Effects of bariatric surgery on retinal microvascular architecture in obese patients.

STUDY AIM: Retinal microvasculature changes reflect systemic small vessel damage from obesity. The impact of bariatric surgery induced weight loss on the microvasculature is relatively unknown. We hypothesized that weight loss following bariatric surgery would be associated with improved structural changes in the retinal microvasculature, reflecting an overall improvement in microvascular health. METHODS: The study included 22 obese subjects scheduled for bariatric surgery (laparoscopic Roux-en-Y gastric bypass or a sleeve gastrectomy) and 15 lean, age-matched controls. Ophthalmic examination, including fundus photography, was performed at baseline and 6-months. Retinal microvasculature caliber was analysed quantitatively using a semi-automated computer program and summarized as central retinal artery equivalent (CRAE) and venular equivalent (CRVE). RESULTS: Mean weight loss at 6 months was 26.1 kg ± 8 kg in the bariatric surgery group. Retinal artery caliber increased (136.0 ± 1.4 to 141.4 ± 1.4 µm, p = 0.013) and venular caliber decreased (202.9 ± 1.9 to 197.3 ± 1.9 µm, p = 0.046) in the bariatric surgery group by 6 months, with no change in arteriolar (136.6 ± 1.1 to 134.5 ± 1.2, p = 0.222) or venular (195.1 ± 2.1 to 193.3 ± 2.2, p = 0.550) caliber in the control group. The arteriolar to venular ratio increased in the bariatric surgery group, with no change in the control group at 6 months. CONCLUSIONS: The findings suggest obesity-related microvascular changes are reversible after bariatric surgery-induced weight loss. The capacity for the retinal microvasculature to improve following bariatric surgery suggests plasticity of the human microvasculature early in the disease course.

Retinal vascular geometry and 6 year incidence and progression of diabetic retinopathy.

AIMS/HYPOTHESIS: We aimed to examine prospectively the association between a range of retinal vascular geometric variables measured from retinal photographs and the 6 year incidence and progression of diabetic retinopathy. METHODS: We conducted a prospective, population-based cohort study of Asian Malay individuals aged 40-80 years at baseline (n = 3280) who returned for a 6 year follow-up. Retinal vascular geometric variables (tortuosity, branching, fractal dimension, calibre) were measured from baseline retinal photographs using a computer-assisted program (Singapore I Vessel Assessment). Diabetic retinopathy was graded from baseline and follow-up photographs using the modified Airlie House classification system. Incidence of diabetic retinopathy was defined as a severity of ≥15 at follow-up among those without diabetic retinopathy at baseline. Incidence of referable diabetic retinopathy was defined as moderate or severe non-proliferative diabetic retinopathy, proliferative diabetic retinopathy or diabetic macular oedema at follow-up in participants who had had no or mild non-proliferative diabetic retinopathy at baseline. Progression of diabetic retinopathy was defined as an increase in severity of ≥2 steps at follow-up. Log-binomial models with an expectation-maximisation algorithm were used to estimate RR adjusting for age, sex, diabetes duration, HbA1c level, BP, BMI, estimated GFR and total and HDL-cholesterol at baseline. RESULTS: A total of 427 individuals with diabetes participated in the baseline and 6 year follow-up examinations. Of these, 19.2%, 7.57% and 19.2% developed incidence of diabetic retinopathy, incidence of referable diabetic retinopathy and diabetic retinopathy progression, respectively. After multivariate adjustment, greater arteriolar simple tortuosity (mean RR [95% CI], 1.34 [1.04, 1.74]), larger venular branching angle (RR 1.26 [1.00, 1.59]) and larger venular branching coefficient (RR 1.26 [1.03, 1.56]) were associated with incidence of diabetic retinopathy. Greater arteriolar simple tortuosity (RR 1.82 [1.32, 2.52]), larger venular branching coefficient (RR 1.46 [1.03, 2.07]), higher arteriolar fractal dimension (RR 1.59 [1.08, 2.36]) and larger arteriolar calibre (RR 1.83 [1.15, 2.90]) were associated with incidence of referable diabetic retinopathy. Greater arteriolar simple tortuosity (RR 1.34 [1.12, 1.61]) was associated with diabetic retinopathy progression. Addition of retinal vascular variables improved discrimination (C-statistic 0.796 vs 0.733, p = 0.031) and overall reclassification (net reclassification improvement 18.8%, p = 0.025) of any diabetic retinopathy risk beyond established risk factors. CONCLUSIONS/INTERPRETATION: Retinal vascular geometry measured from fundus photographs predicted the incidence and progression of diabetic retinopathy in adults with diabetes, beyond established risk factors.

Choroidal structures in polypoidal choroidal vasculopathy, neovascular age-related maculopathy, and healthy eyes determined by binarization of swept source optical coherence tomographic images.

PURPOSE: To evaluate quantitatively the choroidal vascularity in polypoidal choroidal vasculopathy (PCV) and neovascular age-related macular degeneration (AMD) patients compared to healthy controls. METHODS: All eyes underwent swept source optical coherence tomography (OCT), and choroidal images were binarized into blood vessels lumen and stroma. The choroidal vascular index (CVI) was defined as the ratio of luminal area (LA) over total choroidal area of the subfoveal region with a width of 1500 µm. RESULTS: The study included 73 patients with neovascular AMD or PCV with mean ± standard deviation (SD) age of 71.8 ± 9.3 years, which was older than the mean age of 65.1 ± 10.8 years of 72 healthy eyes from control group (p < 0.01). The 44 PCV eyes had significantly higher mean SFCT of 214.23 ± 95.21 µm than neovascular AMD eyes (172.74 ± 96.48 µm, p = 0.03) and greater luminal area (0.23 ± 0.09 mm2 vs. 0.19 ± 0.08 mm2, p = 0.05). After adjusting for age, axial length, and gender in multivariate regression analysis, the SFCT of PCV and neovascular AMD eyes were not significantly different from healthy eyes (195.55 ± 93.11 µm), but the CVI of both PCV (64.94 ± 5.43%, p = 0.01) and neovascular AMD (62.54 ± 5.57%, p = <0.01) were significantly lower than control (68.53 ± 5.91%). CONCLUSION: Despite physiological changes of choroidal vasculature due to aging, the choroidal morphology is different in PCV, neovascular AMD and healthy eyes, which has implication on disease pathogenesis.

Development and Validation of a Deep Learning System for Diabetic Retinopathy and Related Eye Diseases Using Retinal Images From Multiethnic Populations With Diabetes.

Importance: A deep learning system (DLS) is a machine learning technology with potential for screening diabetic retinopathy and related eye diseases. Objective: To evaluate the performance of a DLS in detecting referable diabetic retinopathy, vision-threatening diabetic retinopathy, possible glaucoma, and age-related macular degeneration (AMD) in community and clinic-based multiethnic populations with diabetes. Design, Setting, and Participants: Diagnostic performance of a DLS for diabetic retinopathy and related eye diseases was evaluated using 494 661 retinal images. A DLS was trained for detecting diabetic retinopathy (using 76 370 images), possible glaucoma (125 189 images), and AMD (72 610 images), and performance of DLS was evaluated for detecting diabetic retinopathy (using 112 648 images), possible glaucoma (71 896 images), and AMD (35 948 images). Training of the DLS was completed in May 2016, and validation of the DLS was completed in May 2017 for detection of referable diabetic retinopathy (moderate nonproliferative diabetic retinopathy or worse) and vision-threatening diabetic retinopathy (severe nonproliferative diabetic retinopathy or worse) using a primary validation data set in the Singapore National Diabetic Retinopathy Screening Program and 10 multiethnic cohorts with diabetes. Exposures: Use of a deep learning system. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity of the DLS with professional graders (retinal specialists, general ophthalmologists, trained graders, or optometrists) as the reference standard. Results: In the primary validation dataset (n = 14 880 patients; 71 896 images; mean [SD] age, 60.2 [2.2] years; 54.6% men), the prevalence of referable diabetic retinopathy was 3.0%; vision-threatening diabetic retinopathy, 0.6%; possible glaucoma, 0.1%; and AMD, 2.5%. The AUC of the DLS for referable diabetic retinopathy was 0.936 (95% CI, 0.925-0.943), sensitivity was 90.5% (95% CI, 87.3%-93.0%), and specificity was 91.6% (95% CI, 91.0%-92.2%). For vision-threatening diabetic retinopathy, AUC was 0.958 (95% CI, 0.956-0.961), sensitivity was 100% (95% CI, 94.1%-100.0%), and specificity was 91.1% (95% CI, 90.7%-91.4%). For possible glaucoma, AUC was 0.942 (95% CI, 0.929-0.954), sensitivity was 96.4% (95% CI, 81.7%-99.9%), and specificity was 87.2% (95% CI, 86.8%-87.5%). For AMD, AUC was 0.931 (95% CI, 0.928-0.935), sensitivity was 93.2% (95% CI, 91.1%-99.8%), and specificity was 88.7% (95% CI, 88.3%-89.0%). For referable diabetic retinopathy in the 10 additional datasets, AUC range was 0.889 to 0.983 (n = 40 752 images). Conclusions and Relevance: In this evaluation of retinal images from multiethnic cohorts of patients with diabetes, the DLS had high sensitivity and specificity for identifying diabetic retinopathy and related eye diseases. Further research is necessary to evaluate the applicability of the DLS in health care settings and the utility of the DLS to improve vision outcomes.

Microvascular network alterations in the retina of patients with Alzheimer's disease.

BACKGROUND: Although cerebral small-vessel disease has been implicated in the development of Alzheimer's disease (AD), the cerebral microcirculation is difficult to visualize directly in vivo. Because the retina provides a noninvasive window to assess the microcirculation, we determined whether quantitatively measured retinal microvascular parameters are associated with AD. METHODS: We conducted a case-control study (case:control matching ≈ 1:2). Retinal photographs were analyzed using a computer program, and a spectrum of quantitative retinal microvascular parameters (caliber, fractal dimension, tortuosity, and bifurcation) were measured. Logistic regression models were used to compute the odds ratio (OR) and 95% confidence interval for AD adjusting for age, gender, ethnicity, smoking, hypertension, diabetes, hypercholesterolemia, and history of myocardial infarction. RESULTS: We included 136 demented patients with AD and 290 age-gender-race-matched controls. Persons with narrower venular caliber (OR per standard deviation [SD] decrease, 2.01 [1.27-3.19]), decreased arteriolar and venular fractal dimension (OR per SD decrease 1.35 [1.08-1.68], 1.47 [1.17-1.84], respectively) and increased arteriolar and venular tortuosity (OR per SD increase, 1.84 [1.40-2.31], 1.94 [1.48-2.53], respectively) were more likely to have AD. These associations still persisted when only AD cases without a history of cerebrovascular disease were included. CONCLUSIONS: Patients with AD have altered microvascular network in the retina (narrower retinal venules and a sparser and more tortuous retinal vessels) compared with matched nondemented controls. These changes in retinal microvasculature may reflect similar pathophysiological processes in cerebral microvasculature in the brains of patients with AD.

Retinal vascular fractal dimension is associated with cognitive dysfunction.

Fractal analysis is a method used to quantify the geometric branching complexity and density of retinal vessels. This study examined the relationship of retinal vascular fractal dimension and other retinal vascular parameters with cognitive dysfunction in an older Asian population. Subjects aged 60 years and older from the Singapore Malay Eye Study were selected for analysis. Retinal vascular fractal dimension (Df) and other quantitative retinal vascular parameters (branching angle, tortuosity, and caliber) were measured based on a standardized grading protocol from photographs of the retinal fundus using a computer-assisted program. Qualitative retinal signs were also assessed from photographs. Cognitive dysfunction was defined as a locally validated Abbreviated Mental Test (AMT) score ≤6/10 in participants with 0-6 years of formal education and an AMT score ≤8/10 in those with more than 6 years of formal education. Cognitive dysfunction was identified in 262 of the 1202 participants (21.8%). Decreased retinal vascular Df was significantly associated with lower AMT score (P = .019). In multivariate logistic regression analysis, participants with lower retinal vascular Df values were more likely to have cognitive dysfunction (odds ratio, 1.71; 95% confidence interval, 1.03-2.82, comparing the lowest and highest Df quintiles). In subgroup analysis stratified for cardiovascular risk factors, this association was present in participants with hypertension and current smokers. Other retinal vascular signs were not associated with cognitive dysfunction. Decreased retinal vascular Df is associated with cognitive dysfunction in older persons. Rarefaction of the retinal vasculature may reflect similar changes in the cerebral microvasculature that may contribute to cognitive deterioration.

Retinal microvascular changes and risk of stroke: the Singapore Malay Eye Study.

BACKGROUND AND PURPOSE: To examine the relationship between retinal microvascular measures and incident stroke in an Asian Malay population. METHODS: We conducted a prospective, population-based cohort study of Asian Malay persons 40 to 80 years at baseline. Retinal microvascular signs were assessed from baseline retinal photographs including quantitative retinal microvascular parameters (caliber, branching angle, tortuosity, and fractal dimension) and qualitative retinopathy signs. Incident stroke cases were identified during the follow-up period. Cox proportional-hazards regression and incremental usefulness analysis (calibration, discrimination, and reclassification) were performed. RESULTS: A total of 3189 participants were free of prevalent stroke at baseline. During the follow-up (median, 4.41 years), 51 (1.93%) participants had an incident stroke event. In Cox proportional-hazards models adjusting for established stroke predictors (age, sex, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, smoking, glycosylated hemoglobin, and antihypertensive medication), retinopathy (hazard ratio, 1.94; 95% confidence interval, 1.01-3.72) and larger retinal venular caliber (hazard ratio, 3.28; 95% confidence interval, 1.30-8.26, comparing fourth versus first quartiles) were associated with risk of stroke. Compared with the model with only established risk factors, the addition of retinal measures improved the prediction of stroke (C-Statistic 0.826 versus 0.792; P=0.017) and correctly reclassified 5.9% of participants with incident stroke and 3.4% of participants with no incident stroke. CONCLUSIONS: Retinal microvascular changes are related to an increased risk of stroke in Asian Malay, consistent with data from white populations. Retinal imaging improves the discrimination and stratification of stroke risk beyond that of established risk factors by a significant but small margin.

Retinal vascular geometry and glaucoma: the Singapore Malay Eye Study.

