Smartphone-Based Fundus Imaging: A Systematic Review and Gap Analysis

Abstract

Smartphone-based fundus imaging (SBFI) has emerged as a cost-effective and portable alternative to conventional fundus cameras, particularly in low-resource and underserved areas. Traditional retinal imaging systems are often expensive, bulky, and require trained specialists, limiting access to early detection of diseases such as diabetic retinopathy and glaucoma. Although smartphones offer a practical solution due to their accessibility and compatibility with telemedicine, challenges remain in terms of image quality, standardization, and the applicability of artificial intelligence (AI) models.
This study utilized a systematic review and gap analysis of existing literature on smartphone-based fundus imaging. Relevant studies were collected from electronic databases including PubMed, Scopus, Web of Science, Google Scholar, and selected Philippine medical journals. A purposive sampling approach was applied to select studies focusing on SBFI systems, imaging techniques, and AI applications. Data were extracted using a standardized format and analyzed through descriptive and thematic methods to identify patterns, performance outcomes, and research gaps.
The findings show that SBFI can achieve diagnostic performance comparable to conventional fundus cameras, with several studies reporting high sensitivity and specificity in detecting retinal diseases. Low-cost imaging solutions, such as 3D-printed adapters and handheld lenses, demonstrated feasibility in both clinical and community settings. However, variability in image quality, limited field of view, and dependence on operator skill were consistently observed. Most studies relied on manual grading, with limited use of AI for automated analysis. Key barriers include lack of standardized imaging protocols, inconsistent image quality, and limited validation of AI models for smartphone images.
SBFI is a promising tool for expanding retinal screening in low-resource settings. However, improvements in image standardization, AI adaptability, and implementation strategies are necessary to support wider adoption.