Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jul 18:6:0197.
doi: 10.34133/research.0197. eCollection 2023.

Machine Learning in Predicting Printable Biomaterial Formulations for Direct Ink Writing

Hongyi Chen  1   2 , Yuanchang Liu  1   1 , Stavroula Balabani  1   3 , Ryuji Hirayama  2 , Jie Huang  1
Affiliations

Machine Learning in Predicting Printable Biomaterial Formulations for Direct Ink Writing

Hongyi Chen et al. Research (Wash D C). .

Abstract

Three-dimensional (3D) printing is emerging as a transformative technology for biomedical engineering. The 3D printed product can be patient-specific by allowing customizability and direct control of the architecture. The trial-and-error approach currently used for developing the composition of printable inks is time- and resource-consuming due to the increasing number of variables requiring expert knowledge. Artificial intelligence has the potential to reshape the ink development process by forming a predictive model for printability from experimental data. In this paper, we constructed machine learning (ML) algorithms including decision tree, random forest (RF), and deep learning (DL) to predict the printability of biomaterials. A total of 210 formulations including 16 different bioactive and smart materials and 4 solvents were 3D printed, and their printability was assessed. All ML methods were able to learn and predict the printability of a variety of inks based on their biomaterial formulations. In particular, the RF algorithm has achieved the highest accuracy (88.1%), precision (90.6%), and F1 score (87.0%), indicating the best overall performance out of the 3 algorithms, while DL has the highest recall (87.3%). Furthermore, the ML algorithms have predicted the printability window of biomaterials to guide the ink development. The printability map generated with DL has finer granularity than other algorithms. ML has proven to be an effective and novel strategy for developing biomaterial formulations with desired 3D printability for biomedical engineering applications.

PubMed Disclaimer

Figures

Fig. 1.
Fig. 1.
A summary of 3D printing results of (A) hydrogel-based F127/LP inks and (B) polymer-based PCL/nHA inks with DCM as solvent. The number in the ink name refers to wt% for F127/LP inks and w/v% for PCL/nHA inks. The Barus effect (expansion of the inks exiting the nozzle) was observed at low-viscosity inks. The viscosity of inks increases with the concentrations of fillers LP and nHA; at too high viscosity, inks were unextrudable or clogged nozzle, and the printed filaments were prone to fracture. A higher SF of 3D printed filaments and 4-layer 0°/90° scaffold structure was achieved for 15F/8LP and PCL/20nHA inks.
Fig. 2.
Fig. 2.
Data summary of the distribution of (A) biomaterials, (B) 2 ink systems, and (C) outputs in terms of the number of usages in the data. The input features include biomaterials and solvents used for the 3D printing process while the output is the printability of the inks.
Fig. 3.
Fig. 3.
The optimization process of DT, RF, and DL for predicting printability: (A) Effect of the MNL and MNF for splitting (5, 10, 15, and 20) on the performance of the DT algorithm. (B) Effect of both MNL (10, 15, 20, 25, and 30) in DTs and the number of DTs in the RF algorithm on its performance. (C) Effect of the number of nodes in hidden layers and number of hidden layers in the DL algorithm on its performance.
Fig. 4.
Fig. 4.
Evaluation metrics of the optimized ML models for predicting printability of biomaterial formulations including (A) prediction accuracy, precision, recall, and F1 score of the printable class (B) and unprintable class (C).
Fig. 5.
Fig. 5.
Printability map of both hydrogel-based and polymer-based inks predicted by ML algorithms. (A to C) The printability map of F127/LP hydrogel nanocomposites predicted by DT, RF, and DL, respectively. (D to F) The printability map of PCL/nHA in DCM polymer nanocomposite inks predicted with DT, RF, and DL, respectively. The green triangles and red crosses mark the printable and unprintable formulations, respectively, in the training data. The green and red areas are the predicted printable and unprintable regions, respectively, mapped from the testing results.
Fig. 6.
Fig. 6.
The working principles of the machine learning algorithms including DT, RF, and DL. (A) Example of a dataset of printability of hydrogel formulations. Each item (ink) is associated with several features (hydrogel type/name) and one output (printability). Schematics of the internal structures of machine learning algorithms include (B) DT, (C) RF, and (D) DL after learning from the dataset.
Fig. 7.
Fig. 7.
The confusion matrix for ink printability classification. Each row of the matrix represents predicted instances of each class, while each column represents the actual instances of each class. (A) The confusion matrix with the printable class as the positive class and (B) the confusion matrix with the unprintable class as the positive class.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Lai J, Wang C, Wang M. 3D printing in biomedical engineering: Processes, materials, and applications. Appl Phys Rev. 2021;8(2): Article 021322.
    1. Murphy SV, Atala A. 3D bioprinting of tissues and organs. Nat Biotechnol. 2014;32(8):773–785. - PubMed
    1. Choong YYC, Tan HW, Patel DC, Choong WTN, Chen C-H, Low HY, Tan MJ, Patel CD, Chua CK. The global rise of 3D printing during the COVID-19 pandemic. Nat Rev Mater. 2020;5(9):637–639. - PMC - PubMed
    1. Jariwala SH, Lewis GS, Bushman ZJ, Adair JH, Donahue HJ. 3D printing of personalized artificial bone scaffolds, 3D Print Addit Manuf. 2015;2(2):56–64. - PMC - PubMed
    1. Pinargote N, Smirnov A, Nikita P, Seleznev A, Peretyagin P. Direct ink writing technology (3D printing) of graphene-based ceramic nanocomposites: A review. Nano. 2020;10: Article 1300. - PMC - PubMed

LinkOut - more resources