Related Experiment Video
Updated: Sep 16, 2025
05:51
Smartphone Fundus Photography
Published on: July 6, 2017
39.3K
Advancing AMD screening with an offline, AI-powered smartphone-based fundus camera: A prospective, real-world
Kalpa Negiloni1, Prabu Baskaran2, Divya Parthasarathy Rao3
1Remidio Innovative Solutions Pvt Ltd, Bengaluru, India. kalpa.n@remidio.com.
Eye (London, England)
|July 11, 2025
Summary
A new AI algorithm for age-related macular degeneration (AMD) screening shows high accuracy using smartphone fundus images. This technology offers a promising, accessible tool for detecting referable AMD.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Medical Imaging
Background:
- Age-related macular degeneration (AMD) is a leading cause of vision loss.
- Early detection and screening are crucial for managing AMD.
- Current screening methods can be resource-intensive.
Purpose of the Study:
- To evaluate a novel offline, AI-driven screening algorithm for age-related macular degeneration (AMD).
- To compare the AI algorithm's performance against fundus image-only grading and standard care (SD-OCT and fundus images).
Main Methods:
- Prospective study at a South Asian tertiary eye hospital.
- Used a smartphone-based non-mydriatic fundus camera for image capture.
- Compared Medios AI detection of referable AMD against expert grading of fundus images and combined SD-OCT and fundus images.
Main Results:
- The AI demonstrated strong performance with sensitivity ranging from 88.48% to 90.62% and specificity from 85.41% to 87%.
- Inter-grader agreement among specialists was strong (Cohen's Kappa: 0.81-0.84).
- False negatives were mainly intermediate AMD, and false positives were often early AMD.
Conclusions:
- The AI algorithm integrated with a smartphone fundus camera shows robust performance in identifying referable AMD.
- This technology has the potential to be an affordable and accessible screening solution for AMD.
