Artificial intelligence techniques in inherited retinal diseases: A review

TL;DR

The review addresses the need to bridge existing gaps in focused studies on AI’s role in IRDs, offering a structured analysis of current AI techniques and outlining future research directions, and concludes with an overview of the challenges and opportunities in deploying AI for IRDs.

Abstract

Inherited retinal diseases (IRDs) are a diverse group of genetic disorders that lead to progressive vision loss and are a major cause of blindness in working-age adults. The complexity and heterogeneity of IRDs pose significant challenges in diagnosis, prognosis, and management. Recent advancements in artificial intelligence (AI) offer promising solutions to these challenges. However, the rapid development of AI techniques and their varied applications have led to fragmented knowledge in this field. This review consolidates existing studies, identifies gaps, and provides an overview of AI's potential in diagnosing and managing IRDs. It aims to structure pathways for advancing clinical applications by exploring AI techniques like machine learning and deep learning, particularly in disease detection, progression prediction, and personalized treatment planning. Special focus is placed on the effectiveness of convolutional neural networks in these areas. Additionally, the integration of explainable AI is discussed, emphasizing its importance in clinical settings to improve transparency and trust in AI-based systems. The review addresses the need to bridge existing gaps in focused studies on AI's role in IRDs, offering a structured analysis of current AI techniques and outlining future research directions. It concludes with an overview of the challenges and opportunities in deploying AI for IRDs, highlighting the need for interdisciplinary collaboration and the continuous development of robust, interpretable AI models to advance clinical applications.

Figures (5)

• Figure 1: (A) Fundus photo demonstrating an en-face view of the optic nerve and macula. (B) Optical coherence tomography scan showing a cross-section of the macula (through green horizontal line in Figure 1A). (C) Wide-field image of the retina. (D) Short-wave autofluorescence image of the fundus.
• Figure 2: Schematic diagram illustrating the U-Net architecture. The network consists of an encoder path with convolutional and max-pooling layers, and a decoder path with up-convolutions. Skip connections between corresponding layers in the encoder and decoder allow for the combination of detailed spatial information and contextual features, enhancing segmentation accuracy. White dotted boxes represent concatenation of previous layers in the encoder path with up-convoluted layers at various stages of the decoder path.
• Figure 3: Schematic diagram illustrating the architecture of Inception V3. The network employs Inception modules with parallel convolutions using various filter sizes, enabling the extraction of multi-scale features. Pooling layers reduce the dimensionality of feature maps, while batch normalization and auxiliary classifiers improve training stability and classification accuracy.
• Figure 4: Hyper- and hypoautofluorescent flecks in Stargardt disease as visualized by (A) near-infrared autofluorescence and (B) short wave autofluorescence.
• Figure 5: Hyperautofluorescent ring in retinitis pigmentosa visualized by (A) near-infrared autofluorescence and (B) short wave autofluorescence. Abnormal pigmentation is more easily visualized with short wave autofluorescence.