FDA Guidance on AI-Enabled Medical Devices: Lifecycle Management & Marketing Submissions

On January 6, 2025, the FDA released draft guidance for AI-enabled medical devices, highlighting a Total Product Lifecycle (TPLC) approach. AI-enabled devices use AI device software functions (AI-DSFs) to analyze medical data for diagnosis, monitoring, or treatment.

The guidance emphasizes documentation, risk management, performance validation, and cybersecurity to ensure safety and effectiveness. Manufacturers should include:

• Quality Management System (QMS): Document design controls, risk management, and software development lifecycle (SDLC) processes.
• Device Description: Clearly define intended use, AI model function, patient population, hardware/software dependencies, and environment.
• Data Management: Detail collection, processing, storage, and measures to avoid bias.
• Model Development: Explain AI architecture, learning methods, hyperparameter tuning, and performance metrics.
• Performance Validation: Test AI outputs in real-world conditions, address model drift, and ensure patient safety.
• User Interface & Labeling: Provide clear instructions, capabilities, limitations, and warnings.
• Post-Market Monitoring: Track device performance, detect model drift, and update AI models safely.
• Cybersecurity: Implement encryption, authentication, access control, and incident response protocols.

This guidance provides a roadmap for developing safe, reliable, and compliant AI-enabled medical devices, helping manufacturers streamline approvals while maintaining patient trust.