Computational Analysis of Histological Images of Tissue Engineered Cartilage for Evaluation of Scaffold Cell Migration

Abstract

Human chondrocytes were seeded on porcine collagen scaffolds and cultivated for up to six weeks in a cartilage bioreactor. To evaluate the influence of cultivation parameters on the proliferation and migration of the cells into the scaffolds, microscopic images from histological and immunohistochemical stainings were taken and digitalized. For evaluation of these pictures, image processing algorithms have been developed that enable quantitative conclusions with regards to aggrecan and collagen type I concentrations as well as the number of cell nuclei within the scaffold and respective migration depths. Furthermore, the number of scaffold lacunae and their orientation relative to the scaffold´s surface can be determined. A total of 85 images of different cultivations under various conditions were processed and the results evaluated by an expert. Additionally, the findings were related to results of available conventional biochemical laboratory results. The outcomes showed very few minor flaws but were valid in most cases. Some findings - as the distribution of the total cell number between cells on the surface and inside the scaffold - are superior to conventional laboratory methods that do not give this insight. A further advantage compared to the established common expert evaluation of these images, is that this approach is faster and less dependent on the judgement of the individual expert and offers quantitative results. The software development will be continued and applied for further optimizing of cartilage culture conditions.