Automated Classification of Inherited Retinal Diseases in Optical Coherence Tomography Images Using Few-shot Learning

作     者：ZHAO Qi MAI Si Wei LI Qian HUANG Guan Chong GAO Ming Chen YANG Wen Li WANG Ge MA Ya LI Lei PENG Xiao Yan ZHAO Qi;MAI Si Wei;LI Qian;HUANG Guan Chong;GAO Ming Chen;YANG Wen Li;WANG Ge;MA Ya;LI Lei;PENG Xiao Yan

作者机构：Department of OphthalmologyBeijing Tongren Eye CenterBeijing Tongren HospitalCapital Medical UniversityBeijing Key Laboratory of Ophthalmology and Visual SciencesBeijing 100730China Department of Computer ScienceRutgersThe State University of New JerseyNew Brunswick 08901USA Department of Computer Science and EngineeringUniversity at BuffaloBuffalo 14260USA Beijing Institute of OphthalmologyBeijing Tongren HospitalCapital Medical UniversityBeijing Ophthalmology and Visual Science KeyLaboratoryBeijing 100730China 

出 版 物：《Biomedical and Environmental Sciences》 (生物医学与环境科学（英文版）)

年 卷 期：2023年第36卷第5期

页      面：431-440页

核心收录：

学科分类：0830[工学-环境科学与工程（可授工学、理学、农学学位）] 1004[医学-公共卫生与预防医学(可授医学、理学学位)] 1002[医学-临床医学] 100212[医学-眼科学] 10[医学] 

基　　金：supported by National Natural Science Foundation of China [No.82171073]。 

主　　题：Few-shot learning Student-teacher learning Knowledge distillation Transfer learning Optical coherence tomography Retinal degeneration Inherited retinal diseases 

摘      要：Objective To develop a few-shot learning(FSL) approach for classifying optical coherence tomography(OCT) images in patients with inherited retinal disorders(IRDs).Methods In this study, an FSL model based on a student–teacher learning framework was designed to classify images. 2,317 images from 189 participants were included. Of these, 1,126 images revealed IRDs, 533 were normal samples, and 658 were control samples.Results The FSL model achieved a total accuracy of 0.974–0.983, total sensitivity of 0.934–0.957, total specificity of 0.984–0.990, and total F1 score of 0.935–0.957, which were superior to the total accuracy of the baseline model of 0.943–0.954, total sensitivity of 0.866–0.886, total specificity of 0.962–0.971,and total F1 score of 0.859–0.885. The performance of most subclassifications also exhibited advantages. Moreover, the FSL model had a higher area under curves(AUC) of the receiver operating characteristic(ROC) curves in most subclassifications.Conclusion This study demonstrates the effective use of the FSL model for the classification of OCT images from patients with IRDs, normal, and control participants with a smaller volume of data. The general principle and similar network architectures can also be applied to other retinal diseases with a low prevalence.