Publications by authors named "Carol Y Cheung"

179 Publications

Understanding Visual Acuity Outcomes After Retinal Detachment Repair by Assessing Photoreceptor Integrity on Spectral-Domain Optical Coherence Tomography.

JAMA Ophthalmol 2021 Apr 22. Epub 2021 Apr 22.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong Special Administrative Region, China.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1001/jamaophthalmol.2021.0818DOI Listing
April 2021

Characterization of macular choroid in normal-tension glaucoma: a swept-source optical coherence tomography study.

Acta Ophthalmol 2021 Mar 6. Epub 2021 Mar 6.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, Hong Kong SAR, China.

Purpose: To investigate the features of the small-to-medium (choriocapillaris and Sattler's layer) and large (Haller's layer) sized vessel layers of the macular choroid in normal-tension glaucoma (NTG) patients using swept-source optical coherence tomography (SS-OCT).

Methods: We conducted an observational cross-sectional study in 234 NTG eyes from 134 patients, and 203 normal eyes from 109 non-glaucomatous control subjects. We used Image J to segment the choroidal layer and then a Python script to measure the average macular choroidal thickness (MCT) and choroidal vascular index (CVI) of two vessel layers at five different locations. Quantitative optical coherence tomography angiography (OCTA) metrics for the superficial capillary plexus (SCP) and deep capillary plexus (DCP) at macular region were also measured by a customized MATLAB program. Generalized estimating equations (GEE) models were performed to determine ocular and demographic factors associated with the choroidal metrics, adjusting for inter-eye correlation.

Results: Significant average MCT thinning was found in NTG eyes at all five locations of the two layers, in comparison with controls (all p ≤ 0.05). In addition, compared with control eyes, significant decrease in CVI was found in NTG eyes at all five sections of the large sized vessel layer: 500 µm nasal and temporal to macula (p = 0.002), 1500 µm nasal (p < 0.001), 2500 µm nasal (p = 0.001), 1500 µm temporal (p < 0.001) and 2500 µm temporal (p = 0.004). In contrast, no significant CVI difference was detected in the small-to-medium sized vessel layer. In the comparison of OCTA metrics of SCP and DCP at macular region between NTG and normal eyes, there were no significant difference of parafoveal vessel density (VD), foveal avascular zone (FAZ) area, FAZ circularity and fractal dimension (FD) in both layers.

Conclusion: We found significant alterations in macular choroidal vascularity (reduced CVI and thinner layer) in NTG patients. Such alterations are more pronounced in the Haller's layer, rather than the choriocapillaris & Sattler's layer, in NTG. Choroidal layer may be more related to vasculature changes at macular region in NTG.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/aos.14829DOI Listing
March 2021

Retinal microvascular signs and risk of diabetic kidney disease in asian and white populations.

Sci Rep 2021 Mar 1;11(1):4898. Epub 2021 Mar 1.

Singapore Eye Research Institute, Singapore National Eye Centre, 11 Third Hospital Avenue, Singapore, 168751, Singapore.

The objective was to examine prospectively the association between retinal microvascular signs and development of diabetic kidney disease (DKD) in Asian and White populations. We analysed two population-based cohorts, composing of 1,221 Asians (SEED) and 703 White (WESDR) adults with diabetes. Retinal microvascular signs at baseline included vascular caliber (arteriolar-CRAE, and venular-CRVE) and diabetic retinopathy (DR). Incident cases of DKD were identified after ~ 6-year. Incident cases were defined based on eGFR in SEED and proteinuria or history of renal dialysis in WESDR. The incidence of DKD were 11.8% in SEED and 14.0% in WESDR. Wider CRAE in SEED (OR = 1.58 [1.02, 2.45]) and wider CRVE (OR = 1.69 [1.02, 2.80)) in WESDR were associated with increased risk of DKD. Presence of DR was associated with an increased risk of DKD in both cohorts (SEED: OR = 1.91 [1.21, 3.01] in SEED, WESDR: OR = 1.99 [1.18, 3.35]). Adding DR and retinal vascular calibers in the model beyond traditional risk factors led to an improvement of predictive performance of DKD risk between 1.1 and 2.4%; and improved classification (NRI 3 between 9%). Microvascular changes in the retina are longitudinally associated with risk of DKD.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-021-84464-7DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7921402PMC
March 2021

Detection of Diabetic Retinopathy from Ultra-Widefield Scanning Laser Ophthalmoscope Images: A Multicenter Deep Learning Analysis.

Ophthalmol Retina 2021 Feb 1. Epub 2021 Feb 1.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China. Electronic address:

Purpose: To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO).

Design: Observational, cross-sectional study.

Participants: A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina.

Methods: All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions.

Main Outcome Measures: Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR.

Results: For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection.

Conclusions: The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.oret.2021.01.013DOI Listing
February 2021

Longitudinal Changes in Macular Optical Coherence Tomography Angiography Metrics in Primary Open-Angle Glaucoma With High Myopia: A Prospective Study.

Invest Ophthalmol Vis Sci 2021 Jan;62(1):30

State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China.

Purpose: To characterize longitudinal changes in macular microvasculature as quantified from optical coherence tomography angiography (OCTA) metrics in primary open-angle glaucoma (POAG) eyes with and without high myopia.

Methods: In total, 63 and 61 POAG eyes with and without high myopia, respectively, underwent swept-source OCTA imaging in at least four follow-up visits at an ophthalmic center, with a scanning protocol of 3- × 3-mm centered at the fovea. The foveal avascular zone (FAZ) area, FAZ circularity, and vessel density (VD) in both the superficial (SCP) and deep capillary plexuses (DCP) were measured. The rate of change in macular OCTA metrics over time was estimated using linear mixed-effects models in both groups of POAG eyes.

Results: The mean follow-up time and number of visits were 27.72 ± 8.57 months and 8.5 (8 to 13) times, and 30.95 ± 10.19 months and 10 (8‒13) times in POAG eyes with and without high myopia, respectively. VD in the DCP reduced significantly more quickly in POAG eyes with high myopia than in those without high myopia (-5.14%/year vs. -3.71%/year, P = 0.008). Moreover, lower baseline VD in the DCP was significantly associated with faster VD reduction in POAG with high myopia eyes (P < 0.001). Conversely, the VD reduction rate in the SCP, FAZ area, and FAZ circularity in both the SCP and DCP were similar in both groups (all Ps > 0.05).

Conclusions: VD in DCP reduced significantly more quickly in POAG eyes with high myopia over time. Density in the DCP reduced more quickly when baseline VD was low.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/iovs.62.1.30DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7846949PMC
January 2021

Independent and Synergistic Effects of High Blood Pressure and Obesity on Retinal Vasculature in Young Children: The Hong Kong Children Eye Study.

J Am Heart Assoc 2021 Feb 26;10(3):e018485. Epub 2021 Jan 26.

Department of Ophthalmology and Visual Sciences The Chinese University of Hong Kong Hong Kong SAR China.

Background High blood pressure (BP) and obesity are becoming increasingly prevalent among children globally. Although prior studies have shown their adverse impacts on macrovascular health, less is known about their effects on microvascular heath. This study aims to evaluate the independent and synergistic effects of hypertensive BP and obesity on retinal vasculature in young children. Method and Results 1006 children aged 6 to 8 years were recruited from the Hong Kong Children Eye Study. Quantitative retinal vascular parameters, including central retinal arteriolar and venular equivalents and retinal arteriolar and venular fractal dimensions, were measured from retinal photographs following a standardized protocol. BP and body mass index were categorized according to reference values from American Academy of Pediatrics and International Obesity Task Force guidelines respectively. Children with hypertensive systolic BP had the narrowest central retinal arteriolar equivalents compared with children with either elevated or normotensive systolic BP (162.4, 164.6, and 167.1 µm; -trend <0.001). Increased standardized systolic BP was associated with narrower central retinal arteriolar equivalents (β=-2.276 µm, <0.001), wider central retinal venular equivalents (1.177, =0.007), and decreased arteriolar fractal dimensions (β=-0.004, =0.034). Children with obesity had the smallest arteriolar fractal dimensions compared with children with overweightness and normal weight (1.211, 1.234, and 1.240; -trend=0.004). Children with both hypertensive BP and either overweightness or obesity had the narrowest central retinal arteriolar equivalents and smallest arteriolar (-trend<0.001 and -trend=0.007). Conclusions Our findings demonstrate the potential synergistic or additive effects for both hypertensive BP and obesity on retinal vasculature in children.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1161/JAHA.120.018485DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7955451PMC
February 2021

Artificial intelligence and machine learning for Alzheimer's disease: let's not forget about the retina.

Br J Ophthalmol 2021 May 25;105(5):593-594. Epub 2021 Jan 25.

Vitreo-retinal Department, Singapore National Eye Centre, Singapore

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bjophthalmol-2020-318407DOI Listing
May 2021

Vision, vision-specific functioning and mobility, and their relationship with clinically assessed cognitive impairment.

Age Ageing 2021 Jan 22. Epub 2021 Jan 22.

Singapore Eye Research Institute, Singapore National Eye Centre, Singapore.

Background: The relationship between self-reported visual disability and cognitive impairment in older individuals is unclear.

Objective: To determine the relationship of vision-specific functioning (VSF), vision-specific mobility (VSM) and visual acuity (VA) with clinically assessed cognitive impairment in the Epidemiology of Dementia in Singapore study.

Design: Cross-sectional.

Setting: Population-based.

Subjects: Eight hundred and seventy-four adults aged ≥60 years at higher risk of possible cognitive impairment by the Abbreviated Mental Test and progressive forgetfulness question.

Methods: VSF and VSM were measured using Rasch-transformed continuous scores of two Impact of Vision Impairment questionnaire domains. Cognitive impairment was objectively determined using detailed neuropsychological testing and defined as no cognitive impairment (NCI), mild cognitive impairment-no dementia (CIND), moderate CIND only and moderate CIND or dementia. Associations were assessed using multinomial logistic regression models.

Results: Of the 874 participants (49.0% males, mean age (SD) 65.5 (7.0) years), 277, 281 and 316 had NCI, mild CIND and moderate CIND or dementia, respectively. Compared to NCI, the odds of moderate CIND, and moderate CIND or dementia increased for every SD worsening in VSF (OR: 1.44, 95% CI 1.14-1.82, and OR: 1.52, 95%CI 1.19-1.94, respectively) and VSM (OR: 1.42, 95%CI 1.11-1.81, and OR: 1.50, 95%CI 1.15-1.95). Similarly, the odds of mild CIND (OR: 1.62, 95%CI 1.19-2.22), moderate CIND (OR: 1.93, 95%CI 1.45-2.58), and moderate CIND or dementia (OR: 2.25, 95%CI 1.62-3.11) increased significantly with every SD worsening of VA.

Conclusions: Our results emphasise the importance of interventions to prevent vision loss and improve quality of life to reduce likelihood of age-related cognitive decline.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1093/ageing/afaa276DOI Listing
January 2021

Non-invasive structural and metabolic retinal markers of disease activity in non-proliferative diabetic retinopathy.

Acta Ophthalmol 2021 Jan 8. Epub 2021 Jan 8.

Department of Ophthalmology, Odense University Hospital, Odense, Denmark.

Purpose: Metabolic and structural microvascular retinal alterations are essential components in diabetic retinopathy (DR). The present study aimed to measure changes at different stages of non-proliferative DR (NPDR) and to explore interactions of imaging-based metrics.

Methods: This cross-sectional, cohort study included 139 eyes from 80 diabetic patients. Each patient underwent dilated fundal examinations including colour fundus photography, retinal oximetry and optical coherence tomography angiography (OCTA), analysed by semi-automated and automated software. Diabetic retinopathy (DR) severity was classified according to the International Clinical Diabetic Retinopathy (ICDR) Severity Scale, ranging from no DR to severe NPDR (level 0-3). Retinal metabolism was evaluated by oximetry as retinal arteriolar (raSatO ) and venular oxygen saturation (rvSatO ), and macular microvascular structure was measured by OCTA as the area of foveal avascular zone (FAZ), vessel density (VD), vessel diameter index (VDI), FAZ circularity and fractal dimension (FD) in the superficial and deep retinal capillary plexus.

Results: A trend for increasing rvSatO was found with increasing DR severity (51.3%, 53.3%, 54.2%, 59.8%, p = 0.02). Increasing severity of DR associated with decreasing FD in the superficial and deep plexus (p < 0.001 and p = 0.014), and in the superficial plexus decreasing VD (p < 0.001), increasing VDI (p = 0.003) and decreasing FAZ circularity (p = 0.006). A few interactions were identified between raSatO , rvSatO and VDI, but only in the deep capillary plexus (p < 0.01 and p < 0.01).

Conclusion: Alterations of the venular retinal vascular oxygen saturation and microvascular structural abnormities were found continuously throughout the DR-spectrum. Given the sparse correlations between metabolic and structural abnormalities, it seems that these occur independently in DR.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/aos.14761DOI Listing
January 2021

A deep learning algorithm to detect chronic kidney disease from retinal photographs in community-based populations.

Lancet Digit Health 2020 06 12;2(6):e295-e302. Epub 2020 May 12.

Singapore Eye Research Institute, Singapore National Eye Centre, Singapore; Ophthalmology and Visual Sciences Academic Clinical Program, Duke-NUS Medical School, Singapore. Electronic address:

Background: Screening for chronic kidney disease is a challenge in community and primary care settings, even in high-income countries. We developed an artificial intelligence deep learning algorithm (DLA) to detect chronic kidney disease from retinal images, which could add to existing chronic kidney disease screening strategies.

Methods: We used data from three population-based, multiethnic, cross-sectional studies in Singapore and China. The Singapore Epidemiology of Eye Diseases study (SEED, patients aged ≥40 years) was used to develop (5188 patients) and validate (1297 patients) the DLA. External testing was done on two independent datasets: the Singapore Prospective Study Program (SP2, 3735 patients aged ≥25 years) and the Beijing Eye Study (BES, 1538 patients aged ≥40 years). Chronic kidney disease was defined as estimated glomerular filtration rate less than 60 mL/min per 1·73m. Three models were trained: 1) image DLA; 2) risk factors (RF) including age, sex, ethnicity, diabetes, and hypertension; and 3) hybrid DLA combining image and RF. Model performances were evaluated using the area under the receiver operating characteristic curve (AUC).

