Yuki Kikuchi1,Yuki Okuhashi1,Hiroaki Ishihata1,Misato Kashiba1,Satoshi Sasaki1
Tokyo Univ of Technology1
Yuki Kikuchi1,Yuki Okuhashi1,Hiroaki Ishihata1,Misato Kashiba1,Satoshi Sasaki1
Tokyo Univ of Technology1
Materials that cause apoptosis of cancer cells should potentially be candidates for future anti-cancer drug. Apoptosis cell classification using AI will be an effective method to evaluate the effect. Characteristics of apoptosis cells are effectively identified from the fluorescence-stained images. The effectiveness will increase if the classification could be performed without stains.<br/>γ secretase inhibitor (GSI) is reported to cause apoptosis of K562, a human chronic myelogenous leukemia cell line. We found that AI-based classification results of phase-contrast images of the cells were quite similar to those of fluorescence images. Factors that AI identified, but naked eyes could not identify, were examined. Biochemical change inside cells were studied in connection with morphological characteristics.