Probabilistic Evaluation for Treatments Plans in Radiation Therapy

This repository is the base module for performing probabilistic evaluation of optimized treatment plans in radiation therapy. The goal is to provide a clear visualisation and analysis of uncertainties affect the dose distribution and the clinical goals. This library is not made for optimization purposes, but rather for evaluation of already optimized plans to allow clinicians to make better-informed decisions on trade-offs between organ-at-risk sparing and target coverage under uncertainty.

This repository is not intended for clinical use.

Installation

pip install RT-Prob-E

Optional: Streamlit-based UI tools

Two interactive, browser-based tools live in Probabilistic_Evaluation.ui and are available as an optional extra, since they depend on Streamlit (the results viewer also uses streamlit-aggrid for its tables):

pip install RT-Prob-E[ui]

Clinical goals editor — add, copy, delete and edit the goals in a clinical goals JSON file:

from Probabilistic_Evaluation.ui import edit_clinical_goals

edit_clinical_goals("clinical_goals.json")  # opens the editor in a browser tab

or, from a shell:

rt-eval-goals clinical_goals.json

Results viewer — browse a folder of probabilistic evaluation result JSON files, one browser tab per file, with a draggable, color-coded table of probabilistic objectives (drag rows to reorder, save/load that order) and a read-only table of non-probabilistic ones:

from Probabilistic_Evaluation.ui import view_results

view_results("Results")  # opens one tab per *.json file in that folder

or, from a shell:

rt-view-results Results

Calling any of these without the ui extra installed raises a clear error telling you to install it, rather than an import traceback.