THURSDAY, June 25, 2026 (HealthDay News) -- Integration of smartphone-based imaging and artificial intelligence (AI)-driven diagnostics provides an effective strategy for screening for rare ocular malignancies, according to a study published online June 4 in JAMA Ophthalmology.
Ruixin Wang, M.D., Ph.D., from Sun Yat-sen University in China, and colleagues developed and validated a smartphone-based, media-facilitated AI system for proactive self-screening of ocular surface malignancies. A deep learning model was trained and validated using 12 years of multicenter slitlamp images. The system was optimized for smartphone-based photography, then deployed through a mobile application. Images of suspected lesions were captured using the CaptureTumor standardized smartphone application, incorporating real-time AI-guided photography instructions. Immediate binary (benign versus malignant) and multiclass risk stratification was provided by the application; high-risk cases were triaged for expedited clinical referral.
A total of 256,053 individuals were reached through multimedia outreach, with 614 completing at-home self-screening through the app. The researchers found that the smartphone-based CaptureTumor achieved an area under the receiver operating characteristic curve (AUC) of 0.905 after optimizing image quality, comparable with performance of the slitlamp-based model (AUC, 0.945). Twenty malignancies were pathologically confirmed during real-world screening, with 19 of 20 newly diagnosed (95 percent); no cases required enucleation. CaptureTumor had an AUC of 0.977 at the population level, with sensitivity and specificity of 89.3 and 95.9 percent, respectively.
"This mobile health model offers a potentially scalable, accessible, and affordable strategy for early detection of rare, vision- and life-threatening diseases," the authors write.
