This study presents AIEgen-Deep, a novel classification program that integrates AIEgen fluorescent dyes, deep learning algorithms, and the Segment Anything Model (SAM) for precise cancer cell identification. The approach reduces manual annotation efforts by 80%-90% and achieves a 75.9% accuracy rate in recognizing cancer cell morphology across 26,693 images of eight cell types. In distinguishing healthy from cancerous cells, it attains an accuracy of 88.3% and a recall rate of 79.9%. The model effectively differentiates healthy cells (fibroblast and WBC) from various cancer cells (breast, bladder, and mesothelial), with accuracies of 89.0%, 88.6%, and 83.1%, respectively. AIEgen-Deep's applicability across different cancer types is expected to enhance early detection and improve patient survival rates.
For more info, DOI: 10.1016/j.bios.2024.116086
Comments