Evaluation of an Automated Proton Planning Solution

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• Intensity-modulated proton therapy (IMPT) is increasingly used but faces challenges in treatment planning complexity and variability.
• Automated knowledge-based planning (KBP) solutions aim to improve efficiency and consistency in complex treatment modalities.

Purpose of the Study:

• To detail and evaluate RapidPlan™PT, a pre-clinical, proton-specific, automated knowledge-based planning solution.
• To assess the efficiency and plan quality of RapidPlan™PT-generated IMPT plans compared to manual IMPT plans.

Main Methods:

• A RapidPlan™PT model was built using IMPT plans from 50 head-and-neck cancer patients.
• The model generated knowledge-based plans (KBPs) for 10 evaluation patients.
• Model accuracy was assessed via OAR regression plots and predicted vs. achieved mean dose differences; KBP quality was compared to manual plans for PTV homogeneity and OAR sparing.

Main Results:

• The model demonstrated good quality (average R² of 0.85) and high predictive accuracy for OAR mean dose (<3 Gy difference for 88/109 OARs).
• KBPs were comparable to manual IMPT plans, with similar PTV homogeneity (<0.6% difference) and OAR sparing (only 2/109 OARs >3 Gy higher mean dose).
• KBP generation was efficient, with DVH prediction at 0.7 minutes and optimization/calculation at 4.1 minutes.

Conclusions:

• RapidPlan™PT offers an efficient and consistent approach to IMPT, producing plans comparable to manual methods.
• While automated planning shows promise, manual plan checks remain crucial, especially for OAR sparing.
• This automated solution can aid clinical trials and streamline IMPT learning curves.