PURPOSE: To determine the associations of geometric measurements (tortuosity, branching angle, and fractal dimension) of retinal vessels with glaucoma. DESIGN: Population-based, cross-sectional study. PARTICIPANTS: Persons aged 40 to 80 years who participated in the Singapore Malay Eye Study (n=3280; 78.7% response rate). METHODS: Quantitative retinal vascular parameters (tortuosity, branching angle, and fractal dimension) were measured from digital retinal fundus photographs using a computer-assisted program following a standardized grading protocol. Glaucoma was diagnosed according to the International Society of Geographic and Epidemiological Ophthalmology classification system. MAIN OUTCOME MEASURES: The associations among retinal vascular parameters with glaucoma, the main glaucoma subtype primary open-angle glaucoma (POAG), and ocular hypertension (OHT). RESULTS: A total of 123 persons (4.4% of the 2789 participants) had glaucoma in the final analysis, 87 (70.7%) of whom were diagnosed with POAG. After adjusting for age, sex, body mass index, diabetes, hypertension, smoking, axial length, and intraocular pressure (IOP), decreased retinal arteriolar tortuosity (odds ratio [OR], 1.73; 95% confidence interval [CI], 1.38-2.18, comparing lowest vs. highest quartiles), decreased retinal venular tortuosity (OR, 1.59; 95% CI, 1.29-1.97), and narrower retinal venular branching angle (OR, 1.22; 95% CI, 1.00-1.48) were associated with glaucoma. Similar associations were found between these retinal vascular parameters and POAG. Decreased retinal vascular fractal dimension was associated with OHT (OR 1.37; 95% CI, 1.04-1.82). CONCLUSIONS: Certain features of retinal vascular geometry are associated with glaucomatous optic neuropathy independently of vascular risk factors and IOP. FINANCIAL DISCLOSURE(S): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Can the retinal microvasculature offer clues to cardiovascular risk factors in early life?

UNLABELLED: Adulthood cardiovascular diseases (CVD) burdening in adulthood have been suggested to have possible predisposition early in life. We aimed to systematically review the literatures on retinal microvasculature among paediatric population in relation to a range of environmental and systemic changes. Papers with quantitative retinal vascular assessment and epidemiological design were reviewed. CONCLUSION: This review supports the theory that major vascular diseases have pathological predisposition early in life by providing morphological evidence on retinal microcirculation in vivo.

Recent Advances in Clinical Applications of Imaging in Retinal Diseases.

Many diseases that cause visual impairment, as well as systemic conditions, manifest in the posterior segment of the eye. With the advent of high-speed, high-resolution, reliable, and noninvasive imaging techniques, ophthalmologists are becoming more dependent on ocular imaging for disease diagnosis, classification, and management in clinical practice. There are rapid advances on the indications of multimodal retinal imaging techniques, including the application of ultra-widefield fundus angiography, fundus autofluorescence, optical coherence tomography, as well as optical coherence tomography angiography. This review summarizes and highlights the clinical applications, latest indications, and interpretations of multimodal imaging in age-related macular degeneration, polypoidal choroidal vasculopathy, diabetic macular edema, central serous chorioretinopathy, diabetic retinopathy, retinal vein occlusion, and uveitis.

Diagnostic assessment of glaucoma and non-glaucomatous optic neuropathies via optical texture analysis of the retinal nerve fibre layer.

The clinical diagnostic evaluation of optic neuropathies relies on the analysis of the thickness of the retinal nerve fibre layer (RNFL) by optical coherence tomography (OCT). However, false positives and false negatives in the detection of RNFL abnormalities are common. Here we show that an algorithm integrating measurements of RNFL thickness and reflectance from standard wide-field OCT scans can be used to uncover the trajectories and optical texture of individual axonal fibre bundles in the retina and to discern distinctive patterns of loss of axonal fibre bundles in glaucoma, compressive optic neuropathy, optic neuritis and non-arteritic anterior ischaemic optic neuropathy. Such optical texture analysis can detect focal RNFL defects in early optic neuropathy, as well as residual axonal fibre bundles in end-stage optic neuropathy that were indiscernible by conventional OCT analysis and by red-free RNFL photography. In a diagnostic-performance study, optical texture analysis of the RNFL outperformed conventional OCT in the detection of glaucoma, as defined by visual-field testing or red-free photography. Our findings show that optical texture analysis of the RNFL for the detection of optic neuropathies is highly sensitive and specific.

Evaluation of retinal nerve fiber layer progression in glaucoma: a comparison between spectral-domain and time-domain optical coherence tomography.

