3D-printed barriers with machine learning powered image analysis for enhanced wound healing assays

Wound healing assay is a standard method enabling investigation of cell proliferation and migration through a cell-free gap in a cell monolayer. Despite very common, it shows several weaknesses: lack of reproducibility and manual and time-based image analysis. Based on novel approach founded on innovative materials and AI-assisted processing of biological images, a promising automated barrier-wound healing assay is realized, achieving consistent results while retaining cells integrity. To increase assay accuracy, biocompatible 3D-printed resin inserts have been developed, facilitating precise control over shape and size of the wound. In parallel, a new image-detection algorithm powered by Deep Learning models was developed to identify cell-free area during the healing process, exceeding limitations of manual analysis. 3D-resin inserts combined with automated image analysis allowed the elimination of subjective errors and provided reproducible quantification of cell-free areas across multiple experiments.

Moreover, a dataset to train a Convolutional Neural Network for monitor healing over time was developed.

As proof of concept, this algorithm was tested on a cancer cell line stimulated by TGF-β, a drug stimulating cell migration.

Innovative design of biocompatible materials combined with Deep Learning for automatically processing high-throughput data enables standardized wound healing assay, increasing efficiency, reliability, and accuracy of results.