Skip to content

Commit

Permalink
Pages site from /docs
Browse files Browse the repository at this point in the history
  • Loading branch information
@mdevans
mdevans committed Jul 17, 2024
1 parent 9cca39d commit 85a5928
Show file tree
Hide file tree
Showing 39 changed files with 561 additions and 22 deletions.
61 changes: 61 additions & 0 deletions docs/1_overview.html
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,61 @@
<!DOCTYPE html>
<!-- List of HTML Tags supported by tkhtmlview:
see https://github.com/bauripalash/tkhtmlview?tab=readme-ov-file#html-support -->
<html>
<body>
<div style="color: #014F8F; font-size: 14px;">
<!-- <h1 style="text-align: center;">RSNA DICOM Anonymizer Overview</h1> -->

<h2>Introduction</h2>
<p>
The RSNA DICOM Anonymizer is a cross platform, stand-alone de-identification utility.
DICOM image files are imported either via network transfers or directly from the local file system.
When a DICOM file is received, it is de-identified and then stored in a user-defined local storage directory.
Anonymized images can be exported either to a remote DICOM server or an AWS S3 Object store such as the RSNA Imaging Archive.
</p>

<h2>Design Guidelines</h2>
<ol>
<li>Open Source</li>
<li>Stand-alone GUI written in Python 3</li>
<li>No internet connection required</li>
<li>Suitable for users with limited DICOM knowledge who are keen to curate and submit medical imaging datasets for AI model development</li>
<li>Platforms supported: Windows, OSX (Intel & ARM), Linux platforms</li>
<li>Ongoing development, build processes and release distribution via RSNA public github account</li>
<li>Backward compatible with Java version</li>
<li>PACS, Cloud PACS and VNA compatible</li>
<li>Reflect all errors to user, do not rely on log analysis</li>
<li>Project based with centralized settings management</li>
<li>Model, View, Controller (MVC) software architecture</li>
<li>Handle 1 million patients</li>
<li>Anonymize 10 files per second</li>
</ol>

<h2>Previous Java Anonymizer</h2>
<p>This version of the Anonymizer is based on the Medical Imaging Resource Centre (MIRC) Clinical Trial Processor (CTP)
<a href="https://mircwiki.rsna.org/index.php?title=The_MIRC_DICOM_Anonymizer">Java Anonymizer V16</a></p>
<h4>Differences from Java Version</h4>
<ol>
<li>Written in Python for Version 3.11</li>
<li>Implements operation subset to handle the default anonymizer script: {@remove, @empty, @hashdate, @round, ptid, acc, uid}</li>
<li>Fields with ptid in the operation will be replaced by a sequentially generated Patient ID starting at 1</li>
<li>Fields with acc in the operation will be replaced by a sequentially generated Accession Number starting at 1</li>
<li>Fields with uid in the operation will be replaced by a sequentially generated UID of form {UIDRoot.SiteID.N} N = 1,2,3...</li>
<li>No built-in editor for anonymizer script</li>
<li>No filter script, uses DICOM association contexts to restrict which storage classes & transfer syntaxes are imported</li>
<li>No internal DICOM image viewer</li>
<li>No internal log viewer</li>
<li>Project based: multiple separate projects with sand boxed settings can be created and managed on same machine</li>
<li>Project defines allowed modalities, storage classes, transfer syntaxes, network timeouts and logging levels</li>
<li>Simplified DICOM client & server configuration</li>
<li>Incoming files are not written to a temporary local disk directory for delayed processing; instead they are kept in memory queue.</li>
<li>Export to AWS S3 Object Store with AWS Cognito user authentication</li>
<li>Create Patient Lookup / Index in csv format with Number of Series and Number of Instances appended to each study row</li>
<li>Allows concurrent import and export processes</li>
<li>Provides a reliable study retrieval method at all levels (Study, Series, Instance) to handle remote DICOM servers which implement asynchronous C-MOVE (eg. VNA)</li>
<li>Fast in-memory datastore for Anonymizer Data Model</li>
</ol>

</div>
</body>
</html>
90 changes: 90 additions & 0 deletions docs/2_deidentification protocol.html
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,90 @@
<!DOCTYPE html>
<!-- List of HTML Tags supported by tkhtmlview:
see https://github.com/bauripalash/tkhtmlview?tab=readme-ov-file#html-support -->
<html>
<body>
<div style="color: #014F8F;font-size: 14px;">

<h3>DICOM Standard De-identification Protocol</h3>
<p>
In <a href="https://dicom.nema.org/medical/dicom/2023b/output/chtml/part15/chapter_E.html">PS 3.15 Appendix E</a>,
the DICOM standard specifies a set of de-identification rules for use in various situations.
The rules are grouped into a Basic Application Confidentiality Profile, which removes all protected health information (PHI),
and a set of options which relax certain classes of PHI that may be retained in certain cases.
The sections below describe the de-identification protocol implemented by the Anonymizer with reference to the DICOM Profile and its options.
</p>

<h4>113100: Basic Application Confidentiality Profile</h4>
<p>
The RSNA DICOM Anonymizer program implements the Basic Profile.
The program either imports the site identifier (SiteID and UIDRoot) from an Index Excel file exported by the Java Anonymizer or auto-generates a new unique SiteID and uses a default UIDRoot.
The PatientName (0010,0010) and PatientID (0010,0020) elements are modified by assigning sequential integers to unique PHI PatientIDs as they are encountered.
The sequential integer is appended to the SiteID parameter, producing a string in the form [SiteID]-[sequential integer], which is used to populate both elements.
If the PHI PatientID is missing or blank, the program assigns [SiteID]-000000 to the PatientID and PatientName elements.
The program maintains a database to ensure that all studies of the same patient are de-identified in the same way.
</p>

