Using measurable dosimetric quantities to characterize the inter-structural tradeoff in inverse planning

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• Traditional inverse planning uses weighting factors, which are time-consuming to determine manually.
• Balancing conflicting dose requirements for different structures is a key challenge in treatment planning.

Purpose of the Study:

• To develop a novel inverse planning framework using physically meaningful quantities to parameterize dosimetric tradeoffs.
• To simplify the search for clinically sensible treatment plans by replacing manual weighting factors.

Main Methods:

• Introduced the dosimetric variation-controlled model (DVCM) using permissible dose or DVH variation ranges.
• Developed a two-phase procedure (TPP) to solve the DVCM, first for feasibility, then for optimization.
• Applied the TPP to prostate and head-and-neck cancer cases, comparing results with CVaR and moment-based methods.

Main Results:

• The TPP successfully generated clinically sensible plans with minimal trial and error for all tested cases.
• TPP plans were competitive with or superior to those from conventional CVaR and moment-based optimization schemes.
• One head-and-neck case showed over 40% reduction in spinal cord fractional volume receiving >20 Gy with TPP, maintaining PTV coverage.

Conclusions:

• The proposed technique significantly reduces the need for manual parameter adjustment in inverse planning.
• Physically meaningful modeling of inter-structural tradeoffs enhances treatment planning efficiency and effectiveness.
• This new formalism offers opportunities for incorporating prior knowledge into treatment planning.