Overview
This project is a collaborative work within the THRIVE centre between biotherapeutic team and AI team to investigate the feasibility of using AI to predict/optimise the physicochemical and mechanical properties of 3D printed IPC scaffolds, thereby reducing the time and data needed for experimentations.
Retinal tissue engineering represents a new approach to treat retinal diseases through the development of a biological substitute: the so-called “scaffold.” A tissue scaffold is a three-dimensional (3D) structure with interconnected pores which are used to deliver cells, drugs and genes into the local tissue. Scaffolds provide a suitable space in which retinal pigment epithelium can grow and generate their own extracellular matrix (ECM). Collagen is the most abundant structural protein and plays a critical role in maintaining the ECM, which has been widely used to fabricate scaffolds for tissue engineering because of its biodegradability, superior biocompatibility and weak antigenicity. However, some limitations exist with cross-linked collagens. These include the use of leachable toxic crosslinkers, a short duration of action, activation of T-cells and difficult injection of the polymerised networked materials. To address these limitations, an in-situ polymerizable collagen (IPC), which is fibrilised without uses of a chemical cross linker, is proposed to use as a scaffold for this project.
To fabricate 3D IPC scaffolds with specific shapes and sizes, different methods will be applied, such as freeze-drying and 3D bioprinting. Multi-objective optimisations are, however, required to identify suitable freeze-drying and printing conditions and material composition (percentage of cells to IPC) to achieve optimal mechanical and porous constructions properties. This requires extensive experimentation time that is resource demanding. Artificial intelligence and machine learning (ML) methods have already been applied in tissue engineering and have been shown to be transformative resources to support researchers in the field of regenerative medicine. The ML-based framework takes the material composition and the printing parameters as input to predict printing parameters to optimise the structure’s properties, assess the quality of the prints and optimise the printability of the material.
Aims
The key innovation lies in combining AI modelling with bioprinting and freeze-drying techniques to optimise the scaffold design process. Instead of relying on trial-and-error experimentation, we use machine learning to:
- Predict printability, mechanical strength, and porosity of IPC scaffolds
- Optimise freeze-drying and printing parameters for target tissue integration
- Model the relationship between cell concentration, material composition, and construct quality
- Accelerate scaffold customisation for patient-specific applications
Significance
Scaffolds fabricated through this AI-aided framework aim to:
- Enable minimally invasive ocular implantation
- Improve scaffold reproducibility and reduce immunogenic response
- Shorten development timelines in regenerative ophthalmology
This work advances precision biomaterial design for vision science and sets a precedent for integrating AI in scaffold fabrication workflows.
Collaborators
This is an in-house collaborative initiative at the THRIVE Centre, led by:
- The Biotherapeutics Team, specialising in collagen engineering and regenerative pharmacology
- The AI & Imaging Team, focusing on machine learning for optimisation and predictive modelling
Technology Stack
- Scaffold Fabrication: 3D bioprinting, freeze-drying, in-situ polymerisation
- AI/ML Tools: Supervised learning for regression and optimisation, multi-objective modelling
- Material Evaluation: Mechanical testing, porosity characterisation, live-cell assays
This project exemplifies THRIVE’s integrated approach to translational research—fusing biomaterials science with artificial intelligence to create next-generation ocular implants.