Implementing a PCCP & CI/CD for AI/ML-Enabled Medical Devices
60
minutes
This browser does not support HTML video.
Click for sound
Understanding PCCP in a CICD Environment
Overview of Predetermined Change Control Plans (PCCP)
IEC 62304 and Software Lifecycle Management
Challenges in AI Software Development
Metrics and Validation in PCCP
Steps to Validate AI Systems
Model Validation and Acceptance Criteria
Integrating PCCP and CICD at Scale
Designing Systems for Regulatory Compliance
Challenges in Implementing PCCP
Ensuring Traceability and Risk Management
Understanding Product Architecture
Project Milestones Overview
Exploring System Traceability
Reviewing System Architecture
Finalizing the Release Approval
Risk Management Insights
Strategies for Rapid Change
AI Program Governance Framework
Closing Remarks and Invitation for Engagement
The FDA's Predetermined Change Control Plan (PCCP) allows medical manufacturers to update their device without repeated approvals. With this regulatory strategy, top medical device companies are incorporating agile methodologies into their work, including Continuous Integration and Continuous Delivery (CI/CD), so they can bring products to market faster while still ensuring safety and regulatory compliance.
What you'll learn
To help you start using CI/CD and PCCPs for your AI/ML-enabled products, this webinar will examine:
- Practical and systematic approaches to designing, constructing, and validating a successful PCCP with examples of model creation, data training, bias assessment, testing, and best-in-class infrastructure
- Processes for post-market ML software releases that enable continuous validation in production including ingesting real-world data, modification protocols, and metrics for ML models
- Strategies for implementing CI/CD to fulfill a PCCP for a SaMD project to accelerate product release cycles