Findings: In the SEED validation dataset, the AUC was 0·911 for image DLA (95% CI 0·886 -0·936), 0·916 for RF (0·891-0·941), and 0·938 for hybrid DLA (0·917-0·959). Corresponding estimates in the SP2 testing dataset were 0·733 for image DLA (95% CI 0·696-0·770), 0·829 for RF (0·797-0·861), and 0·810 for hybrid DLA (0·776-0·844); and in the BES testing dataset estimates were 0·835 for image DLA (0·767-0·903), 0·887 for RF (0·828-0·946), and 0·858 for hybrid DLA (0·794-0·922). AUC estimates were similar in subgroups of people with diabetes (image DLA 0·889 [95% CI 0·850-0·928], RF 0·899 [0·862-0·936], hybrid 0·925 [0·893-0·957]) and hypertension (image DLA 0·889 [95% CI 0·860-0·918], RF 0·889 [0·860-0·918], hybrid 0·918 [0·893-0·943]).

Interpretation: A retinal image DLA shows good performance for estimating chronic kidney disease, underlying the feasibility of using retinal photography as an adjunctive or opportunistic screening tool for chronic kidney disease in community populations.

Funding: National Medical Research Council, Singapore.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/S2589-7500(20)30063-7DOI Listing
June 2020

Retinal microvasculature dysfunction is associated with Alzheimer's disease and mild cognitive impairment.

Alzheimers Res Ther 2020 12 4;12(1):161. Epub 2020 Dec 4.

Singapore Eye Research Institute, Singapore National Eye Centre, 20 College Road, The Academia, Level 6, Discovery Tower, Singapore, 169856, Singapore.

Background: The retina and brain share many neuronal and vasculature characteristics. We investigated the retinal microvasculature in Alzheimer's disease (AD) and mild cognitive impairment (MCI) using optical coherence tomography angiography (OCTA).

Methods: In this cross-sectional study, 24 AD participants, 37 MCI participants, and 29 controls were diagnosed according to internationally accepted criteria. OCTA images of the superficial and deep capillary plexus (SCP, DCP) of the retinal microvasculature were obtained using a commercial OCTA system (Zeiss Cirrus HD-5000 with AngioPlex, Carl Zeiss Meditec, Dublin, CA). The main outcome measures were vessel density (VD) and fractal dimension (FD) in the SCP and DCP within a 2.5-mm ring around the fovea which were compared between groups. Perfusion density of large vessels and foveal avascular zone (FAZ) area were additional outcome parameters.

Results: Age, gender, and race did not differ among groups. However, there was a significant difference in diabetes status (P = 0.039) and systolic blood pressure (P = 0.008) among the groups. After adjusting for confounders, AD participants showed significantly decreased VD in SCP and DCP (P = 0.006 and P = 0.015, respectively) and decreased FD in SCP (P = 0.006), compared to controls. MCI participants showed significantly decreased VD and FD only in SCP (P = 0.006 and P < 0.001, respectively) and not the DCP (P > 0.05) compared with controls. There was no difference in the OCTA variables between AD and MCI (P > 0.05). Perfusion density of large vessels and FAZ area did not differ significantly between groups (P > 0.05).

Conclusions And Relevance: Eyes of patients with AD have significantly reduced macular VD in both plexuses whereas MCI participants only showed reduction in the superficial plexus. Changes in the retinal microvasculature and capillary network may offer a valuable insight on the brain in AD.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s13195-020-00724-0DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7718666PMC
December 2020

Improved Automated Foveal Avascular Zone Measurement in Cirrus Optical Coherence Tomography Angiography Using the Level Sets Macro.

Transl Vis Sci Technol 2020 11 13;9(12):20. Epub 2020 Nov 13.

Joint Shantou International Eye Center, Shantou University and The Chinese University of Hong Kong, Shantou, China.

Purpose: To evaluate automated measurements of the foveal avascular zone (FAZ) using the Level Sets macro (LSM) in ImageJ as compared with the Cirrus optical coherence tomography angiography (OCTA) inbuilt algorithm and the Kanno-Saitama macro (KSM).

Methods: The eyes of healthy volunteers were scanned four times consecutively on the Zeiss Cirrus HD-OCT 5000 system. The FAZ metrics (area, perimeter, and circularity) were measured manually and automatically by the Cirrus inbuilt algorithm, the KSM, and the LSM. The accuracy and repeatability of all methods and agreement between automated and manual methods were evaluated.

Results: The LSM segmented the FAZ with an average Dice coefficient of 0.9243. Compared with the KSM and the Cirrus inbuilt algorithm, the LSM outperformed them by 0.02 and 0.19, respectively, for Dice coefficients. Both the LSM (intraclass correlation coefficient [ICC] = 0.908; coefficient of variation [CoV] = 9.664%) and manual methods (ICC ≥ 0.921, CoV ≤ 8.727%) showed excellent repeatability for the FAZ area, whereas the other methods presented moderate to good repeatability (ICC ≤ 0.789, CoV ≥ 15.788%). Agreement with manual FAZ area measurement was excellent for both the LSM and KSM but not for the Cirrus inbuilt algorithm (LSM, ICC = 0.930; KSM, ICC = 0.928; Cirrus, ICC = 0.254).

Conclusions: The LSM exhibited greater accuracy and reliability compared to the KSM and inbuilt automated methods and may be an improved and accessible option for automated FAZ segmentation.

Translational Relevance: The LSM may be a suitable automated and customizable tool for FAZ quantification of Cirrus HD-OCT 5000 images, providing results comparable to those for manual measurement.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/tvst.9.12.20DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7671870PMC
November 2020

Global assessment of arteriolar, venular and capillary changes in normal tension glaucoma.

Sci Rep 2020 11 5;10(1):19222. Epub 2020 Nov 5.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China.

Microcirculatory insufficiency has been hypothesized in glaucoma pathogenesis. There is a scarcity of data to comprehensively examine the changes in retinal microvasculature and its role in normal tension glaucoma (NTG). We conducted a cross-sectional case-control study and included 168 eyes from 100 NTG patients and 68 healthy subjects. Quantitative retinal arteriolar and venular metrics were measured from retinal photographs using a computer-assisted program. Radial peripapillary capillary network was imaged with OCT-A and quantitative capillary metrics (circumpapillary vessel density (cpVD) and circumpapillary fractal dimension (cpFD)) were measured with a customized MATLAB program. We found that NTG was associated with decreased arteriolar and venular tortuosity, arteriolar branching angle, cpVD and cpFD. Decreased venular caliber, arteriolar and venular branching angles, cpVD and cpFD were associated with thinner average RNFL thickness. Decreased arteriolar and venular branching angles, cpVD and cpFD were also associated with worse standard automated perimetry measurements (mean deviation and visual field index). Compared with retinal arteriolar and venular metrics, regression models based on OCT-A capillary metrics consistently showed stronger associations with NTG and structural and functional measurements in NTG. We concluded that NTG eyes showed generalized microvascular attenuations, in which OCT-A capillary metrics attenuations were more prominent and strongly associated with NTG.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-020-75784-1DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7644687PMC
November 2020

Exposure to Secondhand Smoke in Children is Associated with a Thinner Retinal Nerve Fiber Layer: The Hong Kong Children Eye Study.

Am J Ophthalmol 2021 03 29;223:91-99. Epub 2020 Oct 29.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong SAR, China; Department of Ophthalmology and Visual Sciences, Prince of Wales Hospital, Hong Kong; Hong Kong Eye Hospital, Hong Kong SAR, China; Department of Ophthalmology, Hong Kong Children's Hospital SAR, China. Electronic address:

Purpose: We sought to assess the effects of exposure to secondhand smoke (SHS) on peripapillary retinal nerve fiber layer (p-RNFL) thickness in children.

Design: Cross-sectional study.

Methods: Children 6-8 years of age were consecutively recruited from the population-based Hong Kong Children Eye Study. All participants received comprehensive ophthalmic examinations and p-RNFL thickness was measured by spectral-domain optical coherence tomography. SHS data were derived from a validated questionnaire. Associations between p-RNFL thickness and SHS exposure status, number of smokers in the family, and quantity of smoking in the family were determined by multivariate linear regression after adjusting for potential confounders.

Results: Among the Hong Kong Children Eye Study cohort (n = 3,103), approximately one-third of children were exposed to SHS (35.4%, n = 1,097). Compared to those without exposure to SHS, children exposed to SHS had similar age (P = .83), gender (P = .17), body mass index (P = .44), birth weight (P = .23), and axial length (P = .34), but had lower family income (P < .001) and lower parental education level (P < .001). After adjusting for all the above factors, exposure to SHS was associated with a thinner global p-RNFL by 4.4 μm (P < .001). Reduced p-RNFL was also associated with increased numbers of smokers in the family (β = -3.40, P < .001) and increased quantity of SHS (β = -0.22, P < .001).

Conclusions: Exposure to SHS in children was associated with a thinner p-RNFL. A thinner p-RNFL may increase the risk of irreversible visual impairment in the future. Our results provide evidence to recommend that children avoid exposure to SHS.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ajo.2020.10.016DOI Listing
March 2021

Comparison of Peripapillary Vessel Density of Acute Nonarteritic Anterior Ischemic Optic Neuropathy and Other Optic Neuropathies With Disc Swelling Using Optical Coherence Tomography Angiography: A Pilot Study.

J Neuroophthalmol 2020 Oct 26. Epub 2020 Oct 26.

Hong Kong Eye Hospital (JCC, JKHL, CKMC), Hong Kong, China; Department of Ophthalmology and Visual Sciences (JCC, NCYC, CYC, JKHL, CKMC), the Chinese University of Hong Kong, Hong Kong, China; and Department of Ophthalmology and Visual Sciences (NCYC), the Prince of Wales Hospital and Alice Ho Miu Ling Nethersole Hospital, Hong Kong, China.

Background: The purpose of this study is to quantitatively compare the peripapillary vessel density (PPVD), measured with optical coherence tomography angiography (OCT-A), between acute nonarteritic anterior ischemic optic neuropathy (NAION) and other causes of disc swelling ("others").

Methods: In this prospective comparative case series, patients with unilateral disc swelling due to acute NAION (n = 7) and "others" (n = 7) underwent OCT-A scanning of the optic nerve head with a swept-source OCT (Triton DRI-OCT), in addition to functional assessment. OCT-A images were analyzed using an automated customized MATLAB program. Comparison was made between total and 6 sectoral PPVD (radial peripapillary capillary [RPC] and choroid layers) of affected and fellow eyes; and between the 2 groups' affected eyes. Five NAION patients had repeated assessments at 1, 3, and 6 months.

Results: Acute NAION eyes had a significantly lower total and superonasal PPVD (both layers) compared to fellow eyes. No such difference was observed in "others" group for the RPC layer. NAION eyes also had significantly lower total RPC PPVD than affected eyes in the "others" group. Over 6 months, NAION eyes had persistently lower RPC PPVD compared to fellow eyes but the reduced choroidal PPVD resolved by 1 month.

Conclusion: The study demonstrated reduced superonasal and total RPC PPVD in acute NAION, which persisted over 6 months. Because there is currently no single diagnostic test for NAION, use of OCT-A images to analyze RPC PPVD may potentially help distinguish acute NAION from other causes of disc swelling by quantitatively demonstrating capillary dropout in the RPC layer.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1097/WNO.0000000000001106DOI Listing
October 2020

Optical coherence tomography angiography in diabetic retinopathy: an updated review.

Eye (Lond) 2021 Jan 24;35(1):149-161. Epub 2020 Oct 24.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong Special Administrative Region, China.

Diabetic retinopathy (DR) is a common microvascular complication of diabetes mellitus. Optical coherence tomography angiography (OCTA) has been developed to visualize the retinal microvasculature and choriocapillaris based on the motion contrast of circulating blood cells. Depth-resolved ability and non-invasive nature of OCTA allow for repeated examinations and visualization of microvasculature at the retinal capillary plexuses and choriocapillaris. OCTA enables quantification of microvascular alterations in the retinal capillary network, in addition to the detection of classical features associated with DR, including microaneurysms, intraretinal microvascular abnormalities, and neovascularization. OCTA has a promising role as an objective tool for quantifying extent of microvascular damage and identify eyes with diabetic macular ischaemia contributed to visual loss. Furthermore, OCTA can identify preclinical microvascular abnormalities preceding the onset of clinically detectable DR. In this review, we focused on the applications of OCTA derived quantitative metrics that are relevant to early detection, staging and progression of DR. Advancement of OCTA technology in clinical research will ultimately lead to enhancement of individualised management of DR and prevention of visual impairment in patients with diabetes.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41433-020-01233-yDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7852654PMC
January 2021

Correction: Deep learning in glaucoma with optical coherence tomography: a review.

Eye (Lond) 2021 Jan;35(1):357

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, Hong Kong.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41433-020-01244-9DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7852565PMC
January 2021

A deep-learning system for the assessment of cardiovascular disease risk via the measurement of retinal-vessel calibre.

Nat Biomed Eng 2020 Oct 12. Epub 2020 Oct 12.

Singapore Eye Research Institute, Singapore National Eye Center, Singapore, Singapore.

Retinal blood vessels provide information on the risk of cardiovascular disease (CVD). Here, we report the development and validation of deep-learning models for the automated measurement of retinal-vessel calibre in retinal photographs, using diverse multiethnic multicountry datasets that comprise more than 70,000 images. Retinal-vessel calibre measured by the models and by expert human graders showed high agreement, with overall intraclass correlation coefficients of between 0.82 and 0.95. The models performed comparably to or better than expert graders in associations between measurements of retinal-vessel calibre and CVD risk factors, including blood pressure, body-mass index, total cholesterol and glycated-haemoglobin levels. In retrospectively measured prospective datasets from a population-based study, baseline measurements performed by the deep-learning system were associated with incident CVD. Our findings motivate the development of clinically applicable explainable end-to-end deep-learning systems for the prediction of CVD on the basis of the features of retinal vessels in retinal photographs.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41551-020-00626-4DOI Listing
October 2020

Deep learning in glaucoma with optical coherence tomography: a review.

