A machine learning-based multiscale model to predict bone formation in scaffolds

Material Design of Porous Hydroxyapatite Ceramics via Inverse Analysis of an Estimation Model for Bone-Forming Ability Based on Machine Learning and Experimental Validation of Biological Hard Tissue Responses

Hydroxyapatite and β-tricalcium phosphate have been clinically applied as artificial bone materials due to their high biocompatibility. The development of artificial bones requires the verification of safety and efficacy through animal experiments; however, from the viewpoint of animal welfare, it is necessary to reduce the number of animal experiments. In this study, we utilized machine learning to construct a model that estimates the bone-forming ability of bioceramics from material fabrication conditions, material properties, and in vivo experimental conditions. We succeeded in constructing two models: 'Model 1', which predicts material properties from their fabrication conditions, and 'Model 2', which predicts the bone-formation rate from material properties and in vivo experimental conditions. The inclusion of full width at half maximum (FWHM) in the feature of Model 2 showed an improvement in accuracy. Furthermore, the results of the feature importance showed that the FWHMs were the most important. By an inverse analysis of the two models, we proposed candidates for material fabrication conditions to achieve target values of the bone-formation rate. Under the proposed conditions, the material properties of the fabricated material were consistent with the estimated material properties. Furthermore, a comparison between bone-formation rates after 12 weeks of implantation in the porcine tibia and the estimated bone-formation rate. This result showed that the actual bone-formation rates existed within the error range of the estimated bone-formation rates, indicating that machine learning consistently predicts the results of animal experiments using material fabrication conditions. We believe that these findings will lead to the establishment of alternative animal experiments to replace animal experiments in the development of artificial bones.

Geometric Mismatch Promotes Anatomic Repair in Periorbital Bony Defects in Skeletally Mature Yucatan Minipigs

Porous tissue-engineered 3D-printed scaffolds are a compelling alternative to autografts for the treatment of large periorbital bone defects. Matching the defect-specific geometry has long been considered an optimal strategy to restore pre-injury anatomy. However, studies in large animal models have revealed that biomaterial-induced bone formation largely occurs around the scaffold periphery. Such ectopic bone formation in the periorbital region can affect vision and cause disfigurement. To enhance anatomic reconstruction, geometric mismatches are introduced in the scaffolds used to treat full thickness zygomatic defects created bilaterally in adult Yucatan minipigs. 3D-printed, anatomically-mirrored scaffolds are used in combination with autologous stromal vascular fraction of cells (SVF) for treatment. An advanced image-registration workflow is developed to quantify the post-surgical geometric mismatch and correlate it with the spatial pattern of the regenerating bone. Osteoconductive bone growth on the dorsal and ventral aspect of the defect enhances scaffold integration with the native bone while medio-lateral bone growth leads to failure of the scaffolds to integrate. A strong positive correlation is found between geometric mismatch and orthotopic bone deposition at the defect site. The data suggest that strategic mismatch >20% could improve bone scaffold design to promote enhanced regeneration, osseointegration, and long-term scaffold survivability.

Nondeterministic multiobjective optimization of 3D printed ceramic tissue scaffolds

Despite significant advances in the design optimization of bone scaffolds for enhancing their biomechanical properties, the functionality of these synthetic constructs remains suboptimal. One of the main challenges in the structural optimization of bone scaffolds is associated with the large uncertainties caused by the manufacturing process, such as variations in scaffolds' geometric features and constitutive material properties after fabrication. Unfortunately, such non-deterministic issues have not been considered in the existing optimization frameworks, thereby limiting their reliability. To address this challenge, a novel multiobjective robust optimization approach is proposed here such that the effects of uncertainties on the optimized design can be minimized. This study first conducted computational analyses of a parameterized ceramic scaffold model to determine its effective modulus, structural strength, and permeability. Then, surrogate models were constructed to formulate explicit mathematical relationships between the geometrical parameters (design variables) and mechanical and fluidic properties. The Non-Dominated Sorting Genetic Algorithm II (NSGA-II) was adopted to generate the robust Pareto solutions for an optimal set of trade-offs between the competing objective functions while ensuring the effects of the noise parameters to be minimal. Note that the nondeterministic optimization of tissue scaffold presented here is the first of its kind in open literature, which is expected to shed some light on this significant topic of scaffold design and additive manufacturing in a more realistic way.

