Automatic feathering algorithm for VMAT craniospinal irradiation: A comprehensive comparison with other VMAT planning...

Area of Science:

• Radiation Oncology
• Medical Physics
• Radiotherapy Planning

Background:

• Craniospinal irradiation (CSI) is sensitive to setup uncertainties, potentially causing overdoses/underdoses in the planning target volume.
• Volumetric-modulated arc therapy (VMAT) strategies have been developed to mitigate setup uncertainties in CSI.

Purpose of the Study:

• To compare a novel treatment planning system-integrated automatic feathering (AF) algorithm against other VMAT strategies for plan robustness in CSI.
• To evaluate the impact of overlapping lengths and simulated setup uncertainties on plan robustness.

Main Methods:

• Retrospective analysis of three patients treated with CSI.
• Comparison of five VMAT techniques: overlap (O), staggered overlap (SO), gradient optimization (GO), O with AF (O-AF), and SO with AF (SO-AF).
• Simulation of setup uncertainties using isocenter shifts (±1 mm, ±3 mm, ±5 mm) and evaluation of plan robustness via dose differences (ΔD_2%, ΔD_98%).

Main Results:

• The standard overlap technique showed the highest dose differences (ΔD_2%=15.4%, ΔD_98%=15.0%).
• O-AF and SO-AF demonstrated comparable robustness to GO, with improved performance at larger overlapping lengths (10 cm).
• At 10-cm overlap and 5-mm shift, GO, O-AF, and SO-AF achieved better robustness (ΔD_2% < 6.0%, ΔD_98% < 7.6%).

Conclusions:

• The automatic feathering (AF) algorithm significantly enhances plan robustness in craniospinal irradiation, particularly when used with larger overlapping lengths.
• AF integration makes VMAT plans as robust as those generated by more complex optimization methods like gradient optimization.
• AF offers a practical and effective solution for improving the safety and accuracy of VMAT in CSI by reducing sensitivity to setup uncertainties.