Eye (Lond) 2021 Jan 7;35(1):188-201. Epub 2020 Oct 7.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, Hong Kong.

Deep learning (DL), a subset of artificial intelligence (AI) based on deep neural networks, has made significant breakthroughs in medical imaging, particularly for image classification and pattern recognition. In ophthalmology, applying DL for glaucoma assessment with optical coherence tomography (OCT), including OCT traditional reports, two-dimensional (2D) B-scans, and three-dimensional (3D) volumetric scans, has increasingly raised research interests. Studies have demonstrated that using DL for interpreting OCT is efficient, accurate, and with good performance for discriminating glaucomatous eyes from normal eyes, suggesting that incorporation of DL technology in OCT for glaucoma assessment could potentially address some gaps in the current practice and clinical workflow. However, further research is crucial in tackling some existing challenges, such as annotation standardization (i.e., setting a standard for ground truth labelling among different studies), development of DL-powered IT infrastructure for real-world implementation, prospective validation in unseen datasets for further evaluation of generalizability, cost-effectiveness analysis after integration of DL, the AI "black box" explanation problem. This review summarizes recent studies on the application of DL on OCT for glaucoma assessment, identifies the potential clinical impact arising from the development and deployment of the DL models, and discusses future research directions.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41433-020-01191-5DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7852526PMC
January 2021

Different Effect of Media Opacity on Vessel Density Measured by Different Optical Coherence Tomography Angiography Algorithms.

Transl Vis Sci Technol 2020 07 13;9(8):19. Epub 2020 Jul 13.

Joint Shantou International Eye Center, Shantou University and the Chinese University of Hong Kong, Shantou, China.

Purpose: Several studies show that media opacity reduces vessel density (VD) measured by image processing algorithms of optical coherence tomography angiography (OCTA). However, different models of OCTA designed their own algorithms and computational methods, which may have different effects of opacity on VD. This study is aimed to investigate the impact of a simulated model of media opacity on quantitative measurement of two OCTA devices.

Methods: A spectral-domain based OCTA (Cirrus 5000; Carl Zeiss Meditec) and a swept-source based OCTA (Triton DRI-OCT, Topcon Inc.) were used to image retinal microvasculature at the macula using 3 × 3 mm scanning protocol from 22 eyes of 22 healthy subjects. Media opacity was simulated with neutral-density filters (optical density (OD) ranges 0.10-0.48 in Cirrus; OD ranges 0.15-0.51 in Triton). The filters were placed in front of each study eye, and signal strength (SS) or signal strength intensity (SSI) was recorded during imaging. The parafoveal VD of superficial capillary plexus was then measured using the built-in software from the two devices. The correlations among OD, SS/SSI, and VD were analyzed.

Results: Increased OD was significantly correlated with decreased SS and SSI ( = -0.576 and -0.922, respectively, both < 0.001) in Cirrus and Triton, respectively. Although increased OD was significantly correlated with decreased VD in Cirrus ( = -0.539, < 0.001), there was no significant correlation between OD with VD in Triton ( = -0.143, = 0.137).

Conclusions: The effect of media opacity on quantitative measurement of VD is different between different Cirrus and Triton OCTA devices.

Translational Relevance: This study demonstrates that the effect of media opacity on VD measurement is different among different OCTA devices, suggesting that caution must be taken when interpreting VD measurement on OCTA, particularly among individuals with media opacity.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/tvst.9.8.19DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7422875PMC
July 2020

A 3D Deep Learning System for Detecting Referable Glaucoma Using Full OCT Macular Cube Scans.

Transl Vis Sci Technol 2020 02 18;9(2):12. Epub 2020 Feb 18.

Byers Eye Institute, Stanford University, Palo Alto, CA, USA.

Purpose: The purpose of this study was to develop a 3D deep learning system from spectral domain optical coherence tomography (SD-OCT) macular cubes to differentiate between referable and nonreferable cases for glaucoma applied to real-world datasets to understand how this would affect the performance.

Methods: There were 2805 Cirrus optical coherence tomography (OCT) macula volumes (Macula protocol 512 × 128) of 1095 eyes from 586 patients at a single site that were used to train a fully 3D convolutional neural network (CNN). Referable glaucoma included true glaucoma, pre-perimetric glaucoma, and high-risk suspects, based on qualitative fundus photographs, visual fields, OCT reports, and clinical examinations, including intraocular pressure (IOP) and treatment history as the binary (two class) ground truth. The curated real-world dataset did not include eyes with retinal disease or nonglaucomatous optic neuropathies. The cubes were first homogenized using layer segmentation with the Orion Software (Voxeleron) to achieve standardization. The algorithm was tested on two separate external validation sets from different glaucoma studies, comprised of Cirrus macular cube scans of 505 and 336 eyes, respectively.

Results: The area under the receiver operating characteristic (AUROC) curve for the development dataset for distinguishing referable glaucoma was 0.88 for our CNN using homogenization, 0.82 without homogenization, and 0.81 for a CNN architecture from the existing literature. For the external validation datasets, which had different glaucoma definitions, the AUCs were 0.78 and 0.95, respectively. The performance of the model across myopia severity distribution has been assessed in the dataset from the United States and was found to have an AUC of 0.85, 0.92, and 0.95 in the severe, moderate, and mild myopia, respectively.

Conclusions: A 3D deep learning algorithm trained on macular OCT volumes without retinal disease to detect referable glaucoma performs better with retinal segmentation preprocessing and performs reasonably well across all levels of myopia.

Translational Relevance: Interpretation of OCT macula volumes based on normative data color distributions is highly influenced by population demographics and characteristics, such as refractive error, as well as the size of the normative database. Referable glaucoma, in this study, was chosen to include cases that should be seen by a specialist. This study is unique because it uses multimodal patient data for the glaucoma definition, and includes all severities of myopia as well as validates the algorithm with international data to understand generalizability potential.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1167/tvst.9.2.12DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7347026PMC
February 2020

Effect of weight loss on the retinochoroidal structural alterations among patients with exogenous obesity.

PLoS One 2020 9;15(7):e0235926. Epub 2020 Jul 9.

Department of Ophthalmology, Advanced Eye Center, Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, India.

Purpose: To evaluate the changes in the retinochoroidal vasculature in patients with exogenous obesity using swept-source optical coherence tomography (SS-OCT) and OCT angiography (OCTA).

Methods: In this prospective study, 60 patients diagnosed with obesity (47 males) (mean age: 46.47±10.9 years) were included, of which 30 patients underwent bariatric surgery (Group A), and 30 patients underwent conservative management (exercise/diet) (Group B). Parameters including choroidal thickness (CT), choroidal vascularity index (CVI) and retinal capillary density index (CDI) and arteriovenous ratio (AVR) were measured at the baseline and three months follow up. 30 eyes (30 age and gender-matched) of normal participants were included for comparison.

