DVH Analytics: A DVH database for clinicians and researchers

Area of Science:

• Medical Physics
• Radiation Oncology
• Health Informatics

Background:

• Radiation therapy planning generates complex DICOM data.
• Analyzing this data is crucial for quality assurance and outcome prediction.
• Current tools may be vendor-specific or lack comprehensive analytical capabilities.

Purpose of the Study:

• To develop a vendor-agnostic, open-source software application for analyzing non-image-based DICOM RT files.
• To provide tools for managing, editing, viewing, and downloading radiation therapy planning data.
• To enable advanced analytics for plan evaluation, comparison, benchmarking, and outcome prediction.

Main Methods:

• Developed DVH Analytics using Python with libraries for DICOM parsing, database management (PostgreSQL), statistical analysis, and web UI creation.
• Imported DICOM RT Dose, RT Structure, and RT Plan files into a SQL database.
• Implemented analytical tools for DVH computation, plan evaluation, and predictive modeling.

Main Results:

• Successfully created a proof-of-concept database with over 3,000 DVHs.
• Observed significant differences in means, correlations, and temporal trends in organ-at-risk (OAR) doses.
• Demonstrated a predictive model for brainstem maximum dose based on planning parameters.

Conclusions:

• DVH Analytics is a free, open-source tool for parsing, organizing, and analyzing radiation oncology DICOM data.
• The software facilitates quality control and the development of statistical models for outcome prediction.
• This tool supports data-driven improvements in radiation therapy planning and patient care.