OBJECTIVE: To compare the performance of a spectral-domain optical coherence tomography (OCT) device and a time-domain OCT device to detect retinal nerve fiber layer (RNFL) progression in glaucoma patients. DESIGN: Prospective study. PARTICIPANTS: One hundred twenty-eight eyes of 81 glaucoma patients. METHODS: Patients were followed up at 4-month intervals for at least 24 months for RNFL imaging and visual field examination. Both eyes were imaged by the Cirrus HD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA) and the Stratus OCT (Carl Zeiss Meditec, Inc.) and underwent visual field testing at the same visit. Linear regression analyses between circumpapillary RNFL measurements (average, superior, and inferior RNFL thicknesses), visual field index (VFI), and follow-up time were performed. RNFL progression and RNFL improvement were identified when a significant negative or positive trend was detected, respectively. The agreement between the OCT instruments for progression detection was analyzed with κ statistics. MAIN OUTCOME MEASURES: Number of eyes with RNFL progression and improvement, agreement of progression detection between RNFL measurements and VFI, and rate of change of average RNFL thickness. RESULTS: Twenty-two eyes (19 patients) and 4 eyes (4 patients) had progression, and 0 and 5 eyes (5 patients) had improvement detected by the Cirrus HD-OCT and the Stratus OCT average RNFL measurements, respectively. The agreement for detection of RNFL progression was poor between the 2 OCT instruments (κ = 0.188, 0.027, and 0.267 for average, superior, and inferior RNFL thicknesses, respectively). The respective agreement between VFI and average RNFL thickness progression determined by the Cirrus HD-OCT and the Stratus OCT was 0.125 and 0.047. The rate of average RNFL thickness progression ranged between -1.52 µm/year and -5.03 µm/year for the Cirrus HD-OCT and between -2.22 µm/year and -7.60 µm/year for the Stratus OCT. CONCLUSIONS: The Cirrus HD-OCT outperformed the Stratus OCT in detecting more eyes with RNFL progression and fewer eyes with RNFL improvement. Because of reduced measurement variability, the Cirrus HD-OCT could detect changes in RNFL thickness sooner than the Stratus OCT. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found after the references.

Evaluation of retinal nerve fiber layer progression in glaucoma a prospective analysis with neuroretinal rim and visual field progression.

OBJECTIVE: To evaluate the performance of progression detection and the rate of change of retinal nerve fiber layer (RNFL), neuroretinal rim, and visual field measurements in glaucoma. DESIGN: Prospective study. PARTICIPANTS: One hundred eight eyes of 70 glaucoma patients. METHODS: Patients were followed up every 4 months for at least 2.9 years (median, 3.2 years) for measurement of RNFL thickness with the Stratus optical coherence tomograph (OCT) (Carl Zeiss Meditec, Dublin, CA), neuroretinal rim area with the Heidelberg Retinal Tomograph (HRT 3; Heidelberg Engineering, GmbH, Dossenheim, Germany), and visual field with the Humphrey Field Analyzer II (Carl Zeiss Meditec). Linear regression analyses were performed between visual field index (VFI), RNFL, and neuroretinal rim measurements and age, with progression defined when a significant negative trend was detected. The agreement among structural and functional measurements was evaluated with κ statistics. The mean rate of change was estimated with linear mixed modeling. MAIN OUTCOME MEASURES: The agreement on progression detection and the rate of change of RNFL, neuroretinal rim, and VFI measurements. RESULTS: A total of 1105 OCT, 1062 HRT, and 1099 visual field measurements were analyzed. The agreement of progression detection among the 3 investigations was poor (κ≤0.09). Ten eyes (9.3%; 9 patients) showed progression by average RNFL thickness, 16 (14.8%; 14 patients) by global neuroretinal rim area, and 35 (32.4%; 31 patients) by VFI. Only 1 eye (0.9%) had progression detected by all 3 methods. There were large variations in the rate of change of VFI, average RNFL thickness, and global neuroretinal rim area, with a range between -0.63% and -4.97% per year, -2.32% and -10.12% per year, and -0.61% and -8.48% per year, respectively. The respective mean rate estimates were -1.15% per year (95% confidence interval [CI], -1.56% to -0.73%), -0.70% per year (95% CI, -1.19% to -0.21%), and -1.06% per year (95% CI, -1.56% to -0.55%). CONCLUSIONS: The agreement of progression detection among RNFL, neuroretinal rim, and visual field measurements was poor, and the rate of RNFL, neuroretinal rim, and visual field progression varied considerably within and between subjects. Given this variability, interpretation of RNFL, neuroretinal rim, and VFI progression always should be evaluated on an individual basis. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found after the references.

Retinal vascular tortuosity, blood pressure, and cardiovascular risk factors.

