DICOM PS3 2019c - Release Notes​ Page 1​

2019/06/14​

The following changes have been made relative to the previously published PS3 2019b release of the standard, by incorporating the​
changes specified in the supplements and correction items.​

The Final Text of all applied Supplements and Correction Proposals is available at ftp://medical.nema.org/medical/dicom/final/​

Production Notes​
The DocBook XML files are the source format, and all other formats are rendered from it.​

The PDF format is rendered from the DocBook XML, and remains the "official" (authoritative) form of the standard. The PDF contains​
hyperlinks to sections, figures and tables both within and between parts (which in the latter case work if you are reading the PDF in​
a tool that supports linking to other parts).​

The two HTML formats are provided for the convenience of those who find them easier to navigate within a browser, and though the​
appearance and organization is different, the content is the same. One form consists of entire parts in one very large HTML page,​
and the other consist of chunks of sections with navigation elements. Both forms are hyper-linked within and between parts. The figures​
in the HTML are SVG, so a browser that supports SVG is required (most contemporary browsers do).​

All paragraphs (<p/> elements) in the HTML files of this release, are uniquely identified with a hypertext anchor (<a/> element), each​
of which has an id attribute (derived from the source DocBook <para/> element xml:id attribute). These unique identifiers will remain​
stable in subsequent releases, so they may be reliably used as the persistent targets of hyperlinks relative to the current release base​
URL, and are more specific than the existing anchors for entire sections or tables. Unlike the section and table anchors, there is no​
semantic significance to the syntax of the identifiers (i.e., they are UUIDs, rather than being derived from the section or table numbering​
pattern). Subsequent releases will add new identifiers for new paragraphs and text split out of existing paragraphs into new paragraphs,​
and will, if possible, empty, rather than entirely remove, existing paragraphs that are retired (in order to avoid dead links).​

The chunked HTML format includes navigation elements in the header and footer, as well as a hyperlink to the current release of that​
page, in case the user happens to find or be using an older release of the page.​

The DOCX (for Word) and ODT (for OpenOffice or LibreOffice) formats are provided for the convenience of future Supplement and​
CP editors. Their main claim to fame is that they exist at all, and though they are viewable and editable, they are lacking many features​
of the Word source of previous release, for example the use of styles for section headings. They do contain embedded hyperlinks,​
and these are also present in the table of contents, even though the page numbers rendered in the table of contents may be mean-​
ingless. To reiterate, the intent of these files is to provide a source to cut and past into new Word documents, and not to be functional​
documents in their own right. Since Word does not support SVG, all figures embedded in the DOCX files have been rasterized to a​
fixed resolution and are adequate for position only and are not editable and are not intended to be a substitute for the SVG figures.​

The rendering pipeline used to produce these files is available but requires some expertise to use it. It is not supported. To achieve​
quality rendering, the use of some commercial tools was necessary, to supplement the many open source tools that were also used.​
Oxygen (commercial) was used as the XML editor since it supports a WYSIWG authoring mode. OpenOffice (open source) was used​
as the equation editor. The DocBook (open source, version docbook-xsl-ns-1.78.1) style sheets were used to create the HTML and​
intermediate FO form used to created the PDF and DOCX. MathML equations were converted to SVG using pMML2SVG (open​
source, version pMML2SVG-0.8.5). RenderX XEP (commercial) was used to produce the PDF, and XMLmind FO-Converter (com-​
mercial) was used to produce the DOCX. The difference files were produced using DeltaXML DocBook Compare (commercial). The​
PDF files were post-processed with qpdf to generate object streams to reduce the size of the tagged PDF and improve searching for​
strings that span lines within tables and to linearize the files for streamed web page viewing.​

Some characteristics of the DocBook XML may be of interest to those performing automated processing or extraction:​

•​ Zero width spaces (U+200B) are used in some places to allow long words (such as PS3.6 keywords and UIDs) to break within table​
columns and avoid tables becoming too wide to fit on a page. These need to be filtered out before using these words literally.​

•​ Enumerated values and defined terms are formalized in PS3.3 as DocBook variablelist elements with a title identifying them as​
such, to facilitate their automated detection and extraction.​

•​ Template and context group tables in PS 3.16 are preceded by variablelist elements defining whether or not they are extensible,​
etc., again to enable automated extraction.​

- Standard -​
 DICOM PS3 2019c - Release Notes​ Page 2​

•​ Hyperlinks (xref and link elements) are used extensively but may obscure the identifier of what is being linked to from the perspective​
of automated extraction. It may be useful to consult the olink targetdb files that are included in the package to "look up" the target​
of such links, rather than reinventing this mechanism, which is used by the DocBook stylesheets for cross-document linking. E.g.,​
one can look up "sect_TID_300" in "output/html/targetdb/PS3_16_target.db" to determine that it has a "number" of "TID 300" and​
a "ttl" of "Measurement", etc.​

Changes to Parts​
General Changes​
•​

PS3.1​
PS3.2​
•​ Remove reference to 3.0 version of DICOM​

PS3.3​
•​ Correct Projection Calibration figure titles​

•​ Make mouse strain example consistent wrt. presence of data elements for code sequence items​

•​ CP 1839​

•​ CP 1854​

•​ CP 1856​

•​ CP 1864​

PS3.4​
•​ Update incorrect links to PS3.18 caused by Sup 183​

PS3.5​
•​ Remove reference to 3.0 version of DICOM​

•​ Make consistent definition of which private groups are not allowed (include FFFF) Sections 7.1 and 7.8​

PS3.6​
•​ CP 1858​

PS3.7​
PS3.8​
PS3.10​
PS3.11​
PS3.12​
•​ Remove reference to 3.0 version of DICOM​

- Standard -​
 DICOM PS3 2019c - Release Notes​ Page 3​

PS3.14​
PS3.15​
•​ CP 1856​

•​ CP 1860​

PS3.16​
•​ Make TID 15400 Row 1 DT rather than EV per CP 1835 FT​

•​ Update CID 33 table title and leading paragraph and note to match CP 1853​

•​ CP 1856​

•​ CP 1857​

•​ CP 1858​

•​ CP 1859​

•​ CP 1874​

PS3.17​
PS3.18​
•​ Reapply CP 1816 that was not included in Sup 183​

PS3.19​
PS3.20​
PS3.21​

Supplements Incorporated​
Correction Items Incorporated​
CP 1839​ Incorrect Ophthalmic QC Reference tag​

CP 1854​ Change Visual Field Test Point Normals Sequence’s Type from 1C to 2C​

CP 1856​ Unique device identification consistency and de-identification​

CP 1857​ Factor out algorithm identification common to multiple observations​

CP 1858​ Allow single modifier for categorical observations in TID 1500​

CP 1859​ Allow single modifier for categorical observations in TID 1500​

CP 1860​ Content items for de-identification with Clean Structured Content Option​

CP 1864​ Add Anatomic Region Sequence to RT Structure Set​

CP 1874​ Modality-specific Image Library Entry Descriptors should be mutually exclusive​

CP 1875​ Two Dimensional Measurement Graph in SR has no concept for data values​

- Standard -​