Artificial Intelligence in Glaucoma: Advances in Diagnosis, Progression Forecasting, and Surgical Outcome Prediction.

Journal: International journal of molecular sciences
Published Date:

Abstract

Glaucoma is a leading cause of irreversible blindness, with challenges persisting in early diagnosis, disease progression, and surgical outcome prediction. Recent advances in artificial intelligence have enabled significant progress by extracting clinically relevant patterns from structural, functional, and molecular data. This review outlines the current applications of artificial intelligence in glaucoma care, including early detection using fundus photography and OCT and disease progression prediction using deep learning architectures such as convolutional neural networks, recurrent neural networks, transformer models, generative adversarial networks, and autoencoders. Surgical outcome forecasting has been enhanced through multimodal models that integrate electronic health records and imaging data. We also highlight emerging AI applications in omics analysis, including transcriptomics and metabolomics, for biomarker discovery and individualized risk stratification. Despite these advances, key challenges remain in interpretability, integration of heterogeneous data, and the lack of personalized surgical timing guidance. Future work should focus on transparent, generalizable, and multimodal AI models, supported by large, well-curated datasets, to advance precision medicine in glaucoma.

Authors

  • Chiao-Hsin Lan
    School of Medicine, National Defense Medical Center, Taipei 11490, Taiwan.
  • Ta-Hung Chiu
    Department of General Medicine, Tri-Service General Hospital, National Defense Medical Center, Taipei 114202, Taiwan.
  • Wei-Ting Yen
    Department of Ophthalmology, Tri-Service General Hospital, National Defense Medical Center, Taipei 114202, Taiwan.
  • Da-Wen Lu
    Department of Ophthalmology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan.