Accelerated iterative beam angle selection in IMRT.

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• Intensity-modulated radiation therapy (IMRT) planning involves selecting optimal beam angles for effective dose delivery.
• Iterative beam angle selection (BAS) methods sequentially add beam directions to an ensemble.
• Traditional BAS requires computationally intensive fluence map optimization (FMO) for each candidate beam.

Purpose of the Study:

• To evaluate two novel methods for accelerating iterative BAS using surrogates for the FMO objective function.
• To assess the impact of these accelerated methods on treatment plan quality and computation time.
• To investigate the utility of BAS with noncoplanar beams in IMRT.

Main Methods:

• Candidate beams were selected based on early FMO iterations (5 iterations) or projected gradients (1st iteration).
• Objective function surrogates were employed to approximate the full FMO objective function value.
• A treatment planning study included intracranial, pancreas, and prostate cases, with additional evaluation for noncoplanar beam arrangements.

Main Results:

• BAS utilizing objective function surrogates achieved comparable objective function values and dose statistics to naïve BAS.
• Computation time was reduced by one to two orders of magnitude through early FMO stopping and projected gradients.
• Optimized ensembles with few noncoplanar beams approached the quality of fully noncoplanar ensembles.

Conclusions:

• Iterative BAS combined with objective function surrogates offers a viable approach for automated BAS.
• This accelerated method achieves clinically acceptable computation times.
• The findings support the integration of surrogate-based BAS for efficient IMRT planning.