Predicting calcification risk in prosthetic aortic valves: a hybrid physics-based and machine learning approach

Calcific degeneration remains the foremost failure mode of bioprosthetic aortic valves and a mounting concern for polymeric substitutes. Six prosthetic valve configurations, pairing two leaflet scallop geometries (V-shaped D1 and U-shaped D2) with three constitutive materials (elastomer, bovine pericardium and porcine tissue), are simulated throughout systole within a patient-specific curved aorta reconstructed from clinical CT, using a high-fidelity fluid-structure interaction framework. An incremental formulation of Finite-Time Lyapunov Exponents (FTLEs), applied to the leaflet deformation gradients, is introduced as a descriptor of leaflet structural point coherence. With wall shear stress (WSS) derived features, the FTLE field feeds an unsupervised k -means clustering that partitions each leaflet into four calcification risk classes with no fitting to experimental data. Validated against an experimentally calibrated micro-CT calcification intensity map of explanted bovine pericardial leaflets, the combined FTLE and WSS classification attains a Spearman correlation ≥0.9 under near-equal weighting (γ ≈ 0.5). Porcine tissue concentrates strain into localised high-gradient zones whereas the isotropic elastomer spreads strain evenly yet sustains pronounced shear fluctuations through leaflet flutter. The D2 U geometry with bovine pericardium proves least susceptible across both risk dimensions, furnishing a transferable, model-free tool for ranking next-generation valve designs.