Beyond Accuracy: A Metrological Framework for the Reliability of Artificial Intelligence-Based Biomedical Measurement Systems

This review proposes the Biomedical AI Metrological Reliability Framework (BAMRF), a 12-domain framework for evaluating the reliability of artificial intelligence-enabled biomedical measurement systems. The paper integrates concepts from medical metrology, measurement uncertainty, traceability, clinical validation, AI reporting guidelines, regulatory science, and risk management. It argues that algorithmic accuracy alone is insufficient to establish reliable clinical measurement and proposes a comprehensive framework covering the complete measurement chain, from sensor calibration to lifecycle surveillance. The review aims to support researchers, regulators, manufacturers, conformity assessment bodies, and healthcare professionals involved in trustworthy biomedical AI systems.