Results: Baseline CT was lower in 60 participants with obesity compared to controls. Compared with normal subjects, subjects with obesity had higher mean CVI (0.66±0.02 versus 0.63±0.04; p<0.01), smaller FAZ area (0.26±0.07 versus 0.45±0.32; p<0.01), higher CDI (superficial plexus: 0.7±0.04 versus 0.68±0.06; p = 0.04, deep plexus: 0.38±0.02 versus 0.35±0.06; p = 0.01), and lower AVR (0.68±0.05 versus 0.70±0.03 versus; p<0.01). At 3-month after intervention, CT showed a significant increase in participants from Group A (329.27±79μm; p<0.01) but not in Group B from baseline. No significant change was noted in CVI or CDI at 3-month in either group compared to baseline. AVR significantly increased in Group B (p = 0.03).

Conclusion: Subclinical changes in retinochoroidal vasculature occurs in participants with exogenous obesity compared to healthy subjects. Surgical intervention (bariatric surgery) may have a favorable outcome on the choroidal thickness in these patients.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0235926PLOS
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7347179PMC
September 2020

Association of Optical Coherence Tomography Angiography Metrics With Detection of Impaired Macular Microvasculature and Decreased Vision in Amblyopic Eyes: The Hong Kong Children Eye Study.

JAMA Ophthalmol 2020 08;138(8):858-865

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong.

Importance: Microvascular abnormalities in amblyopia are becoming evident with high-resolution imaging, such as optical coherence tomography angiography (OCT-A); however, to our knowledge, the clinical significance and use of these findings are unknown.

Objective: To assess changes in quantitative OCT-A metrics in amblyopic eyes and explore their association with visual acuity in children.

Design, Setting, And Participants: This population-based nested case-control study included children aged 6 to 8 years who were consecutively recruited between January 2016 and July 2017 from the population-based Hong Kong Children Eye Study (HKCES) at the Chinese University of Hong Kong Eye Centre. All participants underwent OCT-A with a swept-source OCT and detailed ophthalmic investigations. Macular microvasculature of the superficial capillary plexus was quantified by a customized automated image analysis program. A multivariable linear regression was conducted to evaluate the differences in OCT-A metrics between amblyopic and nonamblyopic eyes after adjustment for all known confounders. Data analysis was conducted from September to November 2018.

Main Outcomes And Measures: Differences in OCT-A metric (foveal avascular zone [FAZ]) area, FAZ circularity, vessel density, vessel diameter index, and fractal dimension between amblyopic and nonamblyopic eyes.

Results: There were 30 participants with amblyopia (mean [SD] age, 7.57 [1.2] years; 16 girls [53.3%]) and 1045 controls (mean [SD] age, 7.65 [1.0] years; 580 girls [55.5%]) in this cohort. Compared with control eyes, amblyopic eyes had decreased FAZ circularity (-0.058; 95% CI, -0.096 to -0.021, P = .002), decreased fractal dimension (-0.014; 95% CI, -0.024 to -0.003; P = .01), and increased vessel diameter index (0.002; 95% CI, 0.002 to 0.003; P < .001). A difference was not identified between FAZ area and vessel density. LogMAR visual acuity was associated with FAZ circularity (sβ, -0.133; P < .001) and vessel diameter index (sβ, 0.097; P = .001) but not with vessel density nor FAZ area.

Conclusions And Relevance: The results of this population-based study in children supports the presence of macular microvascular abnormalities in amblyopic eyes. Such changes as measured by OCT-A metrics are associated with visual acuity, inferring retinal involvement in the development of amblyopia and suggesting a potential role of quantitative OCT-A metrics in the diagnosis and recognition of amblyopia.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1001/jamaophthalmol.2020.2220DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7317658PMC
August 2020

Assessment of retinal neurodegeneration with spectral-domain optical coherence tomography: a systematic review and meta-analysis.

Eye (Lond) 2021 May 24;35(5):1317-1325. Epub 2020 Jun 24.

Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, SAR, China.

Objectives: To comprehensively assess diabetic retinopathy neurodegeneration (DRN) as quantified by retinal neuronal and axonal layers measured with spectral-domain optical coherence tomography (SD-OCT) in subjects with diabetes mellitus (DM).

Methods: Articles on the topic of examining macular ganglion cell-inner plexiform layer (m-GCIPL), macular retinal nerve fibre layer (m-RNFL), macular ganglion cell complex (m-GCC), and peripapillary RNFL (p-RNFL) measured with SD-OCT in DM subjects without DR (NDR) or with non-proliferative DR (NPDR) were searched in PubMed and Embase up to November 31, 2019. Standardized mean difference (SMD) as effect size were pooled using random-effects model.

Results: Thirty-six studies searched from online databases and the CUHK DM cohort were included in the meta-analysis. In the comparison between NDR and control, macular measures including mean m-GCIPL (SMD = -0.26, p = 0.003), m-RNFL (SMD = -0.26, p = 0.046), and m-GCC (SMD = -0.28; p = 0.009) were significantly thinner in the NDR group. In the comparison between NPDR and NDR, only mean p-RNFL was significantly thinner in the NPDR group (SMD = -0.27; p = 0.03), but not other macular measures.

Conclusions: Thinning of retinal neuronal and axonal layers at macula as measured by SD-OCT are presented in eyes with NDR, supporting DRN may be the early pathogenesis in the DM patients without the presence of clinical signs of DR. In the future, these SD-OCT measures may be used as surrogates of DRN to stratify DM patients with a high risk of DR, and may be used as a therapeutic target if neuroprotection treatment for DR is available.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41433-020-1020-zDOI Listing
May 2021

Association of Retinal Microvascular Signs with Incident Atrial Fibrillation: The Multi-Ethnic Study of Atherosclerosis.

Ophthalmol Retina 2021 01 19;5(1):78-85. Epub 2020 Jun 19.

Department of Preventive Medicine, Northwestern University, Chicago, Illinois.

Purpose: Microvascular diseases may contribute to the occurrence of atrial fibrillation (AF). Retinal microvascular signs that are similar to other microvasculature in the body and can be visualized directly via ophthalmoscopy may provide insights into such a relationship.

Design: Prospective, longitudinal, multiethnic study.

Participants: We examined the association between retinal microvascular signs and incident AF in 4994 participants 47 to 86 years of age and free of prior AF who underwent fundus photography from 2002 through 2004 and were followed up through 2015 in the Multi-Ethnic Study of Atherosclerosis (MESA).

Methods: Retinal microvascular signs evaluated include central retinal arteriolar equivalent and central retinal venular equivalent (CRVE) and presence of any retinopathy signs (e.g., retinal microaneurysms or hemorrhages). A multivariate Cox regression analysis was used to determine the relationship while adjusting for traditional risk factors, alcohol intake, body mass index, diabetes status, chronic kidney disease status, hemoglobin A1c level, C-reactive protein level, medications, and prevalent cardiovascular diseases or heart failure.

Main Outcome And Measures: Incident AF events were identified using 12-lead electrocardiographic findings, hospital discharge records, and Medicare claims data.

Results: During a median follow-up of 14.1 years, 643 AF events were identified. No association was found between any retinal microvascular signs and incident AF except for retinal focal arteriolar narrowing (hazard ratio, 1.75; 95% confidence interval, 1.06-2.87) in the overall population. However, in the subgroup analyses by gender, wider CRVE was associated with a higher risk of incident AF in women, but not in men (hazard ratio for every 10-μm increase in CRVE, 1.08 [95% confidence interval, 1.01-1.15] and 0.97 [95% confidence interval, 0.92-1.03], respectively; P = 0.041 for interaction).