Identification and ranking biomaterials for bone scaffolds using machine learning and PROMETHEE

Purpose

Bones have a complex hierarchical structure that supports their diverse chemical, biological, and mechanical functions. High rates of bone susceptibility to fractures and injury have attracted extensive research interest to find alternate biomaterials for bone scaffolds. Natural bone healing is only successful if the defect is very small and when a defect exceeds 1 cm3 then bone grafting is required. Large bone defects or injuries are very serious problems in orthopedics as they bring great harm to health and normal function of daily life routine. A scaffold should have good strength to maintain its own structure after implantation in a load bearing environment and without being stiff that shields surrounding bone from the load. Therefore, mechanical properties of bone scaffolds should match those of the host tissue and should be part of the natural environment of the body without any harm or further damage.

Methods

In this paper, we present two main contributions. First, we investigate the use of machine learning models in identifying biomaterials that are suitable for bone scaffolds. Second, we rank the best materials for biomedical scaffold applications using the multi-criteria decision analysis methods, the Preference Ranking Organization METhod for the Enrichment of Evaluations (PROMETHEE). Machine learning models investigated are AdaBoost, artificial neural network (ANN), Naïve Bayes (NB), Decision tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN). Mechanical properties such as comprehensive strength, tensile strength, and Young’s modulus with the cortical bone are used as the standard reference for classification.

Results

The results show that the ANN outperforms the other machine learning models in identifying the biomaterials suitable for bone tissue engineering, while the ranking results using PROMETHEE show that Brushite and Titanium alloy are the best appropriate biomaterials for the cancellous and cortical bones, respectively.

Conclusion

Brushite and Titanium alloy are the best biomaterials for the cancellous and cortical bones, respectively.

Machine Learning in Tissue Engineering

Machine learning (ML) and artificial intelligence have accelerated scientific discovery, augmented clinical practice, and deepened fundamental understanding of many biological phenomena. ML technologies have now been applied to diverse areas of tissue engineering research, including biomaterial design, scaffold fabrication, and cell/tissue modeling. Emerging ML-empowered strategies include machine-optimized polymer synthesis, predictive modeling of scaffold fabrication processes, complex analyses of structure-function relationships, and deep learning of spatialized cell phenotypes and tissue composition. The emergence of ML in tissue engineering, while relatively recent, has already enabled increasingly complex and multivariate analyses of the relationships between biological, chemical, and physical factors in driving tissue regenerative outcomes. This review highlights the novel methodologies, emerging strategies, and areas of potential growth within this rapidly evolving area of research. Impact statement Machine learning (ML) has accelerated scientific discovery and augmented clinical practice across multiple fields. Now, ML has driven exciting new paradigms in tissue engineering research, including machine-optimized biomaterial design, predictive modeling of scaffold fabrication, and spatiotemporal analysis of cell and tissue systems. The emergence of ML in tissue engineering, while relatively recent, has already enabled increasingly complex analyses of the relationships between biological, chemical, and physical factors in driving tissue regenerative outcomes. This review highlights the novel methodologies, emerging strategies, and areas of potential growth within this rapidly evolving area of research.

Influence of the parameters of the convolutional neural network model in predicting the effective compressive modulus of porous structure

In recent years, the convolutional neural network (CNN) technique has emerged as an efficient new method for designing porous structure, but a CNN model generally contains a large number of parameters, each of which could influence the predictive ability of the CNN model. Furthermore, there is no consensus on the setting of each parameter in the CNN model. Therefore, the present study aimed to investigate the sensitivity of the parameters in the CNN model for the prediction of the mechanical property of porous structures. 10,500 samples of porous structure were randomly generated, and their effective compressive moduli obtained from finite element analysis were used as the ground truths to construct and train a CNN model. 8,000 of the samples were used to train the CNN model, 2000 samples were used for the cross-validation of the CNN model and the remaining 500 new structures, which did not participate in the CNN training process, were used to test the predictive power of the CNN model. The sensitivity of the number of convolutional layers, the number of convolution kernels, the number of pooling layers, the number of fully connected layers and the optimizer in the CNN model were then investigated. The results showed that the optimizer has the largest influence on the training speed, while the fully connected layer has the least impact on the training speed. Additionally, the pooling layer has the largest impact on the predictive ability while the optimizer has the least impact on the predictive ability. In conclusion, the parameters of the CNN model play an important role in the performance of the CNN model and the parameter sensitivity analysis can help optimize the CNN model to increase the computational efficiency.