Attention Based Deep Learning System for Automated Diagnoses of Age‐Related Macular Degeneration in Optical Coherence Tomography Images

2021 ◽  
Author(s):  
Yan Yan ◽  
Kai Jin ◽  
Zhiyuan Gao ◽  
Xiaoling Huang ◽  
Fanyi Wang ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Deep Learning ◽  
Macular Degeneration ◽  
Learning System ◽  
Age Related Macular Degeneration ◽  
Optical Coherence ◽  
Age Related

scholarly journals Detection of features associated with neovascular age-related macular degeneration in ethnically distinct data sets by an optical coherence tomography: trained deep learning algorithm

2020 ◽  
pp. bjophthalmol-2020-316984
Author(s):  
Tyler Hyungtaek Rim ◽  
Aaron Y Lee ◽  
Daniel S Ting ◽  
Kelvin Teo ◽  
Bjorn Kaijun Betzler ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Deep Learning ◽  
Macular Degeneration ◽  
External Validation ◽  
Age Related Macular Degeneration ◽  
Data Sets ◽  
Optical Coherence ◽  
Data Set ◽  
Saliency Maps ◽  
Age Related

BackgroundThe ability of deep learning (DL) algorithms to identify eyes with neovascular age-related macular degeneration (nAMD) from optical coherence tomography (OCT) scans has been previously established. We herewith evaluate the ability of a DL model, showing excellent performance on a Korean data set, to generalse onto an American data set despite ethnic differences. In addition, expert graders were surveyed to verify if the DL model was appropriately identifying lesions indicative of nAMD on the OCT scans.MethodsModel development data set—12 247 OCT scans from South Korea; external validation data set—91 509 OCT scans from Washington, USA. In both data sets, normal eyes or eyes with nAMD were included. After internal testing, the algorithm was sent to the University of Washington, USA, for external validation. Area under the receiver operating characteristic curve (AUC) and precision–recall curve (AUPRC) were calculated. For model explanation, saliency maps were generated using Guided GradCAM.ResultsOn external validation, AUC and AUPRC remained high at 0.952 (95% CI 0.942 to 0.962) and 0.891 (95% CI 0.875 to 0.908) at the individual level. Saliency maps showed that in normal OCT scans, the fovea was the main area of interest; in nAMD OCT scans, the appropriate pathological features were areas of model interest. Survey of 10 retina specialists confirmed this.ConclusionOur DL algorithm exhibited high performance for nAMD identification in a Korean population, and generalised well to an ethnically distinct, American population. The model correctly focused on the differences within the macular area to extract features associated with nAMD.

A Comparison of Radial and Horizontal Macular Optical Coherence Tomography on Detection of Macular Pathology: A Prospective Study at an Academic Institution

2021 ◽  
pp. 247412642199705
Author(s):  
Halward M.J. Blegen ◽  
Samuel D. Hobbs ◽  
Reggie Taylor ◽  
Andrew L. Plaster ◽  
Paul M. Drayna
Keyword(s):  
Optical Coherence Tomography ◽  
Prospective Study ◽  
Macular Degeneration ◽  
Epiretinal Membrane ◽  
Central Serous Chorioretinopathy ◽  
Age Related Macular Degeneration ◽  
Optical Coherence ◽  
Age Related ◽  
Macular Diseases ◽  
A Prospective Study

Purpose: Optical coherence tomography (OCT) is useful in diagnosing and monitoring retinal pathology such as age-related macular degeneration, diabetic macular edema (DME), central serous chorioretinopathy, and epiretinal membrane, among others. This study compared the ability of horizontal (H) 25-, 13-, and 7-cut macular OCT vs 24-, 12-, and 6-cut radial (R) macular OCT in identifying various macular pathology. Methods: This was a prospective study of 161 established patients evaluated at Wilford Hall Eye Center Retina Clinic between September and October of 2019. Pathology included age-related macular degeneration, central serous chorioretinopathy, DME, and epiretinal membrane, among others. Patients obtained 25-, 13-, and 7-cut H raster OCT as well as 24-, 12-, and 6-cut R OCT. Primary outcomes were sensitivity in detecting macular fluid and each macular abnormality. Results: The 24-cut radial (R24) OCT equally or out-performed the H25 (horizontal 25-cut OCT) in detecting macular fluid across all pathological groups. Generally, a higher number of cuts correlated with better detection of fluid. In detecting any macular abnormalities, H25, R24, and R12 had 100% sensitivity. R6 OCT had near 100% sensitivity across all groups, except for DME (95%). Overall, R OCT had better sensitivity (0.960) than H OCT (0.907) in detecting macular pathology. Conclusions: R outperformed H macular OCT in detecting fluid and other abnormalities. Clinically, both scanning patterns can be used by ophthalmologists in diagnosis and management of commonly encountered macular diseases. Technicians may be able to use a variety of these scans to screen for pathology prior to physician evaluation.

scholarly journals The Spectrum of Central Choriocapillaris Abnormalities on Swept-Source Optical Coherence Tomography Angiography in the Fellow Eye of Unilateral Exudative Age-Related Macular Degeneration Patients: From Flow Deficits to Subclinical Non-Exudative Neovascularization

2021 ◽  
Vol 10 (12) ◽  
pp. 2658
Author(s):  
Alexis Khorrami Kashi ◽  
Eric Souied ◽  
Selim Fares ◽  
Enrico Borrelli ◽  
Vittorio Capuano ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Macular Degeneration ◽  
Optical Coherence Tomography Angiography ◽  
Clinical Findings ◽  
Age Related Macular Degeneration ◽  
Optical Coherence ◽  
Swept Source ◽  
Exudative Amd ◽  
Age Related ◽  
Fellow Eyes

We evaluated the spectrum of choriocapillaris (CC) abnormalities in the fellow eyes of unilateral exudative age-related macular degeneration (AMD) patients using swept-source optical coherence tomography angiography (SS-OCTA). Fellow eyes of unilateral exudative AMD patients were prospectively included between May 2018 and October 2018. Patients underwent a multimodal imaging including a SS-OCTA. Demographics and clinical findings were analyzed. The estimated prevalence of macular neovascularization (MNV) was computed. Number and size of flow deficits (FDs) and percentage of flow deficits (FD%) were computed on the compensated CC flow images with the Fiji software. We included 97 eyes of 97 patients (mean age was 80 ± 7.66 years, 39 males, 58 females). The prevalence of MNV in the studied eyes was 8.25% (8/97 eyes). In the 89 non-neovascular eyes, FD% averaged 45.84% ± 11.63%, with a corresponding total area of FDs of 4.19 ± 1.12 mm2. There was a higher prevalence of drusenoid pigment epithelial detachment in eyes with subclinical neovascularization (p = 0.021). Fellow eyes with unilateral exudative AMD encompassed a series of CC abnormalities, from FDs of the aging CC to subclinical non-exudative MNV.

Sign in / Sign up

Export Citation Format

Share Document