A Visualization and Radiation Treatment Plan Quality Scoring Method for Triage in a Population-Based Context.

Area of Science:

• Medical Physics
• Radiation Oncology
• Health Informatics

Background:

• Radiation treatment planning is complex, requiring objective quality assessment.
• Current methods for evaluating treatment plan quality can be subjective and time-consuming.
• A need exists for an intuitive and statistically sound system to evaluate radiation therapy plans.

Purpose of the Study:

• To develop a clinically intuitive scoring method for radiation treatment plan quality.
• To utilize statistical metrics for visual determination of treatment plan quality.
• To establish a percentile-based scoring system for objective evaluation.

Main Methods:

• Developed a percentile-based scoring system using data from 111 head and neck cancer patients.
• Visualized scores using a daisy plot for plan-by-plan and objective-by-objective evaluation.
• Validated the system with 6 physicians assessing 60 plans using a Likert scale and Spearman correlation, ROC analysis.

Main Results:

• Clinically acceptable plans (n=8) scored at the 89.0th ± 14.5 percentile.
• Plans needing edits or unacceptable had lower percentile scores (62.6th ± 25.1 and 35.6th ± 25.7, respectively).
• Spearman correlation (0.53, P < .001) and ROC analysis (AUC=0.76) confirmed scoring system's validity and ability to discern plan quality.

Conclusions:

• The developed scoring system provides intuitive visual feedback on radiation treatment plan quality.
• The system demonstrates a statistically significant correlation with physician ratings.
• This method offers utility in the quality assurance process for radiation therapy plans.