Evaluation of dose accumulation methods and workflows utilising cone beam computed tomography images

Area of Science:

• Medical Physics
• Radiation Oncology

Background:

• Adaptive radiation therapy (ART) methods lack standardized protocols.
• Comparing dose mapping (DM) and Monte Carlo recalculation (MCR) for automated dose accumulation in ART is crucial.

Purpose of the Study:

• To compare the accuracy of DM versus MCR in automated ART dose accumulation workflows using cone beam computed tomography (CBCT) images.
• To evaluate the impact of deformable image registration (DIR) and workflow on dose accumulation accuracy.

Main Methods:

• Retrospective analysis of 38 cancer patient treatment plans (prostate and head and neck).
• Dose accumulation performed using DM and MCR on CBCT images, deformably registered to planning CT (DR-pCT).
• Comparison of DM and MCR doses using Intraclass Correlation (ICC) and Bland-Altman analysis.

Main Results:

• DM and MCR demonstrated good agreement (ICC > 0.9, within ±5% difference) on CBCT images.
• Agreement significantly decreased when doses were deformed back to the planning CT, indicating workflow-related inaccuracies.
• Over 67% of ICC values exceeded 0.9, suggesting good correlation between DM and MCR on CBCT.

Conclusions:

• Workflow adaptations are necessary to optimize algorithm performance in automated ART dose accumulation.
• Manual user intervention, improved protocols, and software enhancements are recommended to improve the reliability of ART dose accumulation.