Artificial Intelligence Algorithms for Analysis of Geographic Atrophy: A Review and Evaluation

Abstract

Purpose The purpose of this study was to summarize and evaluate artificial intelligence (AI) algorithms used in geographic atrophy (GA) diagnostic processes (e.g. isolating lesions or disease progression). Methods The search strategy and selection of publications were both conducted in accordance with the Preferred of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed and Web of Science were used to extract literary data. The algorithms were summarized by objective, performance, and scope of coverage of GA diagnosis (e.g. lesion automation and GA progression). Results Twenty-seven studies were identified for this review. A total of 18 publications focused on lesion segmentation only, 2 were designed to detect and classify GA, 2 were designed to predict future overall GA progression, 3 focused on prediction of future spatial GA progression, and 2 focused on prediction of visual function in GA. GA-related algorithms reported sensitivities from 0.47 to 0.98, specificities from 0.73 to 0.99, accuracies from 0.42 to 0.995, and Dice coefficients from 0.66 to 0.89. Conclusions Current GA-AI publications have a predominant focus on lesion segmentation and a minor focus on classification and progression analysis. AI could be applied to other facets of GA diagnoses, such as understanding the role of hyperfluorescent areas in GA. Using AI for GA has several advantages, including improved diagnostic accuracy and faster processing speeds. Translational Relevance AI can be used to quantify GA lesions and therefore allows one to impute visual function and quality-of-life. However, there is a need for the development of reliable and objective models and software to predict the rate of GA progression and to quantify improvements due to interventions.

Clinically relevant deep learning for detection and quantification of geographic atrophy from optical coherence tomography: a model development and external validation study

“…Although not directly comparable, greater DSCs were observed than those reported in a systematic review from 2020 of segmentation algorithms for geographic atrophy (range 0•68-0•89). 15 Currently, slowing growth rate of geographic atrophy and thus vision loss is an accepted clinical trial endpoint. The segmentation approach presented here can support rapid, standardised, and scalable assessment of geographic atrophy over time.…”

“…10 A promising advance is the use of artificial intelligence to develop models that automatically process OCT images, quantifying each of their anatomical constituents in three dimensions and providing quantitative OCT parameters. [11][12][13][14][15] Automated quantitative OCT generates rapidly accessible, objective data and is thus ideally suited to standardise OCT measurements across institutions.…”

Deep Learning Applied to Automated Segmentation of Geographic Atrophy in Fundus Autofluorescence Images

“…The cohort of images and patients was quite large and diverse as compared to others in the GA-AI segmentation space. 11 The cohort consisted of 99 eyes, 49 left eyes (49.5%) and 50 right eyes (50.5%). A total of 359 images were for the left eye and 343 images were for the right eye.…”

“…Several groups have, however, developed artificial intelligence (AI)-based automation methods for isolation of GA lesions. [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] These studies applied a process known as semantic segmentation-the labeling of each pixel within an image and the precise extraction of regions of interest. The range of algorithms tested included regiongrowing, interactive segmentation using watershed transform, level set approach, geometric active contour model, Fuzzy c-means, k-nearest neighbor (kNN), Chan-Vese model via local similarity factor, convolutional neural networks (CNN), sparse autoencoder deep networks, and an offline/self-learning model.…”

Modeling of atrophy size trajectories: variable transformation, prediction and age-of-onset estimation

“…These areas enlarge with time and lead to irreversible loss of visual function [ 7 ]. A relevant clinical measure of disease progression is the eye-specific size of GA which can be quantified based on imaging techniques including color fundus photography, spectral domain optical coherence tomography imaging, or fundus autofluorescence (FAF) imaging [ 8 , 9 ].…”