Installation

The package can be installed via pip:

pip install ethoPyRT

In addition, dcm2niix must be installed (see https://github.com/rordenlab/dcm2niix) and the executable must be visible from the python environment.

The pyradiomics (https://pyradiomics.readthedocs.io/en/latest/installation.html) executable must also be visible.

General remarks

This package helps analyze DICOM data exported from the Varian ETHOS system. The Ethos treatment interface allows per-session export of data, i.e.

<dcmpath>/<patname>/Session Export/<PID>/<TreatmentIntent>/Session_<n>

where dcmpath is manually selected, <patname> is manually created and the subfolders are automatically created during export.

Usage

General

Load the package with

import ethoPyRT

The ethoPyRT package provides the RTMetrics class, which is instantiated per patient using its ID pid (correspoding to above):

rtm = RTMetrics(dcmpath, basepath, pid, analysisDir)

dcmpath is the path where the exported data from Ethos is stored. basepath is a folder to store transformed intermediate data generated with ethoPyRT (i.e. nifti files, see below) and can be freely selected, as well as as <analysisDir>, where results are stored.

QC & Preprocessing

First, completeness of data is checked via

rtm.checkFiles() 

File conversions to help manual inspection of data is performed with

rtm.createNifti()

and RT structures are extracted with

rtm.extractRTStruct()

Data is stored in

`<basepath>/<TreatmentIntent>/<PID>/Session_<n>/`

in the nifti/ and RTSTRUCT/ subfolders.

Data loading

RT doses are loaded with

rtm.loadDoses()

and plans are assigned (scheduled / adapted) with

rtm.assignPlanType()

RTStructures are loaded with

rtm.loadMasks()

Data analysis

DVHs are calculated with

rtm.calcDVH()

Scheduled vs adapted vs treated plans are compared using

rtm.compare()

Difference DVHs are computed with

rtm.diffDVH()

and plotted with

rtm.plotDVH()

Dose metrics are calculated with

rtm.calcMetrics()

Radiomics features are computed with

rtm.calcRadiomics()

The number of adapted / scheduled plans.

rtm.countAdpSched()

Registration of CBCTs (currently only tested for prostate)

rtm.registerCBCT()

and application:

rtm.applyTransform(type="CBCT", ref="CBCT")