MII IG Kerndatensatz-Modul Molekulargenetischer Befundbericht - Downloaded Version 2027.0.0-ballot.1 See the Directory of published versions

Migration STU2 to STU3

Migration from Clinical Genomics STU2 to STU3

​Version 2026.0.0 of this module moved from Clinical Genomics Reporting STU2 to STU3. This page states what changed and how to move existing data across. It addresses sites that already produce data against the 2025 line.

There are no StructureMaps here, deliberately. Data integration at a Data Integration Center runs on its own ETL, and a FHIR mapping engine is rarely part of it. A StructureMap would describe a transformation most sites could not execute. What follows is the mapping stated declaratively — implementable in whatever technology your pipeline already uses.

Profiles: what became of what

Note the two levels on the biomarkers: MSI and Mutationslast do not derive from the STU3 molecular-biomarker directly. They derive from this module's MII_PR_MolGen_MolekularerBiomarker, and that is what derives from STU3. The practical consequence is the same — see the category requirement below — but a profile binding written against the STU3 canonical will not match.

Components in Variante

Slice names changed; in an instance you identify a component by its code, so most of this is invisible in the data:

Three slices gained a representative- prefix — coding-hgvs, transcript-ref-seq and protein-hgvs. This matters only where you reference slices by name, for example in your own derived profiles or in FHIRPath; instance data is unaffected.

The added category on MSI and Mutationslast

Because both now sit under a biomarker profile, they require a second category alongside laboratory:

"category": [
  { "coding": [{ "system": "http://terminology.hl7.org/CodeSystem/observation-category",
                 "code": "laboratory", "display": "Laboratory" }] },
  { "coding": [{ "system": "http://hl7.org/fhir/uv/genomics-reporting/CodeSystem/tbd-codes-cs",
                 "code": "biomarker-category",
                 "display": "A characterization of a given biomarker observation." }] }
]

Note the second system: the biomarker category comes from the Clinical Genomics tbd-codes-cs, not from observation-category. STU3 requires both slices at 1..1, so an instance carrying only laboratory fails validation. This is the single most common failure when moving 2025-line data forward.

Path 1 — UntersuchteRegion becomes GenomicStudy

STU2 recorded which regions were examined as components on one Observation. STU3 replaces this with a Procedure pair: a GenomicStudy that orchestrates, and one or more GenomicStudyAnalysis steps carrying the detail.

Before:

{
  "resourceType": "Observation",
  "meta": { "profile": ["…/mii-pr-molgen-untersuchte-region"] },
  "component": [{
    "code": { "coding": [{ "code": "48018-6" }] },
    "valueCodeableConcept": { "coding": [{ "code": "HGNC:1097", "display": "BRAF" }] }
  }]
}

After — the studied region moves into the genomic-study-analysis-regions extension of an analysis step:

{
  "resourceType": "Procedure",
  "meta": { "profile": ["…/mii-pr-molgen-genomic-study-analysis"] },
  "extension": [{
    "url": ".../genomic-study-analysis-regions",
    "extension": [{
      "url": "studied",
      "valueCodeableConcept": { "coding": [{ "code": "HGNC:1097", "display": "BRAF" }] }
    }]
  }]
}

The minimum viable migration is exactly this: one analysis step per former UntersuchteRegion, carrying the same gene codes. Everything else GenomicStudyAnalysis offers — sequencer and pipeline as Device, method type, read depth and coverage, specimen, input and output files — is optional, and worth adding only where your source system actually holds it.

Note that regions-studied is what lets a query tell not found from not looked for. If your STU2 data recorded examined regions, carry them across; dropping them loses that distinction.

Path 2 — ErgebnisZusammenfassung becomes conclusionCode

Before: a separate Observation referenced from DiagnosticReport.result.

{ "resourceType": "DiagnosticReport",
  "result": [{ "reference": "Observation/ergebnis-zusammenfassung-1" }] }

{ "resourceType": "Observation", "id": "ergebnis-zusammenfassung-1",
  "valueCodeableConcept": { "coding": [{ "code": "LA6576-8", "display": "Positive" }] } }

After: the code moves onto the report itself and the Observation disappears.

{ "resourceType": "DiagnosticReport",
  "conclusionCode": [{ "coding": [{ "code": "LA6576-8", "display": "Positive" }] }] }

Remember to remove the now-dangling result entry, not just to add the conclusionCode.

Path 3 — renamed slices in Variante

Only relevant if you reference slice names. In instance data nothing changes, because components are identified by code:

Checklist

Before you start

• Pull the current package and read the Changelog for the version you are moving to.
• Establish which profiles you actually produce. Most sites use a fraction of the module.
• Find every instance carrying untersuchte-region or ergebnis-zusammenfassung in meta.profile — those are the two that cannot be migrated mechanically.

While migrating

• Add the second category to MSI and Mutationslast instances.
• Split each UntersuchteRegion into a GenomicStudy plus at least one GenomicStudyAnalysis, preserving the studied regions.
• Move each ErgebnisZusammenfassung value into DiagnosticReport.conclusionCode and delete the Observation and its result reference.
• Move any molecular consequence out of Variante into its own MolekulareKonsequenz.

Afterwards

• Validate against the published package; the profiles this module declares and the expectation attached to each are on the Capability Statements page.
• Search your output for remaining STU2 canonicals — untersuchte-region, ergebnis-zusammenfassung, region-studied — before declaring the migration done.