapy treatment planning algorithms demand substantial computational resources, limiting clinical workflow efficiency.
• High computational costs hinder the adoption of advanced proton therapy techniques such as robust optimization, arc therapy, and adaptive therapy.

Purpose of the Study:

• Introduce a novel beamlet-free algorithm to overcome computational limitations in proton therapy.
• Integrate Monte Carlo dose calculation and optimization, eliminating the need for a dose influence matrix.

Main Methods:

• The beamlet-free algorithm simulates proton dose using randomly selected spots and assesses their objective function impact.
• Spot weights are iteratively updated based on approximated gradients to refine the spot probability distribution.
• Comparative analysis against a conventional beamlet-based algorithm was performed using brain and prostate treatment cases.

Main Results:

• The beamlet-free approach yielded comparable plan quality to conventional methods.
• Demonstrated a significant reduction in computation time and memory requirements, independent of the number of spots.
• The algorithm's performance was validated on both brain and prostate cancer treatment scenarios.

Conclusions:

• The beamlet-free algorithm is a feasible and effective alternative for proton therapy treatment planning.
• Achieves comparable plan quality to existing methods while offering substantial computational efficiency.
• Its low spot dependence and resource efficiency position it as a promising tool for complex proton therapy applications.