Artificial Intelligence is revolutionizing medical technology. However, EU regulators expect stable, quantified proof of safety and performance for the intended use.
Under MDR Annex XIV and the MDCG 2025-6 guidance, requirements such as accuracy, robustness, and cybersecurity needs to be met, which are essential aspects of performance as well as specifically requires testing of high-risk MDAI against prior defined metrics and probabilistic thresholds to ensure that high-risk MDAI perform consistently for their intended purpose.
The clinical evaluation must show that your device performs at least as safely and effectively as comparable technologies. That means planning your evidence generation from the Clinical Evaluation Plan (CEP) to the Clinical Evaluation Report (CER) and Post-Market Clinical Follow-up (PMCF).
Both MDR and IVDR require manufacturers to validate MDAI outputs through rigorous testing in the form of clinical or performance evaluation.
How do you prove that your AI device is clinically safe?
By defining benchmarks that make safety measurable. These quantitative benchmark values are derived from the current state of the art, such as clinical literature, registries, or equivalent technologies. Benchmarks are the bridge between innovation and approval. By defining them early, manufacturers can align every step of their clinical strategy, from study design to data analysis and documentation.
At WQS, we help manufacturers develop benchmark-driven clinical strategies that integrate MDR and MDCG requirements into a single, coherent framework. Our team connects data, validation, and documentation while defining clear clinical performance metrics aligned with Annex XIV, MDCG 2025-6, and Good Machine Learning Practice (GMLP).
Contact us for the clinical evaluation of your AI device at us.wqs.de.
