Machine learning to mechanically assess 2D and 3D biomimetic electrospun scaffolds for tissue engineering applications - Between the predictability and the interpretability

Aug 5, 2026

Upload by Aurora

Currently, autografts remain the gold standard for replacing damaged biological tissues, despite drawbacks such as donor-site morbidity, long rehabilitation, and limited availability. This study evaluates how machine learning can predict and interpret the mechanical performance of 2D and 3D polyvinyl alcohol (PVA) electrospun biomimetic scaffolds for tissue engineering use. Crosslinked and non-crosslinked scaffolds are tensile-tested in dry/wet and longitudinal/transverse loading. Twenty-eight ML models relate four exogenous variables to two endogenous mechanical targets, identifying ligament-, skin-, oral/nasal-, and renal-like structures. CART models are interpretable, while Cubist and SVM deliver the highest accuracy, enabling manufacturing optimization.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Abstract

Introduction

Autografts as the current gold standard and their limitations

Design premises for tissue-engineered grafts

Target mechanical properties and tissue-specific benchmarks

Methods

Scaffold fabrication and crosslinking conditions

Mechanical testing setup and loading regimes

Machine learning workflow and variables

Results and Discussion

Model performance in predicting Young’s modulus and ultimate tensile strength

Identification of biomimetic scaffold structures and testing conditions

Comparison of interpretability versus predictability across models

Conclusion

Implications for optimizing scaffold manufacturing

Please cite the Published Version Roldán, Elisa , Reeves, Neil D , Cooper, Glen and Andrews, Kirstie (2024) Machine learning to mechanically assess 2D and 3D biomimetic electrospun scaffolds for tissue engineering applications: between the predictability and the interpretability. Journal of the Mechanical Behavior of Biomedical Materials, 157 . 106630 ISSN 1751-6161 DOI: [https://doi.org/10.1016/j.jmbbm.2024.106630](https://doi.org/10.1016/j.jmbbm.2024.106630) Publisher: Elsevier BV Version: Published Version Downloaded from: [https://e-space.mmu.ac.uk/634892/](https://e-space.mmu.ac.uk/634892/) Usage rights: Creative Commons: Attribution 4 .0 Additional Information: This is an open access article which ﬁrst appeared in Journal of the Mechanical Behavior of Biomedical Materials Data Access Statement: Data will be made available on request. Enquiries: If you have questions about this document, contact [openresearch@mmu.ac.uk](openresearch@mmu.ac.uk. Please)[. Please](openresearch@mmu.ac.uk. Please) include the URL of the record in e-space. If you believe that your, or a third party's rights have been compromised through this document please see our Take Down policy (available from [https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines](https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines)) journal of the mechanical behavior of biomedical materials 157 (2024) 106630 Contents lists available at ScienceDirect Journal of the Mechanical Behavior of Biomedical Materials journal [homepage:](homepage: www.elsevier.com/locate/jmbbm)[ www.elsevier.com/locate/jmbbm](homepage: www.elsevier.com/locate/jmbbm) | Machine learning to mechanically assess 2D and 3D biomimetic electrospun scaffolds for tissue engineering applications: Between the predictability and the interpretability
Elisa Rold´ana, * , Neil D. Reeves b, c , Glen Cooper d , Kirstie Andrewsa
a Department of Engineering, Faculty of Science & Engineering, Manchester Metropolitan University, Manchester, M1 5GD, UK b Department of Life Sciences, Faculty of Science & Engineering, Manchester Metropolitan University, Manchester, M1 5GD, UK c Lancaster Medical School, Faculty of Health and Medicine, Lancaster University, Lancaster, LA1 4YW, UK
d School of Engineering, University of Manchester, Manchester, M13 9PL, UK | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Keywords:
Machine learning Decision trees Electrospinning
PVA
Mechanical characterisation Tissue engineered implants Biomimetic scaffolds Ligament
Human tissue | | Currently, the use of autografts is the gold standard for the replacement of many damaged biological tissues. However, this practice presents disadvantages that can be mitigated through tissue-engineered implants. The aim of this study is to explore how machine learning can mechanically evaluate 2D and 3D polyvinyl alcohol (PVA) electrospun scaffolds (one twisted filament, 3 twisted filament and 3 twisted/braided filament scaffolds) for their use in different tissue engineering applications. Crosslinked and non-crosslinked scaffolds were fabricated and mechanically characterised, in dry/wet conditions and under longitudinal/transverse loading, using tensile testing. 28 machine learning models (ML) were used to predict the mechanical properties of the scaffolds. 4 exogenous variables (structure, environmental condition, crosslinking and direction of the load) were used to predict 2 endogenous variables (Young’s modulus and ultimate tensile strength). ML models were able to identify 6 structures and testing conditions with comparable Young’s modulus and ultimate tensile strength to ligamentous tissue, skin tissue, oral and nasal tissue, and renal tissue. This novel study proved that Classification and Regression Trees (CART) models were an innovative and easy to interpret tool to identify biomimetic electrospun structures; however, Cubist and Support

FAQs

What mechanical properties were predicted for the PVA electrospun scaffolds?

The study predicts Young’s modulus and ultimate tensile strength using machine learning models trained on experimental data.

What inputs (exogenous variables) were used to make the predictions?

Four exogenous variables were used: scaffold structure, environmental condition, crosslinking status, and direction of the load.

Which machine learning models were most accurate, and what trade-off was observed?

Cubist and Support Vector Machine (SVM) models were the most accurate for ultimate tensile strength and Young’s modulus, respectively, while Classification and Regression Trees (CART) provided an easier-to-interpret tool for identifying biomimetic structures.

Aug 5, 2026

Upload by Aurora