HIC.BadMedicine.Dicom 0.1.0

dotnet add package HIC.BadMedicine.Dicom --version 0.1.0                
NuGet\Install-Package HIC.BadMedicine.Dicom -Version 0.1.0                
This command is intended to be used within the Package Manager Console in Visual Studio, as it uses the NuGet module's version of Install-Package.
<PackageReference Include="HIC.BadMedicine.Dicom" Version="0.1.0" />                
For projects that support PackageReference, copy this XML node into the project file to reference the package.
paket add HIC.BadMedicine.Dicom --version 0.1.0                
#r "nuget: HIC.BadMedicine.Dicom, 0.1.0"                
#r directive can be used in F# Interactive and Polyglot Notebooks. Copy this into the interactive tool or source code of the script to reference the package.
// Install HIC.BadMedicine.Dicom as a Cake Addin
#addin nuget:?package=HIC.BadMedicine.Dicom&version=0.1.0

// Install HIC.BadMedicine.Dicom as a Cake Tool
#tool nuget:?package=HIC.BadMedicine.Dicom&version=0.1.0                

BadMedicine.Dicom

NuGet Badge Build, test and package CodeQL

The purpose of BadMedicine.Dicom is to generate large volumes of complex (in terms of tags) dicom images for integration/stress testing ETL and image management tools.

There are a number of public sources of Dicom clinical images e.g. TCIA . The difficulty with using these for integration/stress testing is that they often:

  • Are anonymised (majority of tags have been removed)
  • Do not represent the breadth of Modalities/Tags found in a live clinical PACS.
  • Take up a lot of space

BadMedicine.Dicom generates dicom images on demand based on an anonymous aggregate model of tag data found in scottish medical imaging. It is an extension of SynthEHR which generates traditional EHR records.

Usage

BadDicom is available as a nuget package for linking as a library

The standalone CLI (BadDicom.exe) is available in the releases section of Github

Usage is as follows:

BadDicom.exe c:\temp\testdicoms

Generates 10 dicom studies (~700MB)

BadDicom.exe c:\temp\testdicoms 5 10 --NoPixels

Generates 10 dicom studies from a pool of 5 patients without pixel data (~3MB)

You can pass -s to seed the random number generators. Seeding will ensure the same StudyDate, PatientId etc get generated but will not affect UIDs generated (UIDs are always unique)

BadDicom.exe c:\temp\testdicoms 5 10 --NoPixels -s 100

Direct to Database

You can generate DICOM metadata directly into a relational database (instead of onto disk). This can be done by downloading an image template or by creating one yourself.

To turn this mode on rename the file BadDicom.template.yaml to BadDicom.yaml and provide the connection strings to your database e.g.:

Database:
  # The connection string to your database
  ConnectionString: server=127.0.0.1;Uid=root;Pwd=;Ssl-Mode=None
  # Your DBMS provider ('MySql', 'PostgreSql','Oracle' or 'MicrosoftSQL')
  DatabaseType: MySql
  # Contains the table schema (which dicom tags to use for which tables)
  Template: CT.it
  # Database to create/use on the server
  DatabaseName: SynthEHRTestData
  # Setting this deduplicates study/series level schemas (works only if tables do not already exist on server)
  MakeDistinct: true

EHR Datasets

If you want to generate EHR datasets with a shared patient pool with the dicom data (e.g. for doing linkage) you can provided the -s (seed) and use the main SynthEHR application.

BadDicom.exe c:\temp\testdicoms 12 10 -s 100
SynthEHR.exe c:\temp\testdicoms 12 100 -s 100

Generates a pool of 12 patients and 10 Studies (for random patients) then 100 rows of data for each EHR dataset

Library Usage

You can generate test data for your program yourself by referencing the nuget package:

//create a test person
var r = new Random(23);
var person = new Person(r);

//create a generator 
using (var generator = new DicomDataGenerator(r, null, "CT"))
{
    //create a dataset in memory
    DicomDataset dataset = generator.GenerateTestDataset(person, r);

    //values should match the patient details
    Assert.AreEqual(person.CHI,dataset.GetValue<string>(DicomTag.PatientID,0));
    Assert.GreaterOrEqual(dataset.GetValue<DateTime>(DicomTag.StudyDate,0),person.DateOfBirth);

    //should have a study description
    Assert.IsNotNull(dataset.GetValue<string>(DicomTag.StudyDescription,0));   
}

Building

Building requires MSBuild 15 or later (or Visual Studio 2017 or later). You will also need to install the DotNetCore 2.2 SDK.

Csproj files are in the 2017 format and require Visual Studio 2017 or later to run. The following projects are part of the solution:

Project Runtime Purpose Build Output
BadDicom.csproj Dot Net Core 2.2 Command Line Tool BadDicom.exe generated by dotnet publish*
BadMedicine.Dicom.csproj Dot Net Standard 2.0 Library BadMedicine.Dicom.dll. Use BadMedicine.Dicom.nuspec to upload to NuGet
BadMedicine.Dicom.Tests Dot Net Core 2.2 Tests for library None

*Publish an OS specific binary by building BadDicom.csproj then running:

dotnet publish BadDicom.csproj -r win-x64 --self-contained
cd .\bin\Debug\netcoreapp2.2\win-x64\

For Linux, a few extra requirements are needed:

# For Ubuntu
$ sudo apt install libc6-dev libgdiplus

# For CentOS
$ sudo yum install libc6-devel libgdiplus

Tag Data

Basic random patient information (age, CHI etc) are generated by SynthEHR

The following tags are populated in dicom files generated:

Tag Model
PatientID The CHI number of the (random) patient
StudyDate A random date during the patients lifetime
StudyTime Random time of day with hours favoured during the middle of the working day.
SeriesDate Same as StudyDate*
PatientAge Age of patient at SeriesDate e.g. "032Y"
Modality Random modality for which we have at least 1 image locally. (proportionate to modality popularity)
StudyDescription Random description that exists in the modality (proportionate to frequency seen)

*SeriesDate is always the same as Study Date (see Seres constructor), for secondary capture this should/could not be the case (we should look at how this corresponds in the PACS data we have)

Pixel Data

Currently pixel data is written as a black square with the SOP Instance UID written in white.

Product Compatible and additional computed target framework versions.
.NET net8.0 is compatible.  net8.0-android was computed.  net8.0-browser was computed.  net8.0-ios was computed.  net8.0-maccatalyst was computed.  net8.0-macos was computed.  net8.0-tvos was computed.  net8.0-windows was computed. 
Compatible target framework(s)
Included target framework(s) (in package)
Learn more about Target Frameworks and .NET Standard.

NuGet packages

This package is not used by any NuGet packages.

GitHub repositories

This package is not used by any popular GitHub repositories.

Version Downloads Last updated
0.1.0 851 5/30/2024
0.0.16 4,173 10/4/2023
0.0.16-rc1 563 3/16/2023
0.0.15 4,944 10/31/2022
0.0.14 1,738 7/11/2022
0.0.13 839 6/2/2022
0.0.12 695 5/18/2022
0.0.11 864 3/29/2022
0.0.10 1,462 2/17/2022
0.0.9 4,519 3/3/2021
0.0.8 345 3/2/2021
0.0.7 2,543 8/18/2020
0.0.6 1,694 5/20/2020
0.0.5 780 1/31/2020
0.0.4 1,301 10/28/2019
0.0.3 494 10/28/2019