Dataset Information

Feasibility study to improve deep learning in OCT diagnosis of rare retinal diseases with few-shot classification.

ABSTRACT: Deep learning (DL) has been successfully applied to the diagnosis of ophthalmic diseases. However, rare diseases are commonly neglected due to insufficient data. Here, we demonstrate that few-shot learning (FSL) using a generative adversarial network (GAN) can improve the applicability of DL in the optical coherence tomography (OCT) diagnosis of rare diseases. Four major classes with a large number of datasets and five rare disease classes with a few-shot dataset are included in this study. Before training the classifier, we constructed GAN models to generate pathological OCT images of each rare disease from normal OCT images. The Inception-v3 architecture was trained using an augmented training dataset, and the final model was validated using an independent test dataset. The synthetic images helped in the extraction of the characteristic features of each rare disease. The proposed DL model demonstrated a significant improvement in the accuracy of the OCT diagnosis of rare retinal diseases and outperformed the traditional DL models, Siamese network, and prototypical network. By increasing the accuracy of diagnosing rare retinal diseases through FSL, clinicians can avoid neglecting rare diseases with DL assistance, thereby reducing diagnosis delay and patient burden.

SUBMITTER: Yoo TK

PROVIDER: S-EPMC7829497 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Feasibility study to improve deep learning in OCT diagnosis of rare retinal diseases with few-shot classification.

Yoo Tae Keun TK Choi Joon Yul JY Kim Hong Kyu HK

Medical & biological engineering & computing 20210125 2

Deep learning (DL) has been successfully applied to the diagnosis of ophthalmic diseases. However, rare diseases are commonly neglected due to insufficient data. Here, we demonstrate that few-shot learning (FSL) using a generative adversarial network (GAN) can improve the applicability of DL in the optical coherence tomography (OCT) diagnosis of rare diseases. Four major classes with a large number of datasets and five rare disease classes with a few-shot dataset are included in this study. Befo ...[more]

PMID: 33492598

Similar Datasets

Project description:ImportanceRecent studies have demonstrated the successful application of artificial intelligence (AI) for automated retinal disease diagnostics but have not addressed a fundamental challenge for deep learning systems: the current need for large, criterion standard-annotated retinal data sets for training. Low-shot learning algorithms, aiming to learn from a relatively low number of training data, may be beneficial for clinical situations involving rare retinal diseases or when addressing potential bias resulting from data that may not adequately represent certain groups for training, such as individuals older than 85 years.ObjectiveTo evaluate whether low-shot deep learning methods are beneficial when using small training data sets for automated retinal diagnostics.Design, setting, and participantsThis cross-sectional study, conducted from July 1, 2019, to June 21, 2020, compared different diabetic retinopathy classification algorithms, traditional and low-shot, for 2-class designations (diabetic retinopathy warranting referral vs not warranting referral). The public domain EyePACS data set was used, which originally included 88 692 fundi from 44 346 individuals. Statistical analysis was performed from February 1 to June 21, 2020.Main outcomes and measuresThe performance (95% CIs) of the various AI algorithms was measured via receiver operating curves and their area under the curve (AUC), precision recall curves, accuracy, and F1 score, evaluated for different training data sizes, ranging from 5120 to 10 samples per class.ResultsDeep learning algorithms, when trained with sufficiently large data sets (5120 samples per class), yielded comparable performance, with an AUC of 0.8330 (95% CI, 0.8140-0.8520) for a traditional approach (eg, fined-tuned ResNet), compared with low-shot methods (AUC, 0.8348 [95% CI, 0.8159-0.8537]) (using self-supervised Deep InfoMax [our method denoted as DIM]). However, when far fewer training images were available (n = 160), the traditional deep learning approach had an AUC decreasing to 0.6585 (95% CI, 0.6332-0.6838) and was outperformed by a low-shot method using self-supervision with an AUC of 0.7467 (95% CI, 0.7239-0.7695). At very low shots (n = 10), the traditional approach had performance close to chance, with an AUC of 0.5178 (95% CI, 0.4909-0.5447) compared with the best low-shot method (AUC, 0.5778 [95% CI, 0.5512-0.6044]).Conclusions and relevanceThese findings suggest the potential benefits of using low-shot methods for AI retinal diagnostics when a limited number of annotated training retinal images are available (eg, with rare ophthalmic diseases or when addressing potential AI bias).

Project description:Objective: To evaluate the accuracy and feasibility of the auto-detection of 15 retinal disorders with artificial intelligence (AI)-assisted optical coherence tomography (OCT) in community screening. Methods: A total of 954 eyes of 477 subjects from four local communities were enrolled in this study from September to December 2021. They received OCT scans covering an area of 12 mm × 9 mm at the posterior pole retina involving the macular and optic disc, as well as other ophthalmic examinations performed using their demographic information recorded. The OCT images were analyzed using integrated software with the previously established algorithm based on the deep-learning method and trained to detect 15 kinds of retinal disorders, namely, pigment epithelial detachment (PED), posterior vitreous detachment (PVD), epiretinal membranes (ERMs), sub-retinal fluid (SRF), choroidal neovascularization (CNV), drusen, retinoschisis, cystoid macular edema (CME), exudation, macular hole (MH), retinal detachment (RD), ellipsoid zone disruption, focal choroidal excavation (FCE), choroid atrophy, and retinal hemorrhage. Meanwhile, the diagnosis was also generated from three groups of individual ophthalmologists (group of retina specialists, senior ophthalmologists, and junior ophthalmologists) and compared with those by the AI. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were calculated, and kappa statistics were performed. Results: A total of 878 eyes were finally enrolled, with 76 excluded due to poor image quality. In the detection of 15 retinal disorders, the ROC curve comparison between AI and professors' presented relatively large AUC (0.891-0.997), high sensitivity (87.65-100%), and high specificity (80.12-99.41%). Among the ROC curve comparisons with those by the retina specialists, AI was the closest one to the professors' compared to senior and junior ophthalmologists (p < 0.05). Conclusion: AI-assisted OCT is highly accurate, sensitive, and specific in auto-detection of 15 kinds of retinal disorders, certifying its feasibility and effectiveness in community ophthalmic screening.

Dataset Information

Feasibility study to improve deep learning in OCT diagnosis of rare retinal diseases with few-shot classification.

Publications

Feasibility study to improve deep learning in OCT diagnosis of rare retinal diseases with few-shot classification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets