Machine Learning for Predicting Mechanical Properties of Tissue-Mimicking Phantoms of Soft Biological Tissues

Abstract

Hydrogels appear to be the three-dimensional polymer networks, where the chemical or physical crosslinks between macromolecules are acting as the nodes. Currently, polyacrylamide-based hydrogels are the promising materials in obtaining the soft biological tissue-mimicking phantoms. A key advantage of these hydro-gels is their ability to control mechanical properties by varying the crosslinker concentration, i.e., the number of crosslinks, which makes it possible to simulate characteristics of various organs and tissues. However, selecting the crosslinker concentration to create phantoms with the specified characteristics is a complex task requiring significant time and computational resources. The machine learning algorithms are able to determine correlation between the resulting hydro-gel crosslink concentration and the elastic properties. A machine learning model based on two random forest algorithms is implemented being capable of predicting mechanical properties of the tissue-mimicking phantoms. The algorithms are trained using synthetic data obtained from the numerical indentation experiments in the licensed finite element analysis software ANSYS Workbench (Ansys Inc., USA) using the linear-elastic material behavior model, neo-Hookean and Mooney --- Rivlin hyperelasticity models, as well as data from the full-scale experiments. Predictions were validated using the test data, which constituted 30 % of the entire data set and were not used in the algorithm learning, as well as results of the full-scale experiments

The work was performed within the State Assignment of the NRC Kurchatov Institute

Please cite this article in English as:

Semenov L.I., Krupnin A.E., Antipova K.G., et al. Machine learning for predicting mechanical properties of tissue-mimicking phantoms of soft biological tissues. Herald of the Bauman Moscow State Technical University, Series Natural Sciences, 2025, no. 6 (123), pp. 38--59 (in Russ.). EDN: XFPURH

References

[1] Armstrong S.A., Jafary R., Forsythe J.S., et al. Tissue-mimicking materials for ultrasound-guided needle intervention phantoms: a comprehensive review. Ultrasound Med. Biol., 2023, vol. 49, iss. 1, pp. 18--30. DOI: https://doi.org/10.1016/j.ultrasmedbio.2022.07.016

[2] Liu D., Jiang P., Wang Y., et al. Engineering tridimensional hydrogel tissue and organ phantoms with tunable springiness. Adv. Funct. Mater., 2023, vol. 33, iss. 17, art. 2214885. DOI: https://doi.org/10.1002/adfm.202214885

[3] Guntur S.R., Kim S.-C., Choi M.-J. A cost-effective reusable tissue mimicking phantom for high intensity focused ultrasonic liver surgery. Bioengineering, 2022, vol. 9, iss. 12, art. 786. DOI: https://doi.org/10.3390/bioengineering9120786

[4] Payne A.H., Goodrich K.C., Kholmovski E.G., et al. Isolated kidney phantom for development of biothermal vascular models with application to high intensity focused ultrasound therapy. Med. Phys., 2008, vol. 35, iss. 10, pp. 4426--4434. DOI: https://doi.org/10.1118/1.2975226

[5] Filippou A., Louca I., Damianou C. Characterization of a fat tissue mimicking material for high intensity focused ultrasound applications. J. Ultrasound, 2023, vol. 26, no. 2, pp. 505--515. DOI: https://doi.org/10.1007/s40477-022-00746-4

[6] Tajik M., Akhlaqi M.M., Gholami S. Advances in anthropomorphic thorax phantoms for radiotherapy: a review. Biomed. Phys. Eng. Express, 2022, vol. 8, no. 5,art. 052001. DOI: https://doi.org/10.1088/2057-1976/ac369c

[7] Zhong X., Cao Y., Zhou P. Thermochromic tissue-mimicking phantoms for thermal ablation based on polyacrylamide gel. Ultrasound Med. Biol., 2022, vol. 48, iss. 8, pp. 1361--1372. DOI: https://doi.org/10.1016/j.ultrasmedbio.2022.03.021

[8] Wang Y., Tai B.L., Yu H., et al. Silicone-based tissue-mimicking phantom for needle insertion simulation. J. Med. Devices, 2014, vol. 8, iss. 2, art. 021001. DOI: https://doi.org/10.1115/1.4026508

[9] Ali A.M., Hogg P., Johansen S., et al. Construction and validation of a low cost paediatric pelvis phantom. Eur. J. Radiol., 2018, vol. 108, pp. 84--91. DOI: https://doi.org/10.1016/j.ejrad.2018.09.015

[10] Zhao M., Zhou M., Cao X., et al. Stable tissue-mimicking phantoms for longitudinal multimodality imaging studies that incorporate optical, CT, and MRI contrast. J. Biomed. Opt., 2023, vol. 28, iss. 4, art. 046006. DOI: https://doi.org/10.1117/1.JBO.28.4.046006

[11] Negus I.S., Holmes R.B., Jordan K.C., et al. Technical note: development of a 3D printed subresolution sandwich phantom for validation of brain SPECT analysis. Med. Phys., 2016, vol. 43, iss. 9, pp. 5020--5027. DOI: https://doi.org/10.1118/1.4960003

[12] O’Doherty J., Sammut E., Schleyer P., et al. Feasibility of simultaneous PET-MR perfusion using a novel cardiac perfusion phantom. European J. Hybrid Imaging, 2017, vol. 1, no. 1, art. 4. DOI: https://doi.org/10.1186/s41824-017-0008-9

[13] Tan Z., Dini D., Baena F.R., et al. Composite hydrogel: a high fidelity soft tissue mimic for surgery. Mater. Des., 2018, vol. 160, pp. 886--894. DOI: https://doi.org/10.1016/j.matdes.2018.10.018

[14] Sanchez-Cid P., Jimenez-Rosado M., Romero A., et al. Novel trends in hydrogel development for biomedical applications: a review. Polymers, 2022, vol. 14, iss. 15, art. 3023. DOI: https://doi.org/10.3390/polym14153023

[15] Antipova K.G., Krupnin A.E., Khramtsova E.A., et al. Polyacrylamide-based soft biological tissue phantoms: preparation and properties. Geny i kletki [Genes & Cells], 2022, vol. 17, no. 3, p. 13 (in Russ.).

