RadOnc: An R Package for Analysis of Dose-Volume Histogram and Three-Dimensional Structural Data

Abstract

Purpose/Objectives: Dose volume histogram (DVH) data are generally analyzed within the context of a treatment planning system (TPS) on a per-patient basis, with evaluation of single-plan or comparative dose distributions. However, TPS software generally cannot perform simultaneous comparative dosimetry among a cohort of patients. The same limitations apply to parallel analyses of three-dimensional structures and other clinical data. Materials/Methods: We developed a suite of tools ("RadOnc" package) using R statistical software to better compare pooled DVH data and empower analysis of structure data and clinical correlates. Representative patient data were identified among previously analyzed adult (n=13) and pediatric (n=1) cohorts and these data were used to demonstrate the performance and functionality of the RadOnc package. Results: The RadOnc package facilitates DVH data import from the TPS and includes automated methods for DVH visualization, dosimetric parameter extraction, statistical comparison among multiple DVHs, basic three-dimensional structural processing, and visualization tools to enable customizable production of publication-quality images. Conclusions: The RadOnc package provides a potent clinical research tool with the ability to integrate robust statistical software and dosimetric data from cohorts of patients. It is made freely available to the community for their current use and remains under active development.