 |
Ocular
Microbiology and Immunology Group
Back
to OMIG Main Page
< Previous | 2021
Agenda and Abstracts | Next >
2021 OMIG Abstract
Automated Quantification of Activated Dendritic Cells in Central Cornea
Harry Levine1,2, Adam Cohen-Karp1,2, Anat Galor1,2 Brian Goldhagen1,2
1Miami Veterans Administration Medical Center, Miami, Florida; 2 Bascom Palmer Eye Institute,
University of Miami Miller School of Medicine, Miami, Florida
Purpose: In this study, we sought to validate an algorithm quantifying activated dendritic cells (aDCs) in confocal microscopy images of the central cornea.
Methods: Retrospective analysis was conducted on in-vivo confocal microscopy (IVCM) images obtained at the Miami Veterans Hospital. Confocal images from individuals with corneal scarring were excluded from analysis. An automated aDC counter was developed using machine learning and utilized transfer learning with confocal images that were acquired prior to the start date of this study. It was ensured that the validation dataset did not include any individuals whose images may have been used in the development of this algorithm. ADCs were manually quantified based on morphology by reviewers masked to the algorithm findings, and intra-class correlation (ICC) was used to compare automated and manual counts.
Results: 193 non-overlapping images from 110 individuals were included. The mean age was 55.5±18.4 years; 70.0% were male, 58.2% self-identified as White and 24.5% as Hispanic. The mean number of aDCs in the central cornea as quantified manually was 1.21±1.98 cells/image (7.56±12.38 cells/mm2). The algorithm identified a total of 207 aDCs within the dataset, of which 184 were manually verified as aDCs. The automated algorithm performed the aDCs count with 87% accuracy and an ICC of 83% (p<0.01).
Conclusions: The number of aDCs in the central cornea can be successfully estimated with the use of an automated deep learning algorithm, with comparable results to manual quantification. Further assessment is needed before the widespread clinical implementation of an automated aDC counter.
Disclosure: N
Support: Supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Clinical Sciences R&D (CSRD) I01 CX002015 (Dr. Galor) and Biomedical Laboratory R&D (BLRD) Service I01 BX004893 (Dr. Galor), Department of Defense Gulf War Illness Research Program (GWIRP) W81XWH-20-1-0579 (Dr. Galor) and Vision Research Program (VRP) W81XWH-20-1-0820 (Dr. Galor), National Eye Institute R01EY026174 (Dr. Galor) and R61EY032468 (Dr. Galor), NIH Center Core Grant P30EY014801 (Institutional) and Research to Prevent Blindness Unrestricted Grant (Institutional), Consejo Nacional de Ciencia y Tecnología (CONACYT) CVU810654 (H. Levine).
< Previous | 2021
Agenda and Abstracts | Next >
|
|