OBJECTIVE: To examine the relationship of retinal vascular tortuosity to age, blood pressure, and other cardiovascular risk factors. DESIGN: Population-based, cross-sectional study. PARTICIPANTS: A total of 3280 participants aged 40 to 80 years from the Singapore Malay Eye Study (78.7% response rate). METHODS: Retinal arteriolar and venular (vascular) tortuosity were quantitatively measured from fundus images using a computer-assisted program. Retinal vascular tortuosity was defined as the integral of the curvature square along the path of the vessel, normalized by the total path length. Data on blood pressure and major cardiovascular disease (CVD) risk factors were collected from all participants. MEAN OUTCOME MEASURES: Retinal arteriolar and venular tortuosity. RESULTS: A total of 2915 participants contributed data to this study. The mean (standard deviation) and median were 2.99 (1.40) and 2.73 for retinal arteriolar tortuosity (×10(4)), and 4.64 (2.39) and 4.19 for retinal venular tortuosity (×10(4)), respectively. Retinal venules were significantly more tortuous than retinal arterioles (P<0.001). In multivariable-adjusted linear regression models, less arteriolar tortuosity was independently associated with older age, higher blood pressure, higher body mass index (BMI), and narrower retinal arteriolar caliber (all P<0.05); greater venular tortuosity was independently associated with younger age, higher blood pressure, lower high-density lipoprotein (HDL) cholesterol level, and wider retinal venular caliber (all P<0.05). CONCLUSIONS: Retinal arteriolar tortuosity was associated with older age and higher levels of blood pressure and BMI, whereas venular tortuosity was also associated with lower HDL level. The quantitative assessment of retinal vascular tortuosity from retinal images may provide further information regarding effects of cardiovascular risk factors on the retinal vasculature.

Evaluation of retinal nerve fiber layer progression in glaucoma: a comparison between the fast and the regular retinal nerve fiber layer scans.

OBJECTIVE: To compare the performance of the fast (256 A-scans in each scan circle) and the regular (512 A-scans in each scan circle) retinal nerve fiber layer (RNFL) scan protocols for detection of glaucoma progression using the Stratus optical coherence tomography (OCT) device (Carl Zeiss Meditec, Dublin, CA). DESIGN: Retrospective, longitudinal study. PARTICIPANTS: One hundred twenty-nine eyes from 72 glaucoma patients. METHODS: All patients had been followed up for 2.9 to 6.1 years with a median follow-up of 4 months. All eyes had at least 4 serial RNFL measurements obtained with both the fast and the regular RNFL scans. Visual field (VF) assessment was performed on the same day as RNFL imaging. Retinal nerve fiber layer thickness and VF progression were evaluated with linear regression analysis against age. The mean rate of average RNFL thickness reduction was estimated with linear mixed modeling. MAIN OUTCOME MEASURES: The agreement of progression detection and the rate of change of RNFL thicknesses. RESULTS: A total of 1373 fast and 1373 regular RNFL scans and 1236 VF tests were analyzed. With reference to the average RNFL thickness, the fast RNFL scan detected more eyes with progression (21 eyes from 19 patients vs. 15 eyes from 13 patients) than the regular scan at a comparable level of specificity (96.9% vs. 96.1%). More eyes were found to have increasing RNFL thickness with age at individual clock hours (except for 3, 5, 6, and 11 o'clock) when the measurements were obtained with the regular scan. The agreement between the fast and the regular scan for detection of RNFL progression was fair to moderate, with κ values ranging between 0.14 and 0.49. The rate of average RNFL thickness progression was -1.01 µm per year for the fast RNFL scan and -0.77 µm per year for the regular scan. CONCLUSIONS: The choice of scan protocols in the Stratus OCT has a significant impact in the evaluation of RNFL progression. The fast RNFL scan seems to be preferable to follow RNFL damage in glaucoma.

Influence of blood pressure on retinal vascular caliber in young children.

OBJECTIVE: Prior studies have reported that blood pressure (BP) has a significant influence on retinal vascular caliber both in adults and children aged 6 years and older. This study aimed to examine the association between BP and retinal vascular caliber in Singapore Chinese preschoolers 4 to 5 years of age. DESIGN: Population-based, cross-sectional study. PARTICIPANTS: A total of 385 eligible and healthy Singapore Chinese children 4 to 5 years of age who were recruited in The Strabismus, Amblyopia and Refractive Error Study in Singaporean Chinese Preschoolers from May 2006 through October 2008 underwent BP measurements and retinal photography. METHODS: According to standard protocols, BP was measured with an automatic Omron sphygmomanometer (Omron HEM 705 LP, Omron Healthcare, Inc., Bannockburn, IL) and a retinal photograph was obtained with a Canon 45° digital retinal camera (Model CR6-NM45, Canon, Inc., Tokyo, Japan) after pupil dilation. Anthropometric and optical biometric measurements such as height, weight, and axial length were obtained also. Information regarding sociodemographic status and child birth information was supplied by parents in either English or Chinese questionnaires. MAIN OUTCOME MEASURES: The computer imaging program was used to measure the caliber of all retinal arterioles and venules located in zone B. The central retinal arteriolar equivalent and central retinal venular equivalent were estimated by using a revised Knudtson-Parr-Hubbard formula. RESULTS: The mean retinal arteriolar and venular calibers were 156.19 µm and 220.01 µm in boys and 161.97 µm and 224.22 µm in girls. Higher systolic BP was associated with narrower retinal arterioles. After adjusting for age, gender, father's education, body mass index, birth weight, axial length, and caliber of the fellow retinal vessel, each 10-mmHg increase in systolic BP was associated with 2.00 µm (95% confidence interval, 0.39-3.61; P = 0.02) of retinal arteriolar narrowing and 2.51 µm (95% confidence interval, 0.35-4.68; P = 0.02) of retinal venular widening. However, neither diastolic BP nor mean arterial BP was associated with retinal arteriolar or venular caliber. CONCLUSIONS: In very young children 4 to 5 years of age, higher systolic BP was associated with narrower retinal arterioles and wider retinal venules. This suggests that elevated BP may affect the retinal microvasculature from early childhood.

