Implemented Protocol for Data Collection
DICOM API
- DICOM File Reception: The API receives DICOM files from institutional storage systems or through a secure interface where users can upload the files.
Data Anonymization:
- DICOM Header Extraction: Access and read the information from the DICOM header.
- Identification of Sensitive Data: Identify fields with potentially sensitive information that require anonymization.
- Data Deletion or Modification: Personal data is deleted and replaced with randomly generated unique identifiers (16 characters).
Image Conversion:
- Image Processing: Images are extracted from the DICOM file once the header anonymization is completed.
- Conversion to PNG: Medical images are converted from their original DICOM format to PNG, facilitating handling in machine learning models.
Secure Storage (Google Cloud SafeKey):
- Encryption at Rest: Implementation of Client-Managed Encryption (CMEK).
- Encryption in Transit: Use of TLS, VPN, and Private Peering to protect data during transfer.
- Key Management Encryption:
- Google Cloud Key Management Service (KMS) for secure key management.
- Logging and monitoring of key operations.
- Access Controls:
- Implementation of Identity and Access Management (IAM).
- Application of robust security policies.
- Audit and Compliance: Setup of Cloud Audit Logs and regular security and compliance reviews.
API Interface and Access:
- Exposure through Google Cloud: Detail how the API interface is protected, possibly including two-factor authentication or certificates.
Compliance and Audit:
- Regular Compliance Review: Schedule and document regular audits to ensure continuous adherence to security and privacy policies.
Relationship IMSS – EpyDiagnosis
- Data Agreement: Confirm that there is a formal agreement detailing the rights and responsibilities of both parties in relation to the use of the data.
- Consent Verification: Ensure that the consent process has been conducted in accordance with ethical and legal standards.
- AI Ethics: Jointly develop with IMSS a specific ethical protocol for AI.
- Collaborative and Multidisciplinary Review: Establish an ethics review committee and a multidisciplinary team to regularly review the project.