What problem does it solve?
This Skill extracts clinically meaningful report text and study metadata from DICOM files while identifying header attributes that may contain protected health information. It helps prepare imaging data for safe clinical NLP without exposing patient identifiers or overlooking PHI in structured reports and free-text tags.
Core Features & Use Cases
- DICOM Header Inspection: Reads study, series, patient, equipment, and free-text metadata with pydicom while reporting PHI-bearing tags without logging their values.
- DICOM-SR Text Extraction: Recursively traverses Structured Report content trees to reconstruct narrative text, measurements, and relevant coded findings.
- OpenMed Handoff: Sends extracted narrative through OpenMed de-identification and clinical text analysis while keeping DICOM header scrubbing as a separate standards-based step.
- Use Case: When ingesting radiology reports from a PACS, extract the SR impression, flag identifiers such as PatientName and AccessionNumber, de-identify the narrative, and analyze it with OpenMed before exporting sanitized data.
Quick Start
Use the extracting-dicom-metadata skill to read the attached DICOM file, extract its report text, identify present PHI tags, and prepare the narrative for OpenMed de-identification and analysis.
