Public roadmap
Building the Evidence Atlas in public
The system is designed for continuity: automated ingestion, open governance, and replaceable AI modules. This roadmap will update as milestones ship.
Non-negotiable principles
- No patient-identifiable data in early versions.
- Evidence labeled: primary source vs derived vs AI interpretation.
- Nightly automation with documented operations and backups.
- Replaceable AI modules (no single-vendor lock-in).
Phase 0 — Foundations
- Clean schema aligned to ingest code
- ClinicalTrials.gov ingest + change detection
- Trials explorer with stable filters
- Updates feed + cron scheduling
Phase 1 — Evidence types
- PubMed ingest and paper pages
- Drug + biomarker hubs
- Cancer hub pages with linked objects
Phase 2 — Guidelines
- Guideline links (NCCN/ESMO/ASCO) with licensing care
- Change summaries tied to cancer hubs
Phase 3 — Matching + discovery
- Stage / line-of-therapy cues
- Geographic proximity search
- Relevance ranking improvements
Phase 4 — Community participation
- Structured contributions and feedback
- Moderation and provenance tracking
Phase 5 — Research acceleration
- Gap maps and priority lists
- State-of-the-field syntheses
- Funder-ready briefs