AI-Rad Companion Chest CT Extractor

This is a python library and command line tool for extracting the results of Siemens' AI-Rad companion Chest CT software into a relational database.
NOTICE: Command line tool only for use with an appropriate DicomConquest PACS

To install, simply use

pip install airc_extract

This will install the package and the two necessary command line programs to your python environment.

Usage

Python API

If you simply want to use the extraction method stored in the AIRCReport class, you can import and use it in any python script with

from airc_extract.airc_report import AIRCReport
# This is the list of structured report (SR) dicoms that make up the AIRC Chest CT output
report_dicoms = ['dcm1.dcm', 'dcm2.dcm', ...]
report = AIRCReport(report_dicoms)
report.extract_report()  # This is the method to pull the data
print(report.report_data)  # A dictionary containing the results

Command Line

If you are using a DicomConquest server, you can use the two command line tools provided in this package to automatically create and update a SQLite database, storing the AIRC results in 6 tables. More information on these tables can be found here.

To start, you will need to run airc-create-config.

airc-create-config \
--dicom-db /path/to/conquest.db3 \ # Required (usually dicomserver/data/dbase/conquest.db3)
--dicom-data-dir /path/to/conquest/data \ # Required (usually dicomserver/data)
--data-db /path/to/output/database.db3 \ # Required. Can NOT be on a network share.
--log-level-term INFO # Terminal logging level \
--log-level-file DEBUG # Log file logging level \
--log-dir . # Directory for log files to be stored

This will setup a package config.ini file that will be used for extraction, as well as test the connection to the DicomConquest database and the output database. It will create a new output database with all necessary tables at the specified path if it does not exist.

After this, just run

airc-extract

and all the series not found in the output database will be extracted and inserted into the output database.