Screening and identifying hepatobiliary diseases through deep learning using ocular images: a prospective, multicentre study.

BACKGROUND: Ocular changes are traditionally associated with only a few hepatobiliary diseases. These changes are non-specific and have a low detection rate, limiting their potential use as clinically independent diagnostic features. Therefore, we aimed to engineer deep learning models to establish associations between ocular features and major hepatobiliary diseases and to advance automated screening and identification of hepatobiliary diseases from ocular images. METHODS: We did a multicentre, prospective study to develop models using slit-lamp or retinal fundus images from participants in three hepatobiliary departments and two medical examination centres. Included participants were older than 18 years and had complete clinical information; participants diagnosed with acute hepatobiliary diseases were excluded. We trained seven slit-lamp models and seven fundus models (with or without hepatobiliary disease [screening model] or one specific disease type within six categories [identifying model]) using a development dataset, and we tested the models with an external test dataset. Additionally, we did a visual explanation and occlusion test. Model performances were evaluated using the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and F1* score. FINDINGS: Between Dec 16, 2018, and July 31, 2019, we collected data from 1252 participants (from the Department of Hepatobiliary Surgery of the Third Affiliated Hospital of Sun Yat-sen University, the Department of Infectious Diseases of the Affiliated Huadu Hospital of Southern Medical University, and the Nantian Medical Centre of Aikang Health Care [Guangzhou, China]) for the development dataset; between Aug 14, 2019, and Jan 31, 2020, we collected data from 537 participants (from the Department of Infectious Diseases of the Third Affiliated Hospital of Sun Yat-sen University and the Huanshidong Medical Centre of Aikang Health Care [Guangzhou, China]) for the test dataset. The AUROC for screening for hepatobiliary diseases of the slit-lamp model was 0·74 (95% CI 0·71-0·76), whereas that of the fundus model was 0·68 (0·65-0·71). For the identification of hepatobiliary diseases, the AUROCs were 0·93 (0·91-0·94; slit-lamp) and 0·84 (0·81-0·86; fundus) for liver cancer, 0·90 (0·88-0·91; slit-lamp) and 0·83 (0·81-0·86; fundus) for liver cirrhosis, and ranged 0·58-0·69 (0·55-0·71; slit-lamp) and 0·62-0·70 (0·58-0·73; fundus) for other hepatobiliary diseases, including chronic viral hepatitis, non-alcoholic fatty liver disease, cholelithiasis, and hepatic cyst. In addition to the conjunctiva and sclera, our deep learning model revealed that the structures of the iris and fundus also contributed to the classification. INTERPRETATION: Our study established qualitative associations between ocular features and major hepatobiliary diseases, providing a non-invasive, convenient, and complementary method for hepatobiliary disease screening and identification, which could be applied as an opportunistic screening tool. FUNDING: Science and Technology Planning Projects of Guangdong Province; National Key R&D Program of China; Guangzhou Key Laboratory Project; National Natural Science Foundation of China.

Referral for disease-related visual impairment using retinal photograph-based deep learning: a proof-of-concept, model development study.