Conclusions: No consistent pattern of association was found between retinal microvascular signs and incident AF. We observed an association in women, but not in men, of wider retinal venular calibers with incidence of AF. The reasons for a possible interaction are incompletely understood.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.oret.2020.06.019DOI Listing
January 2021

Towards multi-center glaucoma OCT image screening with semi-supervised joint structure and function multi-task learning.

Med Image Anal 2020 07 19;63:101695. Epub 2020 May 19.

Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, China.

Glaucoma is the leading cause of irreversible blindness in the world. Structure and function assessments play an important role in diagnosing glaucoma. Nowadays, Optical Coherence Tomography (OCT) imaging gains increasing popularity in measuring the structural change of eyes. However, few automated methods have been developed based on OCT images to screen glaucoma. In this paper, we are the first to unify the structure analysis and function regression to distinguish glaucoma patients from normal controls effectively. Specifically, our method works in two steps: a semi-supervised learning strategy with smoothness assumption is first applied for the surrogate assignment of missing function regression labels. Subsequently, the proposed multi-task learning network is capable of exploring the structure and function relationship between the OCT image and visual field measurement simultaneously, which contributes to classification performance improvement. It is also worth noting that the proposed method is assessed by two large-scale multi-center datasets. In other words, we first build the largest glaucoma OCT image dataset (i.e., HK dataset) involving 975,400 B-scans from 4,877 volumes to develop and evaluate the proposed method, then the model without further fine-tuning is directly applied on another independent dataset (i.e., Stanford dataset) containing 246,200 B-scans from 1,231 volumes. Extensive experiments are conducted to assess the contribution of each component within our framework. The proposed method outperforms the baseline methods and two glaucoma experts by a large margin, achieving volume-level Area Under ROC Curve (AUC) of 0.977 on HK dataset and 0.933 on Stanford dataset, respectively. The experimental results indicate the great potential of the proposed approach for the automated diagnosis system.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.media.2020.101695DOI Listing
July 2020

Artificial Intelligence in Ophthalmology: Evolutions in Asia.

Asia Pac J Ophthalmol (Phila) 2020 Mar-Apr;9(2):78-84

Department of Ophthalmology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, South Korea.

Artificial intelligence (AI) has been studied in ophthalmology since availability of digital information in ophthalmic care. The significant turning point was availability of commercial digital color fundus photography in the late 1990s, which caused digital screening for diabetic retinopathy (DR) to take off. Automated Retinal Disease Assessment software was then developed using machine learning to detect abnormal lesions in fundus to screen DR. The use of this version of AI had not been generalized because the specificity at 45% was not high enough, although the sensitivity reached 90%. The recent breakthrough in machine learning is the invent of deep learning, which accelerates its performance to be on par with experts. The first 2 breakthrough studies on deep learning for screening DR were conducted in Asia. The first represented collaboration of datasets between Asia and the United States for algorithms development, whereas the second represented algorithms developed in Asia but validated in different populations across the world. Both found accuracy for detecting referable DR of >95%. Diversity and variety are unique strengths of Asia for AI studies. There are many more studies of AI ongoing in Asia not only as prospective deployments in DR but in glaucoma, age-related macular degeneration, cataract, and systemic disease, such as Alzheimer's disease. Some Asian countries have laid out plans for digital health care system using AI as one of the puzzle pieces for solving blindness. More studies on AI and digital health are expected to come from Asia in this new decade.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1097/01.APO.0000656980.41190.bfDOI Listing
March 2021

Artificial Intelligence to Detect Papilledema from Ocular Fundus Photographs.

N Engl J Med 2020 04 14;382(18):1687-1695. Epub 2020 Apr 14.

From the Singapore National Eye Center (D.M., D.T., S.S., C.-Y.C., T.Y.W.), Singapore Eye Research Institute (D.M., R.P.N., D.T., C.V., S.S., C.-Y.C., T.Y.W.), Duke-NUS Medical School (D.M., R.P.N., D.T., S.S., C.-Y.C., T.Y.W.), Institute of High Performance Computing, Agency for Science, Technology, and Research (J.Z., X.X., Y.L.), and Yong Loo Lin School of Medicine, National University of Singapore (S.S., T.Y.W.) - all in Singapore; Farabi Eye Hospital, Tehran University of Medical Science, Tehran, Iran (M.A.F.); the Department of Ophthalmology, Centro Hospitalar e Universitário de Coimbra, and the Coimbra Institute for Biomedical Imaging and Translational Research, University of Coimbra, Coimbra, Portugal (P.F.); the Department of Ophthalmology, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand (K.V.); the Eye Center, Medical Center, University of Freiburg, Freiburg (W.A.L.), and the Department of Ophthalmology, Ruprecht Karl University of Heidelberg, Mannheim (J.B.J.) - both in Germany; IRCCS Istituto delle Scienze Neurologiche di Bologna, Unità Operativa Complessa Clinica Neurologica, and Dipartimento di Scienze Biomediche e Neuromotorie, Università di Bologna, Bologna, Italy (C.L.M.); the Department of Ophthalmology and Visual Sciences, Chinese University of Hong Kong, Hong Kong (C.Y.C.), and Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou (H.Y.) - both in China; the Department of Ophthalmology, Rigshospitalet, University of Copenhagen, Glostrup, Denmark (S.H.); the Department of Ophthalmology, University Hospital of Grenoble-Alpes, and Grenoble-Alpes University, HP2 Laboratory, INSERM Unité 1042, Grenoble (C.C.), Service d'Ophtalmologie, Unité Rétine-Uvéites-Neuro-Ophtalmologie, Hôpital Pellegrin, Centre Hospitalier Universitaire de Bordeaux, Bordeaux (M.-B.R.), the Department of Ophthalmology, Lille Catholic Hospital, Lille Catholic University, and INSERM Unité 1171, Lille (T.T.H.C.), the Department of Ophthalmology, University Hospital Angers, Angers (P.G.), and Rothschild Foundation Hospital, Paris (C.C.-V.) - all in France; the Department of Clinical Neurosciences, Geneva University Hospital, Geneva (N.S.); the Department of Neurology, SUNY Upstate Medical University, Syracuse, NY (L.J.M.); the American Eye Center, Mandaluyong City, Philippines (R.K.); Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, University College London, London (P.Y.-W.-M.), and Cambridge Eye Unit, Addenbrooke's Hospital, Cambridge University Hospitals, and Cambridge Centre for Brain Repair and Medical Research Council Mitochondrial Biology Unit, Department of Clinical Neurosciences, University of Cambridge, Cambridge (P.Y.-W.-M.) - all in the United Kingdom; the Save Sight Institute, Faculty of Health and Medicine, University of Sydney, Sydney (C.L.F.); the Department of Ophthalmology and Neurology, Mayo Clinic, Rochester, MN (J.J.C.); the Department of Neuro-ophthalmology, Sankara Nethralaya, Medical Research Foundation, Chennai, India (S.A.); the Departments of Ophthalmology, Neurology, and Neurosurgery, Johns Hopkins University School of Medicine, Baltimore (N.R.M.); and the Departments of Ophthalmology and Neurology, Emory University School of Medicine, Atlanta (N.J.N., V.B.).

