Histological Image Processing for the Assessment of Tissue Engineered Cartilage

Abstract

In a recently developed perfusion bioreactor, human nasal septum chondrocytes were seeded on decellularized porcine collagen scaffolds and cultivated for 6 weeks. Afterwards, the samples were stained with alcian blue and multiple histological microscope images were taken to assess the migration of the human cells and their redifferentiation. In this paper, we propose an image processing algorithm to objectively quantify the cultivation results. This algorithm performs a white balance, detects cell nuclei and their migration depth, and detects lacunae and their orientation. To evaluate performance, sensitivity and precision are calculated using manually labeled data over 50 images. The detection of cell nuclei has a reasonable precision (73%), but the level of sensitivity (42%) should be further improved. The proposed method demonstrates acceptable level of precision (83%) and sensitivity (86%) for the detection of lacunae.

Keywords

• image analysis
• histological images
• cartilage
• scaffold
• bioreactor
• evaluation