- Industry: Digital Health / Ophthalmic Diagnostics
- Product: Cloud-based AI-powered eye scanning application (Software as a Medical Device)
- Use Case: Retinal image analysis for early detection of diabetic retinopathy and other ophthalmic conditions
- Users: Ophthalmologists, telemedicine providers, and diagnostic labs
A health tech startup developing an eye scanning SaMD platform sought to align with FDA requirements during the premarket submission process and improve postmarket cybersecurity readiness. With sensitive patient imaging and AI-powered diagnostics at stake, robust security was a core regulatory and operational priority.
- Insecure API endpoints for image upload and retrieval
- Inconsistent session handling logic led to session fixation issues
- Imaging data stored temporarily in unencrypted cloud storage
- No security risk management plan in place for SaMD classification
- Conducted structured threat modeling aligned with FDA’s 2022 Cybersecurity Premarket Guidance
-
Defined and documented potential risks to:
• Patient safety from unauthorized model manipulation
• Data confidentiality (PHI in DICOM or image metadata)
• Diagnostic integrity from tampered AI inputs/outputs - Created a Cybersecurity Risk Management File (CRMF) linked to the SaMD’s Design History File (DHF)
- Encrypted imaging data-at-rest using AES-256
- Role-based access control (RBAC) across user types (e.g., ophthalmologist, technician)
- Secure boot and firmware validation for imaging devices
- Used TLS 1.3 encryption and certificate pinning for all image and result transmission
- Conducted static and dynamic application security testing (SAST/DAST)
- Performed third-party penetration testing
- Submitted SBOM (Software Bill of Materials)
- Included a cybersecurity use case traceability matrix
- Deployed a SIEM system
- Defined and tested a coordinated vulnerability disclosure (CVD) process
- Created an incident response plan with FDA reporting thresholds
- Introduced automated vulnerability scanning
- Established patch release cycles
- Subscribed to ISAC health sector threat feeds
- Conducted annual third-party security audits
- Updated risk assessments per FDA Postmarket Cybersecurity Guidance
- Maintained traceability of new threats to software changes
- FDA SaMD submission accepted with cybersecurity risk controls validated
- All critical vulnerabilities mitigated premarket; none discovered during initial 12 months post-launch
- Audit-ready documentation maintained
- Elevated trust from clinics and hospitals adopting the AI-driven diagnostics solution
- A corneal scan-based AI platform
- A low-connectivity mobile eye care unit
- The modular design supports FDA, EU MDR, and Health Canada cybersecurity expectations.
This case illustrates the necessity of building cybersecurity into both the design and lifecycle management of Software as a Medical Device (SaMD). Aiyanaar’s intervention