This study uses validated high-fidelity fluid-structure interaction simulations to assess calcification risk in six prosthetic aortic valve configurations (two leaflet geometries (D1 and D2) and three materials, namely bovine pericardium (BOV), porcine pericardium (PORC) and elastomer (ELA)). Leaflet thickness is held constant (300 μm) across all configurations. Finite-time Lyapunov exponents derived from the Cauchy-Green strain tensor quantify leaflet deformation coherence throughout systole, with the strong fluid-structure coupling handled via variational transfer. Unsupervised k-means clustering assigns spatial risk categories across leaflet surfaces. FTLE patterns for biological tissues correlate with micro-CT calcification maps from explanted valves, supporting the predictive capacity of this descriptor. The clustering ranks calcification propensity of the six configurations unambiguously. Bovine pericardium with geometry D2 exhibits the smallest high-calcification-risk area fraction (10.2%), elastomeric leaflets occupy an intermediate position (14.5-16.6%) and porcine configurations fare worst (20.1-21.1%). Geometry D2 outperforms D1 for biological tissues. FTLE proves more discriminating than haemodynamic shear metrics alone. The framework furnishes an objective basis for comparing valve designs; elastomeric valves emerge as a plausible compromise between durability and manufacturability. Future validation will target calcific deposits in elastomers via true blood in vitro testing.