Vector-model-supported optimization in volumetric-modulated arc stereotactic radiotherapy planning for brain...

Area of Science:

• Medical Physics
• Radiation Oncology
• Artificial Intelligence in Medicine

Background:

• Volumetric-modulated arc stereotactic radiotherapy (VMA-SRT) planning is time-consuming, limiting its clinical application.
• Existing methods require extensive manual input and iterative adjustments, contributing to prolonged treatment preparation.
• Efficient VMA-SRT is crucial for managing complex cases like brain metastases.

Purpose of the Study:

• To develop and evaluate a vector model for optimizing VMA-SRT.
• To assess the model's ability to reduce planning time for brain metastasis cases.
• To compare the quality of VMA-SRT plans generated with and without vector-model support.

Main Methods:

• A reference database of 36 VMA-SRT brain metastasis cases was established.
• A vector model was developed to retrieve similar past cases and their optimization parameters.
• Ten new VMA-SRT cases were planned using both conventional and vector-model-supported optimization.
• Planning time, number of iterations, and plan quality were statistically compared (Wilcoxon signed rank test, pFDR < 0.05).

Main Results:

• Vector-model-supported optimization significantly reduced median planning time by 40% (3.7 to 2.2 hours, p=0.002).
• The number of optimization iterations decreased by 30% (8.5 to 6.0, p=0.006).
• Plan quality was comparable between conventional and vector-model-supported VMA-SRT.

Conclusions:

• Vector-model-supported optimization effectively expedites the VMA-SRT planning process for brain metastases.
• This approach maintains comparable plan quality to conventional methods.
• The vector model shows promise for improving the clinical efficiency of VMA-SRT.