TODO

CDISC ODM follow-up validation

The version-aware ODM 1.3 and 2.0 Snapshot pipeline is implemented for 0.34 with synthetic regression fixtures. Broaden the evidence before removing the experimental label from generic and OpenClinica ODM profiles.

• Add an attributed ODM 2.0 fixture from an official CDISC example and validate it against the corresponding XML schema
• Obtain an anonymized OpenClinica ODM 1.3 Full export with extensions, acknowledge its source, and compare its normalized records with the EDC data
• Exercise additional independently produced ODM snapshots and document any intentionally unsupported extension structures

cBioPortal follow-up

Clinical cBioPortal study input is implemented for 0.34 with directory and ZIP package support, entity-aware BFF output, optional Mapping V2 augmentation, and PXF and OMOP-CDM routes. The attributed DataHub fixture passes the official cBioPortal validator in no-portal mode, and generated target files pass the corresponding BFF and OMOP validators.

• Exercise independently produced study packages and incorporate differences in project-specific clinical attributes
• Add timeline data after defining mappings for relative dates, treatments, procedures, measurements, and specimen events
• Defer mutation, copy-number, expression, fusion, and structural-variant files until BFF genomicVariations, analyses, and runs are supported

mCODE and Dataset-XML follow-up

mCODE 4.0 profile detection and primary cancer stage mapping are implemented inside the FHIR R4 route for 0.34. CDISC Dataset-XML v1.0 input with required Define-XML v2 metadata is also implemented through the shared SDTM mapper.

• Exercise mCODE with independently generated oncology Bundles and assess additional first-class mappings only where BFF has an appropriate target
• Exercise Dataset-XML from multiple generators with Define-XML 2.0 and 2.1
• Evaluate a bounded-memory Dataset-XML reader if multi-million-row studies become a practical use case

Candidate inputs after 0.34

Demand-driven candidates

• SAS XPORT with Define-XML is relevant for regulatory datasets, but binary XPT parsing should use a maintained implementation rather than new parser code in Convert-Pheno
• C-CDA/CCD could supply patient summaries and discharge documents, but its narrative content and template ecosystem make it a high-complexity route
• PCORnet CDM is feasible but overlaps substantially with OMOP-CDM and has a more geographically concentrated user base
• HL7 v2, i2b2, and proprietary EDC APIs should require a concrete user, stable contract, and representative fixture before implementation

MEDIUM PRIORITY

• Pxf to BFF - genomicVariations

LOW PRIORITY

• range_high and range_low OMOP value?