Built-in wavelet-induced smoothness to reduce plan complexity in intensity modulated radiation therapy (IMRT)

Area of Science:

• Medical Physics
• Radiotherapy Planning
• Image Processing

Background:

• Intensity modulated radiation therapy (IMRT) planning aims to optimize radiation dose delivery while minimizing toxicity.
• Current IMRT planning methods often result in complex intensity maps, impacting treatment delivery efficiency and accuracy.
• Achieving fluence smoothness typically relies on implicit soft constraints, requiring careful weight tuning.

Purpose of the Study:

• To develop and evaluate a novel approach for IMRT plan complexity reduction using an incomplete wavelet basis.
• To improve dosimetric accuracy and delivery efficiency in IMRT by introducing built-in smoothness.
• To mitigate the challenges associated with fine-tuning regularization weights in treatment planning.

Main Methods:

• Representing beamlet intensities using an incomplete wavelet basis, excluding high-frequency components in the leaf travel direction (X-direction).
• Utilizing an explicit hard constraint to enforce wavelet-induced smoothness, complemented by a regularization term for Y-direction smoothness.
• Testing the method on paraspinal, lung, and prostate cancer cases using the Varian Eclipse treatment planning system.

Main Results:

• Significant reduction in monitor units (MUs) by 10% (paraspinal), 45% (lung), and 14% (prostate).
• Improved organ-at-risk sparing, with reduced mean doses to the esophagus (7%), lung (13%), and cord (14%) in the paraspinal case.
• Enhanced robustness to regularization weight variations, simplifying the planning process.

Conclusions:

• The proposed wavelet-based method effectively reduces IMRT plan complexity and improves plan quality.
• Explicitly imposing smoothness constraints via wavelets offers a robust alternative to traditional regularization techniques.
• This approach has the potential to enhance the clinical applicability and efficiency of IMRT planning.