Pathology Report Structuring Agent
Converts free-text pathology reports into structured findings for oncology pathways and tumor boards.
Pathology reports are rich in clinical detail but often free-text, making it hard to feed oncology pathways, tumor boards, registries, and decision support
Manual abstraction of diagnosis, grade, stage elements, biomarkers, and margins is slow and inconsistent
Downstream teams re-read reports to find key fields, delaying treatment planning
The Pathology Report Structuring Agent converts free-text pathology reports into structured findings so oncology pathways and tumor boards can act on complete, computable data
The agent ingests final pathology reports, extracts and normalizes key diagnostic and biomarker fields using NLP and ontology mapping, validates required elements for cancer pathways, and publishes structured outputs to EHR oncology modules, tumor board packets, and registries.
Ingest Pathology Reports
- Retrieve final and amended surgical pathology and cytology reports
- Identify report sections (diagnosis, microscopic, synoptic, IHC/molecular)
- Link reports to specimen, encounter, and oncology episode
Extract Structured Findings
- Extract primary diagnosis, histology, grade, and laterality
- Capture margins, lymph nodes, pTNM elements, and specimen adequacy
- Pull biomarkers and molecular results (e.g., ER/PR/HER2, EGFR, MSI)
Normalize and Validate
- Map terms to SNOMED, ICD-O, and CAP synoptic standards
- Validate completeness against cancer pathway and registry requirements
- Flag ambiguities or missing elements for pathologist review
Publish for Pathways and Tumor Boards
- Write structured findings to EHR oncology and discrete data fields
- Assemble tumor board summaries with key pathology highlights
- Feed cancer registry and clinical trial matching workflows