Real-World Evidence

Real-World Evidence (RWE) and Biocompatibility: Implications of ISO 10993-1:2025 for Medical Device Manufacturers

December 11, 2025
|
By Dr. Ebot Eyong

The revised ISO 10993-1 marks a significant shift in biocompatibility assessment by enabling risk-based justifications supported by real-world evidence. This article explores pre-market submissions, manufacturer implications, AI-enabled device challenges, and regulatory pitfalls.

In November 2025, a revised version of ISO 10993-1 was published, marking a significant shift in biocompatibility assessment by formally enabling risk-based justifications, supported by real-world evidence (RWE), rather than relying solely on prescriptive biological testing. This update aligns biocompatibility expectations more closely with modern device development, including software-driven and AI-enabled technologies.

Key Points for pre-Market Submissions

For regulatory submissions (e.g., FDA 510(k), De Novo, PMA, and EU MDR technical documentation), manufacturers may now justify biocompatibility using a structured risk management approach that incorporates existing data, material characterization, clinical experience, post-market surveillance, and RWE. Submissions must clearly document the biological evaluation plan, rationale for omitted tests, data relevance, and residual risk acceptability, all aligned with ISO 14971.

Implications for Manufacturers

The update reduces unnecessary testing, potentially lowering costs and timelines, but increases expectations for scientific rigor and documentation quality. Manufacturers must demonstrate strong data governance, traceability, and lifecycle management, particularly when leveraging RWE across device generations or platforms.

Challenges for AI-Enabled Devices

AI-enabled devices introduce additional complexity, as software-driven changes may alter use conditions, patient interaction, or exposure profiles. Manufacturers must ensure that biocompatibility assumptions remain valid as AI functionality evolves, especially when supported by Predetermined Change Control Plans (PCCPs).

Regulatory Pitfalls

Regulators face challenges assessing the sufficiency, relevance, and bias of RWE, as well as understanding indirect safety impacts from adaptive AI behavior. Ensuring consistency across reviews while supporting innovation remains a key risk.

For more information, visit https://eemedicals.com/

Explore More Publications

Continue exploring Dr. Ebot Eyong’s professional insights on healthcare regulation, FDA submissions, AI-enabled medical devices, quality systems, and global compliance strategy.

AI & Digital Health

Regulatory Challenges and Solutions for AI-Enabled Medical Devices

February 9, 2025
|
By Dr. Ebot Eyong

Regulatory agencies worldwide are encountering significant obstacles in managing and supervising AI-enabled medical devices. This article explores AI/ML regulatory oversight, ethical concerns, bias, performance degradation, transparency, cross-site deployment, post-market surveillance, cybersecurity risks, and standardized validation.

Read Article

Cybersecurity & Cloud Systems

Cybersecurity Documentation for Premarket Approval (PMA) Submissions

February 9, 2024
|
By Dr. Ebot Eyong

Cybersecurity documentation is vital for PMA submissions to the FDA, ensuring device safety and effectiveness. This article outlines cybersecurity risk assessment, security controls, SBOM, vulnerability disclosure, incident response planning, labeling, monitoring, and documentation requirements.

Read Article

Software & SaMD

The Hidden Risk: Gaps in Software Remediation for Medical Devices

June 10, 2025
|
By Dr. Ebot Eyong

As medical devices evolve into software-driven, connected systems, software remediation has become a critical driver of patient safety, regulatory compliance, and commercial success. Yet current frameworks remain fragmented, reactive, and poorly integrated, creating systemic risk and a major opportunity to build regulatory-aware software infrastructure.

Read Article

AI & Digital Health

Challenges in Medical Device AI Validation

June 18, 2026
|
Dr. Ebot Eyong

The most difficult aspect is ensuring that an AI system remains reliable after deployment as patient populations, clinical practices, and data characteristics evolve over time. This has led regulators to shift from a one-time validation model toward a lifecycle-based approach involving continuous monitoring and revalidation.

Read Article