P253: Revolutionizing Computer System Validation (CSV) with Artificial Intelligence: A Risk-Based Approach for Quality Assurance
Poster Presenter
Terry Katz
Senior Director, Biostatistics and DM Planning and Functional Excellence
Daiichi Sankyo United States
Objectives
Summarize regulatory expectations for validation of computerized system used in clinical trials, Identify opportunities to use AI-enabled technologies to facilitate software validation, and provide an example of a workflow.
Method
FDA’s draft guidance on Computer Software Assurance (CSA)(2022) marked a paradigm shift promoting a risk-based approach and incorporating flexible testing methods, such as unscripted testing and error guessing. Automation including SaaS and AI has allowed industry to take better prescribed risks.
Results
AI offers opportunities to enhance validation processes by automating risk assessment, generating and optimizing test scripts, executing tests, and continuously monitoring systems. By leveraging machine learning algorithms, AI can potentially identify high-risk areas, model impact of software changes, and help to achieve compliance with regulatory requirements through enhanced accuracy and efficiency.
Conclusion
AI systems may have the potential to increase self-validation through iterative or continuous improvements. As organizations adopt these technologies, AI-driven CSV might not only accelerate validation timelines but also strengthen compliance, and, if implemented on a large scale, e.g., on product and study portfolios, reduce costs. This integration underscores the need to develop robust AI tools, reliable training datasets, and effective risk-based strategies to fully realize the potential of AI in modern quality assurance and validation.
