Speaker
Description
Background
Biofabrication technologies are advancing rapidly, offering novel therapeutic strategies for regenerating complex tissue defects. The LUMINATE project, co-funded under the Horizon Europe programme, focuses on the development of an AI-powered bioprinter for intraoperative repair of osteochondral tissue. The device combines high-precision material deposition and AI-powered process control to enable point-of-care clinical interventions. Current EU regulatory frameworks,such as the Medical Device Regulation (EU) 2017/745 and Artificial Intelligence Act, provide both regulatory requirements and a methodological blueprint for ensuring safety, performance, and transparency. Within LUMINATE, these frameworks have guided the implementation of robust traceability and verifiability strategies that underpin both regulatory compliance and clinical effectiveness and guide the design process.
A structured approach to traceable and verifiable design inputs
Within the LUMINATE project, we propose a systematic regulatory methodology for capturing and managing design inputs in a high-risk, AI-enabled, biofabrication medical device. Our approach is based on compliance to ISO 13485 and applies design control principles.
We propose an Integrated design traceability matrix to enhance traceability and verifiability. It systematically lists core design inputs such as clinical requirements, usability requirements, safety requirements, user needs. This explicit mapping then connects the listed design inputs to design outputs, such as biomaterial specifications, device parameters, and AI functionalities, ensuring traceability of decisions.
Structured documentation regarding verification testing enables verifiability, allowing confirmation of risk mitigation strategies, supporting MDR Annex II and AI Act Article 9.
We recommend establishing comprehensive documentation protocols for all AI modules within the LUMINATE suite. This includes detailed records of datasets and records of the design process (model architectures, training-validation processes, and generalization assessments), in alignment with AI Act Articles 10 and 11. Such documentation enhances verifiability by enabling independent assessment of dataset quality, model validity, and reproducibility of performance of the model, consistently meeting clinical and safety expectations.
We advocate systematic integration of international standards to support robust, interoperable system design. These standards enhance traceability through harmonized formats and procedures across hardware, software, and data workflows. They also ensure verifiability by providing benchmarks and tests enabling external reviewers to assess device performance reliably. This approach supports compliance with AI Act Article 15. All planned verification tests are designed to demonstrate conformity with pre-identified standards.
To further improve effectiveness of the design process, we propose embedding iterative co-design processes involving clinicians, biomaterials experts, engineers, and regulators from early development stages. This strategy operationalizes AI Act Article 14 on human oversight, ensuring that human control and interpretability are integrated into the design of AI-assisted bioprinting workflows.Stakeholder engagement improves traceability by linking design to clinical needs and verifiability through documented, validated user input.
Conclusion
The LUMINATE project proposes a replicable methodology to establish traceable, verifiable, and regulatory-compliant design inputs in AI-powered medical devices intended for intraoperative application. By aligning with EU MDR, MDCG guidances, and the AI Act, our approach offers a pathway to de-risk clinical translation and facilitate regulatory approvals. By advancing systematic traceability and verifiability frameworks, we aim to strengthen the reliability and accountability of next-generation biofabrication solutions.
85410419705