<ol>
<li>UIDs are replaced by those generated with the format: [UIDRoot].[SiteID].[sequential integer]</li>
<li>All UIDs are removed in groups.</li>
<li>All elements in groups 0032 through 4008 are removed.</li>
<li>All private groups are removed.</li>
<li>The PatientIdentityRemoved element (0012,0062) is assigned the value 'YES'.</li>
<li>The DeIdentificationMethod element (0012,0063) is assigned the value 'RSNA DICOM ANONYMIZER'.</li>
<li>113100 is appended to the DeIdentificationMethodCodeSequence element (0012,0064).</li>
<li>Private Block: 0x0013 is created and assigned the value "RSNA", the ProjectName and SiteID are added to this block.</li>
</ol>


<h4>113101: Clean Pixel Data Option</h4>
<p>Not implemented.</p>
<h4>113102: Clean Recognizable Visual Features Option</h4>
<p>Not implemented</p>
<h4>113103: Clean Graphics Option</h4>
<p>Not implemented but curve and overlay groups (5XXX and 6XXX) are removed.</p>
<h4>113104: Clean Structured Content Option</h4>
<p>Not implemented however project settings can allow/disallow SR (Structured Report) related storage classes.</p>
<h4>113105: Clean Descriptors Option</h4>
<p>This option is partially implemented. Only the following elements defined by the option are kept:</p>
<ul>
<li>(0008,1030) StudyDescription </li>
<li>(0008,103e) SeriesDescription</li>
</ul>
<h4>113106: Retain Longitudinal Temporal Information Full Dates Option</h4>
<p>Not Implemented</p>
<h4>113107 Retain Longitudinal Temporal Information Modified Dates Option</h4>
<p>This option is partially implemented.</p>
<ul>
<li>All dates are modified by calculating the MD5 hash of PHI PatientID.</li>
<li>The integer value of the hash is divided by the number of days in 10 years (3652) and the remainder is used to offset the date being modified.</li>
<li>This retains the temporal relationships between studies of a single patient while using different offsets for different patients.</li>
<li>Time (time of day) elements are not modified.</li>
<li>113107 is appended to the DeIdentificationMethodCodeSequence element (0012,0064).</li>
</ul>
<h4>113108: Retain Patient Characteristics Option</h4>
<p>This option is partially implemented. Only the following elements defined by the option are kept:</p>
<ul>
<li>(0010,0040) PatientSex</li>
<li>(0010,1010) PatientAge</li>
<li>(0010,1020) PatientSize</li>
<li>(0010,1030) PatientWeight</li>
<li>(0010,2160) EthnicGroup</li>
<li>(90010,21a0) SmokingStatus</li>
<li>113108 is appended to the DeIdentificationMethodCodeSequence element (0012,0064).</li>
</ul>
<h4>113109: Retain Device Identity Option</h4>
<p>This option is partially implemented. Only the following elements defined by the option are kept:</p>
<ul>
<li>(0008,0070) Manufacturer</li>
<li>(0008,1090) ManufacturerModelName</li>
<li>113109 is appended to the DeIdentificationMethodCodeSequence element (0012,0064).</li>
</ul>
<h4>113110: Retain UIDs Option</h4>
<p>Not Implemented</p>
<h4>113111: Retain Safe Private Option</h4>
<p>Not Implemented</p>


</div>
</body>
</html>
45 changes: 45 additions & 0 deletions docs/3_managing anonymization projects.html
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
<!DOCTYPE html>
<!-- List of HTML Tags supported by tkhtmlview:
see https://github.com/bauripalash/tkhtmlview?tab=readme-ov-file#html-support -->
<html>
<body>
<div style="color: #014F8F;font-size: 14px;">

<h2>Managing Anonymization Projects</h2>

<h3>Creating a New Project</h3>
<img src="assets/locales/en_US/html/images/NewProject.png"/>
<ul>
<li>From the Menu select <i>File/New Project</i> and the <i>New Project Settings</i></a> dialog will appear.</li>
<li>Refer to <i>Help/Project Settings</i> for full details.</li>
</ul>

<h3>Closing a Project</h3>
<img src="assets/locales/en_US/html/images/CloseProject.png"/>
<ul>
<li>From the Menu, select <i>File/Close Project</i> or simply close the window with the <b>X</b> button.</li>
</ul>

<h3>Re-opening a Project</h3>
<img src="assets/locales/en_US/html/images/OpenRecent.png"/>
<ul>
<li>From the Menu, select <i>File/Open Recent</i> to see a list of previously open projects.</li>
<li>Click on the project name to re-open it.</li>
<li>At startup the Anonymizer will re-open the last project if it was not closed when the Anonymizer was shutdown.</li>
</ul>

<h3>Cloning a Project</h3>
<img src="assets/locales/en_US/html/images/CloneProject.png"/>
<ul>
<li>From the Menu, select <i>File/Clone</i> to clone the current project's settings into a clean new project.</li>
<li>Select the new storage directory for the clone.</li>
<li>A <i>New Project Settings</i> dialog box is then presented allowing any setting editing, including new project name and UID Root.</li>
<li>Only the current project settings via the file <i>Project.pkl</i> will be copied over, no images files.</li>
<li>The Site ID will be preserved.</li>
<li><b>Warning:</b> ensure the UID Root is unique for all your projects to prevent UID clashes.</li>

</ul>

</div>
</body>
</html>
Loading

0 comments on commit 85a5928

Please sign in to comment.