Smartphone App Screens for Rare Ocular Surface Malignancies
Smartphone App Screens for Rare Ocular Surface Malignancies
Publish Date: 2026-06-05 00:00:00
Source Domain: www.optometryadvisor.com
-
Development and Validation of CaptureTumor App: A smartphone-based AI tool known as the CaptureTumor app was developed to screen for rare ocular surface malignancies. The app uses a deep learning model trained on over a decade of slit-lamp images to identify and differentiate between benign and malignant lesions.
-
Efficient Case Detection: The app was found to be up to 5 times more efficient than conventional referral pathways. It projected an increase in case detection from 8 to 40 per center per year, making the screening process more accessible.
-
Wide Participation and Usability: Through various platforms, 256,053 individuals were educated about ocular lesions, and eventually, 614 participants (out of which 805 images were analyzed) used the app for self-screening, demonstrating high engagement and accessibility.
-
Comparable Accuracy to Specialist Exams: Although the AI model had an area under the curve (AUC) of 0.905 compared to the expert slit-lamp model’s 0.945, the app showed high sensitivity and specificity (89.3% and 95.9%, respectively) in diagnosing ocular surface malignancies.
-
Reduced Specialist Referrals: Users of the app needed significantly fewer referrals (1.02 vs 3.69) to see a specialist, showcasing operational efficiency and reducing healthcare load.
-
High Diagnostic Efficiency and Economic Viability: With over 200,000 individuals diagnosed with 20 malignant cases within 6 months, the tool provides a cost-effective and efficient screening method for rare ocular conditions.
-
Cross-Population Generalizability: While the trial was mostly conducted in China, it included a diverse group of patients suggesting potential effectiveness across different populations.
-
Successful Implementation and Minimal Complications: The app ensured high usability, with nearly all users correctly following instructions. It successfully diagnosed malignant cases without requiring enucleations or orbital exenterations, demonstrating safe and correct diagnosis.
