Automated Identification of Referable Retinal Pathology in Teleophthalmology Setting

Gao, Qitong; Amason, Joshua; Cousins, Scott W.; Pajić, Miroslav; Hadziahmetovic, Majda

doi:10.1167/tvst.10.6.30

Cited by 4 publications

(4 citation statements)

References 38 publications

(94 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For external testing, we provided only central foveal B-scans in the OCT volume to the model. Although this approach is consistent with prior studies, 5 , 51 – 53 there may be pathologies that are not captured in a central foveal B-scan. 54 Indeed, the only case of drusen our model missed was absent in the central foveal image supplied to RobOCTNet, and the human experts also did not detect any pathology when reviewing this image ( Supplementary Fig.…”

Section: Discussionsupporting

confidence: 83%

RobOCTNet: Robotics and Deep Learning for Referable Posterior Segment Pathology Detection in an Emergency Department Population

Song,

Lusk,

Roh

et al. 2024

Trans. Vis. Sci. Tech.

View full text Add to dashboard Cite

Purpose To evaluate the diagnostic performance of a robotically aligned optical coherence tomography (RAOCT) system coupled with a deep learning model in detecting referable posterior segment pathology in OCT images of emergency department patients. Methods A deep learning model, RobOCTNet, was trained and internally tested to classify OCT images as referable versus non-referable for ophthalmology consultation. For external testing, emergency department patients with signs or symptoms warranting evaluation of the posterior segment were imaged with RAOCT. RobOCTNet was used to classify the images. Model performance was evaluated against a reference standard based on clinical diagnosis and retina specialist OCT review. Results We included 90,250 OCT images for training and 1489 images for internal testing. RobOCTNet achieved an area under the curve (AUC) of 1.00 (95% confidence interval [CI], 0.99–1.00) for detection of referable posterior segment pathology in the internal test set. For external testing, RAOCT was used to image 72 eyes of 38 emergency department patients. In this set, RobOCTNet had an AUC of 0.91 (95% CI, 0.82–0.97), a sensitivity of 95% (95% CI, 87%–100%), and a specificity of 76% (95% CI, 62%–91%). The model's performance was comparable to two human experts’ performance. Conclusions A robotically aligned OCT coupled with a deep learning model demonstrated high diagnostic performance in detecting referable posterior segment pathology in a cohort of emergency department patients. Translational Relevance Robotically aligned OCT coupled with a deep learning model may have the potential to improve emergency department patient triage for ophthalmology referral.

show abstract

Section: Discussionsupporting

confidence: 83%

RobOCTNet: Robotics and Deep Learning for Referable Posterior Segment Pathology Detection in an Emergency Department Population

Song,

Lusk,

Roh

et al. 2024

Trans. Vis. Sci. Tech.

View full text Add to dashboard Cite

show abstract

“…This approach may be particularly useful in areas with a shortage of FFA instruments or experienced examiners. With a clear CFP, patients with BRVO can immediately receive a precise segmentation of NPA and be guided with further laser photocoagulation ( 35 ). This study is an in-depth study from computer-aided diagnosis to treatment, and the potential use of this DL algorithm will be an outcome measure in clinical trials and a decision tool in clinical practice, which will be the theoretical basis for the application of intelligent guided laser ( 34 , 36 ).…”

Section: Discussionmentioning

confidence: 99%

Deep Learning Models for Segmenting Non-perfusion Area of Color Fundus Photographs in Patients With Branch Retinal Vein Occlusion

Miao

Zou

et al. 2022

Front. Med.

