Fast online Monte Carlo-based IMRT planning for the MRI linear accelerator

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Technology

Background:

• MRI-guided radiotherapy integrates magnetic resonance imaging (MRI) with linear accelerators (LINACs) for real-time anatomical tracking.
• Continuous anatomical updates necessitate rapid online re-optimization of intensity-modulated radiotherapy (IMRT) plans.
• The presence of magnetic fields in MRI-LINACs may affect radiotherapy planning and delivery.

Purpose of the Study:

• To develop and evaluate a fast IMRT optimization system for online replanning in MRI-guided radiotherapy.
• To assess the impact of a 1.5 T magnetic field on IMRT plan quality and optimization speed.
• To demonstrate the feasibility of generating tightly conformal IMRT plans in the presence of magnetic fields.

Main Methods:

• Development of a GPU-based Monte Carlo dose calculation engine for rapid online beamlet generation.
• Implementation of a fast inverse dose optimization algorithm for IMRT replanning.
• Testing the system on four phantom and two clinical (cervix, kidney) cases at 0 T and 1.5 T magnetic fields.

Main Results:

• The beamlet generation and optimization routines demonstrated sufficient speed for online IMRT planning.
• No significant influence of the 1.5 T magnetic field on IMRT plan quality or complexity was observed.
• Nearly identical dose distributions were achieved at 0 T and 1.5 T with equivalent optimization constraints.

Conclusions:

• The developed fast IMRT optimization system is suitable for online replanning in MRI-guided radiotherapy.
• The 1.5 T magnetic field does not compromise IMRT plan quality or optimization efficiency.
• This technology facilitates adaptive radiotherapy by enabling rapid, accurate treatment adjustments during delivery.