Artificial Intelligence in Oculoplastics: A Review

Innovations in artificial intelligence (AI) have paved the way for automated analysis and detection of various anterior and posterior segment diseases in ophthalmology. The goal of this study is to review the literature surrounding the utilization of AI for oculoplastic diseases or conditions.

A literature search was conducted on Pubmed and Embase using a combination of terms including AI and oculoplastics, blepharoptosis, eyelid ptosis, brow lift, blepharoplasty, eyelid aesthetics, entropion, eyelid reconstruction, eyelid malposition, periocular cancer, periorbital melanoma, eyelid tumor, eyelid cancer, mohs reconstruction, orbital inflammation, orbital decompression, thyroid eye disease, thyroid ophthalmopathy, giant cell arteritis, lacrimal gland or lacrimal duct, dacryoadenitis, orbital fracture, orbital infections. AI large language models were excluded.

This review included 19 studies on eyelid and brow ptosis, 16 on periocular and intraocular cancer, 9 on orbital trauma and tumors, 17 on autoimmune, infectious, and inflammatory conditions of the orbit and malformations, and 30 on thyroid eye disease, with focus on clinical diagnosis, severity, and prediction of therapeutic outcomes.

Oculoplastics has seen an expansion of potential uses for AI, focusing on diagnosis of ptosis and brow ptosis, detection of eyelid cancer from pathology slides and clinical photographs, detection and diagnosis of orbital tumors and orbital processes, evaluation of radiographic images, and detection and monitoring of thyroid eye disease. Across applications, results may be promising, however, there are significant challenges such as data privacy, ethical considerations, acquisition and analysis of nonstandardized imaging data, and the need for continuous validation of AI algorithms in real-world clinical settings.