Related Experiment Video
Updated: Jul 25, 2025
08:58
Artificial Intelligence Approaches to Assessing Primary Cilia
Published on: May 1, 2021
3.5K
Automated identification and quantification of activated dendritic cells in central cornea using artificial
Harry Levine1, Arianna Tovar1, Adam K Cohen1
1Bascom Palmer Eye Institute, University of Miami Miller School of Medicine, Miami, FL, 33136, USA; Miami Veterans Administration Medical Center, 1201 NW 16th St, Miami, FL, 33125, USA.
The Ocular Surface
|June 29, 2023
Summary
An automated algorithm effectively quantifies activated dendritic cells (aDCs) in the cornea using in-vivo confocal microscopy (IVCM) images. This AI-driven approach shows excellent agreement with manual counting, offering a reliable method for research and clinical applications.
Area of Science:
- Ophthalmology
- Immunology
- Medical Imaging
Background:
- Activated dendritic cells (aDCs) play a role in ocular surface inflammation.
- Accurate quantification of aDCs is crucial for understanding and managing conditions like Dry Eye (DE).
- In-vivo confocal microscopy (IVCM) allows for real-time imaging of corneal cells.
Purpose of the Study:
- To validate an automated algorithm for quantifying activated dendritic cells (aDCs) using in-vivo confocal microscopy (IVCM) images.
- To compare the accuracy of automated aDC quantification with manual counting methods.
- To assess the algorithm's performance across different subtypes of Dry Eye (DE).
Main Methods:
- Retrospective analysis of 173 IVCM images from 86 individuals.
- Quantification of aDCs performed using both an automated algorithm and manual counting.
- Comparison of automated and manual counts using Intra-class-correlation (ICC) and Bland-Altman plots.
- Secondary analysis grouped participants by DE subtype: aqueous-tear deficiency (ATD) and evaporative DE (EDE).
Main Results:
- The automated algorithm showed excellent agreement with manual aDC counts (ICC = 0.80).
- A small difference was observed between automated and manual quantification (0.19, p < 0.01).
- Similar high agreement was found across DE subtypes: ATD (ICC=0.75), EDE (ICC=0.80), and controls (ICC=0.82).
Conclusions:
- An automated, machine learning-based algorithm can reliably quantify aDCs in the central cornea using IVCM.
- AI-driven analysis demonstrates comparable results to manual quantification.
- Further longitudinal research in diverse populations is recommended to validate these findings.
