Improving IMRT delivery efficiency with reweighted L1-minimization for inverse planning

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Imaging

Background:

• Intensity modulated radiation therapy (IMRT) inverse planning aims to optimize radiation dose delivery.
• Current methods for simplifying fluence-maps in IMRT can increase complexity, leading to inefficient treatment delivery.
• First-order total-variation (TV) minimization using L1-norm offers some simplification but struggles with complex dose sparing requirements.

Purpose of the Study:

• To develop and validate an improved technique for simplifying fluence-maps in IMRT inverse planning.
• To reduce plan complexity and enhance dose delivery efficiency without compromising plan quality.
• To explore the application of iteratively reweighted L1-minimization for sparse signal recovery in IMRT.

Main Methods:

• Employed iteratively reweighted L1-minimization, inspired by L0-norm's ideal sparsity, to overcome L1-norm's limitations.
• Integrated the reweighted L1-minimization with TV minimization to further enhance fluence-map sparsity.
• Validated the proposed method by comparing treatment plans against quadratic minimization and conventional TV minimization using clinical data from five patients.

Main Results:

• The proposed reweighted TV minimization technique generated simpler fluence-maps compared to existing methods.
• Achieved significant reductions in the number of segments (10-35) and modulation index (20-60%) relative to conventional techniques.
• Demonstrated a notable decrease in total treatment time (12-80 seconds) due to reduced multileaf collimator (MLC) travel time.

Conclusions:

• Iteratively reweighted L1-minimization is a promising approach for simplifying fluence-maps in IMRT inverse planning.
• The technique effectively improves delivery efficiency by reducing segments and treatment time.
• Plan quality, including target conformity and critical structure sparing, is maintained with the proposed method.