"Image-based analysis of cells is a powerful modality to measure and record the high content information which reflects the cellular status during their culture. Although morphology has been known for a long to contain significant information to monitor the transitions of cellular status, their analysis has been limited to experience-based interpretations. However, by the recent rapid development of image processing and AI technologies, we now have tools and platforms to tackle for understanding the real-time cellular status in a more quantitative manner. However, image-based analysis has been extremely limited to analyzing “labeled-images,” and rare applications have been challenged to measure and utilize the heterogeneity information of cells for predicting cellular quality.
Our group has been reporting the label-free morphology-based analysis approaches for developing enabling technology for cell quality monitoring and control for maximizing the efficiency and reproducibility of cellular research and manufacturing [1-3]. Practically, from the time-course microscopic images, our data processing extracts not only the morphological descriptors of individual cells, but also the populational transition information, and tags them to the experimentally obtained cell quality ground-truth data. The key point of this technology is to combine the right combination of imaging hardware, automation technology, image processing, and data processing for objective performance. It should be noted that such a deep neural network structure is not always the best solution for such AI-based quantitative analysis.
In this talk, we will show the practical successful examples to apply such image-based label-free analysis for cell quality maintenance applications in (1) mesenchymal stem cell allogenic cell bank establishment, (2) single-cell morphological analysis for detecting senescence in mesenchymal stem cells, (3) novel optical challenge for label-free evaluation of spheroids. Our results show a high potential of image-based morphological analysis to enhance the quantitative understanding of the status of cells and their culture conditions. We also discuss the limitations and technological difficulties of AI-based image analysis compared to other image-based achievements in other fields for sharing the key points to enable the image-based cell evaluation successful.
- Sasaki, H., et al., PLoS One 9(4), e93952 (2014)
- Kato R., et al., Sci. Rep., 6, 34009, (2016)
- Imai Y., et al., Inflamm. Regen. 42(1), 8-20 (2022)"