Improving workflow for adaptive proton therapy with predictive anatomical modelling: A proof of concept

Area of Science:

• Radiation Oncology
• Medical Physics
• Cancer Treatment

Background:

• Adaptive intensity-modulated proton therapy (IMPT) is crucial for head and neck cancer treatment.
• Anatomical changes during radiotherapy can compromise treatment efficacy.
• Predictive modeling offers a potential solution to proactively adapt treatment plans.

Purpose of the Study:

• To evaluate a predictive anatomical modeling approach for enhancing the clinical workflow of adaptive IMPT.
• To compare the dosimetry outcomes of predicted adaptive plans versus standard reactive replanning strategies.

Main Methods:

• A retrospective study included 10 nasopharyngeal cancer patients undergoing IMPT.
• Predicted weekly CTs from an anatomical model were generated.
• Two adaptive strategies were compared: predicted plan adaptation and standard reactive replanning, with adaptation triggered by a >3 Gy(RBE) mean parotid dose increase.

Main Results:

• No statistically significant differences in accumulated dose were found between predicted and standard adaptive IMPT strategies for CTVs and OARs (p > 0.05).
• Mean differences in D95 for CTVs were minimal across no adaptation, standard, and predicted plan adaptation (-1.20% to -1.25%).
• Predicted plan adaptation resulted in a slightly lower average accumulated parotid mean dose compared to standard replanning (0.31 Gy(RBE) lower).

Conclusions:

• Prediction-based replanning shows potential for seamless adaptive therapy delivery, avoiding treatment gaps.
• While not statistically significant in this study, predictive adaptation may improve clinical workflow efficiency.
• Further research is warranted to fully elucidate the benefits of predictive adaptive IMPT.