BACKGROUND: In current approaches to vision screening in the community, a simple and efficient process is needed to identify individuals who should be referred to tertiary eye care centres for vision loss related to eye diseases. The emergence of deep learning technology offers new opportunities to revolutionise this clinical referral pathway. We aimed to assess the performance of a newly developed deep learning algorithm for detection of disease-related visual impairment. METHODS: In this proof-of-concept study, using retinal fundus images from 15 175 eyes with complete data related to best-corrected visual acuity or pinhole visual acuity from the Singapore Epidemiology of Eye Diseases Study, we first developed a single-modality deep learning algorithm based on retinal photographs alone for detection of any disease-related visual impairment (defined as eyes from patients with major eye diseases and best-corrected visual acuity of <20/40), and moderate or worse disease-related visual impairment (eyes with disease and best-corrected visual acuity of <20/60). After development of the algorithm, we tested it internally, using a new set of 3803 eyes from the Singapore Epidemiology of Eye Diseases Study. We then tested it externally using three population-based studies (the Beijing Eye study [6239 eyes], Central India Eye and Medical study [6526 eyes], and Blue Mountains Eye Study [2002 eyes]), and two clinical studies (the Chinese University of Hong Kong's Sight Threatening Diabetic Retinopathy study [971 eyes] and the Outram Polyclinic Study [1225 eyes]). The algorithm's performance in each dataset was assessed on the basis of the area under the receiver operating characteristic curve (AUC). FINDINGS: In the internal test dataset, the AUC for detection of any disease-related visual impairment was 94·2% (95% CI 93·0-95·3; sensitivity 90·7% [87·0-93·6]; specificity 86·8% [85·6-87·9]). The AUC for moderate or worse disease-related visual impairment was 93·9% (95% CI 92·2-95·6; sensitivity 94·6% [89·6-97·6]; specificity 81·3% [80·0-82·5]). Across the five external test datasets (16 993 eyes), the algorithm achieved AUCs ranging between 86·6% (83·4-89·7; sensitivity 87·5% [80·7-92·5]; specificity 70·0% [66·7-73·1]) and 93·6% (92·4-94·8; sensitivity 87·8% [84·1-90·9]; specificity 87·1% [86·2-88·0]) for any disease-related visual impairment, and the AUCs for moderate or worse disease-related visual impairment ranged between 85·9% (81·8-90·1; sensitivity 84·7% [73·0-92·8]; specificity 74·4% [71·4-77·2]) and 93·5% (91·7-95·3; sensitivity 90·3% [84·2-94·6]; specificity 84·2% [83·2-85·1]). INTERPRETATION: This proof-of-concept study shows the potential of a single-modality, function-focused tool in identifying visual impairment related to major eye diseases, providing more timely and pinpointed referral of patients with disease-related visual impairment from the community to tertiary eye hospitals. FUNDING: National Medical Research Council, Singapore.

A new method to measure peripheral retinal vascular caliber over an extended area.

OBJECTIVE: To describe a new computer-assisted method to measure retinal vascular caliber over an extended area of the fundus. METHODS: Retinal photographs taken from participants of the Singapore Malay Eye Study (n = 3280) were used for this study. Retinal vascular caliber was measured and summarized as central retinal artery equivalent (CRAE) and central retinal vein equivalent (CRVE) using a new semi-automated computer-based program. Measurements were made at the Standard zone (from 0.5 to 1.0 disk diameter) and an Extended zone (from 0.5 to 2.0 disk diameter). RESULTS: Reliability of retinal vascular caliber measurement was high for the new Extended zone (intraclass correlation coefficients >0.90). Associations of CRAE with blood pressure were identical between the Extended and Standard zones (linear regression coefficient -2.53 vs. -2.61, z-test between the two measurements, p = 0.394). Associations of CRAE and CRVE with other cardiovascular risk factors were similar between measurements in the two zones. The R² of regression models for the Extended zone was slightly higher than that for the Standard zone for both CRAE (R², 0.324 vs. 0.288) and CRVE (R², 0.325 vs. 0.265). CONCLUSIONS: The new measures from Extended zone are comparable with the previous measures, and also more representative of retinal vascular caliber.

Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography a study on diagnostic agreement with Heidelberg Retinal Tomograph.

OBJECTIVE: To evaluate and compare the diagnostic agreement and performance for glaucoma detection between a confocal scanning laser ophthalmoscope and a spectral-domain optical coherence tomograph (OCT). DESIGN: Prospective, cross-sectional study. PARTICIPANTS: One hundred fifty-five subjects (79 glaucoma and 76 normal subjects). METHODS: One eye from each individual was selected randomly for optic disc and retinal nerve fiber layer (RNFL) imaging by the Heidelberg Retinal Tomograph (HRT; Heidelberg Engineering, GmbH, Dossenheim, Germany) and the Spectralis OCT (Heidelberg Engineering), respectively. Glaucoma was defined based on the presence of visual field defects with the Humphrey visual field analyzer (Carl Zeiss Meditec, Dublin, CA). The agreement of the categorical classification ("within normal limits," "borderline," and "outside normal limits") at the temporal, superotemporal, superonasal, nasal, inferonasal and inferotemporal sectors of the optic disc were evaluated (kappa statistics). The diagnostic sensitivity and specificity between optic disc and RNFL assessment were compared (McNemar's statistics). Area under the receiver operating characteristic curve (AUC) of OCT RNFL and HRT optic disc parameters were computed after adjustment of age, axial length, and optic disc area. MAIN OUTCOME MEASURES: Agreement of categorical classification, AUC of optic disc, and RNFL parameters. RESULTS: The agreement of categorical classification between HRT and Spectralis OCT were fair to moderate (kappa ranged between 0.30 and 0.53) except for global (kappa = 0.63) and inferotemporal (kappa = 0.68) measurements. Defining glaucoma as having "outside normal limits" in the global and/or in >or=1 of the sectoral measurements, the respective sensitivities of Spectralis OCT and HRT were 91.1% and 79.8% (P = 0.012) at a similar level of specificity (97.4% and 94.7%). The AUC of OCT global RNFL thickness (0.978) was greater than those of HRT global rim area (0.905), vertical cup-disc ratio (0.857), rim-disc area ratio (0.897), and multivariate discriminant analysis (0.880-0.925; all with P