Annotation of antibody-dependent cellular phagocytosis video images by ImageJ plugin provides ground truth for artificial intelligence development 3372

Le résumé fourni par la source

Abstract Description Monoclonal antibodies (mAb) have significantly improved outcomes in patients with hematological malignancies such as chronic lymphocytic leukemia (CLL). Clearance of circulating CLL cells by mAb is primarily by antibody-dependent cellular phagocytosis (ADCP). In vitro time-lapse video microscopy of ADCP with macrophages and mAb-coated CLL cells recapitulates the clinical rapid loss of mAb effectiveness within the first hour of therapy. The data generated by video imaging is substantial, requiring commercial microscopy software for expeditious analysis. However, significant frame-by-frame variability is generated. To improve ADCP video imaging analysis, manually annotated “ground truth” video image data is required to determine the error in commercial software methods and to train improved software, including Artificial Intelligence methods. We developed the ImageJ plugin Seg2Tracks for rapid semi-automated image annotation of ADCP macrophage imaging data. High content microscopic video imaging files of cells undergoing ADCP were collected using a Nikon Eclipse Ti-E Live Cell Imaging System and converted to *.tif files for annotation. Seg2Tracks allowed each annotator to produce rapid consistent annotations with minimal variability, providing ground truth data that highlighted the need for improvement in our NIS-Elements software method. Ground truth data obtained with Seg2Tracks will enable the development/training of improved software to enhance ADCP video imaging analysis. Funding Sources We thank the University of Rochester Schwartz Discover Grant, the University of Rochester Medical Center Robert I. Weed Summer Fellowship, the Cadregari Foundation at the University of Rochester, gifts from Ms. Elizabeth Aaron, and gifts from Mr. Lawrence Halpern for funding that enabled this research. Topic Categories Technological Innovations in Immunology (TECH)

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.