Helmut Koenig
Co-Chair DICOM WG20
• Tagging text contents
• Mechanism specified by CDA R2 and used in
CDA Diagnostic Imaging Report IG
– Content ID in narrative block is referenced by structured entry
– Structured entry includes code to define semantics of tagged content (Observation.code,
Observation.value Concept Descriptor used depending on type of observation)
<section>
…
<text>
Narrative Block
Content Identifier
< content ID="F1" >Hilar mass …<content>
</text>
<entry>
<observation classCode="OBS" moodCode="EVN">
…
<originalText>
< reference value="#F1" />
</originalText>
…
</observation>
…
</section>
Structured Entry
Reference
Findings in RSNA Templates can be represented by the following types of CDA observation events:
• Quantity Measurement Observation
• Coded Observation
<observation classCode="OBS" moodCode="EVN">
<templateId root=" 2.16.840.1.113883.10.20.6.2.14
"/>
<code code="246120007" codeSystem="2.16.840.1.113883.6.96" codeSystemName="SNOMED" displayName="Nodule size">
Registered Quantity
Measurement Obs.
Template
<originalText>
<reference value="#Diam2"/>
</originalText>
</code>
The unit of measure is specified in the Unified Code for Units of
<effectiveTime value="20060823223912"/>
Measure (UCUM)
<value xsi:type="PQ" value="45" unit="mm" >
<translation code="mm" codeSystem="2.16.840.1.113883.6.8" codeSystemName="UCUM" codeSystemVersion="1.5"/>
</value>
…
</observation>
• Observation.code used to convey coded concept
• Observation.value uses Physical Quantity (PQ) Data Type
<observation classCode="OBS" moodCode="EVN">
…
<code code="ASSERTION" codeSystem="2.16.840.1.113883.5.4"/>
<originalText>
<reference value="#F1"/>
</originalText>
SNOMED CT
<value xsi:type="CD" code="309530007” codeSystem="2.16.840.1.113883.6.96“>
<displayName value="Hilar Mass"/>
</value>
…
</observation>
• „Assertion“ used as Observation.code
• Observation.value is the place for coded concept
• CD (Concept Descriptor) data type allows for use of multiple coding systems (translation codes) and post-coordination of codes
• We need to define the scope of coded observations and get consensus on essential patterns needed in that context
• We should avoid post-coordination of codes wherever possible
– RadLex compositional grammar?
• Allow for capturing of temporal aspects, i.e. clinically or operationally relevant time of observations (effectiveTime)