Mobile Device-Enabled Proactive Self-Assessment for Ocular Surface Cancers: A Nonrandomized Clinical Investigation

To develop and validate a smartphone-based screening system for ocular surface tumors in China.

Study Design: A prospective, 3-stage study was conducted to create an AI-driven diagnostic system called CaptureTumor (CaT) for smartphone deployment.
Stages of Development: Stage 1 involved training the deep-learning model using clinical slitlamp images. Stage 2 adapted the model for smartphone photography. Stage 3 assessed the real-world implementation of the mobile application.
Diagnostic Taxonomy: A publicly oriented classification of pigmented ocular lesions was developed, including malignant and benign tumors.

CaptureTumor (CaT) integrates public education, AI-facilitated self-screening, and structured referral pathways.
The study received ethical approval and was registered at ClinicalTrials.gov.

The integrated framework of CaT aims to improve access to specialist care for rare ocular conditions through a closed-loop mobile health ecosystem.

The study is nonrandomized and may have biases inherent to observational studies.
The effectiveness of the mobile application in diverse populations and settings remains to be fully evaluated.

The CaT system represents a strategy to enhance resource allocation and equity in managing rare ocular diseases.