View full text Add to dashboard Cite

PurposeTo develop artificial intelligence (AI)-based deep learning (DL) models for automatically detecting the ischemia type and the non-perfusion area (NPA) from color fundus photographs (CFPs) of patients with branch retinal vein occlusion (BRVO).MethodsThis was a retrospective analysis of 274 CFPs from patients diagnosed with BRVO. All DL models were trained using a deep convolutional neural network (CNN) based on 45 degree CFPs covering the fovea and the optic disk. We first trained a DL algorithm to identify BRVO patients with or without the necessity of retinal photocoagulation from 219 CFPs and validated the algorithm on 55 CFPs. Next, we trained another DL algorithm to segment NPA from 104 CFPs and validated it on 29 CFPs, in which the NPA was manually delineated by 3 experienced ophthalmologists according to fundus fluorescein angiography. Both DL models have been cross-validated 5-fold. The recall, precision, accuracy, and area under the curve (AUC) were used to evaluate the DL models in comparison with three types of independent ophthalmologists of different seniority.ResultsIn the first DL model, the recall, precision, accuracy, and area under the curve (AUC) were 0.75 ± 0.08, 0.80 ± 0.07, 0.79 ± 0.02, and 0.82 ± 0.03, respectively, for predicting the necessity of laser photocoagulation for BRVO CFPs. The second DL model was able to segment NPA in CFPs of BRVO with an AUC of 0.96 ± 0.02. The recall, precision, and accuracy for segmenting NPA was 0.74 ± 0.05, 0.87 ± 0.02, and 0.89 ± 0.02, respectively. The performance of the second DL model was nearly comparable with the senior doctors and significantly better than the residents.ConclusionThese results indicate that the DL models can directly identify and segment retinal NPA from the CFPs of patients with BRVO, which can further guide laser photocoagulation. Further research is needed to identify NPA of the peripheral retina in BRVO, or other diseases, such as diabetic retinopathy.

show abstract

“…The derived model was shown to have high accuracy and sensitivity to identify the presence of retinal disease (yes or no) with a relatively low false-negative rate. Using data obtained in this study, we have developed an automated, deep learning-based image interpretation model 41 with high accuracy and sensitivity to recognize referable retinal disease. Such an automated image analysis tool, combined with remote retinal imaging, could potentially deliver on the promise of a low-cost, fast, and effective retinal disease screening.…”

Section: Discussionmentioning

confidence: 99%

Incidence of Referable Retinal Disease in Diabetic Patients at a Primary Care Practice

Lee

Amason

Risco

et al. 2021

Journal of VitreoRetinal Diseases

Self Cite

View full text Add to dashboard Cite

Purpose: This work tests the feasibility of remote ophthalmic imaging to identify referable retinal abnormalities and assesses the effectiveness of color fundus photography (CFP) vs optical coherence tomography (OCT) for this purpose. Methods: This prospective, nonrandomized study included 633 patients with diabetes at Duke Primary Care. Undilated patients underwent screening with CFP and OCT camera (MaestroCare, Topcon). Images were graded independently for interpretability and the presence of predetermined retinal disease. Retinal disease was classified as diabetic retinopathy (DR) referable to a retina specialist or incidental findings referable to either a retina specialist or a general ophthalmologist, depending on severity. Results: Mean (SD) age of screened patients was 66 (13) years, and 49% were women. The average glycated hemoglobin A1c level was 7.6 % (SD, 1.7%), and 30% of the patients were on insulin. The average duration of diabetes was 5.9 (SD, 7.3) years. Remote images from OCT were significantly more interpretable than CFP (98% vs 83%, respectively; P < .001). Referral rates were 9% for DR and 28% for incidental findings. Among patients with DR, OCT and CFP were helpful in 58% and 87% of cases, respectively ( P < .001). Conclusions: Remote diagnosis of ophthalmic imaging at the point of service may allow for early identification of retinal disease and timely referral and treatment. Our approach showed that OCT had significantly better interpretability, while CFP was more helpful in identifying DR. These findings may be important when choosing the screening device in a specific context.

show abstract

Automated Identification of Referable Retinal Pathology in Teleophthalmology Setting

Cited by 4 publications

References 38 publications

RobOCTNet: Robotics and Deep Learning for Referable Posterior Segment Pathology Detection in an Emergency Department Population

RobOCTNet: Robotics and Deep Learning for Referable Posterior Segment Pathology Detection in an Emergency Department Population

Deep Learning Models for Segmenting Non-perfusion Area of Color Fundus Photographs in Patients With Branch Retinal Vein Occlusion

Incidence of Referable Retinal Disease in Diabetic Patients at a Primary Care Practice

Contact Info

Product

Resources

About