Beyond biomimicry: Mechanobiological qualification of human-relevant NAMs

Human-relevant new approach methodologies (NAMs) increasingly reproduce tissue architecture, multicellular organization and dynamic physical environments, yet biomimetic complexity does not by itself establish human relevance. Mechanical inputs such as matrix properties, flow, stretch, pressure and confinement are transformed by device geometry, materials and living tissues before they reach cells. The resulting tissues then remodel that same environment. We therefore introduce mechanobiological qualification as a fit-for-purpose process for determining whether decision-relevant mechanical variables are specified, quantitatively controlled, causally linked to biological responses and, where required, shown to improve human prediction. We distinguish four components of a mechanical exposure: programmed input, experienced mechanical state, biological output and state history. Current evidence is then organized on an M0-M4 ladder that progresses from mechanical incorporation to incremental human predictive value. The framework links four operations – Define, Measure, Perturb and Predict – with control of confounders, uncertainty and reporting. It also clarifies how qualified mechanical states can support disease-mechanism discovery, mechanotherapeutic target identification and mechanics-conditioned drug testing. We argue that the decisive comparison is not between a simple culture and a feature-rich platform, but between otherwise matched models with and without a defined, qualified mechanical state. The next generation of NAMs should therefore be judged less by the number of physiological features they contain than by whether the minimum mechanical determinants required for a specified decision have been identified, tested and human-anchored.