Predicting and understanding arterial elasticity from key microstructural features by bidirectional deep learning. (15th July 2022)

• Record Type: Journal Article
• Authors: Linka, Kevin; Cavinato, Cristina; Humphrey, Jay D.; Cyron, Christian J.
• Is Part Of: Acta biomaterialia. Volume 147(2022)
• Page Start: 63
• Page End: 72
• Publication Date: 2022-07-15
• Subjects: hybrid modeling -- arterial tissues -- explainable AI -- tissue maturation; Biomedical materials -- Periodicals; 610.28
• DOI: 10.1016/j.actbio.2022.05.039
• Languages: English
• ISSNs: 1742-7061

Abstract

Graphical abstract: Abstract: Microstructural features and mechanical properties are closely related in all soft biological tissues. Both yet exhibit considerable inter-individual differences and are affected by factors such as aging and disease and its progression. Histological analysis, modern in situ imaging, and biomechanical testing have deepened our understanding of these complex interrelations, yet two key questions remain: (1) Given the specific microstructure, can one predict the macroscopic mechanical properties without mechanical testing? (2) Can one quantify individual contributions of the different microstructural features to the macroscopic mechanical properties in an automated, systematic and largely unbiased way? Here we propose a bidirectional deep learning architecture to address these two questions. Our architecture uses data from standard histological analyses, two-photon microscopy and biaxial biomechanical testing. Its capabilities are demonstrated by predicting with high accuracy ( R 2 = 0.92 ) the evolving mechanical properties of the murine aorta during maturation and aging. Moreover, our architecture reveals that the extracellular matrix composition and organization are the most prominent factors governing the macroscopic mechanical properties of the tissues studied herein. Statement of significance: We present a physics-informed machine learning architecture that can predict macroscopic mechanical properties of arterial tissue with high accuracy (R2=0.92) from the tissue microstructure (characterized by imaging data). For the first time, this architecture enables also a fully automatic and largely unbiased quantification of the relevance of different microstructural features (such as collagen volume fraction and fiber straightness) for the macroscopic mechanical properties. This approach opens up unprecedented ways to predictive mechanical modeling of soft biological tissues. Moreover, it provides quantitative insights into the relation between tissue microstructure and its macroscopic properties that promise to play an important role in future tissue engineering.