Quality Assessment of Mesenchymal Stem Cells Using Deep Learning Based Image Analysis

Abstract

Cell-based therapeutics is a current effective strategy for the potential curing of several human diseases. Mesenchymal stem cells (MSCs) are a heterogeneous group of cells that have been the subject of recent attention because of their clinically relevant therapeutic effects and transformative morphology. The success of MSCs to provide new remedies is dependent on their quality. Their quality can be assessed by examining their physical nature. Morphological evaluation has been a robust method for monitoring culture quality, but standard techniques are either subjective, destructive, or time consuming making real-time monitoring difficult. The goal is to develop an automated image analysis algorithm using deep learning to assess the viability of MSCs. An algorithm using Keras and TensorFlow libraries in Python will be our main method for phase contrast microscope images of MSCs. The cell images are first preprocessed and then given to the U-Net architecture model for the segmentation of cells in the images. Results were validated using the manual outlining of cells by MSC culture experts as the ground truth. . In summary, the proposed technique shows the potential to be incorporated into automated MSC quality control processes.