Background: Nonophthalmologist physicians do not confidently perform direct ophthalmoscopy. The use of artificial intelligence to detect papilledema and other optic-disk abnormalities from fundus photographs has not been well studied.

Methods: We trained, validated, and externally tested a deep-learning system to classify optic disks as being normal or having papilledema or other abnormalities from 15,846 retrospectively collected ocular fundus photographs that had been obtained with pharmacologic pupillary dilation and various digital cameras in persons from multiple ethnic populations. Of these photographs, 14,341 from 19 sites in 11 countries were used for training and validation, and 1505 photographs from 5 other sites were used for external testing. Performance at classifying the optic-disk appearance was evaluated by calculating the area under the receiver-operating-characteristic curve (AUC), sensitivity, and specificity, as compared with a reference standard of clinical diagnoses by neuro-ophthalmologists.

Results: The training and validation data sets from 6779 patients included 14,341 photographs: 9156 of normal disks, 2148 of disks with papilledema, and 3037 of disks with other abnormalities. The percentage classified as being normal ranged across sites from 9.8 to 100%; the percentage classified as having papilledema ranged across sites from zero to 59.5%. In the validation set, the system discriminated disks with papilledema from normal disks and disks with nonpapilledema abnormalities with an AUC of 0.99 (95% confidence interval [CI], 0.98 to 0.99) and normal from abnormal disks with an AUC of 0.99 (95% CI, 0.99 to 0.99). In the external-testing data set of 1505 photographs, the system had an AUC for the detection of papilledema of 0.96 (95% CI, 0.95 to 0.97), a sensitivity of 96.4% (95% CI, 93.9 to 98.3), and a specificity of 84.7% (95% CI, 82.3 to 87.1).

Conclusions: A deep-learning system using fundus photographs with pharmacologically dilated pupils differentiated among optic disks with papilledema, normal disks, and disks with nonpapilledema abnormalities. (Funded by the Singapore National Medical Research Council and the SingHealth Duke-NUS Ophthalmology and Visual Sciences Academic Clinical Program.).
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1056/NEJMoa1917130DOI Listing
April 2020

UD-MIL: Uncertainty-Driven Deep Multiple Instance Learning for OCT Image Classification.

IEEE J Biomed Health Inform 2020 12 4;24(12):3431-3442. Epub 2020 Dec 4.

Deep learning has achieved remarkable success in the optical coherence tomography (OCT) image classification task with substantial labelled B-scan images available. However, obtaining such fine-grained expert annotations is usually quite difficult and expensive. How to leverage the volume-level labels to develop a robust classifier is very appealing. In this paper, we propose a weakly supervised deep learning framework with uncertainty estimation to address the macula-related disease classification problem from OCT images with the only volume-level label being available. First, a convolutional neural network (CNN) based instance-level classifier is iteratively refined by using the proposed uncertainty-driven deep multiple instance learning scheme. To our best knowledge, we are the first to incorporate the uncertainty evaluation mechanism into multiple instance learning (MIL) for training a robust instance classifier. The classifier is able to detect suspicious abnormal instances and abstract the corresponding deep embedding with high representation capability simultaneously. Second, a recurrent neural network (RNN) takes instance features from the same bag as input and generates the final bag-level prediction by considering the individually local instance information and globally aggregated bag-level representation. For more comprehensive validation, we built two large diabetic macular edema (DME) OCT datasets from different devices and imaging protocols to evaluate the efficacy of our method, which are composed of 30,151 B-scans in 1,396 volumes from 274 patients (Heidelberg-DME dataset) and 38,976 B-scans in 3,248 volumes from 490 patients (Triton-DME dataset), respectively. We compare the proposed method with the state-of-the-art approaches, and experimentally demonstrate that our method is superior to alternative methods, achieving volume-level accuracy, F1-score and area under the receiver operating characteristic curve (AUC) of 95.1%, 0.939 and 0.990 on Heidelberg-DME and those of 95.1%, 0.935 and 0.986 on Triton-DME, respectively. Furthermore, the proposed method also yields competitive results on another public age-related macular degeneration OCT dataset, indicating the high potential as an effective screening tool in the clinical practice.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/JBHI.2020.2983730DOI Listing
December 2020

Profiles of Ganglion Cell-Inner Plexiform Layer Thickness in a Multi-Ethnic Asian Population: The Singapore Epidemiology of Eye Diseases Study.

Ophthalmology 2020 08 8;127(8):1064-1076. Epub 2020 Feb 8.

Singapore Eye Research Institute, Singapore National Eye Centre, Singapore; Ophthalmology and Visual Sciences Academic Clinical Program (Eye ACP), Duke-NUS Medical School, Singapore; Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore. Electronic address:

Purpose: To examine the normative profile and determinants of macular ganglion cell-inner plexiform layer (GCIPL) thickness based on spectral-domain OCT (SD-OCT) in a nonglaucoma, multi-ethnic Asian population.

Design: Population-based, cross-sectional study.

Participants: Ethnic Chinese, Malay, and Indian adults aged ≥40 years recruited from the Singapore Epidemiology of Eye Diseases Study.

Methods: All participants underwent standardized examinations. The GCIPL thickness was measured using Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA). Participants with glaucoma or poor-quality scans were excluded. Eye-specific data were used. Associations of ocular and systemic factors with GCIPL thickness parameters were investigated using multivariable linear regression with generalized estimating equation models to account for correlation between both eyes.

Main Outcome Measures: GCIPL thickness.

Results: A total of 4464 participants (7520 eyes) consisting of 1625 Chinese, 1212 Malay, and 1627 Indian adults contributed to this analysis. Average GCIPL thickness was 82.6±6.1 μm in Chinese, 81.5±6.8 μm in Malays, and 78.0±6.9 μm in Indians (P < 0.001 by analysis of variance). The 5th percentile limit of average GCIPL thickness was 72 μm in Chinese, 70 μm in Malays, and 67 μm in Indians. In multivariable analysis adjusting for age, gender, axial length, presence of cataract, OCT signal strength, disc area, hypertension, diabetes, and hyperlipidemia, eyes of Indians were observed to have 3.43 μm thinner GCIPL on average compared with Chinese (P < 0.001) and 3.36 μm thinner GCIPL compared with Malays (P < 0.001). In addition, older age (per decade; β = -2.51), female (β = -1.57), longer axial length (per mm; β = -1.54), and presence of chronic kidney disease (β = -1.49) were significantly associated with thinner average GCIPL (all P ≤ 0.008). Larger optic disc area (per mm; β = 0.78; P < 0.001) was associated with thicker GCIPL. These factors were consistently observed to be significant for superior and inferior hemisphere GCIPL thickness.

Conclusions: GCIPL thickness profiles were significantly thinner in Indians compared with Chinese and Malays. Our findings further highlight the need of a more refined, ethnic-specific normative database for GCIPL thickness, which in turn may improve the detection and diagnosis of glaucoma in Asians.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ophtha.2020.01.055DOI Listing
August 2020