2D Light scattering images analyzed by deep learning algorithm for label-free differentiation of dead and live colonic adenocarcinoma cells. Issue 1 (May 2021)

• Record Type: Journal Article
• Authors: Li, Shuaiyi; Li, Ya; Yao, Jianning; Chen, Bing; Song, Jiayou; Xue, Qi; Yang, Xiaonan
• Is Part Of: Journal of physics. Volume 1914:Issue 1(2021)
• Publication Date: 2021-05
• Subjects: Label-free -- Colon cancer cells -- 2D Light scattering -- Deep learning; Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1914/1/012007
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: The detection of cell viability or the detection of the percentage of live and dead cells in a sample of cells is an important parameter. At present, the common methods for cell viability determination mainly rely on the responses to cell dyes. However, the additional need for cell staining will consequently cause time-consuming and laborious efforts. Furthermore, the determination of cell viability by cell staining is invasive and may damage the internal structure of cells. In this work, we proposed a label-free method to classify live and dead colonic adenocarcinoma cells by 2D light scattering combined with deep learning algorithm. The deep convolutional network of YOLO-v3 was used to identify and classify light scattering images of live and dead HT29 cells. This method achieved an excellent sensitivity (92.16%), specificity (94.23%), and accuracy (93.2%). The results show that the combination of 2D light scattering images and deep neural network may provide a new label-free method for cellular analysis.