This repository houses software to automatically evaluate Dicom data for different MRI Quality Assuracne protocols such as common ones by the Americal College of Radiology (ACR) to the fMRI QA protocol by Friedman and Glover (according to DOI 10.1002/jmri.20583). On top of this is the analysis for a custom protocol included, the so called OrQA protocol. The goal of designing this protocol is to provide a similar phantom to the ACR one but with a modular design and smaller size, since this protocol was designd with small pTx coils present e.g. at 7T scanners in mind.
The Open regular Quality Assurance (OrQA) Protocol is inspired by the ACR Protocol and comes with a custom, 3D-printed phantom. Key considerations are a matching size for all scanners, ease of use for operators in scanning and evaluating the data and furthermore modularity. The last point plays a key role, because it means that improvements can be made fast and without compromising the whole phantom, which will be saving money and especially printing time allowing for rapid prototyping and easy parts replacement.
The OrQA Phantom consists of a cylindrical case which measures 160mm in outer diameter and 142mm in height, not including the lid. The inner part of the phantom is 148mm in diameter and 132mm in height. The inner part houses four rails at the side of the walls. This case houses the modules and is filled with saline solution which is MR-visible. Earlier versions were filled with pure water but had problems with signal outages at 7T caused by suboptimal electric conductivity. The used materials used for the phantom are polyethylene terephthalate glycol (PETG) and polylactide (PLA) which are MR-opaque and have a similar susceptibility to water.
The software is written in Python and runs inside a Conda environment. The environment is provided under conda_env.yml and is created by this command
conda env create -f conda_env.yml
The environments name is karo and is activated by
conda activate karo
To run the software choose one of the start scripts in the directory.
Two filepaths are defined in the first lines:
- path_to_dicom is the path the dicom data is stored at.
- savedir is the path where the reports will be saved.
Two objects are initialized.
dicomFolderScanner prepares the Dicom data and produces a 4D array (time, slice, Y, X).
- Choose_scan_via_menu() shows menu and operator can choose a scan. With the flag True, the latest scan in the folder will be chosen automatically.
- dfs.list_scans can be used alternativaly to show a list of all scans inside a folder and a specific scan be used via dfs.choose_scan("SEQUENCE_NAME") .
- get_data() creates the 4D array (time, slice, y, x) with metadata stored at dfs.metadata .
{francis|glover|acr}Analyzer analyzed the dicom data according to the three proposed protocols *.
*The OrQA protocol is currently referred by the name Francis in the software for historical reasons
- runall runs all necessary methods in order to evaluate the data and outputting results in the following way:
