1- Assistant Professor, Department of Health Information Sciences, Faculty of Management and Medical Information Sciences, Kerman University of Medical Sciences, Kerman, Iran
2- Master of Sciences Student in Nursing, Student Research Committee, Fasa University of Medical Sciences, Fasa, Iran
3- Assistant Professor, Department of Health Information Thechnology, Noncommunicable Diseases Research Center, Fasa University of Medical Sciences, Fasa, Iran
2- Master of Sciences Student in Nursing, Student Research Committee, Fasa University of Medical Sciences, Fasa, Iran
3- Assistant Professor, Department of Health Information Thechnology, Noncommunicable Diseases Research Center, Fasa University of Medical Sciences, Fasa, Iran
Abstract: (15 Views)
Background and Aim: Cataracts are recognized as the cause of 51% of blindness worldwide. Following the promising initial results of artificial intelligence systems in eye diseases, AI algorithms have been applied in the diagnosis of cataracts, grading the severity of cataracts, intraocular lens calculations, and even as an assistive tool in cataract surgery. This study presents a systematic review of AI techniques in the management of cataract disease.
Materials and Methods: This systematic review study was conducted to investigate artificial intelligence techniques to manage cataract disease until November 11, 2023, and based on PRISMA guidelines. We retrieved all relevant articles published in English through a systematic search of PubMed, Scopus, and Web of Science online databases.
Results: In our initial search, 192 records were identified in the databases, and eventually, 23 articles were selected for review. The results indicated that convolutional neural network algorithms (6 articles), recurrent neural networks (1 article), deep convolutional networks (1 article), support vector machines (2 articles), transfer learning (1 article), decision trees (4 articles), random forests (4 articles), logistic regression (3 articles), Bayesian algorithms (3 articles), XGBoost (3 articles), and K-nearest neighbors clustering algorithms (2 articles) were the artificial neural network and machine learning techniques and algorithms utilized. These techniques were employed in the studies for the diagnosis (70%), management (17%), and prediction (13%) of cataract disease.
Conclusion: Various artificial intelligence and machine learning techniques and algorithms can be effective and efficient in diagnosing, grading, managing, and predicting cataracts with high accuracy. In this study, deep learning techniques and convolutional neural networks have made the greatest contribution to cataract diagnosis. Deep learning techniques, decision trees, and Bayesian algorithms were involved in cataract management. Machine learning algorithms such as logistic regression, random forest, artificial neural network, decision tree, K1-nearest neighbor, XGBoost, and adaptive boosting also played a role in cataract prediction. Just as early prediction, diagnosis, and timely referral can reduce future complications of the disease, the use of systems based on artificial intelligence models that have acceptable accuracy can be effective in supporting the decision-making process of physicians and managing this disease.
Materials and Methods: This systematic review study was conducted to investigate artificial intelligence techniques to manage cataract disease until November 11, 2023, and based on PRISMA guidelines. We retrieved all relevant articles published in English through a systematic search of PubMed, Scopus, and Web of Science online databases.
Results: In our initial search, 192 records were identified in the databases, and eventually, 23 articles were selected for review. The results indicated that convolutional neural network algorithms (6 articles), recurrent neural networks (1 article), deep convolutional networks (1 article), support vector machines (2 articles), transfer learning (1 article), decision trees (4 articles), random forests (4 articles), logistic regression (3 articles), Bayesian algorithms (3 articles), XGBoost (3 articles), and K-nearest neighbors clustering algorithms (2 articles) were the artificial neural network and machine learning techniques and algorithms utilized. These techniques were employed in the studies for the diagnosis (70%), management (17%), and prediction (13%) of cataract disease.
Conclusion: Various artificial intelligence and machine learning techniques and algorithms can be effective and efficient in diagnosing, grading, managing, and predicting cataracts with high accuracy. In this study, deep learning techniques and convolutional neural networks have made the greatest contribution to cataract diagnosis. Deep learning techniques, decision trees, and Bayesian algorithms were involved in cataract management. Machine learning algorithms such as logistic regression, random forest, artificial neural network, decision tree, K1-nearest neighbor, XGBoost, and adaptive boosting also played a role in cataract prediction. Just as early prediction, diagnosis, and timely referral can reduce future complications of the disease, the use of systems based on artificial intelligence models that have acceptable accuracy can be effective in supporting the decision-making process of physicians and managing this disease.
Type of Study: Review |
Subject:
Health Information Technology
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
