International Journal of Research and Innovation in Applied Science (IJRIAS) |Volume VIII, Issue I, January 2023|ISSN 2454-6194
Integrating Imaging Bioinformatics in Ophthalmology
Hadi Khazaei*, Danesh Khazaei, Kaneez Abbas, Davin Ashraf, John D Ng.
Casey Eye Institute, Oregon Health & Science University, Portland, Oregon, 97239,
*Corresponding author
Abstract: Imaging informatics collates the multitude of information into data; allowing research to occur, driving data quality, and ultimately improving patient care. Imaging informatics increases the efficiency of imaging workflows by enhancing productivity and making information accessible to multiple users simultaneously. Consistency of critical data is essential for marrying information together through the process, to save the radiologist time, for consistency, billing, and research.
I. Introduction
The advancement of computer science and the availability of big data has enabled the emergence of artificial intelligence, which has led to a technological revolution significantly affecting many aspects of our daily life [1–4]. The application of AI in the field of medicine is expanding rapidly [5], mainly due to the advancement of machine learning (ML) that can be utilized for the analysis of medical images and patient data, diagnosis of diseases, and prediction of treatment outcomes [6]. ML is a paradigm of AI that systematically allows computer algorithms to adapt according to a large amount of raw input data and make predictions or determinations using the learned patterns [1,7,8]. The method can be roughly divided into conventional machine learning (CML) and deep learning (DL) [8]. CML algorithms, such as the support vector machine (SVM), random forest (RF), decision tree (DT), and linear regression and logistic regression, generally do not involve large neural networks [8] and have been applied for the construction of predictive algorithms for the diagnosis or classification of diseases based on data from medical records or population-based studies [9]. DL has usually been applied for the analysis of multimedia datasets, including images, sound, and videos [7,8], and involves large neural networks composed of multiple neuron-like layers of algorithms, such as artificial neural networks (ANNs), recurrent neural networks (RNNs), and convolutional neural networks (CNNs) [7,8].
In ophthalmology, AI has initially been applied for the analysis of fundus photographs and optical coherence tomography (OCT) images; thus, previous studies have mostly focused on the integration of AI into the diagnostic approach of posterior segment diseases, such as diabetic retinopathy, glaucoma, macular degeneration, and retinopathy of prematurity [5,10–14]. However, as DL algorithms can be utilized for the analysis of imaging the data of anterior segment structures, such as anterior segment photographs (ASPs), anterior segment OCT (AS-OCT) images, specular microscopy, corneal topography, in vivo confocal microscopy (IVCM), infrared meibography, and tear interferometry [2], AI is also expected to assist in the diagnosis and monitoring of various anterior segment diseases. Recently, many studies have been conducted on the application of AI in various anterior segment diseases [2].