[16] Reinhards-Hervas C., Rico A., Rodriguez J. Crosslinker concentration effect on the poroviscoelastic relaxation of polyacrylamide hydrogels using depth-sensing indentation. Polym. Test., 2021, vol. 100, art. 107265. DOI: https://doi.org/10.1016/j.polymertesting.2021.107265

[17] Xiao Y., Chen Q., Yang Z., et al. Asymmetric and skin-mimicking hydrogels with wide temperature tolerance and superior elasticity for high-performance strain sensors. ACS Omega, 2023, vol. 8, iss. 49, pp. 46676--46684. DOI: https://doi.org/10.1021/acsomega.3c05779

[18] Hariri A., Palma-Chavez J., Wear K.A., et al. Polyacrylamide hydrogel phantoms for performance evaluation of multispectral photoacoustic imaging systems. Photo-acoustics, 2021, vol. 22, art. 100245. DOI: https://doi.org/10.1016/j.pacs.2021.100245

[19] Avci Yayla I.I., Bilal M., Salmaslioglu A., et al. Ultrasound phantom with solids mimicking cancerous tissue for needle breast biopsy. Turk. J. Chem., 2022, vol. 46, no. 5, pp. 1493--1503. DOI: https://doi.org/10.55730/1300-0527.3454

[20] Upadhyay K., Subhash G., Spearot D. Hyperelastic constitutive modeling of hydrogels based on primary deformation modes and validation under 3D stress states. Int.J. Eng. Sci., 2020, vol. 154, art. 103314. DOI: https://doi.org/10.1016/j.ijengsci.2020.103314

[21] Lee D., Zhang H., Ryu S. Elastic modulus measurement of hydrogels. In: Mondal M. (eds). Cellulose-Based Superabsorbent Hydrogels. Polymers and Polymeric Composites: a Reference Series. Cham, Springer, 2018, pp. 1--21. DOI: https://doi.org/10.1007/978-3-319-76573-0_60-1

[22] Vila Pouca M.C.P., Cerqueira M.R.G., Ferreira J.P.S., et al. Simulating 3D printing on hydrogel inks: a finite element framework for predicting mechanical properties and scaffold deformation. Finite Elem. Anal. Des., 2024, vol. 230, art. 104098. DOI: https://doi.org/10.1016/j.finel.2023.104098

[23] Abdolazizi K.P., Linka K., Sprenger J., et al. Concentration-specific constitutive modeling of gelatin based on artificial neural networks. Proc. Appl. Math. Mech., 2021, vol. 20, iss. 1, art. e202000284. DOI: https://doi.org/10.1002/pamm.202000284

[24] Zhu J.-A., Jia Y., Lei J., et al. Deep learning approach to mechanical property prediction of single-network hydrogel. Mathematics, 2021, vol. 9, iss. 21, art. 2804. DOI: https://doi.org/10.3390/math9212804

[25] Shokrollahi Y., Dong P., Gamage P., et al. Finite element-based machine learning model for predicting the mechanical properties of composite hydrogels. Appl. Sci., 2022, vol. 12, iss. 21, art. 10835. DOI: https://doi.org/10.3390/app122110835

[26] Belkin A.E., Dashtiev I.Z., Kostromitskikh A.V. Determining polyurethane elastic parameters at large strains using torsion and tensile test results. Izvestiya vysshikh uchebnykh zavedeniy. Mashinostroenie [BMSTU Journal of Mechanical Engineering], 2016, no. 8, pp. 3--10 (in Russ.). DOI: https://doi.org/10.18698/0536-1044-2016-8-3-10

[27] Liu J., Zheng H., Poh P., et al. Hydrogels for engineering of perfusable vascular networks. Int. J. Mol. Sci., 2015, vol. 16, no. 7, pp. 15997--16016. DOI: https://doi.org/10.3390/ijms160715997

[28] Bootsma K., Dimbath E., Berberich J., et al. Materials used as tissue phantoms in medical simulation. In: Studies in Mechanobiology, Tissue Engineering and Biomaterials. Berlin, Heidelberg, Springer, 2017, pp. 1--48. DOI: https://doi.org/10.1007/8415_2016_1

[29] Antipova C.G., Krupnin A.E., Zakirov A.R., et al. A comprehensive mechanical testing of polyacrylamide hydrogels: the impact of crosslink density. Polymers, 2025, vol. 17, iss. 6, art. 737. DOI: https://doi.org/10.3390/polym17060737

[30] Breiman L. Random forests. Mach. Learn., 2001, vol. 45, no. 1, pp. 5--32. DOI: https://doi.org/10.1023/A:1010933404324