Research Progress in Artificial Intelligence for Central Serous Chorioretinopathy: A Systematic Review

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Central serous chorioretinopathy (CSCR) diagnosis and treatment can be improved with advanced technologies.
• Artificial intelligence (AI) offers promising avenues for enhancing diagnostic accuracy and therapeutic strategies in CSCR.

Purpose of the Study:

• To review and synthesize current advancements in AI applications for CSCR.
• To identify challenges and future research directions for personalized CSCR diagnostics and therapeutics.

Main Methods:

• A systematic literature search was performed in Web of Science using AI and CSCR keywords.
• 73 original research studies were selected from 698 records based on predefined inclusion criteria.

Main Results:

• AI models demonstrate superior performance in CSCR classification and lesion segmentation compared to clinical experts, utilizing multimodal data fusion (OCT, OCTA, FFA).
• Clinical translation is hindered by infrastructural barriers and lack of physician trust due to "black box" AI models; explainable AI (XAI) is emerging as a solution.
• Limitations include single-center data, annotation variability, and static frameworks unable to capture dynamic lesion progression.

Conclusions:

• AI significantly enhances the efficiency of CSCR diagnosis and subtyping.
• Future research priorities include multicenter data integration, dynamic visualization, standardized guidelines, and explainable AI to improve clinical adoption and physician trust.
• Federated learning and prospective trials are crucial for developing personalized treatment strategies and real-time insights into lesion progression.