This case study demonstrates how a radiology clinic can leverage Yoctobe’s AI-powered integration engine to efficiently manage medical images, extract patient metadata using DICOM, and integrate this information with their Radiology Information System (RIS) and local storage.

Requirements

1. Extract images and patient metadata from various imaging devices using DICOM
2. Organize and save images in a locally shared directory (intranet)
3. Register metadata and image locations in a MySQL database
4. Integrate data with the Radiology Information System (RIS)
5. Ensure efficiency, accuracy, compliance, security, and real-time data availability

System Setup

1. Imaging Devices: Multiple DICOM-compatible imaging devices (e.g., CT scanners, MRI machines, X-ray systems) are connected to the clinic’s secure network.
2. Yoctobe Integration Engine: Deployed on a dedicated server or cloud instance with high availability and scalability features.
3. Local Shared Storage: A high-performance Network Attached Storage (NAS) system accessible via the clinic’s intranet.
4. MySQL Database: A robust, properly sized MySQL server to handle concurrent connections and transactions.
5. Radiology Information System (RIS): The clinic’s existing RIS, accessible via API endpoints.

Real-time Data Flow

1. Image Acquisition

• A patient undergoes an imaging procedure on one of the clinic’s DICOM-compatible devices.
• The imaging device generates a DICOM file containing both the image data and associated metadata.

2. DICOM Transmission

• The imaging device immediately initiates a DICOM association with the Yoctobe Integration Engine.
• The DICOM file is transmitted in real-time over the secure clinic network to the integration engine.

3. Data Reception and Validation

• Yoctobe’s Integration Engine receives the incoming DICOM data stream.
• The engine performs immediate validation checks:
  • Verifies the DICOM file structure
  • Checks for required metadata fields
  • Validates data types and formats

4. Image Extraction and Storage

• Upon successful validation, the integration engine extracts the image data from the DICOM file.
• Simultaneously, it generates a unique filename based on patient ID, study date, and modality.
• The image is immediately written to the appropriate directory in the local shared storage.
  • Example path: /shared_storage/patientID_12345/20230724/CT/image_001.dcm

5. Metadata Extraction and Database Insertion

• Concurrently with image storage, the engine extracts relevant metadata from the DICOM file.
• It prepares an SQL insert statement with the extracted metadata and the newly created file path.
• The engine establishes a connection to the MySQL database and executes the insert statement.
• The database transaction is committed, making the metadata immediately available for querying.

6. RIS Update

• Upon successful database insertion, the integration engine prepares a data payload for the RIS.
• It maps the DICOM metadata to the format expected by the RIS API.
• The engine makes an API call to the RIS, updating it with the new study information.
• It waits for a success response from the RIS before proceeding.

7. Confirmation and Logging

• After successful completion of all steps, the integration engine logs the entire process.
• It includes timestamps for each step, allowing for performance monitoring and auditing.
• If configured, it sends a confirmation message back to the originating imaging device.

8. Error Handling

• At each step, if an error occurs:
  • The engine logs the error with detailed information.
  • It attempts to roll back any partially completed steps (e.g., deleting partially written files).
  • If configured, it sends alerts to the IT team for immediate attention.
  • For non-critical errors, it may queue the operation for retry.

9. Real-time Monitoring

• Throughout the process, the integration engine updates its internal status metrics.
• These metrics are available in real-time through a monitoring dashboard, showing:
  • Current processing rate
  • Queue lengths
  • Error rates
  • System resource utilization

10. Data Availability

The study information is visible in the RIS for radiologists and administrative staff.

As soon as the process completes:

• The image is available on the shared storage for viewing by authorized personnel.
• The metadata is queryable in the MySQL database for analytics or other systems.