Optimizing leaf widths for a multileaf collimator.

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Technology

Background:

• Modern linear accelerators (linacs) utilize multileaf collimators (MLCs) for shaping radiation fields.
• MLC-shaped fields exhibit stepwise boundaries, leading to discrepancies (under- or over-blocking) compared to desired target conformities.
• The extent of these discrepancies is directly influenced by the widths of the MLC leaves.

Purpose of the Study:

• To develop an optimization model for determining MLC leaf widths.
• To minimize the total area of discrepancy between MLC-shaped fields and ideal target conformities.
• To investigate the impact of optimized leaf widths on radiation field shaping accuracy.

Main Methods:

• An optimization model was formulated with leaf widths as variables and total discrepancy area as the objective function, constrained by total leaf width.
• The model was solved using a hybrid approach combining simulated annealing (ASA) and a gradient technique (DONLP2).
• Performance was evaluated on 634 target fields from a treatment planning system database.

Main Results:

• Optimal leaf width distribution was determined for MLCs with 28, 40, and 60 leaf pairs.
• Optimal leaf widths showed a non-linear relationship with distance from the central line, decreasing slightly then increasing.
• MLCs with optimized leaf widths reduced total discrepancy area by 11.1% to 28.6% compared to conventional MLCs.

Conclusions:

• Optimizing MLC leaf widths significantly improves the conformity of shaped radiation fields to treatment targets.
• This optimization allows for enhanced field conformity without increasing the number of leaf pairs.
• Alternatively, it enables a reduction in leaf pairs while maintaining or improving field conformity, potentially simplifying linac design.