Iterative dataset optimization in automated planning: Implementation for breast and rectal cancer radiotherapy

Area of Science:

• Radiation Oncology
• Medical Physics
• Computational Biology

Background:

• Radiotherapy planning is complex and time-consuming.
• Automated solutions can improve efficiency and consistency.
• Accurate prediction of dose distributions is crucial for optimal treatment.

Purpose of the Study:

• To develop an automated treatment planning solution for breast and rectal cancer radiotherapy.
• To improve the accuracy of dose volume histogram (DVH) prediction for organs at risk (OARs).
• To automatically generate clinically acceptable and high-quality treatment plans.

Main Methods:

• Iterative optimization of a training dataset from 40 patient plans.
• Utilizing a two-dimensional kernel density estimation (KDE) algorithm for DVH prediction.
• Employing Pinnacle^3 Auto-Planning module with predicted DVH-derived objectives for re-planning.

Main Results:

• The automated strategy, combining iterative dataset selection and KDE, enhances DVH prediction accuracy.
• Automatically generated plans achieve improved OAR dose sparing without compromising overall plan quality.
• Comparison with manually optimized plans demonstrates the effectiveness of the automated approach.

Conclusions:

• The developed automated treatment planning solution efficiently evaluates and enhances radiotherapy plan quality.
• This solution is applicable to intensity-modulated radiation therapy for breast and rectal cancers.
• It offers potential for improved consistency and quality in clinical radiotherapy planning.